docCalnuc Binds to LRP9 and Affects its
download 
Report Transcription
Calnuc Binds to LRP9 and Affects its
Université de Sherbrooke Rôle de Calnuc dans le triage endosomial des récepteurs lysosomiaux et implication potentielle dans les maladies du lysosome Par Heidi Larkin Programme de pharmacologie Thèse présentée à la Faculté de médecine et des sciences de la santé en vue de l’obtention du grade de philosophiae doctor (Ph.D.) en pharmacologie Sherbrooke, Québec, Canada 21 janvier 2016 Membres du jury d’évaluation Dre. LAVOIE Christine Directrice de recherche, Pre. Département de pharmacologie, Sherbrooke Dr. PSHEZHETSKY Alexey Juge externe, Pr. Université de Montréal, Montréal Dr. ROUCOU Xavier Juge interne, Pr. Département de biochimie, Sherbrooke Dr. GRANDBOIS Michel Directeur de programme, Pr. Département de pharmacologie, Sherbrooke © Heidi Larkin, 2015 RÉSUMÉ FRANÇAIS Rôle de Calnuc dans le triage endosomial des récepteurs lysosomiaux et implication potentielle dans les maladies du lysosome Par Heidi Larkin Programme de doctorat en pharmacologie Thèse présentée à la Faculté de médecine et des sciences de la santé en vue de l’obtention du diplôme de philosophiae doctor (Ph.D.) en pharmacologie à la Faculté de médecine et des sciences de la santé, Université de Sherbrooke, Sherbrooke, Québec, Canada, J1H 5N4 Calnuc est une protéine ubiquitaire qui lie le calcium et qui est présente au réseau transgolgien (TGN) ainsi qu'aux endosomes. Notre groupe a précédemment mis en évidence le rôle de Calnuc dans le transport de Low density lipoprotein receptor-related protein 9 (LRP9), un récepteur aux lipoprotéines de faible densité qui cycle entre le TGN et les endosomes. Les récepteurs lysosomiaux au mannose-6-phosphate (MPR) et Sortiline sont bien caractérisés et empruntent également cette voie. À l'image de LRP9, nous avons montré que Calnuc prévient leur dégradation aux lysosomes en participant à leur recyclage à partir des endosomes vers le TGN. En fait, Calnuc est importante pour l'activation et l'association membranaire de Rab7, une petite protéine G qui recrute ensuite le complexe Rétromère responsable du transport rétrograde des récepteurs. La glycoprotéine lysosomiale Ceroid lipofuscinosis neuronal 5 (CLN5) est également impliquée dans ce processus. La structure et la fonction de cette dernière n'étant pas clairement définies, nous avons établi qu'elle est synthétisée sous forme d’une glycoprotéine transmembranaire de type II, mais son domaine N-terminal cytoplasmique et son segment transmembranaire sont rapidement éliminés suivant le clivage du peptide signal de manière à former une protéine CLN5 mature fortement associée à la membrane par une hélice amphipathique (AH). La compréhension des propriétés de base de CLN5 est particulièrement pertinente puisque la protéine est impliquées dans certaines variantes de céroïdeslipofuscinoses neuronales (NCL), une maladie neurodégénérative rare causée par une surcharge des lysosomes. D'ailleurs, nos données indiquent que les mutants pathologiques de CLN5 dépourvus de cette AH perdent leur association membranaire, sont retenus au réticulum endoplasmique et sont rapidement dégradés. En raison de la similitude des fonctions de Calnuc et de CLN5 au niveau du triage endosomial, nous avons exploré le lien entre les deux protéines. Calnuc cytosolique et CLN5 luminale semblent former un complexe, par l'intermédiaire de la protéine transmembranaire CLN3, de façon à influencer l'activité de Rab7. CLN3 étant aussi associée aux NCL, nous avons finalement exploré la potentielle implication de Calnuc dans la maladie. L'absence de Calnuc entraîne des phénotypes cellulaires typiques des NCL comme un engorgement des lysosomes, une accumulation de matériel autofluorescent et une augmentation de l'autophagie. Les niveaux protéiques de Calnuc sont diminués dans toutes les lignées de fibroblastes de patients atteints de NCL disponibles ce qui indique que Calnuc pourrait être impliquée dans certains types de NCL. La présente thèse couvre donc la découverte de la fonction de Calnuc dans le transport intracellulaire, jusqu'à son implication potentielle dans les NCL, de même qu'une étude topologique de CLN5. Mots clés : Calnuc, CLN3, CLN5, MPR, Sortiline, Rab, Rétromère, lipofuscinose RÉSUMÉ ANGLAIS Calnuc fonction in endosomal sorting of lysosomal receptors and potential implication in lysosomal diseases Par Heidi Larkin Programme de doctorat en pharmacologie Thèse présentée à la Faculté de médecine et des sciences de la santé en vue de l’obtention du diplôme de philosophiae doctor (Ph.D.) en pharmacologie à la Faculté de médecine et des sciences de la santé, Université de Sherbrooke, Sherbrooke, Québec, Canada, J1H 5N4 Calnuc is a ubiquitous Ca2+-binding protein present on the trans-Golgi network (TGN) and endosomes. We previously highlighted the role of Calnuc in the transport of Low density lipoprotein receptor-related protein 9 (LRP9), a low density lipoprotein (LDL) receptor that cycles between the TGN and endosomes. Lysosomal receptors mannose-6-phosphate receptor (MPR) and Sortilin are well-characterized and also use the TGN-to-endosome trafficking pathway. Similarly to LPR9, we showed that Calnuc prevent their degradation in lysosomes by acting in their recycling from endosomes to the TGN. In fact, Calnuc is a important for the activation and the membrane association of Rab7, a small G protein which then recruit the Retromer complex known to be responsible for the retrograde transport of receptors. Lysosomal glycoprotein Ceroid lipofuscinosis neuronal 5 (CLN5) is also involved in this process. Because its structure and function have not yet been clearly defined, we established that it is synthesized as a type II transmembrane (TM) glycoprotein, but its cytoplasmic N-terminus and TM segment are rapidly removed following signal-peptide cleavage to generate mature CLN5 which is tightly associated to membrane through an amphipathic helix (AH). The understanding of the basic properties of CLN5 is particularly important given that CLN5 is involved in some variants of neuronal ceroid lipofuscinosis (NCL), a rare neurodegenerative disease caused by lysosomal overload. Moreover, our data indicate that CLN5 pathological mutants deprived of AH lose their membrane association, are retained in the endoplasmic reticulum, and are rapidly degraded. Based on the similarity featured by Calnuc and CLN5 in endosomal sorting, we explored the link between these two proteins. Cytosolic Calnuc and luminal CLN5 seem to form a complex, through the transmembrane protein CLN3, in order to influence the activity of Rab7. As CLN3 is also associated with NCL, we finally explored the potential involvement of Calnuc in this disease. Canuc depletion leads to typical NCL phenotypes such as lysosome enlargement, accumulation of autofluorescent material and of an increased of autophagy induction. Canuc's levels are decreased in all fibroblasts cell lines of NCL patients available indicating that Calnuc could be involved in some types of NCL. This thesis thus covers the discovery of the function of Calnuc in intracellular transport up to its potential involvement in the NCL, as well as a topological study CLN5. Mots clés : Calnuc, CLN3, CLN5, MPR, Sortiline, Rab, rétromère, lipofuscinosis iii À tous les étudiants qui passeront à travers ce tumultueux périple Aux Débrouillards qui influencèrent une génération de scientifiques Mais surtout, En mémoire de mon oncle Paul que je n'ai que trop brièvement connu La science ne sert guère qu'à nous donner une idée de l'étendue de notre ignorance. Félicité Robert de Lamennais TABLE DES MATIÈRES RÉSUMÉ FRANÇAIS .........................................................................................................II RÉSUMÉ ANGLAIS ......................................................................................................... III LISTE DES FIGURES ................................................................................................... VIII LISTE DES TABLEAUX .................................................................................................. IX LISTE DES ABRÉVIATIONS ........................................................................................... X INTRODUCTION ................................................................................................................ 1 1 CALNUC ....................................................................................................................... 1 1.1 Une nouvelle protéine d’intérêt .......................................................................... 1 1.2 Structure et localisation cellulaire ..................................................................... 2 1.3 Homologues ........................................................................................................ 6 1.3.1 Cab45 .............................................................................................................. 7 1.3.2 Calréticuline .................................................................................................... 7 1.3.3 NUCB2 ........................................................................................................... 8 1.4 Partenaires d’interaction, fonctions et maladies associées.............................. 12 1.4.1 Calcium et minéralisation ............................................................................. 14 1.4.2 Protéines G et sécrétion régulée.................................................................... 16 1.4.3 ADN et lupus ................................................................................................ 19 1.4.4 Cancer, apoptose et Caspases ....................................................................... 20 1.4.5 COX et inflammation.................................................................................... 21 1.4.6 Activité de protéase à sérine ......................................................................... 23 1.4.7 ATF6 et réponse au stress réticulaire ............................................................ 23 1.4.8 Necdin ........................................................................................................... 24 1.4.9 APP et Alzheimer ......................................................................................... 25 1.4.10 LRP10 et transport intracellulaire ................................................................. 26 2 TRANSPORT INTRACELLULAIRE ................................................................................. 27 2.1 Transport antérograde ...................................................................................... 27 2.1.1 Transport dépendant de M6P - MPR ............................................................ 28 2.1.2 Transport indépendant de M6P - Sortiline.................................................... 29 2.1.3 Transport du Golgi aux endosomes .............................................................. 31 2.1.4 Système endolysosomial ............................................................................... 32 2.2 Transport rétrograde ........................................................................................ 35 2.2.1 Transport des endosomes précoces au Golgi ................................................ 36 2.2.2 Transport des endosomes tardifs au Golgi .................................................... 41 2.3 Rab comme principaux régulateurs du transport intracellulaire ..................... 42 2.3.1 Cycle d'activation des Rab ............................................................................ 42 2.3.2 Localisation des Rab ..................................................................................... 44 2.3.3 Fonctions des Rab dans le transport entre le Golgi et les endosomes .......... 45 3 MALADIES ASSOCIÉES AUX LYSOSOMES .................................................................... 49 3.1 Lysosomes ......................................................................................................... 49 3.2 Maladies de surcharge lysosomiale .................................................................. 51 3.3 Céroïdes-lipofuscinoses neuronales ................................................................. 52 3.3.1 Causes génétiques et protéines associées...................................................... 55 3.3.2 CLN3 ............................................................................................................ 56 3.3.3 CLN5 ............................................................................................................ 61 3.3.4 Interactosome des CLN ................................................................................ 65 3.3.5 Aspects cliniques .......................................................................................... 66 4 PROBLÉMATIQUE, HYPOTHÈSE ET OBJECTIFS ............................................................. 69 4.1 Problématique ................................................................................................... 69 4.2 Hypothèse.......................................................................................................... 69 4.3 Objectifs ............................................................................................................ 70 ARTICLE 1 ......................................................................................................................... 71 RÉSUMÉ – ARTICLE 1 .................................................................................................... 72 ARTICLE 2 ....................................................................................................................... 111 RÉSUMÉ – ARTICLE 2 .................................................................................................. 112 ARTICLE 3 ....................................................................................................................... 146 RÉSUMÉ – ARTICLE 3 .................................................................................................. 147 DISCUSSION .................................................................................................................... 181 CONCLUSION ................................................................................................................. 196 REMERCIEMENTS ........................................................................................................ 197 LISTE DES PUBLICATIONS ........................................................................................ 200 ANNEXE 1......................................................................................................................... 231 ANNEXE 2......................................................................................................................... 254 ANNEXE 3......................................................................................................................... 284 vii LISTE DES FIGURES FIGURE 1 - STRUCTURE DE CALNUC. ....................................................................................... 3 FIGURE 2 - COMPARAISON ENTRE CALNUC (NUCB1) ET NUCB2. ........................................ 11 FIGURE 3 - DISTRIBUTION DE CALNUC ET DE SES DIVERS PARTENAIRES D'INTERACTION. ...... 13 FIGURE 4 - PRINCIPAUX MÉCANISMES DE TRANSPORT ANTÉROGRADE ET RÉTROGRADE. ....... 33 FIGURE 5 - LES RAB COMME PRINCIPAUX RÉGULATEURS DU TRANSPORT INTRACELLULAIRE. ...................................................................................................................................... 43 FIGURE 6 - FORMATION ET ULTRASTRUCTURES DE LA LIPOFUSCINE. ..................................... 54 FIGURE 7 - SCHÉMA DE LA STRUCTURE DE CLN3. ................................................................. 58 FIGURE 8 - SCHÉMA DE LA STRUCTURE DE CLN5. ................................................................. 62 FIGURE 9 - CALNUC INTERAGIT AVEC FURINE, TRKA ET CD-MPR, MAIS PAS AVEC P75/NTR OU EGFR. .................................................................................................................... 183 FIGURE 10 - LA DÉPLÉTION DE CALNUC INFLUENCE LES NIVEAUX DE CI-MPR, FURINE, CDMPR ET EGFR. ........................................................................................................... 184 FIGURE 11 - RÔLES POTENTIELS DE CALNUC DANS LE CYCLE D'ACTIVATION DES RAB. ....... 185 FIGURE 12 - MODÈLE DU COMPLEXE CALNUC-CLN3-CLN5 INFLUENÇANT LE RECRUTEMENT DE RAB7 ET DU COMPLEXE RÉTROMÈRE. ..................................................................... 192 LISTE DES TABLEAUX TABLEAU 1 - PATHOLOGIE DES NCL. .................................................................................... 56 TABLEAU 2 - PRINCIPALES MUTATIONS DANS CLN5. ............................................................ 65 LISTE DES ABRÉVIATIONS A aa, Acide aminé AA, Acide arachidonique AAK1, Adaptor-associated protein kinase 1 ACTH, Hormone adrénocorticotropine ADN, Acide désoxyribonucléique AICD, Domaine intracellulaire d’APP AP, Adaptator protein ApoE, Apolipoprotéine E APP, Protéine précurseur de la β-amyloïde AR, Région acide (Acidic region) Arf-1, ADP-ribosylation factor-binding protein 1 ARN, Acide ribonucléique ARNm, ARN messager ARTS-1, Aminopeptidase regulator of TNRF1 shedding ASM, Sphingomyélinase ATF6, Activating transcription factor 6 ATP13A2, Probable cation-transporting ATPase 13A2 Aβ, Peptide β-amyloïde B BACE1, ß-sécrétase BAR, Bin-Amphiphysine-Rvs Btn, Battenine BCMP84, Breast cancer membrane protein BDNF, Brain derived neurotrophic factor BR, Région basique (Basic region) C Ca2+, Calcium Cab45, 45 kDa calcium-binding protein Casp, Caspase CCDC53, Coiled-coil domain-containing protein 53 CD-MPR, Cation-dependant MPR CerS, Ceramide synthase CI-MPR, Cation-independant MPR CL, Curviligne CLC-7, Chloride channel 7 CLEAR, Coordinated Lysosomal Expression And Regulation CLN, Ceroid lipofuscinosis neuronal CMH, Complexe majeur d'histocompatibilité COG, Conserved oligomeric Golgi CORVET, Class C core vacuole/endosome tethering COX, Cyclooxygénase CPY, Carboxypeptidase Y Creg1, E1A-stimulated genes 1 CRH, Corticolibérine CRT, Caréticuline CSP, Cysteine string protein CTF, Fragment C-terminal CTSB, Cathepsine B CTSD, Cathepsine D CTSF, Cathepsine F CUB, Complement C1r/C1s, Uegf, Bmp1 D DIC, Chaîne intermédiaire de la Dynéine DMT1, Divalent metal transporter 1 DNAJC5, DnaJ homolog subfamily C member 5 E E1A, Adenovirus early region 1A protein EEA1, Early endosome antigen 1 EF, Main EF (EF-hand) EHD1, EH domain-containing protein 1 ENT1, Equilibrative nucleoside transporter 1 ERC55, Endoplasmic reticulum calcium-binding protein of 55 kDa ERSE, Élément de réponse au stress réticulaire EUR, Européenne F FAAH, Amidohydrolase des acides gras Fe65, Amyloid beta A4 precursor protein-binding family B member 1 FinM, Finlandaise majeure Finm, Finlandaise mineure FP, Empreinte digitale (fingerprint) FYCO1, FYVE and coiled-coil domain–containing protein 1 G GAE, Gamma-adaptin ear GAP, Protéine d’accélération de la GTPase GAPvD1, GTPase activating protein and VPS9 domains 1 GARP, Golgi-associated retrograde protein GAT, GGA and TOM GDF, Facteur de déplacement de GDI xi GDI, Inhibiteur de la séparation des nucléotides de guanine GDP, Guanosine diphosphate GEF, Facteur d’échange des nucléotides de guanine GGA, Golgi-localized, gamma-ear containing Arf binding protein GGT, Geranylgeranyl transférase GM, Granulocyte/monocyte colony stimulation factor GM130, Golgi matrix protein 130 GPCR, Récepteur couplé aux protéines G GPI, Glycosylphosphatidyl inositol GROD, Granular osmiophilic deposit GRN, Granulins precursor GTP, Guanosine triphosphate H HOPS, Homotypic fusion and vacuole protein sorting Hpse, Héparanase Hsc70, Heat shock cognate protein 70 I IGF-2, Insulin-like growth factor-2 IgG, Immunoglobuline G J JNK, Jun N-terminal kinase K KCTD7, Potassium channel tetramerisation domain containing 7 Kd, Constante de dissociation KIF3A, Kinesin Family Member 3A L LAMP, Lysosome-associated membrane protein LC3, Microtubule-associated protein 1A/1B-light chain 3 LD, Gouttelette lipidique (Lipid droplet) LDL-A, Répétition de classe A LDLR, Récepteur aux lipoprotéines de faible densité LIMP, Lysosomal integral membrane protein Lpl, Lipoprotein lipase LRP, Low density lipoprotein receptor-related protein LSD, Maladie de surcharge lysosomiale (Lysosomal storage disease) LYNUS, Lysosome nutrient sensing LysoNaATP, canal sodique ATP-sensible LZ, Fermeture/Glissière en leucine (Leucine zipper) M M6P, Mannose-6-phosphate Man2b1, α-Mannosidase B1 xii MFSD8, Major facilitator superfamily domain-containing protein 8 Mg2+, Magnésium MPR, Récepteur au mannose-6-phosphate (Mannose-6-phosphate receptor) mTORC1, Mammalian target of rapamycin complex 1 N NCL, Céroïde-lipofuscinose neuronale (Neuronal ceroid lipofuscinose) Necdin, Neurally differentiated embryonal carcinoma-derived protein NEFA, DNA binding, EF-hand, acidic region Nesfatin, NEFA/NUCB2-encoded satiety- and fat-influencing protein NFL, Terre-Neuve NGF, Nerve growth factor NLP, Nesfatin-1-like peptide NLS, Signal de localisation nucléaire (Nuclear localisation signal) NPC1, Niemann-Pick C1 protein NSF, N-éthylmaleimide-sensitive-factor NT, Neurotensine NTR3, Neurotensin receptor 3 NUC, Nucleobindin NUCB1, Calnuc O OCRL1, Lowe oculocerebrorenal syndrome protein ou Inositol polyphosphate 5phosphatase ORP1L, Oxysterol‐binding protein P PAI-2, Inhibiteur 2 de l’activateur du Plasminogène (Plasminogen activator inhibitor-2) PAT, Perilipine, Adipophiline, TIP47 PC, Prohormone convertase PDZ, Postsynaptic density 95/discs large/zonus occludens-1 PG, Prostaglandine PI(3)P, Phosphatidylinositol-3-phosphate PI(4,5)P2, Phosphatidylinositol-4,5-bisphosphate PI, Phosphatidylinositol PI3K, Phosphatidylinositol-3-kinase PPAR-g, Peroxisome proliferator-activated receptor gamma pRB, Retinoblastoma protein Protéine G, Guanosine nucleotide-binding protein Psap, Prosaposine PX, Phox homology R RAB11FIP2, RAB11 family-interacting protein 2 REP, Protéine escorte de Rab xiii RhoBTB3, Rho-related BR-C, ttk and bab (BTB) domain-containing protein 3 RL, Rectiligne RILP, Rab7-interacting lysosomal protein RT-PCR, Transcription inverse combinées à des réactions en chaîne par polymérase S S1P, Site-1-protease SAP, Saposine ou Sphingolipid activator protein SBDS, Shwachmann-Bodian-Diamond syndrome protein SCMAS, Sous-unité C de l'ATP synthase mitochondriale SDF4, Stromal cell derived factor SERCA, Sarco/endoplasmic reticulum Ca2+-ATPase shRNA, Petits ARN en épingle à cheveux siRNA, Petits ARN interférents SLC29, Solute carrier family 29 SNAP, NSF attachment protein SNARE, Soluble N-éthylmaleimide-sensitive-factor (NSF) attachment protein (SNAP) receptor SNX, Sorting Nexin SorLA, Sorting protein-related receptor containing LDLR class A repeats SP, Peptide signal (Signal peptide) sp1, Specificity protein 1 SRP, Particule de reconnaissance du signal (Signal recognition particule) SWE, Suédoise SV40-TAg, Simian Vacuolating Virus 40 large T antigen SWIP, Strumpellin and WASH-interacting protein T TAA, Antigène associé aux tumeurs TBC1D5, Tre2-Bub2-Cdc16-domain-containing protein TFEB, Facteur de transcription EB TfR, Transferrin receptor TGN, Réseau trans-golgien (trans Golgi network) TIP47, Tail interacting protein of 47 kDa Tip60, 60 kDa Tat-interactive protein TNF, Tumor necrosis factor TNFR1, Tumor necrosis factor receptor 1 TPP1, Tripeptidyl peptidase I TRAPP, Transport protein particle TRPML1, Transient receptor potential cation channel, mucolipin subfamily, member 1 U uPAR, Récepteur de l'Urokinase UPR, Unfold protein response UPRE-like, UPR element-like xiv V VAMP, Vesicule-associated membrane protein VHS, Vps27, Hrs, Stam Vps, Protéine de triage vacuolaire (Vacuolar protein sorting) W WASH, Wiskott–Aldrich Syndrome protein (WASP) and Suppressor of cAR (cAMP receptor) (SCAR) homologue WASP, Wiskott–Aldrich Syndrome protein Z ZBD, Domaine de liaison au zinc (Zinc-binding domain) Zn2+, Zinc Autres β2-AR, Récepteur bêta-adrénergique Ø, Résidu hydrophobe encombrant Φ, Résidu aromatique Ψ, Résidu aliphatique xv INTRODUCTION Le transport des hydrolases acides jusqu’aux lysosomes nécessite leur triage, dans le réseau trans-golgien (TGN), par des récepteurs tels que le Récepteur au mannose-6phosphate (MPR) et Sortiline. Ces derniers permettent l’empaquetage des enzymes dans des vésicules de clathrine puis leur transport antérograde du TGN vers les endosomes. Dans l’environnement acide des endosomes, les hydrolases se dissocient des récepteurs et sont acheminées aux lysosomes pour assurer la dégradation des macromolécules dispensables à la cellule tandis que les récepteurs sont redirigés vers le TGN pour un nouveau cycle de transport. Ce transport rétrograde requiert de nombreuses protéines dont le complexe Rétromère et la protéine Ceroid lipofuscinosis neuronal 5 (CLN5). L’ensemble des mécanismes de régulation du triage endosomial est essentiel pour maintenir la fonction lysosomiale. En effet, un défaut à ce niveau influence la livraison des hydrolases aux lysosomes, causant une accumulation intracellulaire de métabolites qui sont caractéristiques des maladies de désordres lysosomiaux, comme les céroïdes-lipofuscinoses neuronales (NCL). Nous avons récemment démontré l’implication de Calnuc dans le triage rétrograde de Low density lipoprotein receptor-related protein 9 (LRP9), un récepteur peu caractérisé qui cycle entre le TGN et les endosomes. Le but de la présente étude est, dans un premier temps, de démontrer que Calnuc agit comme régulateur général du transport rétrograde des récepteurs des enzymes lysosomiales, CI-MPR et Sortiline. D’autre part, nous examinerons le rôle de Calnuc dans les désordres lysosomiaux de type NCL. 1 Calnuc 1.1 Une nouvelle protéine d’intérêt Calnuc a été isolée en 1986 à partir d’un surnageant de lymphocytes B issu d’un modèle murin (MRL/lpr) de lupus érythémateux disséminé, une maladie systémique autoimmune chronique (Kanai et al., 1986). Originalement nommée Nucleobindin (NUC) en raison de sa capacité à lier l’acide désoxyribonucléique (ADN) in vitro (Miura et al., 1992), Calnuc a alors été identifiée alors comme une protéine soluble de 55 kDa promouvant la formation d’anticorps dirigés contre l’ADN (Kanai et al., 1990; Kanai et al., 1986). Les études subséquentes ont permis de démontrer que Calnuc est hautement répandue dans le règne animal étant présente de Ciona intestinalis à Homo sapiens (Kanuru et al., 2009). L’analyse de sa structure primaire révèle que les différents domaines sont bien conservés à l’intérieur d’espèces rapprochées d’un point de vue évolutif indiquant que la protéine doit maintenir ses fonctions. Aussi, sa séquence est davantage conservée chez les organismes supérieurs qui présentent deux isoformes. Ces derniers, codés par des loci géniques indépendants, proviendraient de la duplication d’un gène ancestral, modifié au travers de l’évolution. Par contre, malgré une structure protéique conservée, le gène codant Calnuc est considérablement variable, suivant la complexité de l’organisme, en ce qui concerne son organisation ainsi que sa quantité d’exons et d’introns (Aradhyam et al., 2010). Chez l’humain, le gène qui code NUCB1 est localisé au niveau du chromosome 19q13.2-q13.4 et est composé de 13 exons dont 12 codants (Miura et al., 1996). La région en amont contient des promoteurs communs aux gènes ubiquitaires nécessaires pour le maintien de la fonction cellulaire de base comme des sites de liaison pour le facteur de transcription Specificity protein 1 (sp1), une boîte CCAAT, une boîte TATA et plusieurs sites d’initiation (Miura et al., 1996). D’ailleurs, Calnuc est exprimée dans tous les tissus et cellules humains investigués jusqu’à présent (Lin et al., 1998; Miura et al., 1996). Le caractère ubiquitaire de Calnuc laisse croire qu’elle aurait un rôle important dans la maintenance de la machinerie cellulaire. 1.2 Structure et localisation cellulaire La forme humaine de Calnuc compte 461 acides aminés arrangés en structure principalement hélicoïdale (Kanuru et al., 2009; Miura et al., 1994). La partie N-terminale est compacte et globulaire alors que la partie C-terminale est allongée (Kapoor et al., 2010). 2 Calnuc contient, dans sa séquence, plusieurs motifs et domaines (figure 1) (Ren et al., 2009), dont certains qui influencent sa localisation cellulaire faisant d’elle une protéine modulaire multicompartimentale. Figure 1 - Structure de Calnuc. Représentation des différents éléments structuraux de Calnuc à l'échelle. SP (peptide signal), BR (région basique), AR (région acide), LZ (fermeture en leucine), ZBD (domaine de liaison au zinc), NLS (signal de localisation nucléaire), EF (main EF). Figure crée avec DOG 2.0, DOG 1.0 : Illustrator of Protein Domain Structures. Ren et al. Cell Research (2009) 19:271-273. Entre autres, elle possède un signal bipartite de localisation nucléaire (NLS) (figure 1) (Dingwall and Laskey, 1991) suggérant sa présence au noyau (figure 3A) (Miura et al., 1994; Wang et al., 1994). En effet, ce motif basique est reconnu par des Importines et permet le transport à travers les pores du noyau. Il se trouve d’ailleurs dans 50% des protéines nucléaires contre seulement 5% des protéines non nucléaires (Petersson et al., 2004). Quoiqu’allant de pair avec la capacité de Calnuc à lier l’ADN, cette localisation est encore aujourd’hui controversée. Aussi, la séquence de Calnuc débute par un peptide signal (SP) de 26 acides aminés (figure 1). Cette structure tripartite au cœur hydrophobe se lie à la Particule de reconnaissance du signal (SRP) ce qui permet l’adressage au réticulum endoplasmique lors de la traduction (figure 3A). Calnuc est donc synthétisée au niveau de la lumière du réticulum endoplasmique, à travers le Translocon, et subit le clivage de son SP par une peptidase signal. Rapidement, la protéine est transportée au Golgi (figure 3A), probablement via un récepteur encore inconnu et par un mécanisme dépendant de sa proline 3 en position 28 (figure 1) (Tsukumo et al., 2009). À cet endroit, Calnuc va subir des modifications post-traductionnelles, dont des O-glycosylations, des sialylations et des sulfatations, la faisant passer de 60 à 63 kDa (Lavoie et al., 2002; Tsukumo et al., 2007). La particularité principale de Calnuc, est de posséder deux mains EF (figure 1) qui lui permettent de lier le calcium (Ca2+). Ces motifs sont constitués d’une hélice alpha (notée ‘E’), d’une boucle et d’une seconde hélice α (notée ‘F’). L’ion est maintenu par la boucle qui forme une pochette de liaison composée d’atomes d’oxygène de la chaîne latérale (positions 1, 3, 5, 12), la chaîne principale (position 7) et une molécule d’eau (position 9). Un résidu hydrophobe (position 8) et une glycine (position 6) sont également nécessaires (Kretsinger, 1987). La première main EF (EF1) de Calnuc présente ce profil idéal et lui confère une forte affinité pour le calcium avec une constante de dissociation (Kd1) de 6,3 µM. Quant à elle, la seconde main (EF2) possède une arginine en position 6 nuisant considérablement à la liaison du calcium (Kd2 = 73,5 µM) (Lin et al., 1999). Globalement, Calnuc lie donc le calcium avec une faible capacité (~1,1 µmol Ca2+/µmol de protéines) et une haute affinité (Kd = 6,6 μM) dans un compartiment comme le Golgi qui est riche en cet ion (0,3 mM). La liaison du calcium entraîne une augmentation du contenu hélicoïdal de Calnuc de 3,3% (Miura et al., 1994) (pouvant potentiellement représenter une hélice allant jusqu'à 15 acides aminés) de même que l’exposition de sites hydrophobes ayant le potentiel d'influencer des interactions avec d’autres protéines. Il est à noter que le magnésium (Mg2+) peut également se lier faiblement au niveau des mains EF (Kd en millimolaire) (Kanuru et al., 2009). En revanche, les changements structuraux engendrés par ce dernier sont deux fois moindres que ceux observés avec le calcium et les deux ions semblent avoir un effet synergique (Kanuru et al., 2009). Il a été proposé que le magnésium aurait pour rôle de maintenir une conformation intermédiaire facilitant la liaison du calcium (Gifford et al., 2007). En 2012, une étude a révélé la présence de deux sites putatifs conservés pour la liaison de zinc (Zn2+) en N-terminal de Calnuc (figure 1) (Kanuru et al., 2013). Ces motifs HXXEXnH (où n correspond à 108-135 résidus) sont similaires à ceux retrouvés chez les protéases de type carboxypeptidase A. Par contre, Calnuc est dénuée d’une telle activité catalytique qui dépend de plus amples résidus. La liaison du zinc est de haute affinité (Kd = 21 nM), permettant une interaction dans les compartiments riches comme certains organites 4 et le milieu extracellulaire (concentration de l'ordre du micromolaire), mais insuffisante pour la faible concentration du cytosol (concentration de l'ordre du picomolaire). Il en résulte un changement de conformation qui expose la surface hydrophobe de Calnuc, de façon indépendante du calcium (Kanuru et al., 2013). Les variations intracellulaires de zinc, mais aussi de calcium et de magnésium, pourraient donc s’avérer primordiales pour modifier la conformation structurale de Calnuc de manière à moduler ses fonctions. En C-terminal, Calnuc possède également une glissière en leucines (LZ) (figure 1) (Kanai and Tanuma, 1992; Miura et al., 1992) caractérisée par une hélice alpha qui présente des leucines à intervalle de sept résidus. Ces dernières sont orientées du même côté de manière à former, avec les résidus non polaires adjacents, une région hydrophobe longitudinale qui peut s’enrouler autour d’une autre hélice de même nature. Cette structure tridimensionnelle permet normalement de lier l’ADN et, par conséquent, est principalement retrouvée au niveau des facteurs de transcription comme Fos et Jun. Quoique Calnuc possède également un signal de localisation nucléaire et une région basique de liaison potentielle à l’ADN (figure 1), aucune évidence n’a démontré son rôle comme régulateur de l’expression génique. La glissière en leucines de Calnuc lui permet de dimériser (Kanuru et al., 2009; Kapoor et al., 2010) tout en constituant un site potentiel pour lier diverses protéines. D'ailleurs, elle a été proposée comme élément permettant à Calnuc de se fixer à la membrane du Golgi via des interactions protéiques. D’autres mécanismes, on été proposés comme étant responsables de la rétention de Calnuc au Golgi. Entre autres, Calnuc possède une région de 15 résidus hydrophobes en Cterminal (figure 1) (Lin et al., 1998) qui pourrait, comme c’est le cas pour Synaptobrevin, suffire à son ancrage (Kutay et al., 1993; Whitley et al., 1996). Le point isoélectrique de 4,9 de la protéine suggère qu’elle est fortement liée à la membrane de l’organite empêchant son extraction par des traitements alcalins. Finalement, en 2001, un groupe a démontré qu’une région riche en leucines et en isoleucines située en N-terminal (acides aminés (aa) 67-120) (figure 1) est essentielle pour le maintien de la protéine au Golgi (Nesselhut et al., 2001). Calnuc est donc retenue au Golgi faisant d’elle une protéine résidente (Lin et al., 1998). C’est d’ailleurs à cet endroit qu’on la retrouve principalement et, de manière plus marquée, 5 au niveau du réseau cis-golgien et du cis-Golgi (Lin et al., 1998). Elle représente environ 0,4% (3,8 g/mg) des protéines golgiennes chez les cellules de foie de rat (Lin et al., 1999) faisant d'elle la protéine soluble la plus abondante du compartiment (Gilchrist et al., 2006). Calnuc va demeurer dans l'organite jusqu’à 24h avant d’être sécrétée par les voies constitutive et constitutive-like (Lavoie et al., 2002) pour se retrouver dans le milieu extracellulaire (figure 3A). Outre sa localisation au niveau de la voie sécrétoire et au niveau extracellulaire, Calnuc se retrouve aussi dans le cytosol, plus particulièrement à la surface du Golgi, de la membrane plasmique, des endosomes et des granules de sécrétion (figure 3A) (Brodeur et al., 2009; Lin et al., 2009). Cette localisation au cytoplasme est quelque peu inhabituelle pour une protéine possédant un peptide signal, mais peut s’expliquer par deux hypothèses. D’abord, Calnuc pourrait subir une rétrotranslocation vers le cytoplasme via le Translocon, comme c’est le cas pour Calréticuline (Afshar et al., 2005). D’autres protéines liant le Ca2+, dont la Phospholipase A2, peuvent transloquer au Golgi en réponse à des variations du Ca2+ cytosolique (Evans et al., 2001). Aussi, il est possible que la reconnaissance du peptide signal de Calnuc fasse parfois défaut ce qui préviendrait l’insertion de Calnuc dans la lumière du RE et mènerait à sa synthèse au niveau du cytosol. C’est d’ailleurs le cas pour l’Inhibiteur 2 de l’activateur du Plasminogène (PAI-2) qui subit une translocation variable (Belin et al., 1996). La différence de mobilité de Calnuc lors de fractionnements suggère l’ajout de modifications post-traductionnelles distinctes en fonction de la localisation ce qui suggère une régulation du peptide signal (Lin et al., 2000). 1.3 Homologues L'homologie entre protéines est d'une grande utilité afin d'identifier des domaines et motifs d'intérêt qui apportent des indices relativement aux fonctions et interactions protéiques. Calnuc possède peu de similarité avec d'autres protéines si ce n'est avec son paralogue NUCB2 ainsi qu'avec deux autres protéines liant le calcium, soit Cab45 et Calréticuline. 6 1.3.1 Cab45 Tout comme Calnuc, 45 kDa calcium-binding protein (Cab45), aussi connu sous le nom de Stromal cell derived factor (SDF4), est une protéine soluble résidente du Golgi qui lie le calcium (Scherer et al., 1996). L'ion est nécessaire à sa rétention au Golgi et ses 362 acides aminés contiennent six mains EF (Koivu et al., 1997). Cab45 lie des cargos de façon dépendante du calcium et permet leur triage au TGN vers la voie de sécrétion (von Blume et al., 2012). Elle pourrait aussi être nécessaire pour l'homéostasie calcique, d'autant plus qu'elle présente une forte homologie avec les protéines Réticulocalbine et Endoplasmic reticulum calcium-binding protein of 55 kDa (ERC55) qui lient le calcium au RE (Scherer et al., 1996). Aussi, Cab45 possède environ 30% de similarité, mais aucune homologie marquée, avec Calnuc à l'exception d'une courte séquence AANXE (E/D) (Lin et al., 1998). 1.3.2 Calréticuline Caréticuline (CRT) est une protéine chaperonne soluble qui interagit avec les protéines mal repliées de manière à prévenir leur export du RE. Malgré l'absence de mains EF dans sa séquence, elle lie également le calcium avec une faible affinité et une haute capacité pour en faire une réserve mobilisable au niveau du même organite. Calnuc a une homologie de séquence considérable de 30% avec CRT de rat (Lin et al., 1998). La similarité la plus importante (43%) se trouve dans la région C-terminale qui présente beaucoup de résidus acides et deux motifs conservés AY(I/A)EE et QRLX(Q/E)E(I/E)E de fonction inconnue (Lin et al., 1998). Caractéristique de la famille des lectines, le domaine P (riche en prolines) de CRT contient trois motifs hélice-boucle-hélice (Michalak et al., 1996) et est semblable (36%) aux mains EF de Calnuc (Lin et al., 1998) alors que la région Nterminale présente moins de similarités (26%) (Lin et al., 1998). Par contre, plusieurs motifs d'importance de CRT sont absents chez Calnuc comme les trois répétitions NPD/E, le motif KPEDWD typique de la famille Calnexine/Calréticuline (Nash et al., 1994) et le motif KDEL de rétention au RE expliquant leur divergence de localisation (Lin et al., 1998). Il est à noter que CRT peut également être sécrétée par les odontoblastes (Somogyi 7 et al., 2003) et inhiberait la minéralisation (St-Arnaud et al., 1995). Comme nous le vérrons ultérieurement, Calnuc aurrait également une fonction au niveau des os. 1.3.3 NUCB2 En 1994, le groupe de Hilschmann a identifié une protéine de 55 kDa hautement homologue à Calnuc et possédant 61,56% d'identité (Barnikol-Watanabe et al., 1994; Karabinos et al., 1996). Nommée alors DNA binding, EF-hand, acidic region (NEFA), cette dernière fut rebaptisée NUCB2 étant probablement un isoforme de Calnuc (NUCB1). Une analyse rapide par le programme 'Lalign' confirme 61,4% d'identité et 81,9% de similarité sur 415 acides aminés des séquences humaines (figure 2A). Les deux protéines coexistent uniquement chez les organismes supérieurs et sont issues de loci géniques différents (19q13.2-q13.4 et 11p15.1-p14) suggérant qu'elles ont évolué à partir d'un gène ancestral commun modifié au travers l’évolution. En effet, les protéines à mains EF sont formées à partir d'un précurseur eucaryotique à un domaine nommé CTER (Nakayama et al., 1992). Ce dernier s'est ensuite transformé en un dérivé à quatre domaines duquel, selon l'analyse phylogénétique, originent Calnuc et NUCB2 (Karabinos et al., 1996). Il serait responsable de la présence de domaines fonctionnels communs dont le SP, le NLS, le LZ, la région acide et évidemment les mains EF (figure 2A). Chez NUCB2, ces dernières lient le calcium avec des kd de 0,08 μM et 0,2 μM permettant une capacité deux fois supérieure à Calnuc (2 μmol Ca2+/μmol protéine) et une augmentation quatre fois supérieure (13%) du contenu hélicoïdal (Kroll et al., 1999). Les deux protéines sont divergentes au niveau des régions N-terminale et C-terminal ainsi qu'au niveau du LZ, mais la principale différence réside dans le fait que, comparée à NUCB2, Calnuc a une extension d'environ 40 résidus (figure 2A). Les variations entre Calnuc et NUCB2 ne sont pas sans rappeler que les deux protéines ont acquis, au travers leurs 830 millions d'années d'évolution indépendante (Gu, 1998), des différences induisant l'apparition de fonctions distinctes. D'ailleurs, jusqu'à présent les études portant sur NUCB2 mettent en lumière des rôles très divergents de ceux attribués à Calnuc. Entre autres, il a été montré que l'association, dépendante du calcium, de 8 NUCB2 avec le Aminopeptidase regulator of TNRF1 shedding (ARTS-1), au niveau des vésicules exosome-like, est nécessaire pour promouvoir la relâche extracellulaire du Tumor necrosis factor (TNF) receptor 1 (TNFR1) (Islam et al., 2006). Ce dernier peut alors moduler l'effet de la cytokine TNF dans une foule de processus tels que l'inflammation, les fonctions immunitaires et l'apoptose, ainsi que dans les pathologies associées comme l'arthrite, le cancer et la sclérose en plaque. Pratiquement au même moment, NUCB2 a été identifiée comme protéine régulée par le récepteur nucléaire Peroxisome proliferator-activated receptor gamma (PPAR-g) qui contrôle l'homéostasie du glucose et des lipides de même que la prise de nourriture (Oh et al., 2006). NUCB2 agit d'ailleurs comme suppresseur de l'appétit. La présence de plusieurs fragments de NUCB2 ainsi que de sites de clivages conservés pour les prohormones convertases (PC) a indiqué aux auteurs de l'étude que NUCB2 pourrait être à l'origine de plusieurs peptides bioactifs (Oh et al., 2006). Basé sur le motif consensus R/K–Xn–R/K↓ (où X est un acide aminé, n est le nombre de résidus espaceurs (soit 0, 2, 4 ou 6) et ↓ est le site de clivage) (Seidah and Chretien, 1999), NUCB2 serait divisé en trois fragments. Suivant la nouvelle fonction de NUCB2, ils ont été nommés NEFA/NUCB2-encoded satiety- and fat-influencing protein (Nesfatin)-1 (résidus 1-82), Nesfatin-2 (résidus 85-163) et Nesfatin-3 (résidus 166-396) (figure 2B) (Oh et al., 2006). Nesfatin-1 a été montré comme étant responsable de l'effet anorexigénique et son clivage, vraisemblablement par PC1/3 et PC2, est nécessaire à cette fonction (Oh et al., 2006). En effet, l'injection centrale (intracérébroventriculaire) ou périphérique de Nesfatin-1 réduit de façon significative la prise de nourriture chez les rongeurs (Goebel et al., 2011; Oh et al., 2006; Stengel et al., 2010) alors que le jeûne diminue l'expression de l'acide ribonucléique messager (ARNm) de NUCB2 dans les glandes endocrines gastriques (Stengel et al., 2009). Le mécanisme qui sous-tend l'effet anorexigénique demeure nébuleux et serait indépendant de la voie de la Leptine (Oh et al., 2006; Shimizu et al., 2009), dépendant de celle de la Mélanocortine (Oh et al., 2006) et potentiellement médié par celles de l'Oxytocine (Maejima et al., 2009) et de la Corticolibérine (CRH). Actuellement, Nesfatin-1 est l'objet d'un engouement pour son potentiel dans le traitement de l'obésité. Son fragment central de 30 résidus (M30) (figure 9 2B), responsable de l'effet anorectique (Shimizu et al., 2009), constitue un modèle pour la conception rationnelle d'analogues thérapeutiques. Nesfatin-1 aurait également un effet anti-hyperglycémique insulino-dépendant chez les souris obèses hyperglycémiques (Su et al., 2010) indiquant un potentiel rôle dans la pathogenèse du diabète. L'expression de Nesfatin-1 dans les tissus adipeux est augmentée par une diète induisant l'obésité alors que la privation de nourriture a l'effet contraire (Ramanjaneya et al., 2010). On observe aussi une diminution des niveaux plasmatique Nesfatin-1 chez les patients à jeun atteints de diabète de type 2 (Li et al., 2010). De plus amples études seront nécessaires pour comprendre la fonction de Nesfatin-1 dans cette maladie. Les voies responsables du contrôle de la masse corporelle sont généralement impliquées dans la puberté et la fertilité. Ces phénomènes ont rapidement été associés à Nesfatin-1 qui est exprimée dans les gonades (Garcia-Galiano et al., 2012). Dans ce sens, l'expression hypothalamique de NUCB2/nesfatin-1 augmente pendant la puberté et sa diminution retarde les signes de puberté et diminue les niveaux circulant de LH chez les rats femelles (Garcia-Galiano et al., 2010; Garcia-Galiano and Tena-Sempere, 2013). Nesfatin-1 fait également partie des neuropeptides qui participent à la création de la réponse au stress (Emmerzaal and Kozicz, ; Merali et al., 2008). L'injection centrale de Nesfatin-1 entraîne une élévation des niveaux circulant de composantes majeures de l'axe hypothalamo-hypophyso-surrénalien, soit l'axe du stress, comme l'hormone Adrénocorticotropine (ACTH) et la Corticostérone (Konczol et al., 2010; Yoshida et al., 2010). On observe également une modulation de la réponse cardiovasculaire, dont de l'hypertension, via la voie Oxytocine-Mélanocortine (Yosten and Samson, 2009; Yosten and Samson, 2010; Yosten and Samson, 2013; Yosten and Samson, 2014). Tout comme Calnuc, NUCB2 est localisée au niveau extracellulaire et intracellulaire, dans le cytosol et au Golgi (Lavoie et al., 2002; Morel-Huaux et al., 2002). Il a été proposé que, tout comme Calnuc (voir section 'Calcium et minéralisation'), NUCB2 formerait une réserve de calcium mobilisable par agoniste (Kalnina et al., 2009). De son côté, Nesfatin-1 est dépourvu de mains EF, mais serait tout de même impliqué dans l'homéostasie calcique en induisant une augmentation intracellulaire de l'ion via les canaux de type L, N ou P/Q 10 Figure 2 - Comparaison entre Calnuc (NUCB1) et NUCB2. (A) Alignement des séquences humaines de NUCB1 et NUCB2 marquant le degré de conservation des acides aminés selon l'échelle allant du rouge au bleu, où rouge (10) 11 représente une forte conservation et bleu (0) une faible conservation. Figure crée avec PRALINE multiple sequence alignment (http://www.ibi.vu.nl/programs/pralinewww). (B) Représentations à l'échelle de NUCB1, tel que présenté en figure 1, et de NUCB2 avec ses différents fragments (incluant le SP). Les domaines de Nesfatin-1 sont également représentés. Figure crée avec DOG 2.0, DOG 1.0 : Illustrator of Protein Domain Structures. Ren et al. Cell Research (2009) 19:271-273. (Brailoiu et al., 2007). Cet effet est bloqué par la toxine pertussique indiquant qu'un récepteur couplé aux protéines Gi/o serait également impliqué. GPCR12, GPCR3 et GPCR6 sont les principaux candidats (Osei-Hyiaman, 2011), mais jusqu'à présent Nesfatin1 demeure considérée comme un peptide bioactif orphelin. Récemment, le groupe d'Unniappan, a démontré que Calnuc contient potentiellement un peptide de 77 acides aminés hautement similaire à Nesfatin-1 et surtout, à la région biologiquement active M30 (identité 76,6%) (Ramesh et al., 2015). Ce dernier, nommé Nesfatin-1-like peptide (NLP), est localisé après le SP et est flanqué de sites potentiels de clivage par des convertases de prohormones (sites Lys-Arg) (Ramesh et al., 2015). Des expériences de transcription inverse combinées à des réactions en chaîne par polymérase (RT-PCR) et des expériences d'immunoréactivités ont révélé que le pancréas, et plus particulièrement les cellules bêta productrices d'insuline dans les îlots de Langerhans, sont des sources de Calnuc et possiblement de NLP (Ramesh et al., 2015; Williams et al., 2014). De plus, l'incubation de ces cellules avec du NLP synthétique (10 et 100 nM) démontre une activité insulinotropique à forte concentration de glucose (Ramesh et al., 2015). Cette étude laisse croire que les deux Nucléobindin sont des précurseurs de peptides biologiquement actifs. 1.4 Partenaires d’interaction, fonctions et maladies associées Calnuc est une protéine multicompartimentale ubiquitaire hautement conservée indiquant qu’elle doit être d’une grande importance pour le maintien des fonctions cellulaires. De plus, elle présente une structure complexe composée de plusieurs domaines modulaires qui constituent des sites d’interactions potentiels pour d’autres protéines. Au travers les années, l’étude de ses différentes composantes et la découverte de partenaires 12 Figure 3 - Distribution de Calnuc et de ses divers partenaires d'interaction. (A) Calnuc, représentée par les points noirs, se localise dans les différents compartiments de la voie de biosynthèse de même qu'au niveau du cytosol et du milieu extracellulaire. À chaque endroit, (B) Calnuc interagit avec différents partenaires. Modifié de Calnuc : Emerging Roles in Calcium Signaling and Human Diseases. Aradhyam et al. IUBMB Life, 62(6): 436–446, June 2010 avec la permission de John Wiley and Sons. (C) Représentation à l'échelle de Calnuc (NUCB1), tel que présenté en figure 1, et des sites d'interactions de 13 ses partenaires. Casp (Caspase), DGLDP (site de clivage par les caspases), D328-H339S378 (triade catalytique attribuée à l'activité de protéase à sérine), GBA (G alpha binding and activating), GEF (facteur d’échange des nucléotides de guanine). Figure crée avec DOG 2.0, DOG 1.0 : Illustrator of Protein Domain Structures. Ren et al. Cell Research (2009) 19:271-273. d’interactions variés ont permis d’associer Calnuc à plusieurs fonctions. Bien que ces rôles restent à être approfondis, ils laissent supposer que la préservation de l’intégrité structurale et fonctionnelle de la protéine est nécessaire pour nous prévaloir de multiples pathologies. 1.4.1 Calcium et minéralisation Le calcium est un élément chimique omniprésent essentiel au bon fonctionnement d'une variété de fonctions biologiques. Il contribue entre autres à la formation osseuse et au maintien du potentiel d'action membranaire tout en agissant comme cofacteur enzymatique et comme second messager dans la transduction de signaux. Au niveau cellulaire, l'efficacité du calcium repose sur sa capacité de lier de nombreuses protéines dont la principale famille est celle des protéines à mains EF. Ces dernières assurent sa compartimentation et sa spécificité d'action agissant d'une part comme tamponneur dans l'homéostasie calcique ou, d'autre part, comme senseur dans la traduction de signaux (da Silva and Reinach, 1991). Les tamponneurs, comme Calbindin D9k et Parvalbumin, chélatent les ions grâce à une haute affinité (Kd <10-7 M) et subissent peu de changements conformationnels (Ikura, 1996). De leur côté, malgré une affinité relativement faible pour le calcium (Kd >10-5 M), les senseurs, comme Calmodulin et Troponin C, subissent des changements conformationnels marqués créant des domaines d’interaction pour une cible à moduler (Ikura, 1996). À l'image de Calbindin D28k (Berggard et al., 2002), Calnuc a été proposée comme étant à la fois un senseur et un tamponneur de calcium (de Alba and Tjandra, 2004; Kanuru et al., 2009) allant de pair avec sa localisation au Golgi et au cytoplasme (figure 3A et 3C). Calnuc comme senseur La fraction cytosolique de Calnuc joue probablement un rôle de senseur permettant la régulation de nombreuses protéines en réponse à une relâche de calcique. Son affinité pour 14 le Ca2+, ainsi que sa sélectivité pour l'ion au détriment du Mg2+, jouent en cette faveur. Lors d'une activation cellulaire, telle une contraction musculaire menant à l'ouverture des canaux Ryanodine, la concentration de calcium dans le cytoplasme passe d'environ 0,1 uM à 1-10 uM (Berggard et al., 2002) entraînant la liaison de l'ion par Calnuc. Cette liaison provoque des changements conformationnels importants dont le repliement des mains EF ainsi que l'exposition des surfaces hydrophobes pouvant servir de sites d'interaction (de Alba and Tjandra, 2004). De multiples interactions décrites plus bas dans cette section sont d'ailleurs influencées par la liaison du calcium laissant croire que Calnuc pourrait agir comme effecteur ou modulateur dans la signalisation. Calnuc comme tamponneur Le calcium est nécessaire pour la cellule mais devient toxique à haute concentration. Il doit donc être emmagasiné dans des compartiments comme le réticulum endoplasmique et le Golgi (Grohovaz et al., 1996) de manière à être rapidement mobilisé au cytosol suivant un stimulus. Ces réserves sont maintenues par des protéines liant le calcium tel que Calréticuline au RE. À l'intérieur du Golgi, Calnuc est la principale protéine liant le calcium (3,8 mg Calnuc/mg protéines dans les cellules NRK, soit 0,4% des protéines) (Lin et al., 1998). En effet, la surexpression de Calnuc au Golgi mène à une augmentation de deux à trois fois de la quantité de calcium emmagasiné et pouvant être mobilisée par l'action d'agonistes ciblant le récepteur IP3 de type 1, comme l'ATP et l'IP3 (Lin et al., 1999). Calnuc agirait donc comme tamponneur à cet endroit grâce à sa main EF1 ainsi qu'à l'action de la pompe Sarco/endoplasmic reticulum Ca2+-ATPase (SERCA) qui assure le mouvement des ions vers la lumière du Golgi (Taylor et al., 1997). Calnuc dans la minéralisation Le tissu osseux forme une réserve calcique d'un tout autre type. En 1995, Calnuc a été identifiée comme étant une constituante mineure de la matrice extracellulaire des os (Wendel et al., 1995), puis de la dentine (Somogyi et al., 2004). En effet, Calnuc est produite par de nombreuses cellules impliquées dans la formation osseuse, soit les ostéoblastes, les ostéocytes (Wendel et al., 1995) et les odontoblastes (Somogyi et al., 2004), mais est indétectable par hybridation in situ dans les cellules connexes comme les 15 chondrocytes (cellules du cartilage), les ostéoclastes et cellules de la moelle (Petersson et al., 2004). Calnuc est sécrétée dans l'ostéoïde, c'est-à-dire la matrice osseuse non minéralisée essentiellement constituée de collagène de type 1. Cette dernière doit donc subir une étape de minéralisation pendant laquelle s'ajoute une matrice minérale composée de phosphate tricalcique (hydroxyapatite), de carbonates de calcium et de phosphates de magnésium. Suivant sa capacité à être sécrétée et à lier l'hydroxyapatite (Wendel et al., 1995), l'implication de Calnuc dans ce procédé a été proposée. Des analyses ultrastructurelles ont mis en évidence une forte présence de Calnuc au niveau des ostéoïdes d'os nouvellement créés. De plus, durant la prolifération et la différenciation cellulaire, Calnuc a été localisée au niveau du noyau alors que pendant la minéralisation Calnuc se retrouve principalement dans le cytoplasme (Petersson et al., 2004). Ce déplacement intracellulaire est accompagné d'une faible expression durant la phase de prolifération qui augmente considérablement durant la différentiation et la maturation de la matrice pour ensuite diminuer après l'initialisation de la minéralisation (Petersson et al., 2004). Ces données suggèrent que Calnuc contribuerait à l'accumulation et au transport des ions calciques et pourrait agir comme modulateur de la maturation avant le dépôt de minéraux. Il est à noter que Calnuc pourrait aussi agir indirectement dans la destruction de l'os via son interaction avec COX-2 qui a pour effet de stimuler la synthèse de PGE2 et la formation d'ostéoclastes. 1.4.2 Protéines G et sécrétion régulée Les Guanosine nucleotide-binding proteins (protéines G), sont d’importants messagers qui transmettent des signaux provenant de récepteurs couplés aux protéines G (GPCR) de la membrane plasmique vers l’intérieur de la cellule. Pour ce faire, le récepteur activé par un ligand catalyse l’échange de guanosine diphosphate (GDP) pour du guanosine triphosphate (GTP) sur la sous-unité Gce qui l'active à son tour. À ce moment, les sousunités Get G associées au récepteur se dissocient pour aller moduler indépendemment une variété d’effecteurs différents. Éventuellement, l'activité GTPase intrinsèque de Gprovoque son inactivation, via l’hydrolyse du GTP, et conséquemment sa réassociation 16 avec le récepteur. Récemment, il a été mis en évidence que les protéines G, ainsi que le dimère G, sont présentes à la membrane de compartiments intracellulaires, comme le Golgi et les endosomes (Wilson et al., 1994), mais leur rôle à ce niveau demeure encore inconnu. En 1995, une expérience de double hybride a permis au groupe d’Insel d’identifier Calnuc comme partenaire d’interaction direct des protéines Gde type Gi2 (Mochizuki et al., 1995). Ultérieurement, Calnuc a été montrée comme étant capable d’interagir, de façon dépendante du Ca2+ et Mg2+ (Lin et al., 2000), avec les types Gi1 et Gi3, ainsi qu’avec les Gi2, Go1, Gz et Gs, mais pas avec les Gq, G12 et G13 (figure 3B) (Lin et al., 1998). L’étude plus approfondie de la liaison avec Gi3 a permis d’établir que les mains EF de Calnuc, ainsi que la région acide très conservée et flanquée entre ces structures, sont essentielles pour l’interaction (figure 3C) (Lin et al., 1998). Les premières expériences visant à établir la fonction de ces interactions ont montré que la surexpression de Calnuc augmente les niveaux de Gi2 (Mochizuki et al., 1995). À ce moment, l’interaction a été montrée comme étant indépendante de l’état d'activation de la protéine G (Lin et al., 1998). Toutefois, une protéine G active augmente la liaison de Ca2+ sur Calnuc alors qu’une protéine G inactive a pour effet d’entraîner une relâche de l’ion par Calnuc (Kanuru et al., 2009). Ce phénomène laisse croire que Calnuc pourrait être agir comme effecteur dans la signalisation des protéines G ou dans la modulation de leurs fonctions. Outre leur localisation à la membrane plasmique, les protéines Gi3 sont retrouvées à la surface du Golgi (Stow et al., 1991) et les Gi1/2 à la surface des vésicules de sécrétion (Wilson et al., 1994). Calnuc a également été observée à ces différents endroits (Lavoie et al., 2002) et est partiellement distribuée avec Gi1/2 et Gi3 à la surface des granules de sécrétion (Lin et al., 2009). Il s’agit également de la seule protéine ayant la capacité de lier à la fois les G et le calcium, deux éléments qui stimulent la sécrétion régulée. Son implication dans ce processus physiologique a donc rapidement été confirmée, dans les cellules neuroendocrines, par une augmentation de la sécrétion régulée d’ACTH dépendante des G (Lin et al., 2009). La surexpression de Calnuc entraîne aussi une 17 redistribution de Gi3 sur les granules de sécrétion et de Gi1/2 à la membrane plasmique sans affecter leur activité. Ces résultats indiquent que Calnuc influencerait la sécrétion régulée en agissant comme modulateur, possiblement sensible au calcium, dans la dynamique de distribution intracellulaire des protéines G. La signalisation des protéines G est dépendante de leur cycle d’activation-inactivation lui-même finement régulée par de nombreuses protéines accessoires. Les facteurs d’échange des nucléotides de guanine (GEF) catalysent l’activation des protéines G via l’échange du GDP pour du GTP. De leur côté, les protéines d’accélération de la GTPase (GAP), dépourvues d’activité catalytique, stimulent l’activité GTPase intrinsèque des protéines G pour mener à leur inactivation. Finalement, les inhibiteurs de la séparation des nucléotides de guanine (GDI) maintiennent les protéines G dans un état d’inactivation. Dernièrement, deux études ont suggéré l’implication de Calnuc dans cette mécanistique en lui attribuant les rôles, quelque peu opposés, de GDI et de GEF. En 2010, le groupe de Sakmar a montré que la forme cytosolique de Calnuc est un dimère physiologique qui peut interagir avec Gαi1-GDP, suivant une affinité de 18,3 μM, mais pas avec Gαi1-GTP (Kapoor et al., 2010). Calnuc a pour effet d’inhiber la relâche du GDP de Gαi1, prévenant ainsi l’activation de la protéine G et l’inhibition de la voie de signalisation subjacente (Kapoor et al., 2010). Ces résultats suggèrent que Calnuc agirait comme une GDI malgré l’absence dans sa structure d’un motif GoLoco normalement associé à cette fonction. La région N-terminale de Calnuc (26-333) serait plutôt en cause (Kapoor et al., 2010). Pratiquement au même moment, le groupe de Farquhar identifiait un motif G alpha binding and activating (GBA) (figure 3C) attribué à l’activité GEF dans plusieurs protéines, dont la G alpha-interacting vesicle-associated protein (GIV) (Garcia-Marcos et al., 2011). Ce motif Ψ-S/T-Φ/Ψ-X-D/E-F-Ψ (où Ψ est un résidu aliphatique et Φ est un résidu aromatique) se trouve conservé dans Calnuc et NUCB2 et leur permet de lier préférentiellement la forme inactive des Gi au niveau de la crevasse entre l’hélice alpha3 et la région Switch II (Garcia-Marcos et al., 2011). Ceci contredit une étude du même groupe qui impliquait l’hélice alpha5 de Gi3 (Lin et al., 2000), une région d’importance impliquée dans la liaison de GDP (Mixon et al., 1995; Sprang, 1997), de GPCR (Lambright 18 et al., 1996) et d’effecteurs (Sunahara et al., 1997). Néanmoins, Gi3 lie davantage de GTP en présence de Calnuc ou NUCB2 indiquant que ces dernières agiraient comme GEF (Garcia-Marcos et al., 2011). La divergence des fonctions apportées par ces articles peut être imputable à l’utilisation de concentrations protéiques différentes (sous les Kd ou affectant les réactions enzymatiques) ou à des lectures artéfactuelles. Par contre, les deux groupes s’entendent pour dire que Calnuc lie uniquement les formes inactives des protéines Get que le calcium abolit l’interaction. En effet, l’ion entraîne le repliement des mains EF réquisitionnant et dissimulant des résidus nécessaires pour l’interaction avec G (GarciaMarcos et al., 2011). D’ailleurs, l’interaction entre Calnuc et Gi3 se produit dans un milieu pauvre en calcium, soit dans le cytosol à la surface du Golgi (Weiss et al., 2001). Calnuc est alors libre d’ions et lie Gseulement s’il est ancré aux membranes par ses ancrages lipidiques (Weiss et al., 2001). De plus amples études seront nécessaires afin d’éclaircir le rôle de Calnuc dans le cycle des protéines G, les régions impliquées ainsi que l’effet de l’élévation physiologique du calcium. 1.4.3 ADN et lupus Calnuc possède une région riche en acides aminés basiques qui lui permet de lier l’ADN (figure 3B et 3C) (Miura et al., 1992) et une fermeture en leucines connue chez plusieurs protéines pour faciliter cette liaison. Une hypothèse stipule que l'interaction entre Calnuc et l'ADN serait importante pour l'induction d'anticorps contre l'ADN, un phénomène en cause dans la physiopathologie du lupus érythémateux disséminé. Il s’agit d’une maladie chronique qui attaque principalement les tissus conjonctifs. Le système immunitaire produit alors des anticorps spécifiques contre des entités normalement présentes dans l'organisme dans le but de les éliminer. À l’origine, Calnuc a été identifiée comme étant un facteur promouvant la formation de tels auto-anticorps ciblant l’ADN double brin, la Ribonucléoprotéine U1, la protéine Lupus La et le facteur rhumatoïde (Kanai et al., 1995b; Kanai et al., 1993). D’ailleurs, l’injection intrapéritonéale de Calnuc cause des symptômes associés au lupus dont une surproduction d’immunoglobuline G (IgG) suivie par l'élévation d'auto-anticorps de même type ce qui accélère la réponse auto19 immune (Kanai et al., 1995b). Les souris injectées présentent aussi une fragmentation d’ADN liée à de l’apoptose dans des tissus lymphoïdes comme le thymus et entraînant une accumulation de segments d’ADN dans le sérum. Ce phénomène pourrait être à l’origine de la production d’anticorps dirigés contre l’ADN et de l’auto-immunité (Kanai et al., 1995a). Des études mécanistiques plus approfondies devront être faites pour comprendre les relations de causalité et élucider le rôle nébuleux de Calnuc dans cette maladie. 1.4.4 Cancer, apoptose et Caspases Le cancer est une maladie caractérisée par une prolifération cellulaire incontrôlée en dépit d'anomalies génétiques. Le système immunitaire peut détecter certains dysfonctionnements cellulaires associés aux tumeurs comme étant immunogéniques. Plusieurs protéines impliquées dans la tumorigenèse servent alors d’antigènes pour la production d’auto-anticorps ce qui n’est pas sans rappeler l’induction d’auto-anticorps par Calnuc dans un contexte de lupus. Dans cette logique, plusieurs groupes ont étudié l’expression antigénique de Calnuc. Dès 1994, une augmentation de Calnuc est décelée dans 10% des adénocarcinomes gastriques et 56% de ceux présentant des métastases aux ganglions lymphatiques (Wang et al., 1994). Également, le niveau d’expression de Calnuc dans les lymphomes non hodgkiniens reflète le degré de prolifération des cellules et donc le grade de la pathologie (Kubota et al., 1998). Aussi, Calnuc est hautement spécifique pour le cancer du côlon puisqu’environ 60% des tissus affectés surexpriment cette protéine en comparaison aux côlons sains qui ne semblent pas l'exprimer en immunohistochimie (Chen et al., 2007). Dans ce même article de 2007, des auto-anticorps contre Calnuc ont été retrouvés dans 11,5% des cancers du côlon contre seulement 1,2% des côlons sains. Calnuc semble donc bel et bien agir comme un antigène associé aux tumeurs (TAA) et contribue favorablement aux outils diagnostiques. De plus amples études seront nécessaires pour établir la fonction exacte de Calnuc et afin de déterminer si la réaction immunitaire est en cause dans le développement du caractère malin de la maladie. En effet, en raison de sa capacité à lier le calcium et les protéines G, Calnuc perturberait l’adhésion et la motilité cellulaire (Mermelstein et al., 2003) de même que la croissance, la prolifération et la migration. Pour les mêmes raisons, Calnuc pourrait être impliquée dans le mécanisme 20 d'action d'autres protéines associées à des cancers. Entre autres, par double hybride, Calnuc a été identifiée comme partenaire d'interaction de la protéine peu caractérisée nommée Breast cancer membrane protein (BCMP84) (Adam et al., 2003). Cette dernière est surexprimée et est déplacée à la membrane plasmique dans les carcinomes du sein et de la vessie. Le cancer est induit par le blocage de l'apoptose, le processus de mort physiologique par lequel une cellule déclenche son auto-destruction. Ce mécanisme mène à l’activation de multiples Caspases, des protéases à cystéine qui clivent nombre de substrats au niveau de sites consensus spécifiques pour entraîner le démantèlement cellulaire. L’expression de Calnuc dans certains cancers de même que l’induction de l’apoptose dans les tissus lymphoïdes lors de son administration exogène (Kanai et al., 1995b) suggèrent l’implication de la protéine à ce niveau. Dans le même ordre d'idées, in vitro, NUCB1 a été identifiée comme substrat potentiel pour les Caspases 3, 6 et 8 alors que son homologue NUCB2 serait clivé par Caspase 3, 6, 7, 8 et 9 (Valencia et al., 2008). Un site général de clivage DXXD est hautement conservé à travers les espèces. Chez la forme humaine de Calnuc, ce motif non optimal DGLDP (figure 3C) est localisé entre le NLS et la première main EF suggérant que les Caspases pourraient découpler les domaines fonctionnels en situation d’apoptose. En effet, le fragment C-terminal comprend les mains EF et la glissière en leucines qui sont essentielles pour la liaison du calcium et la liaison à des partenaires, dont l’ADN. Par contre, il est dépourvu de son domaine riche en leucines et isoleucines ainsi que de son NLS, ce qui signifie qu’il n’est pas retenu au Golgi et qu’il ne transloque pas au noyau. De son côté, le fragment N-terminal contenant les domaines de liaison au zinc est rapidement dégradé. 1.4.5 COX et inflammation Les Cyclooxygénases (COX) sont des enzymes qui permettent la synthèse de prostanoïdes, à partir d’acide arachidonique (AA) de manière à moduler divers effets physiologiques. Alors que la COX-1 est constitutive, la COX-2 est inductible par de multiples facteurs pro-inflammatoires dont le Granulocyte/monocyte colony stimulation 21 factor (GM) et le TNF. En 1996, une étude a démontré que Calnuc peut interagir avec les COX, de manière indépendante du calcium, via une région qui comprend ses 123 résidus en N-terminal (figure 3B et 3C) (Ballif et al., 1996). Malgré une affinité plus forte pour la COX-1 que la COX-2, les deux enzymes, fortement liées à la membrane luminale du RE, diminuent de plus de 80% la sécrétion de Calnuc en condition de surexpression. Elles participeraient ainsi à la rétention intracellulaire de Calnuc (Ballif et al., 1996). Pour faire suite, un groupe québécois s’est concentré sur la COX-2, qui prédomine dans la production de prostaglandines (PG) E2 dans les cellules inflammatoires telles que les neutrophiles. Ils ont démontré que Calnuc interagit et colocalise avec la COX-2 au niveau du RE et du Golgi dans des neutrophiles stimulés au GM/TNF (Leclerc et al., 2008). Calnuc se retrouve également dans les mêmes fractions que les enzymes impliquées dans la génération de prostanoïdes (Microsomal PGE2-synthase-1, Thromboxane-synthase, Calcium-dependent phospholipase A2) (Leclerc et al., 2008). À ce sujet, des essais sur neutrophiles stimulés ont démontré que Calnuc recombinante augmente de 5 fois la synthèse de PGE2 via la COX-2 et de manière dépendante de la dose (Leclerc et al., 2008). Calnuc présente donc des propriétés régulatrices des COX durant l’inflammation alors que ces dernières jouent un rôle dans la rétention intracellulaire de Calnuc. Dans le même ordre d'idées, en 2015, Calnuc a été identifiée comme protéine liant les lipides (Niphakis et al., 2015). Plus de 1000 protéines, dont plusieurs transporteurs, enzymes et protéines peu caractérisées, ont été enrichies avec une sonde semblable à l'acide arachidonique (arachidonoyl) (Niphakis et al., 2015). Il a été montré que le lipide se lie à Calnuc au niveau du domaine d'interaction des COX et que le calcium augmente significativement et sélectivement l'interaction (Niphakis et al., 2015). Calnuc perturbe le métabolisme oxydatif et hydrolytique des endocannabinoïdes et des eicosanoïdes. En effet, l'utilisation de deux ligands sélectifs de Calnuc (MJN228 et KML110) ou de petits acides ribonucléiques (ARN) en épingle à cheveux (shRNA) dirigés contre Calnuc a pour effet d'élever la concentration cellulaire d'acides gras (Niphakis et al., 2015). Ces résultats tendent à démonter que Calnuc faciliterait le métabolisme des acides gras en les transférant à des enzymes comme l'Amidohydrolase des acides gras (FAAH) et la COX2 (Niphakis et al., 2015). 22 1.4.6 Activité de protéase à sérine En 2012, un groupe indien a démontré, par essais enzymatiques, que Calnuc possède une activité de protéase à sérine (Kanuru et al., 2013). Cette activité est attribuable à une triade catalytique en C-terminal composée de l’aspartate328 du motif DTG (séquence consensus de protéases aspartiques (Rawlings and Barrett, 1995)), de l’histidine339 et de la sérine378 du motif GXSXG (séquence consensus de lipases (Wagner et al., 2002)) (figure 3C). En utilisant une variété de substrats différents, l’équipe a établi que Calnuc, à l’image de la trypsine, clive en aval d’une arginine. Aussi, l’interaction avec la sous-unité alpha des protéines G à proximité de la sérine catalytique a pour effet de réduire l’activité protéolytique. La liaison allostérique du zinc produit le même effet inhibiteur possiblement en nuisant à la géométrie du site actif. Cette étude suggère une grande importance des ions dans la régulation des fonctions de Calnuc qui existerait, de manière dynamique, entre autres comme protéase à sérine. 1.4.7 ATF6 et réponse au stress réticulaire La réponse au stress lié à l'accumulation de protéines mal repliées dans la lumière du réticulum endoplasmique (ou UPR, unfold protein response) est un mécanisme activé pour restaurer les fonctions normales de la cellule. Elle agit de façon à mettre fin à la traduction des protéines, à dégrader les protéines mal repliées et à augmenter la production de chaperonnes moléculaires. L'Activating transcription factor 6 (ATF6) est alors transporté du réticulum endoplasmique au Golgi pour être clivé par la Site-1-protease (S1P) et S2P (Ye et al., 2000). Alors sous sa forme courte, p50ATF6 transloque au noyau pour lier et activer des éléments de réponse au stress réticulaire (ERSE) au niveau de promoteur de gènes cibles (Haze et al., 1999). En 2007, il a été démontré que le promoteur de Calnuc contient des séquences telles que ERSEII et UPR element-like (UPRE-like) qui permettent son induction par ATF6 (Tsukumo et al., 2007). Calnuc agit alors comme rétrorépresseur en situation de stress réticulaire. En effet, elle inhibe le clivage d’ATF6 (figure 3B) en réduisant indirectement son interaction avec S1P de façon à prévenir la translocation au noyau. 23 1.4.8 Necdin La Neurally differentiated embryonal carcinoma-derived protein (Necdin) est un suppresseur de croissance exprimé de façon ubiquitaire dans les tissus humains (Jay et al., 1997), mais principalement dans les neurones postmitotiques. Au niveau du noyau de ces cellules (Maruyama et al., 1991), Necdin maintiendrait l'état postmitotique en interagissant avec les facteurs des transcriptions p53 et E2F1 de manière à réprimer leur action (Taniura et al., 1999). Necdin interagit aussi avec des oncoprotéines virales comme l'Adenovirus early region 1A protein (E1A) et le Simian Vacuolating Virus 40 large T antigen (SV40TAg) (Taniura et al., 1999) lui permettant d'exercer un contrôle négatif sur le cycle cellulaire via la Retinoblastoma protein (pRB). Outre sa localisation nucléaire, Necdin est aussi présente au niveau du cytosol (Niinobe et al., 2000) et est sécrétée (BarnikolWatanabe et al., 1994) ce qui suggère d'autres fonctions. En l'an 2000, un groupe a identifié Calnuc ainsi que NUCB2 comme partenaires d'interaction pour Necdin (figure 3B) (Taniguchi et al., 2000). À cause de la forte homologie entre NUCB2 et Calnuc, il a été proposé que les mêmes régions d'interaction et les mêmes fonctions soient attribuées à Calnuc dans le contexte de Necdin (Taniguchi et al., 2000). Il a été démontré que NUCB2 lie la région centrale de Necdin (102-325), nécessaire aux interactions avec E2F1, p53 et SV4-TAg laissant croire que NUCB2 pourrait séquestrer Necdin l'empêchant de lier les facteurs de transcription. Du côté de NUCB2, l'interaction est attribuée à la région Cterminale qui englobe à la fois la région acide et les mains EF (souris 214–358) (figure 3C), mais elle est indépendante du calcium (Taniguchi et al., 2000). Necdin augmente la rétention intracellulaire de NUCB2 et bloque la sécrétion dans le milieu extracellulaire (Taniguchi et al., 2000). Puisque NUCB2 lie le calcium et est principalement localisée au RE, les auteurs ont investigué les relâches calciques de l'organite par stimulation des récepteurs Ryanodine à la caféine. La sortie de calcium est augmentée par la surexpression de NUCB2 et cet effet est exacerbé par la coexpression de Necdin indiquant un rôle dans l'homéostasie calcique (Taniguchi et al., 2000). Necdin a donc un effet modulatoire sur la distribution et la fonction de NUCB2 ce qui permet de spéculer que Calnuc, semblable à NUCB2, pourrait être affecté de façon similaire. 24 1.4.9 APP et Alzheimer L’Alzheimer est une maladie neurodégénérative qui entraîne des pertes de fonctions mentales irréversibles. Au point de vue mécanistique, la pathologie est caractérisée par un dysfonctionnement cellulaire résultant de l’accumulation de deux protéines. Au niveau intracellulaire, la protéine Tau s’agglutine en neurofibrilles et, au niveau extracellulaire, le peptide β-amyloïde (Aβ) forme des plaques amyloïdes. Ce dernier est issu du clivage anormal du précurseur de la protéine précurseur de la β-amyloïde (APP). Plusieurs facteurs impliqués dans ces processus participeraient également à un déséquilibre de l’homéostasie calcique menant à la dysfonction synaptique et à l’apoptose (LaFerla, 2002; Mattson and Chan, 2003; Yang and Cook, 2004). Quoique mécompris, le maintien des niveaux de zinc serait aussi primordial pour se prévaloir de l’Alzheimer (Cuajungco and Lees, 1997a; Cuajungco and Lees, 1997b). Calnuc, par sa capacité à lier à la fois le calcium et le zinc, a logiquement été associée à la maladie. Il a été démontré que Calnuc interagit directement, et de façon sensible au Ca2+, avec le domaine intracellulaire d’APP (AICD) (figure 3B) (Lin et al., 2007). La population cytosolique de Calnuc serait donc impliquée lorsque présente à la surface du Golgi où elle colocalise avec APP (Lin et al., 2007). D’ailleurs, la surexpression de Calnuc réduit l’ARNm d’APP ce qui entraîne une diminution d’APP et de ses produits protéolytiques (APPα, Aβ et le fragment C-terminal (CTF)) (Lin et al., 2007). À l’opposé, la réduction de Calnuc par petits ARN interférents (siRNA) augmente la biosynthèse d’APP (Lin et al., 2007). Calnuc affecterait donc la transcription d’APP possiblement en séquestrant le fragment AICD l’empêchant d’agir en complexe avec d’autres protéines (Amyloid beta A4 precursor protein-binding family B member 1 (Fe65) et 60 kDa Tat-interactive protein (Tip60)) pour réguler l’expression de gènes incluant APP (Cao and Sudhof, 2001; Kim et al., 2003a; von Rotz et al., 2004). Ces résultats sont appuyés par une diminution de Calnuc dans les cerveaux de patients souffrant de la maladie d’Alzheimer (Lin et al., 2007). 25 1.4.10 LRP10 et transport intracellulaire En 2009, notre groupe a voulu identifier de nouveaux partenaires d'interaction pour Calnuc dans le but d'en apprendre davantage sur ses fonctions intracellulaires. Des expériences de double hybride chez la levure ont apporté les premières évidences d'une interaction, qui s'est avérée directe, entre la région N-terminale de Calnuc et la queue cytoplasmique de la Low density lipoprotein receptor (LDLR)-related protein 10 (LRP10) (figure 3B et 3C) (Brodeur et al., 2009). Ce récepteur de type 1, aussi connu sous le nom de LRP9, fait partie d'une sous-famille de récepteurs aux lipoprotéines de faible densité (LDLR) qui inclue également LRP3 et LRP12. De façon générale, les LDLR sont caractérisés par des répétitions de classe A (LDL-A) extracellulaires permettant la fixation de lipoprotéines à internaliser dans un contexte de métabolisme des lipides. La sous-famille de LRP10 se distingue par l'absence de répétitions EGF-like et de motifs YWTD ainsi que par une grande queue cytoplasmique comportant un domaine riche en prolines, des domaines Complement C1r/C1s, Uegf, Bmp1 (CUB) extracellulaires pour la liaison de ligands, mais surtout au moins un motif acide-dileucine (DXXLL). Ce dernier motif lie des adaptateurs de clathrine nommés Golgi-localized, gamma-ear containing Arf binding protein (GGA) et Adaptator protein (AP) qui permettent l'entrée du récepteur dans des vésicules de clathrine et son transport subséquent entre la membrane plasmique, le TGN et les endosomes (Boucher et al., 2008; Doray et al., 2008). LRP10 a été proposée pour participer à l'internalisation de l'Apolipoprotéine E (Sugiyama et al., 2000), mais sa localisation préférentielle au TGN et aux endosomes (Boucher et al., 2008; Doray et al., 2008) indique plutôt une fonction potentielle à ce niveau. À ce sujet, il été démontré que LRP10 agit comme récepteur de transport, interagissant directement avec l'ectodomaine d'APP pour la maintenir au Golgi et prévenir son clivage amyloïdogénique (Brodeur et al., 2012). Au contraire, une diminution de LRP10 a pour effet de redistribuer APP aux endosomes précoces et d'augmenter la production du fragment Aβ par la β-sécrétase qui s'y trouve (Brodeur et al., 2012). Ceci n'est pas sans rappeler qu'une réduction de Calnuc augmente la biosynthèse d’APP (Lin et al., 2007). D'ailleurs, comme c'est le cas pour Calnuc (Lin et al., 2007), une diminution de LRP10 est observée chez les patients souffrants de la maladie d'Alzheimer (Brodeur et al., 2012). 26 L'implication de Calnuc et de LRP10 dans l'Alzheimer supporte un rôle fonctionnel pour leur interaction qui est indépendante du calcium (Brodeur et al., 2009). Dans la queue cytosolique de LRP10, une région riche en arginines, qui forme un segment amphiphilique propice à des interactions, s'est avérée importante, mais pas nécessaire, pour la liaison avec Calnuc (Brodeur et al., 2009). Les deux protéines colocalisent au niveau du TGN et des endosomes et la déplétion de Calnuc, par siRNA, a pour conséquence d'entraîner une diminution des niveaux du récepteur (Brodeur et al., 2009). En effet, ce dernier n'est alors plus recyclé vers le Golgi et est plutôt redistribué aux lysosomes où il est dégradé (Brodeur et al., 2009). Ce phénomène peut être renversé par l'expression d'un mutant strictement cytosolique de Calnuc, impliquant cette fraction protéique dans l'effet observé (Brodeur et al., 2009). De la même manière, Calnuc interagit et affecte le triage de LRP3, un autre LDLR de la même sous-famille que LRP10, qui possède aussi une région riche en arginines (Brodeur et al., 2009). Ces travaux indiquent que Calnuc serait impliquée dans le transport intracellulaire de LRP10 et LRP3 et laisse croire que Calnuc pourrait également affecter d'autres récepteurs transitant entre le TGN et les endosomes. 2 Transport intracellulaire Le transport bidirectionnel intracellulaire permet de recevoir, dissocier et trier les cargos de la membrane plasmique ou de la voie de synthèse tout en maintenant l'identité de chaque compartiment. La voie antérograde assure la distribution de nombreuses protéines du RE, en passant par le Golgi, vers les lysosomes, la membrane plasmique et le milieu extracellulaire. De son côté, la voie rétrograde permet l'internalisation de multiples molécules ainsi que le recyclage de matériel réutilisable jusqu'au Golgi, et même jusqu'au RE. Dans le cadre de la présente étude, seuls les mécanismes d'échanges et de régulation entre le Golgi et le système endolysosomial seront abordés (figure 4A). 2.1 Transport antérograde La voie de transport antérograde achemine des protéines nouvellement synthétisées au RE vers leur lieu d'action. Leur passage à travers l'organite leur permet de se replier 27 adéquatement, d'oligomériser et d'acquérir les premières modifications telles que des glycosylations, des ponts disulfures et des ancrages lipidiques de type Glycosylphosphatidyl inositol (GPI). Elles gagnent ensuite le Golgi, organite décrit en 1898 par l'histologiste italien du même nom, qui modifie davantage les protéines et qui agit comme centre de trie et de distribution. En effet, en traversant les différentes citernes contenant un assortiment d'enzymes variées, les protéines subissent de plus amples altérations. Entre autres, la modification de l'arborescence de glycosylations, combinée à une phosphorylation, engendre la formation d'un groupement mannose-6-phosphate (M6P) (Rohrer and Kornfeld, 2001), agissant comme signal d'adressage vers le lysosome, pour plus de 60 précurseurs d'hydrolases acides (Johannes and Wunder). Ces dernières sont transportées, grâce aux récepteurs au M6P (MPR), vers le système endo-lysosomial pour y subir une activation protéolytique. D'autres récepteurs, comme Sortiline et LIMP2 assurent le transport indépendant des groupements M6P. 2.1.1 Transport dépendant de M6P - MPR Il existe deux types de récepteurs MPR, soit le récepteur dépendant des cations manganèse (CD-MPR) et le récepteur indépendant des cations (CI-MPR). Ces glycoprotéines homodimériques (Dahms and Hancock, 2002) transmembranaires de type-1 sont homologues à seulement 20% et sont les seuls membres de la famille des récepteurs de lectine de type P (Dahms and Hancock, 2002), c’est-à-dire qui ont la capacité de lier des groupements M6P. Ils transportent des populations différentes, mais chevauchantes d’hydrolases acides en utilisant des mécanismes similaires de reconnaissance des carbohydrates (Roberts et al., 1998). Ainsi, CD-MPR a comme cargo la Tripeptidyl peptidase I (TPP1), E1A-stimulated genes 1 (Creg1) et l'Héparanase (Hpse) alors que CIMPR lie préférentiellement l'α-Mannosidase B1 (Man2b1), la Cathepsine D (CTSD) et la Prosaposine (Psap) (Qian et al., 2008). Le monomère de CD-MPR a une masse moléculaire de 46 kDa et possède un peptide signal en N-terminal, suivit d’une région luminale de 159 acides aminés qui contient un motif liant M6P (Lobel et al., 1988). Son domaine cytoplasmique inclut des signaux de triage, tels que des palmitoylations (Schweizer et al., 1996) et des phosphorylations (Meresse et al., 1990; Rosorius et al., 1993), ainsi que des 28 sites d’interactions pour les protéines de transport (protéines adaptatrices et rétromères). De son côté, CI-MPR est conservé chez les mammifères et invertébrés et il est de loin le récepteur le mieux caractérisé de la voie endosomiale. Le monomère indépendant des cations est beaucoup plus massif avec ses 300 kDa. Sa structure générale est la même que celle de CD-MPR, mais sa région luminale compte 15 répétitions contigües de 147 acides aminés séparées par des espaceurs (linkers) de 5-12 résidus (Lobel et al., 1988) qui restreignent les arrangements moléculaires (Brown et al., 2002). On y retrouve quatre motifs de liaison pour des groupements M6P (Olson et al., 2015), un site de liaison au récepteur de l'Urokinase (uPAR), un site pour le Plasminogène et un site pour l'Insulin-like growth factor-2 (IGF-2) (Garmroudi et al., 1996; Hancock et al., 2002; Marron-Terada et al., 2000; Schmidt et al., 1995). En effet, les MPR sont retrouvés en faible quantité à la surface cellulaire (3-10% de la quantité totale des MPR) (Braulke et al., 1987; Breuer et al., 1997; Geuze et al., 1988), où seul CI-MPR internalise plusieurs ligands glycosylés ou non glycosylés. CI-MPR possède donc de multiples fonctions qui dépassent la seule ségrégation d'enzymes lysosomiales. 2.1.2 Transport indépendant de M6P - Sortiline L'étude de la mucolipidose II (Inclusion-cell disease) a rapidement imposé un mode de transport alternatif entre le TGN et les endosomes qui soit indépendant des M6P. En effet, la localisation de nombreuses enzymes n'est pas affectée par l'inhibition de la formation des groupements M6P qui caractérise la pathologie. C'est principalement Sortiline (95kDa), aussi connu sous le nom de Neurotensin receptor 3 (NTR3), qui assure alors le triage de ces cargos, dont la Psap (Lefrancois et al., 2003), la GM2-activator protein (Lefrancois et al., 2003) et la Sphingomyélinase (ASM) (Ni and Morales, 2006). Il appartient à la famille des protéines structurellement liées à la protéine de triage vacuolaire (Vps) 10 de la levure (Vps10p) qui comprend également SorCS1, SorCS2, SorCS3 et Sorting protein-related receptor containing LDLR class A repeats (SorLA) (Marcusson et al., 1994). Au niveau de la région luminale, ils sont caractérisés par un domaine riche en cystéines homologue à celui de Vps10p qui permet la liaison des ligands. Le motif reconnu par ce domaine n'a toujours pas été identifié, mais tous les cargos lysosomiaux possèdent 29 des hélices alpha proéminentes qui pourraient être impliquées (Canuel et al., 2009). Sortiline dispose aussi d'un site de clivage pour la Furine (Munck Petersen et al., 1999) et d'une courte queue cytoplasmique homologue à celle de CI-MPR qui contient des motifs d'interaction pour la machinerie de transport intracellulaire (Nielsen et al., 2001). Tout comme les MPR, Sortiline se retrouve également en faible quantité (10% de la quantité totale de Sortiline) à la surface cellulaire où il participe à l’endocytose d'une variété de molécules, dont la Neurotensine (NT) (Quistgaard et al., 2009). D'ailleurs, Sortiline peut agir en tant que corécepteur, entre autres avec NTR1, pour moduler la signalisation de ce neuropeptide. Sortiline est exprimé surtout au niveau des neurones du système nerveux central et périphérique (Petersen et al., 1997; Sarret et al., 2003) où il agit comme récepteurs pour des molécules proapoptotiques comme le précurseur du Nerve growth factor (NGF) (Nykjaer et al., 2004) et du Brain derived neurotrophic factor (BDNF) (Teng et al., 2005). Sortiline serait donc important pour l'induction de la mort neuronale et aurait aussi un rôle dans la neurodégénération associée à la maladie d'Alzheimer. En effet, Sortiline médie le transport rétrograde de la ß-sécrétase (BACE1), des endosomes précoces vers le TGN, ce qui favorise le clivage amyloïdogénique menant à la formation du fragment Aß42 (Finan et al., 2011). À son tour, ce fragment toxique hautement oligomérisable cause une augmentation des niveaux de Sortiline (Saadipour et al., 2013). Ces données sont appuyées par une surexpression de Sortiline dans les cerveaux de patients atteints d'Alzheimer (Saadipour et al., 2013). D'un autre côté, Sortiline favorise également le clivage non amyloïdogénique d'APP par l'α-sécrétase (Gustafsen et al., 2013). Il médie l'internalisation d'APP soluble et facilite sa dégradation au lysosome (Gustafsen et al., 2013). De plus, Sortiline permet l'internalisation d'Aß neurotoxique séquestré par l'Apolipoprotéine E (ApoE) au niveau extracellulaire et mène à sa dégradation (Carlo et al., 2013). Les souris n'exprimant pas Sortiline montrent d'ailleurs une augmentation d'ApoE due à un défaut de clairance, des niveaux exacerbés d'Aß et la formation de plaques amyloïdes (Carlo et al., 2013). Sortiline serait donc un récepteur neuronal ayant un rôle protecteur en contrant l'accumulation d'Aß dans le cerveau grâce à l'ApoE et formerait un lien mécanistique entre le métabolisme des lipoprotéines et la survie neuronale. Sortiline se retrouve également au niveau de cellules 30 spécialisées comme les hépatocytes et les adipocytes (Morris et al., 1998). Au niveau du foie, il interagit avec les Apolipoprotéines A–V pour contrôler l'export de lipoprotéines vers la circulation sanguine (Kjolby et al., 2010). Finalement, chez les adipocytes, il régulerait le système de transport du glucose en réponse à l'insuline et médierait l'internalisation et la dégradation de la Lipoprotein lipase (Lpl) (Nielsen et al., 1999), une enzyme qui hydrolyse les triglycérides contenus dans les lipoprotéines. Il est à noter qu'en plus de la voie de transport directe indépendante des groupements M6P, certaines enzymes peuvent emprunter une voie indirecte en fonction du type cellulaire. Entre autres, au niveau des fibroblastes, de 5 à 20% des protéines lysosomiales nouvellement synthétisées sont sécrétées. Dans ces cas, suite à la sécrétion il y a recapture par endocytose grâce à des récepteurs (Dittmer et al., 1999). Deux membres de la famille des LDLR, soit Megalin et LRP1 seraient ainsi capables de transporter CTSB (Nielsen et al., 2007b) ainsi que Psap et CTSD (Markmann et al., 2015), respectivement. 2.1.3 Transport du Golgi aux endosomes À la suite de leur passage au Golgi, les diverses protéines, telles les hydrolases acides, atteignent le réseau trans-golgien où elles subissent un tri qui les orientera vers leur destination. Pour ce faire, elles sont reconnues par des récepteurs de triage, comme MPR et Sortiline. Le complexe récepteur-cargo est ensuite séquestré dans un microdomaine correspondant à une vésicule de clathrine bourgeonnante qui s'enrichit de diverses protéines utiles ultérieurement. Au TGN, la formation de ces structures est initiée par l'action de la GTPase ADP-ribosylation factor-binding protein 1 (Arf-1) qui recrute à la membrane des protéines adaptatrices de clathrine nommées GGA et AP (figure 4B) (Cooper and Hausman, 2006; Doray et al., 2002). Chez l'humain, il existe trois types de GGA (1-3) qui assemblent le manteau de clathrine grâce à leur région non structurée (Hinge) et leurs domaines modulaires : Vps27, Hrs, Stam (VHS), GGA and TOM (GAT) et Gamma-adaptin ear (GAE). Le domaine GAT est responsable de la liaison à Arf, le domaine VHS reconnaît le motif dileucine (DXXLL) (Chen et al., 1993; Chen et al., 1997; Johnson and Kornfeld, 1992a; Johnson and Kornfeld, 1992b) de la queue cytoplasmique des récepteurs, la région 31 Hinge lie la clathrine alors que le domaine GAE recrute des protéines accessoires (Bonifacino, 2004). Le modèle actuel veut que GGA interagisse avec AP (Doray et al., 2002), de manière à lui transférer le complexe récepteur-cargo (figure 4B) (Doray et al., 2002). Il existe cinq hétérotétramères AP (1-5) (Hirst et al., 2011) ayant une fonction similaire aux GGA, mais AP-1 serait le principal acteur au niveau du TGN (Hirst et al., 2011), où il reconnaît les motifs [D/E]XXXL[L/I] et YXXØ (où Ø est un résidu hydrophobe encombrant) des récepteurs (Bonifacino and Traub, 2003). Les adaptatrices de clathrine permettent le recrutement de plusieurs protéines à domaine de liaison aux phospholipides ENTH/ANTH, dont Epsin, qui induit la polymérisation du triskèle de clathrine (Rosenthal et al., 1999) et la formation d’une invagination (Ford et al., 2002) (figure 4B). Alors que la vésicule croît, la Dynamine s’autoassemble autour du cou du bourgeon se resserrant jusqu’au détachement de la vésicule (Praefcke and McMahon, 2004). Rapidement, le manteau se disloque suivant l’action de la chaperonne Heat shock cognate protein 70 (Hsc70) (Schlossman et al., 1984) et de son partenaire Auxiline (Prasad et al., 1993) qui fixent la chaîne lourde de clathrine. Dès lors, la vésicule nue expose la machinerie nécessaire à l’attachement et la fusion spécifique de la vésicule avec les endosomes précoces, la première composante du système endolysosomial. À cet endroit, l’environnement acide entraîne la dissociation des complexes récepteur-cargo. Conséquemment, les cargos comme les hydrolases sont acheminés aux lysosomes pour effectuer leur fonction dans la dégradation, tandis que les récepteurs sont redirigés vers le TGN pour un nouveau cycle de transport. 2.1.4 Système endolysosomial Le système endolysososomial est un réseau complexe de compartiments intracellulaires qui assure le tri de molécules dans le but de les réutiliser (recycler) ou de les détruire. Il se compose d'endosomes précoces, d'endosomes tardifs (corps multivésiculaires) et de lysosomes (figure 4A) (Huotari and Helenius, 2011). Jadis perçu comme des compartiments distincts stables échangeant du matériel par l'intermédiaire de vésicules, on sait aujourd'hui qu'il s'agit plutôt de compartiments dynamiques dont la composition hétérogène indique un phénomène de maturation en continuité (Cullen, 2011). 32 Figure 4 - Principaux mécanismes de transport antérograde et rétrograde. (A) Schématisation du système endolysosomial. Des vésicules de clathrine contenant de multiples cargos bourgeonnent du Golgi pour aller fusionner avec les endosomes précoces issus des processus d'internalisation à la membrane plasmique. Ces derniers sont caractérisés par des domaines de triage tubulaire qui permettent le recyclage sélectif de cargos vers le Golgi ou la membrane plasmique. Alors que les endosomes se déplacent le long des microtubules vers la région périnucléaire, ils maturent en endosomes tardifs qui fusionnent avec les lysosomes. Figure modifiée de Endosome maturation, Huotari and Helenius, The EMBO Journal (2011) 30, 3481–3500 avec la permission de John Wiley and Sons. (B) Formation d'une vésicule de clathrine sur le TGN. L'activation d'Arf permet le recrutement de protéines adaptatrices de clathrine qui concentre les cargos dans des microdomaines formant des vésicules qui émergent du Golgi. Arf (ADP-ribosylation factor-binding protein), GEF (facteur d’échange des nucléotides de guanine), GGA (Golgi-localized, gamma-ear containing Arf binding protein) et AP (Adaptator protein). Figure modifiée de The cell : A molecular approach 6e, Figure 10.36 Cooper and Hausman avec la permission de Sinauer Associates. (C) Modèle général du recrutement du complexe Rétromère. Le sous-complexe de Vps est recruté sur l'endosome par Rab7 de manière à ségréger les cargos dans des microdomaines tubulaires créés par l'action du sous-complexe de SNX. L'association de protéines motrices et la polymérisation de l'actine génèrent alors des forces qui entraînent le détachement du transporteur en direction du Golgi. Vps (Vacuolar protein sorting), SNX (Sorting Nexin), WASH (WASP and SCAR homologue). Figure inspirée de A global analysis of SNX27–retromer assembly and cargo specificity reveals a function in glucose and metal ion transport. Steinberg et al. Nat Cell Biol. 2013 May;15(5):461-71 avec la permission de Nature Publishing Group (Macmillan Publishers Ltd). 33 Le système endolysosomial est généré par la fusion de vésicules d'endocytose formant des endosomes précoces localisés en périphérie. Suivant l'action de la pompe à proton de type vacuolaire (type V) (Yamashiro et al., 1983), ces derniers sont légèrement acides (pH 6,8–5,9) ce qui assure la dissociation des complexes récepteur-cargo en provenance de la membrane plasmique ou du TGN. Ils sont constitués d'un domaine sphérique duquel émergent de nombreux tubules (Chia et al., 2013) servant au triage des molécules en transit (Cullen, 2011). Par exemple, les récepteurs permettant l'internalisation de nutriments, comme Transferrin receptor (TfR), sont recyclés vers la membrane plasmique alors que les récepteurs de transport, comme MPR et Sortiline, sont acheminés vers leur point d'origine, soit le TGN. TfR constitue donc un marqueur caractéristique de l'organite de même que les protéines cytosoliques associées comme EEA1, Rab4 et Rab5. L'endosome précoce va graduellement maturer et sa région sphérique va former un endosome tardif. L'endosome tardif est typiquement rond, sans tubules, et contient une grande quantité de vésicules intraluminales, d'où son nom alternatif de corps multivésiculaire. Il se déplace vers le noyau et peut devenir un lysosome par l'acidification graduelle de son contenu, passant de pH 6,0– 4,9 à 5,0-4,5 (Yamashiro and Maxfield, 1987), ou par la fusion avec des lysosomes préexistants. En fusionnant avec les lysosomes, l'endosome tardif suit une voie essentiellement unidirectionnelle puisque la plupart de ses composantes seront dégradées. En effet, les lysosomes sont principalement des compartiments de dégradation où converge le matériel de la voie d'endocytose, de phagocytose, d'autophagie et de synthèse. De façon générale, le lysosome se distingue par une haute teneur en hydrolases, par la présence de nombreuses glycoprotéines protégeant sa membrane, telles les Lysosome-associated membrane proteins (LAMP) et les Lysosomal integral membrane proteins (LIMP), et par l'accumulation de matériel résistant à la dégradation comme la lipofuscine. Par contre, dû à la grande diversité de matériel à dégrader et le degré de cette dégradation, il s'agit de compartiments très hétérogènes tant au niveau de leur composition, de leur morphologie que de leur localisation. Les fonctions et composantes des lysosomes seront plus amplement développées ultérieurement. 34 2.2 Transport rétrograde Les premières évidences de l'existence d'un système de transport rétrograde des endosomes vers le Golgi sont venues de l'étude de la toxine du ricin. Cette glycoprotéine retrouvée dans les graines d'un arbrisseau produit son effet toxique en altérant l'ARN ribosomique de manière à inhiber la synthèse protéique (Olsnes and Pihl, 1972). En 1975, la ricine a été observée au niveau du Golgi indiquant la présence d'un système permettant à la molécule exogène de pénétrer par endocytose pour remonter la voie de synthèse (Gonatas et al., 1975). Cette découverte a ouvert la voie à l'identification de protéines cellulaires, dont MPR, qui opère en cyclant entre le TGN et les endosomes (Duncan and Kornfeld, 1988). On sait aujourd'hui que la voie de transport rétrograde peut remonter jusqu'au RE et peut être empruntée par moult molécules classifiées selon cinq groupes, soit les récepteurs de triage (MPR, Sortiline, LRP9 et Wntless), les protéases membranaires intégrales (β-sécretase et Furine), les protéines médiant la fusion des vésicules, les transporteurs de nutriments et les toxines (toxine Shiga (Sandvig et al., 1992), toxine du choléra, toxine pertussique, toxine du ricin) (Burd, 2011). Elle permet le maintien des fonctions du Golgi en le réapprovisionnant de ces molécules pour leur réutilisation. Plusieurs processus cellulaires dépendent donc du bon fonctionnement de la voie de transport rétrograde. Son dysfonctionnement entraîne, entre autres, une l'élévation du peptide β-amyloïde associé à la maladie d'Alzheimer, un débalancement de l'homéostasie des nutriments/éléments (glucose, fer, cuivre) et la perturbation du gradient de morphogenèse de Wnt, une protéine multifonctionnelle d'importance (Burd, 2011). Depuis l'étude des toxines qui a montré le lien à contresens entre la membrane plasmique, les endosomes, le Golgi et le RE, les mécanismes permettant le transport rétrograde ont été approfondis. On distingue désormais deux voies de recyclage existant en parallèle et servant des cargos différents. Ainsi, les molécules peuvent se rendre au TGN à partir des endosomes précoces ou des endosomes tardifs en utilisant des mécanismes différents. 35 2.2.1 Transport des endosomes précoces au Golgi Les endosomes précoces, ou plus particulièrement les tubules qui émergent de leur domaine sphérique, constituent la principale station de triage de la voie endosomiale. À ce niveau, le transport rétrograde vers le Golgi ou la membrane plasmique est orchestré par l'assemblage d'un complexe qui, à l'image du manteau de clathrine, reconnaît et concentre des cargos dans un transporteur. Ce complexe, nommé Rétromère, agit pendant la maturation de l'endosome de manière à prévenir la dégradation au lysosome de nombreuses protéines dont la fonction dépend de leur retour dans la voie de biosynthèse. Le cargo le mieux caractérisé de cette voie est CI-MPR qui cycle entre les endosomes et le Golgi (Seaman, 2004). Sortiline, le transporteur ionique Divalent metal transporter 1 (DMT1)-II (Tabuchi et al., 2010), Wntless la protéine de transport de Wnt (Eaton, 2008), la protéine de polarité cellulaire nommée Crumbs (Pocha et al., 2011) et SorLA un récepteur d'APP (Nielsen et al., 2007a) empruntent également cette voie de transport rétrograde. Le complexe Rétromère a été découvert en 1997 (Seaman et al., 1997), chez la levure Saccharomyces cerevisiae. L'équipe du Dr. Emr a alors démontré son importance pour le recyclage du récepteur à la Carboxypeptidase Y (CPY) des endosomes vers le Golgi (Seaman et al., 1997; Seaman et al., 1998). Il s'agit d'un hétéropentamère formé par l'association transitoire de deux sous-complexes distincts, soit un dimère de Sorting Nexin (SNX) et un trimère de Vacuolar protein sorting (Vps). Vps et capture du cargo Chez les mammifères, le trimère de Vps est responsable de la reconnaissance et de la sélection des cargos (figure 4C) (Seaman et al., 1998; Steinberg et al., 2013). Il comprend les protéines Vps26, Vps29 et Vps35 qui ont été remarquablement conservées au cours de l'évolution des eucaryotes (Koumandou et al., 2011). La composante centrale, Vps35, se présente sous forme d'un solénoïde α-hélicoïdal (α-helical solenoid) en forme de fer à cheval sur lequel Vps26 et Vps29 s'associent aux extrémités opposées, soit N-terminale et C-terminale respectivement (Hierro et al., 2007; Norwood et al., 2010). Vps35 est également le principal responsable de la reconnaissance des cargos (Nothwehr et al., 2000). Il lie au moins un motif conservé de séquence consensus Trp/Phe-Leu-Met/Val (F/W-L36 M/V) dans la queue cytosolique de CI-MPR (motif WLM) et Sortiline (motif FLV) (Seaman, 2007). Chez CI-MPR, ce dernier chevauche le motif de liaison d'AP-1 ([D/E]XXXL[L/I]) démontrant une potentielle régulation structurelle entre le transport antérograde et rétrograde. Des mécanismes de reconnaissance alternatifs, impliquant les autres Vps, commencent à émerger. Vps26 lie directement un motif hydrophobe FANSHY dans la portion cytoplasmique de SorLA (Fjorback et al., 2012) et servirait à la reconnaissance des cargos. Chez l'humain, Vps26 présente deux paralogues (Vps26A et Vps26B) (Kerr et al., 2005) qui pourraient agir à des endroits distincts et pour des cargos différents, comme suggéré par la préférence de CI-MPR pour les complexes possédant Vps26A (Bugarcic et al., 2011). De son côté, Vps29 serait un régulateur du transport rétrograde. En plus d'un rôle structural (Collins et al., 2005a), il possède une activité phosphatase, et ce malgré un repliement de type métallophosphoestérase (Damen et al., 2006). In vitro, Vps29 a la capacité de déphosphoryler la sérine précédant le motif dileucine antérograde (SDEDLL) de CI-MPR et Sortiline. Cette dernière agit comme signal pour déclencher le transport antérograde et sa déphosphorylation pourrait être importante pour le transport vers le TGN (Damen et al., 2006). De plus, la palmitoylation réversible de certains récepteurs est nécessaire pour l'interaction avec les rétromères et, conséquemment, pour la séquestration des cargos dans des microdomaines de transport vers le TGN (McCormick et al., 2008). Il pourrait s'agir là d'un autre mécanisme de régulation agissant, cette fois, en altérant la position et l'accessibilité du motif de triage. SNX et tubulation des endosomes Chez l'humain, le deuxième sous-complexe de rétromères, soit l'hétérodimère de SNX (figure 4C), est généralement composé de SNX1 ou de SNX2 associée à SNX5 ou SNX6 (Wassmer et al., 2007). Il est responsable de détecter la courbure membranaire et de générer la tubulation de l'endosome (Carlton et al., 2004). Il a été proposé que la formation de tubules permettrait de réduire la diffusion latérale des protéines afin de faciliter leur ségrégation (Cullen and Korswagen). Aussi, cette région de haute courbure serait moins épaisse et contiendrait peu de lipides saturés rigides (cholestérol et sphingomyéline) (Roux et al., 2005) ce qui faciliterait la concentration préférentielle des cargos à courts domaines transmembranaires (Sharpe et al., 2010), tels que SorLA et TfR (Chia et al., 2011). Pour 37 générer des microdomaines tubuleux, les SNX lient le PI3P et le PI(3,5)P2 enrichis dans les endosomes, grâce à leur domaine Phox homology (PX) (Cozier et al., 2002) qui favorise leur autoassemblage (Kurten et al., 2001). La dimérisation s'effectue par le domaine Cterminal Bin-Amphiphysine-Rvs (BAR) et résulte en une structure en arc qui épouse et lie les membranes de courbure positive par des interactions électrostatiques (Peter et al., 2004). Les hélices amphipathiques des SNX, à l'interface entre le cytosol et les lipides, génèrent alors des tensions sur la membrane qui s'y accommode en se déformant (Peter et al., 2004). Ainsi, les SNX agissent à la fois comme détecteur de courbure et comme responsables de la tubulation des endosomes. Toutefois, la famille des SNX est grande et, au cours de l'évolution, ses membres sont de plus en plus nombreux et divergents. Plusieurs formes de complexes Rétromère peuvent donc être formées avec divers SNX de manière à agir à des endroits distincts pour des cargos distincts. Par exemple, il existe un complexe composé des Vps en association avec SNX3, qui ne possède pas de domaine de dimérisation BAR, et qui permet le retrait de Wntless vers le TGN (Harterink et al., 2011). Les SNX-BAR et SNX3 semblent constituer des voies différentes agissant à la transition entre l'endosome précoce et tardif (van Weering et al., 2012) et aux endosomes précoces (Harterink et al., 2011), respectivement. Aussi, ils utilisent des transporteurs morphologiquement distincts, les SNX-BAR se retrouvant au niveau de tubules (Mari et al., 2008) alors que les SNX3 sont associés à des vésicules (Harterink et al., 2011) et à la clathrine (Skanland et al., 2009). D'ailleurs, le rôle de la clathrine dans le transport rétrograde, ainsi que de multiples protéines en relation, comme les adaptateurs EpsinR et AP-1 et l'enzyme de polymérisation Lowe oculocerebrorenal syndrome protein ou Inositol polyphosphate 5-phosphatase (OCRL1) est incertain. Ces dernières pourraient former une région de séquestration précédant l'assemblage de certaines SNX (McGough and Cullen, 2011). Il existe également un complexe composé de l'association entre SNX27 et les Vps qui, de son côté, permet le recyclage à partir de l'endosome vers la membrane plasmique (Temkin et al., 2011). Plus de 70 cargos, dont le récepteur bêta-adrénergique (β2-AR), empruntent ce mécanisme dépendant du motif de triage Postsynaptic density 95/discs large/zonus occludens-1 (PDZ) (Lauffer et al., 2010) ou NPxY/NxxY (Ghai et al., 2013) reconnu directement par SNX27. Bien d'autres SNX émergent comme étant impliquées dans la formation de complexes 38 Rétromère laissant présager une mécanistique complexe optimisée de manière à favoriser le retrait d'un maximum de cargos. Recrutement des rétromères Pendant longtemps, le recrutement du complexe Rétromère aux endosomes a été attribué à la seule association du sous-complexe SNX aux phospholipides membranaires (Rojas et al., 2007). Ce dernier génère des domaines tubulaires puis, via sa région Nterminale non structurée, agirait comme plateforme pour le recrutement du sous-complexe Vps avant de capturer le cargo. Depuis, ce modèle s'est avéré être une version simpliste ne reflétant guère la complexité de la machinerie intracellulaire. En effet, les SNX sont nécessaires, mais insuffisants pour le recrutement des Vps. Elles présentent seulement des interactions faibles avec Vps35 et avec plusieurs sites de Vps29 (Swarbrick et al., 2011), ce qui est peu propice à un phénomène de recrutement (Seaman, 2012). Il est maintenant établi que Rab7-GTP participe aussi au recrutement et à la stabilisation du trimère de Vps (figure 4C) (Liu et al., 2012; Nakada-Tsukui et al., 2005; Rojas et al., 2008). Cette petite protéine G lie Vps35 sur l'endosome tardif (Zelazny et al., 2013) de manière à coordonner l'export du cargo avec la maturation de l'endosome. D'ailleurs, les transporteurs tubulaires bourgeonnent majoritairement à partir des endosomes ayant acquis Rab7 (van Weering et al., 2012). En retour, Vps29 interagit directement avec Tre2-Bub2-Cdc16-domaincontaining protein (TBC1D5), une GAP potentielle de Rab7 (Harbour et al., 2010; Seaman et al., 2009). Ceci mènerait, une fois le transporteur tubulaire généré, à la dissociation de Rab7 et conséquemment à la modification des phosphoinositides ainsi qu'au démantèlement du manteau de rétromères (Seaman et al., 2009). Implication du cytosquelette Afin de permettre le détachement du transporteur tubulaire de l'endosome, les rétromères recrutent une panoplie de protéines additionnelles, dont le complexe Wiskott– Aldrich Syndrome protein (WASP) and Suppressor of cAR (cAMP receptor) (SCAR) homologue (WASH) (figure 4C). L'activité de WASH a été montrée comme étant requise pour le triage dans le recyclage de CI-MPR à partir des endosomes vers le Golgi (Gomez and Billadeau, 2009) ainsi que pour le recyclage de TfR et de β2-AR des endosomes vers la 39 membrane plasmique (Derivery et al., 2009; Temkin et al., 2011). Comme les autres membres de la famille Wiskott–Aldrich Syndrome protein (WASP), WASH est un facteur qui promeut la formation de réseaux d'actine ramifiés (Campellone and Welch, 2010; Rotty et al., 2013). Il est formé par l'assemblage de la protéine WASH avec les protéines régulatrices Strumpelline, FAM21, Strumpellin and WASH-interacting protein (SWIP) et Coiled-coil domain-containing protein 53 (CCDC53) (Derivery and Gautreau, 2010). Quoique de nombreuses interactions stabilisantes soient répertoriées entre les composantes du complexe Rétromère et celles du complexe WASH (Gomez and Billadeau, 2009), c'est FAM21, via sa liaison à Vps35, qui est l'élément clé pour le recrutement à la membrane endosomiale. FAM21 possède 21 copies du motif de liaison reconnu par Vps35 suggérant que la protéine peut s'associer avec de multiples complexes Retromère de manière à les enrichir dans des microdomaines tubulaires (Jia et al., 2012; Puthenveedu et al., 2010). À cet endroit, l'activation locale de WASH entraîne une forte polymérisation de l'actine qui génère une force de poussée longitudinale (figure 4C) (Derivery et al., 2009). De son côté, les SNX5 et SNX6 s'associent avec une composante du complexe moteur de la Dynéine (Hong et al., 2009). En s'éloignant sur le microtubule, cette dernière génère une force de traction sur la région tubulaire de l'endosome (figure 4C) (Hong et al., 2009; Wassmer et al., 2009). Les rétromères, en combinaison avec l'actine et les microtubules, sont donc nécessaires à la génération de ces deux forces qui se combinent pour augmenter la tension sur la membrane et entraîner le détachement du tubule. À ce sujet, la protéine EH domaincontaining protein 1 (EHD1) lie Vps26 et, par sa similitude avec la Dynamine (Daumke et al., 2007), pourrait s'assembler en structures oligomériques pour permettre l’étranglement du tubule en plus de stabiliser sa formation (Gokool et al., 2007). Maladies associées aux rétromères L'importance du transport rétrograde s'illustre par le nombre grandissant de maladies qui lui sont associées. Brièvement, une réduction des niveaux protéiques des rétromères est liée à la maladie d'Alzheimer (Muhammad et al., 2008) puisqu'il y a augmentation du clivage amyloïdogénique d'APP possiblement aux endosomes (Choy et al., 2012; Wen et al., 2011). Aussi, une mutation précise dans Vps35(Asp620Asn) est à l'origine d'une forme de maladie de Parkinson (Zimprich et al., 2011). La perte de l'interaction entre le sous40 complexe de Vps et Rab7, causée par la mutation Rab7(K157N), provoque la maladie de Charcot-Marie-Tooth (De Luca et al., 2008). Toutes ces pathologies sont caractérisées par une atteinte du système nerveux. Il est fort à parier que la découverte de nouveaux joueurs dans le transport rétrograde conduira à l'identification d'autres mutations aboutissant à diverses maladies neurodégénératives. 2.2.2 Transport des endosomes tardifs au Golgi En 1993, le groupe de Pfeffer a suggéré pour la première fois que des récepteurs puissent être recyclés vers le TGN à partir des endosomes tardifs (Lombardi et al., 1993). Rab9 et son effecteur Tail interacting protein of 47 kDa (TIP47) sont les deux principales protéines impliquées dans ce processus. Rab9 colocalise avec MPR au niveau des endosomes tardifs et stimule le transport rétrograde du récepteur (Lombardi et al., 1993) in vitro (Goda and Pfeffer, 1988) alors que l'expression d'un dominant négatif a l'effet inverse in vivo (Riederer et al., 1994). De son côté, TIP47 a été identifiée comme protéine interagissant directement avec les queues cytosoliques des MPR, par l'intermédiaire d'un motif Phe-Trp pour CD‐MPR (Diaz and Pfeffer, 1998; Schweizer et al., 1997) et d'un domaine riche en prolines pour CI‐MPR (Orsel et al., 2000). Sa déplétion empêche le recyclage des récepteurs vers le TGN (Diaz and Pfeffer, 1998) qui sont alors acheminés aux lysosomes (Ganley et al., 2004). TIP47 interagit aussi directement avec Rab9 actif faisant de lui un effecteur. Il augmente l'affinité de TIP47 pour MPR qui à son tour augmente la liaison de TIP47 à Rab9 (Carroll et al., 2001; Hanna et al., 2002). Rab9 n'interagit pas avec les récepteurs, mais facilite le recrutement de TIP47 spécifiquement au niveau des endosomes tardifs et stimule la ségrégation des récepteurs par TIP47 dans la vésicule de transport en formation. À ce moment, TIP47 formerait un hexamère grâce à son domaine d'oligomérisation en N-terminal. Cet assemblage est dispensable pour la reconnaissance des récepteurs, mais nécessaire pour la stimulation du transport rétrograde (Sincock et al., 2003). Rho-related BR-C, ttk and bab (BTB) domain-containing protein 3 (RhoBTB3), une Rho atypique hydrolysant de l'ATP plutôt que du GTP, a été suggéré comme étant responsable de la libération de TIP47 de la vésicule (Espinosa et al., 2009). Elle lie spécifiquement et directement Rab9 ainsi que TIP47 au niveau des membranes (Espinosa et 41 al., 2009). À ce jour, la Furine, une proprotéine convertase, est le seul autre récepteur identifié pouvant utiliser la voie de recyclage dépendante de Rab9 (Chia et al., 2011). Par contre, TIP47 ne lie pas la Furine ainsi que d'autres protéines cyclant entre les endosomes et le TGN comme TGN38 et la Carboxypeptidase D (Hanna et al., 2002). De plus amples études seront nécessaires afin de mettre en lumière tous les détails et les composantes de ce mécanisme. Alors que de nombreuses études soutiennent le rôle de Rab9 et TIP47 dans un transport rétrograde à partir des endosomes tardifs, d'autres contestent cette conclusion. En effet, TIP47 est le plus couramment décrite comme protéine des gouttelettes lipidiques (LD, lipid droplets) (Ducharme and Bickel, 2008; Wolins et al., 2001) et serait recrutée à ce niveau et non aux endosomes. D'ailleurs, l'interaction entre Rab9 et TIP47 ne semble pas reproductible et l'affinité, rapportée comme étant deux fois supérieure (Ganley et al., 2004) pour la forme active (96 nM) par rapport à l'inactive (159 nM), soulève un doute sérieux relatif au rôle suggéré d'effecteur (Bulankina et al., 2009). La déplétion de TIP47 n'affecte pas la distribution de MPR, ni le triage des enzymes lysosomiales, mais bloque plutôt la maturation des LD et l'incorporation de triacylglycérol (Wolins et al., 2001). De surcroît, par sa structure, TIP47 fait partie de la famille des protéines liant les LD PAT (Perilipine, Adipophiline, TIP47) (Brasaemle, 2007) et montre beaucoup d'homologie avec l'ApoE au niveau de la région de liaison des lipides (Bulankina et al., 2009). Il reste à savoir si TIP47 peut avoir des rôles différents dans le triage et le métabolisme des lipides. 2.3 Rab comme principaux régulateurs du transport intracellulaire 2.3.1 Cycle d'activation des Rab Avec plus de 60 membres chez l'humain, les Rab constituent la plus grande sousfamille de petites protéines G Ras (Rojas et al., 2012). Elles sont conservées de la levure aux mammifères (Pereira-Leal and Seabra, 2001) et sont d'expression ubiquitaire, tissu spécifique ou régulée au cours du développement (Leung et al., 2006). Elles servent principalement de régulateurs spatio-temporels lors de multiples processus liés au transport 42 Figure 5 - Les Rab comme principaux régulateurs du transport intracellulaire. (A) Illustration des mécanismes du transport intracellulaire régulés par les Rab, soit le bourgeonnement des vésicules, le désassemblage du manteau de clathrine, la motilité, l'arrimage et la fusion de l'endosome. (B) Cycle d'activation des Rab. Les Rab nouvellement synthétisées lient le GDP et sont escortées par une REP pour subir l'ajout de groupements geranylgeranyl par la GGT. Elles peuvent alors s'ancrer aux membranes et être activées par une GEF pour ensuite agir sur des effecteurs avant qu'une GAP ne provoque leur retour à l'état inactif. Cet état est maintenu par une GDI jusqu'à ce qu'une GDF dicte son recrutement à une membrane précise. GAP (protéine d’accélération de la GTPase), GDF (facteur de déplacement de GDI), GDI (inhibiteur de la séparation des nucléotides de guanine), GEF (facteur d’échange des nucléotides de guanine), GGT (Geranylgeranyl transférase) et REP (protéine escorte de Rab). Figure modifiée de RAB11mediated trafficking in host–pathogen interactions. Guichard et al. Nat Rev Microbiol. 2014 Sep;12(9):624-34. (C) Échange de Rab5 pour Rab7 lors de la maturation de l'endosome précoce en endosome tardif. Rab5-GDP est activé par la GEF Rabex-5 ce qui permet à son effecteur Rabaptin-5 de lier Rabex-5 pour promouvoir l'activation de Rab5. Le complexe Mon1-Ccz1 (GEF de Rab7) lie alors Rab5 et Rabex-5 entraînant leur dissociation et le recrutement de Rab7. Figure modifiée de Endosome maturation, Huotari and Helenius, The EMBO Journal (2011) 30, 3481–3500 avec la permission de John Wiley and Sons. 43 intracellulaire (figure 5A). L'efficacité des Rab repose sur leur cycle d'activation-inhibition (figure 5B) (Guichard et al., 2014) qui leur permet d'agir comme interrupteurs moléculaires auprès de divers effecteurs. Comme toutes les GTPases, les Rab possèdent deux conformations structurelles en fonction de leur liaison au GTP ou au GDP. La première est généralement reconnue par les effecteurs comme étant la forme active alors que la seconde correspond à la forme inactive. Les facteurs d’échange des nucléotides de guanine (GEF) catalysent l’activation des Rab en permettant l’échange du GDP pour du GTP. De leur côté, les protéines d’accélération de la GTPase (GAP) catalysent l'hydrolyse du GTP en GDP de manière à inactiver les Rab (Pan et al., 2006). Cet état est maintenu par des inhibiteurs de la séparation des nucléotides de guanine (GDI) qui préviennent la relâche du GDP (Matsui et al., 1990). La mécanistique des Rab est complexe en raison du nombre croissant d'acteurs impliqués qui agissent en multiples combinaisons (Stenmark, 2009). Entre autres, certains effecteurs sont soumis à plusieurs Rab distinctes. Aussi, l'effet d'une Rab peut être amplifié par le recrutement d'un complexe d'effecteurs contenant une GEF pour la même Rab. Il peut y avoir une activation couplée, si cette GEF active plutôt une autre Rab. Finalement, l'activation couplée peut mener à une conversion de Rab s'il y a recrutement d'une GAP qui inactive la Rab d'origine. 2.3.2 Localisation des Rab La localisation intracellulaire distincte des différents Rab au niveau de compartiments intracellulaires et de microdomaines contribue à la précision de leurs actions (Chavrier et al., 1990). Les Rab nouvellement synthétisées doivent subir l'ajout d'un ancrage lipidique pour permettre leur attachement réversible aux membranes (figure 5B). Pour ce faire, elles sont reconnues par des protéines escortes de Rab (REP) qui les présentent à une Geranylgeranyl transférase (GGT). Cette dernière ajoute des groupements du même nom à un ou deux résidus cystéine de la région C-terminale des Rab suivant les motifs CC ou CXC (Seabra et al., 1992). Les REP escortent alors les Rab jusqu'à une membrane tout en agissant comme GDI (Alory and Balch, 2001). Par la suite, le recrutement d'une Rab à une membrane précise sera contrôlé par un facteur de déplacement de GDI (GDF) associé à cette dernière (figure 5B) (Sivars et al., 2003). Dans le système endolysosomial, 44 l'attachement spécifique des Rab en fait d'importants régulateurs spatio-temporels conférant l'identité fonctionnelle des endosomes. Par exemple, Rab5 définit l'endosome précoce par son action. Rab5 est recrutée à la membrane de l'endosome (figure 5C) (Huotari and Helenius, 2011; Ullrich et al., 1994) à la suite de son activation par la GEF Rabex-5 (Yamashiro and Maxfield, 1987). Ce phénomène est amplifié par l'action de l'effecteur Rabaptin-5 qui se lie à Rabex-5 (Horiuchi et al., 1997) pour favoriser l'activation de Rab5 (Lippe et al., 2001). Rab5 recrute également son effecteur la Phosphatidylinositol-3-kinases (PI3K) de classe 3 (Christoforidis et al., 1999b). Cette enzyme convertit le phosphatidylinositol (PI) membranaire en phosphatidylinositol-3-phosphate (PI(3)P) (Schu et al., 1993) ce qui permet le recrutement de diverses protéines. L'endosome précoce va graduellement maturer pour former un endosome tardif. Cette évolution est marquée par l'échange de Rab5 pour Rab7 (figure 5C). Pour ce faire, les protéines Mon1 (Sand1 chez les mammifères) et Ccz1 (GEF de Rab7) lient Rab5, le PI3P et Rabex-5 entraînant la dissociation de ce dernier et, par le fait même, de Rab5 (Poteryaev et al., 2010). La GAP responsable de mettre fin à l'activité de Rab5 est encore inconnue (Huotari and Helenius, 2011). Le complexe Mon1-Ccz1 permet également le recrutement de Rab7 (Poteryaev et al., 2010) qui assure une production soutenue de PI(3)P en liant la PI3K (Stein et al., 2003). 2.3.3 Fonctions des Rab dans le transport entre le Golgi et les endosomes Le recrutement spécifique des Rab confère l'identité fonctionnelle des membranes. Elles agissent alors comme régulateurs dans plusieurs processus principalement liés au transport vésiculaire, dont la formation, la motilité, l'arrimage et la fusion des vésicules ainsi que le désassemblage de complexes de protéines (figure 5A). Bourgeonnement et désassemblage du manteau protéique La fonction des Rab au niveau du bourgeonnement de vésicules est de loin la plus nébuleuse (figure 5A). Par exemple, Rab5 est une composante des vésicules de clathrine qui serait essentielle pour la formation des puits. Pour ce faire, Rab5 permettrait la séquestration de récepteurs par les protéines adaptatrices (Bucci et al., 1992). L'internalisation de TfR est d'ailleurs directement liée à l'activité de Rab5 (Bucci et al., 45 1992). D'autres Rab auraient un rôle semblable auprès d'autres compartiments dont Rab1 qui agirait au niveau de la formation des vésicules COP-I à partir de l'appareil de Golgi (Alvarez et al., 2003; Peter et al., 1994) et Rab9 qui, comme mentionné antérieurement, agit au niveau des endosomes tardifs. Par la suite, les Rab agissent comme régulateurs du dépouillement de vésicules (figure 5A) de par l'exemple de Rab5. Suivant l'action de sa GEF GTPase activating protein and VPS9 domains 1 (GAPvD1, aussi connue sous le nom de RME-6, GAPex5 (Lodhi et al., 2007) ou RAP6 (Hunker et al., 2006)), Rab5 est activée ce qui lui permet de recruter des effecteurs modifiant les éléments cruciaux à la stabilisation de la protéine adaptatrice AP-2 au niveau des vésicules d'endocytose (Semerdjieva et al., 2008). Dans un premier temps, Rab5 stimule la conversion du phosphatidylinositol-4,5bisphosphate (PI(4,5)P2) en PI(3)P via le recrutement de phosphatases et de kinases et, dans un second temps, Rab5 provoque la déphosphorylation d'AP-2 via le déplacement de la Adaptor-associated protein kinase 1 (AAK1) (Semerdjieva et al., 2008). Motilité vésiculaire Suite à leur dénudement, les vésicules peuvent voyager le long de filaments d'actine et de microtubules afin d'atteindre leurs destinations. Les Rab participent également à ce phénomène de motilité (figure 5A) en recrutant des effecteurs qui agissent comme adaptateurs de protéines motrices. Par exemple, Rab11, par l'intermédiaire de RAB11 family-interacting protein 2 (RAB11FIP2) (Hales et al., 2002), lie la Myosine Vb qui assure des déplacements courts et lents sur l'actine. Quand à eux, les déplacements longs et rapides se produisent le long des microtubules et sont assurés par les protéines motrices Kinésine, qui se dirige vers la périphérie, ou Dynéine, qui se déplace vers le centre d'organisation des microtubules. L'attachement de l'endosome tardif à la Dynéine constitue un exemple particulièrement pertinent dans le cadre de la présente étude. L'assemblage du moteur protéique est initié par l'activation de Rab7 qui lie ses effecteurs Rab7-interacting lysosomal protein (RILP) et Oxysterol‐binding protein (ORP1L) de manière à former un dimère d'hétérotrimères (Johansson et al., 2007). RILP recrute alors le complexe DynéineDynactine qui est transféré au récepteur Spectrin betaIII par ORP1L activant le moteur protéique rétrograde (Johansson et al., 2007). Un mécanisme semblable impliquant Rab7, son effecteur FYVE and coiled-coil domain–containing protein 1 (FYCO1) et Microtubule46 associated protein 1A/1B-light chain 3 (LC3) permet le transport des vésicules autophagiques par la Kinésine (Bento et al., 2013; Pankiv et al., 2010). Arrimage Une fois arrivée à proximité de leur destination, les vésicules doivent s'attacher spécifiquement à la membrane du compartiment cible. Encore une fois, la localisation intracellulaire distincte des Rab au niveau des compartiments intracellulaires en fait un régulateur spatio-temporel de choix. Les différentes Rab facilitent le recrutement de facteurs d'arrimage (figure 5A) précis qui servent en quelque sorte d'ancre. Il existe deux groupes de facteurs d'arrimage (tethering factors), soit les protéines constituées de longues superhélices (long coiled-coil proteins) et les complexes à multiples unités. Le premier groupe englobe des protéines dimériques hydrophiles constituées de deux têtes globulaires suivies de superhélices permettant un premier contact de faible affinité entre deux membranes à grande distance (environ 200 nm) (Brocker et al., 2010). Les golgines qui arriment les vésicules au Golgi, comme Golgi matrix protein 130 (GM130), en font partie de même que Early endosome antigen 1 (EEA1). Ce dernier possède deux sites d'interaction opposés pour Rab5 ce qui permet de lier ensemble deux endosomes précoces arborant la GTPase (Christoforidis et al., 1999a). De son côté, le groupe des complexes à multiples unités compte huit membres : Transport protein particle (TRAPP) I et II, Dsl1, Conserved oligomeric Golgi (COG), Exocyst, Class C core vacuole/endosome tethering (CORVET), Homotypic fusion and vacuole protein sorting (HOPS) et Golgi-associated retrograde protein (GARP). Ils sont composés de trois à dix unités et ils agissent à différents endroits dans la cellule vraisemblablement pour renforcer l'attachement des vésicules lorsqu'elles parviennent à moins de 30 nm de distance. CORVET et HOPS sont des hexamères ayant quatre unités en communs et deux unités spécifiques pour la liaison de Rab distinctes (Plemel et al., 2011). CORVET est un complexe effecteur de Rab5 qui permet l'arrimage entre endosomes précoces alors que HOPS agit plutôt sous l'influence de Rab7, après la maturation de l'endosome, pour l'attachement impliquant les endosomes tardifs et les lysosomes. 47 Fusion Les Rab sont également d'importants régulateurs de la fusion des vésicules avec leur compartiment de destination (figure 5A) à cause de leur effet sur les Soluble Néthylmaleimide-sensitive-factor (NSF) attachment protein (SNAP) receptor (SNARE). Les SNARE sont des protéines membranaires majoritairement intégrales agissant comme principales responsables de la fusion spécifique des membranes (McNew et al., 2000). Chez l'humain, il existe 36 membres, nommés Vesicule-associated membrane protein (VAMP), Syntaxine ou SNAP, qui sont différentiellement répartis sur les diverses membranes intracellulaires (Jahn and Scheller, 2006). Elles sont caractérisées par la présence d'une séquence homologue d'environ 60 acides aminés formant une superhélice (Weimbs et al., 1997). Ce motif SNARE, permet l'assemblage complémentaire entre quatre SNARE pour former un faisceau parallèle entremêlé nommé trans-SNARE (Sutton et al., 1998). Le centre (zéro) de ce complexe est généralement composé d'une arginine (R) et de trois glutamines (Q) entourés d'une fermeture en leucines étanche protégeant l'interaction ionique (Antonin et al., 2002; Sutton et al., 1998). Il existe donc deux catégories de SNARE, soit les R-SNARE et les Q-SNARE qui contribuent à la formation d'un complexe spécifique via leur arginine (R) et leur glutamine (Q), respectivement (Fasshauer et al., 1998). De façon générale, le R-SNARE est fourni par la vésicule (v-SNARE) alors que les Q-SNARE proviennent du compartiment accepteur (target(t)-SNARE). Une fois assemblées, les hélices du trans-SNARE s'enroulent et se resserrent de manière à rapprocher deux membranes opposées. Les SNARE subissent alors un stress associé à leur flexion qui les pousse à s'éloigner du lieu de fusion (Risselada et al., 2011). Il en résulte une réduction de la répulsion entre les membranes et la génération d'une force nécessaire à leur fusion (Risselada et al., 2011). Une fois sa fonction accomplie, le complexe SNARE demeure sur la membrane fusionnée et est désigné cis-SNARE. Il est désassemblé par l'ATPase NSF et son cofacteur SNAP pour permettre sa réutilisation (Marz et al., 2003; Sollner et al., 1993). Les Rab agissent comme régulateurs indirects de la fusion puisque plusieurs de leurs effecteurs sont des partenaires d'interaction des SNARE. Certains sont des facteurs d'arrimage qui facilitent la fusion entre les membranes en régulant l'assemblage des SNARE. Entre autres, EEA1, l'effecteur de Rab5, interagit avec le SNARE Syntaxin-13 pour permettre la fusion entre endosomes précoces (McBride et al., 1999). Le rôle de Rab5 48 est essentiel pour la fusion homotypique comme démontré par l'usage courant du mutant constitutivement actif Rab5QL pour stimuler la fusion et ainsi faciliter l'observation d'endosomes géants. Au niveau de l'endosome tardif, l'association d'un effecteur de Rab7 a également été répertoriée en l'exemple de HOPS qui stimule la formation de complexes impliquant VAMP3 et VAMP7 (Collins et al., 2005b). 3 Maladies associées aux lysosomes 3.1 Lysosomes Le lysosome a été découvert en 1955 par Christian de Duve (De Duve et al., 1955), prix Nobel de physiologie ou de médecine de 1974. Il a alors été décrit comme compartiment cellulaire responsable de la digestion intracellulaire et du recyclage de macromolécules. Il s'agit d'un organite juxtanucléaire et hétérogène morphologiquement qui se définit aujourd'hui comme étant acide et riche en hydrolases, mais exempt de MPR (Kornfeld and Mellman, 1989). La biogenèse du lysosome repose vraisemblablement sur la maturation de l'endosome visant l'exclusion de certaines molécules et l'ajout graduel de composantes caractéristiques (Mullins and Bonifacino, 2001). Ces dernières lui permettent de remplir sa fonction de dégradation et comprennent plus de 50 hydrolases acides, telles que des phosphatases, des nucléases, des glycosidases, des protéases, des sulphatases et des lipases, ainsi que 120 protéines membranaires (Braulke and Bonifacino, 2009), dont au moins 20 transporteurs (Sagne and Gasnier, 2008). La plupart des protéines membranaires décorent la face luminale des lysosomes et sont hautement glycosylées (Fukuda, 1990) formant une couche nommée glycocalyx qui protège l'organite de son pH acide. Les plus abondantes sont LAMP1/2 et LIMP-1/2, les LAMP représentant 50% du protéome membranaire. De son côté, l'acidité lysosomiale (pH 5) est principalement maintenue par la pompe à proton de type vacuolaire (V-ATPase) qui hydrolyse de l'ATP pour permettre le transport ionique au travers la membrane (Grabe and Oster, 2001). Cette activité est cruciale pour la fonction de l'organite puisque de nombreuses hydrolases opèrent préférentiellement en milieu acide. Entre autres, la protéase aspartique Cathepsine D subit des changements de conformation dépendants d'une forte protonation qui libère son site 49 actif pour l'activer (Lee et al., 1998). De plus, cette enzyme doit préalablement être clivée le long de la voie endolysosomiale par d'autres cathepsines dont l'efficacité est optimale à des pH qui leurs sont propres (Laurent-Matha et al., 2006). Le matériel à dégrader aux lysosomes provient de l'internalisation à la membrane plasmique, comme c'est le cas pour certains récepteurs et leurs ligands, de la voie de synthèse ou de la séquestration par autophagie. Certaines protéines membranaires, comme LAMP2A, permettraient également la translocation de substrats cytosoliques précis vers la lumière de l'organite via des chaperonnes (Bandyopadhyay and Cuervo, 2008; Bandyopadhyay et al., 2008), un processus référé sous le nom d'autophagie médiée par des chaperonnes. Les produits générés par la dégradation sont ensuite transportés vers le cytosol par des transporteurs spécifiques. Parmi ces transporteurs, on compte Cystinosine et Sialine, des exportateurs de cystéines et d'acides sialiques, respectivement. Transient receptor potential cation channel, mucolipin subfamily, member 1 (TRPML1) transporte principalement le fer et le calcium (Cheng et al., 2010) alors que, quoique non démontré, Niemann-Pick C1 protein (NPC1) permetterait l'efflux de cholestérol. Par l'intermédiaires de nombreux canaux ioniques, comme l'échangeur chlore/proton ClC-7 (Chloride channel 7), les lysosomes contrôlent également la concentration ionique et le pH du cytosol. Les lysosomes possèdent aussi une lysosome nutrient sensing (LYNUS) machinery qui détecte les variations en acides aminés lysosomiaux via la V-ATPase (Zoncu et al., 2011) et le canal sodique ATP-sensible (lysoNaATP) (Cang et al., 2013). Une forte concentration permet au Mammalian target of rapamycin complex 1 (mTORC1) d'être activé et de s'associer à LYNUS à la surface des lysosomes (Sancak et al., 2010; Zoncu et al., 2011). Alors, mTORC1 séquestre le facteur de transcription EB (TFEB) (Cang et al., 2013) et induit la biosynthèse (Laplante and Sabatini, 2012). Au contraire, en condition de jeûne, mTORC1 est inactivé et relâché de LYNUS libérant TFEB qui transloque au noyau pour favoriser l'autophagie ainsi que le catabolisme de molécules. Ainsi, les lysosomes peuvent détecter la présence de nutriments et déclencher une cascade de signalisation de réponse à leur absence. Par le biais de sa capacité à dégrader, le lysosome est impliqué dans plusieurs autres fonctions qui en font un compartiment central et dynamique. Les protéases qu'il contient sont nécessaires dans l'immunité puisqu'elles traitent des antigènes pour la présentation par 50 le Complexe majeur d'histocompatibilité (CMH) de classe II. Les enzymes peuvent aussi être libérées dans le milieu extracellulaire par exocytose des lysosomes pour inactiver des organismes pathogènes ou participer à la dégradation de composantes de la matrice comme dans le cas d'ostéoclastes. À la suite de la perméabilisation de l'organite, la relâche de cathepsines peut aussi se faire dans le cytosol suivant un processus de mort cellulaire par le lysosome, une forme d'apoptose peu connue (Saftig and Klumperman, 2009). 3.2 Maladies de surcharge lysosomiale Les maladies de surcharge lysosomiale (LSD, Lysosomal storage diseases) regroupent une cinquantaine de troubles héréditaires rares de type autosomal récessif, qui ont en commun un dysfonctionnement des lysosomes. Les LSD affectent principalement les enfants et présente une incidence combinée d'environ 1:5000 (Platt et al., 2012) et une prévalence individuelle allant de 1:50000 à 1:4000000 (Meikle et al., 1999; Meikle et al., 2004). Au point de vue cellulaire, les mutations génétiques responsables des LSD altèrent des protéines membranaires intégrales des lysosomes, des protéines de transport membranaire, des protéines impliquées dans les modifications post-traductionnelles ou, plus fréquemment, l'activité d'enzymes particulières du lysosome (Schultz et al., 2011). Chaque défaut entraîne un ralentissement ou une interruption de voies métaboliques impliquées dans la dégradation de molécules. Il s'en suit une accumulation, aux lysosomes, du substrat généralement directement en lien avec la fonction de l'enzyme fautive. La classification conviviale des LSD repose d'ailleurs sur la nature de ce matériel laissant place à des appellations telles que les lipidoses, les glycogénoses, les céroïdes-lipofuscinoses, les pycnodysostoses, les mucopolysaccharidoses, les oligosaccharidoses et les glycoprotéinoses. Il y a alors formation de lésions histologiques qui constituent la principale caractéristique des maladies de surcharge lysosomiale (Heard et al., 2010). Les inclusions cellulaires observées sont hautement hétérogènes au sein d'une même cellule et selon le type cellulaire. Elles se présentent sous forme de larges structures d'aspect vacuolaire contenant divers types de matériels, dont des débris, des produits clairs, des vésicules, des fragments membranaires, des agrégats denses, des structures multilamellaires et des corps zébrés (Heard et al., 2010). La toxicité, plutôt que la quantité, du matériel est 51 davantage critique pour les fonctions biologiques du lysosome (Heard et al., 2010). De nombreux processus associés à l'organite peuvent être affectés comme la régulation du pH, l'endocytose, l'exocytose, la maturation des vésicules, l'homéostasie du calcium et l'autophagie (Ballabio and Gieselmann, 2009; Bellettato and Scarpa, 2010; Vitner et al., 2010). À ce moment, la pathogenèse ne repose pas uniquement sur un problème de contenu, mais aussi sur un problème de contenant (Heard et al., 2010). En effet, le dysfonctionnement entraîne l'accumulation de produits secondaires qui participent à l'hétérogénéité des inclusions cellulaires (Heard et al., 2010). Ces phénomènes entraînent à la longue la mort cellulaire. Quoique l'ensemble des tissus soit affecté, dans la plupart des cas, le système nerveux central demeure le plus sévèrement touché (Mitchison et al., 2004) ce qui se traduit par une neurodégénération évolutive et polyhandicapante. 3.3 Céroïdes-lipofuscinoses neuronales Parmi les LSD figurent les céroïdes-lipofuscinoses neuronales (NCL) dont le premier cas fut décrit par Dr. Christian Stengel en 1826 (Brean, 2004; Stengel, 1826a). Il s'agissait de quatre enfants norvégiens, issus de parents d'apparence saine, dont les symptômes sont apparus vers l'âge de six ans (Brean, 2004; Stengel, 1826a). Au cours du début du siècle suivant, de nombreux cas semblables ont été rapportés par Batten (Batten, 1903; Batten, 1914) et Spielmeyer (Spielmeyer, 1905a; Spielmeyer, 1905b). En l'honneur du pionnier, les NCL sont d'ailleurs globalement désignées comme Batten disease. Suivant leurs ressemblances cliniques et la présence d'accumulations cellulaires, elles ont originalement été classées comme 'amaurotic family idiocy'. Dès 1969, l'appellation Neuronal ceroid lipofuscinose a été proposée par Zeman et Dyken pour souligner le caractère distinct des lipopigments accumulés (Zeman and Dyken, 1969). En effet, les NCL se distinguent des autres LSD par l'accumulation anormalement forte d'un matériel périnucléaire et surtout lysosomial, nommé lipofusine. Décrite pour la première fois en 1843 par Adolph Hannover (Hannover, 1843), la lipofuscine, ou pigment d'âge, s'est rapidement révélée être une caractéristique naturelle du vieillissement, inversement corrélée à l’espérance de vie d'une cellule (Koneff, 1886). Le terme lipofuscine, proposé par Hueck (Hueck, 1912), provient de l'union du mot grec lipo, soit gras et latin fuscus signifiant brun. De son côté, le mot ceroïde 52 est issu du latin cera et du grecque eidos signifiant 'de type cireux'. Ces termes décrivent donc la lipofuscine qui est un produit brun jaunâtre, présentant une forte densité électronique et une autofluorescence à spectre large (Brunk and Terman, 2002a; Porta, 2002; Terman and Brunk, 1998). Il est en effet possible d'observer la lipofucine par excitation dans l'ultraviolet (330–380 nm), le bleu (450–490 nm) et le vert (510–560 nm) (Porta, 2002). Cette importante propriété proviendrait de réactions entre les carbonyles résultant de la peroxydation de lipides et les composés aminés qui produisent principalement des bases de Schiff (Terman and Brunk, 1998). En fait, les lipofucines sont formées de peptides réticulés (cross-link) et de composés variés fortement oxydés qui résistent à la dégradation. Elles comprennent 30-70% de protéines, 20-50% de lipides (triglycérides, acides gras libres, cholestérols et phospholipides) (Porta, 1991), 4-7% de carbohydrates et 2% de métaux (fer, cuivre, aluminium, zinc, calcium et manganèse) (Jolly et al., 1995; Terman and Brunk, 1998). La respiration mitochondriale est à l'origine de la production de lipofuscine. Elle entraîne la production de radicaux libres qui causent des modifications oxydatives au niveau de macromolécules mitochondriales. Les mitochondries entrent ensuite dans les lysosomes par autophagie et y déversent leur contenu résistant à la dégradation. Le faible pH et la présence de substances réductrices favorisent également la production de radicaux directement au niveau du lysosome causant de plus amples dommages. Par la suite, les lysosomes engorgés se déchirent et libèrent la lipofuscine dans le cytosol. Ce modèle (figure 6A) (Brunk and Terman, 2002b; Jung et al., 2007) explique pourquoi la sous-unité C de l'ATP synthase mitochondriale (SCMAS) a été la première protéine identifiée comme matériel emmagasiné dans les NCL (Palmer et al., 1989). Les protéines lysosomiales Sphingolipid activator proteins (saposines ou SAP) A et D sont également bien représentées (Tyynela et al., 1993). Les saposines sont associées à l'apparition d'ultrastructures granulaires (GROD, granular osmiophilic deposits) (figure 6B) (Haltia, 2003) alors que la SCMAS forme des ultrastructures variées pouvant coexister et dites curvilignes, rectilignes ou en empreintes digitales. Les GROD sont des corps membranaires ronds finement granulaires mesurant 0,5 μm de diamètre et formant des agrégats atteignant jusqu'à 5 μm (Mole et al., 2011). Le profil curviligne (CL) (figure 6B) (Haltia, 2003) est 53 Figure 6 - Formation et ultrastructures de la lipofuscine. (A) Modèle de la formation de la lipofuscine selom l'hypothèse proposée par Brunk et Terman en 2002. L'oxydation entraîne la réticulation des protéines et le dysfonctionnement des mitochondries qui s'en trouve élargies. Ces protéines et ces mitochondries sont acheminées aux lysosomes causant leur engorgement et, éventuellement, leur déchirement et le déversement de leur contenu dans le cytosol. Figure tirée de Lipofuscin: formation, distribution, and metabolic consequences, Jung et al., Annals of the New York Academy of Sciences (2007) 1119, 97-111 avec la permission de John Wiley and Sons. (B) Les ultrastructures associées à la lipofuscine se présentent sous quatres profils différents en microscopie électronique, soit granulaire (GROD) ×10000, curviligne (CL) ×20000, en empreintes digitales (FP) ×30000 et rectiligne (RL) ×15000. Figure tirée de The Neuronal Ceroid-Lipofuscinoses, Haltia, Journal of Neuropathology & Experimental Neurology (2003) 62(1), 1-13 avec la permission de Oxford University Press. 54 composé de citernes lamellaires minces et courtes (1,9-2,4 nm) uniformément courbées alternant du noir au blanc (Mole et al., 2011). Il ne doit pas être confondu avec les ultrastructures rectilignes (RL) (figure 6B) (Haltia, 2003), plus grosses et moins régulières, qui présentent de courtes (2,8-3,8 nm) citernes oligolamellaires légèrement ondulées dont la ligne centrale est proéminente (Mole et al., 2011). Finalement, le profil en empreintes digitales (FP, fingerprint) (figure 6B) (Haltia, 2003) présente des lignes parallèles en paires de 7,6-9,6 nm de large parfois fragmentées ou très denses (Mole et al., 2011). 3.3.1 Causes génétiques et protéines associées Les NCL sont définies comme un groupe de « maladies génétiques dégénératives progressives du cerveau et, dans la plupart des cas, de la rétine, en association avec une accumulation intracellulaire d'un matériel qui est morphologiquement caractérisé comme ou similaire à la céroïde-lipofuscine » (Mole et al., 2011). Leur classification originale était principalement basée sur l'âge à l'apparition des symptômes, la progression de la maladie et les caractéristiques des ultrastructures intracellulaires. On distinguerait alors les formes infantile, infantile-tardive, juvénile et adulte qui ont rapidement présenté nombre de cas atypiques et de variantes. En effet, les NCL se prêtent peu à ce système. Elles sont dites à la fois génétiquement hétérogènes, puisque plusieurs mutations dans différents gènes peuvent causer un phénotype clinique similaire, et hétérogènes au niveau allélique, puisque diverses mutations dans le même gène peuvent engendrer des phénotypes cliniques différents (Mole et al., 2011). Il existe également une hétérogénéité phénotypique qui se traduit par le fait qu'une même mutation peut engendrer des symptômes cliniques et des progressions différentes (Mole et al., 2011). Heureusement, l'évolution de la science a permis une nouvelle classification incontestable et simple qui repose sur les déterminants génétiques. Le premier gène impliqué dans la maladie a été identifié comme étant Palmitoyl-protein thioesterase 1 (PPT1), par Vesa en 1995 (Vesa et al., 1995). Depuis 14 gènes, renommés de façon uniforme Ceroid lipofuscinose neuronal 1-14 (CLN), figurent au tableau des NCL (tableau 1) et des centaines de mutations sont répertoriées pour chacun. De plus amples joueurs seront sans doute découverts puisqu'il existe encore des patients de phénotype orphelin, c'est-à-dire que la cause génétique de leur phénotype est inconnue. De grands 55 efforts ont mené à l'identification de la majorité des protéines associées aux gènes CLN, mais la fonction de chacun et leur rôle dans la pathogenèse restent encore obscurs. On retrouve entre autres des enzymes directement impliquées dans la dégradation au lysosome (PPT1 (ou CLN1), TPP1 (CLN2), CTSD (CLN10), CTSF (CLN13)) et des protéines transmembranaires (CLN3, CLN6, Major facilitator superfamily domain-containing protein 8 (MFSD8) ou CLN7), Probable cation-transporting ATPase 13A2 (ATP13A2) ou CLN12)). La plupart de ces protéines se retrouvent au niveau de la voie endolysosomiale ou du réticulum endoplasmique. Dans le contexte de la présente étude, l'accent sera mis sur les protéines CLN3 et CLN5. Tableau 1 - Pathologie des NCL. Les gènes responsables des NCL, les protéines associées et leurs localisations, les protéines accumulées dans la maladie, les principales ultrastructures observées et le nombre de mutations répertoriées. CL (curviligne), FP (empreinte digitale), RL (rectiligne), GROD (granular osmiophilic deposits), SAP (Sphingolipid activator proteins), SCMAS (Subunit C of mitochondrial ATP synthase). Figure inspirée de Human NCL Neuropathology, Radke et al., et mis à jour à partir de NCL Resource à l'adresse http://www.ucl.ac.uk/ncl/. 3.3.2 CLN3 Le gène CLN3 a été isolé en 1995 par The International Batten Disease Consortium(The International Batten Disease Consortium, 1995). Il code la protéine 56 CLN3, aussi connue sous le nom de Battenine (Btn) en l'honneur de Dr. Frederick Batten, un pionnier dans la découverte des NCL. Il s'agit d'une protéine hautement hydrophobe de 438 acides aminés comprenant six domaines transmembranaires, des extrémités cytoplasmiques (Ezaki et al., 2003; Janes et al., 1996; Kaczmarski et al., 1999; Kyttala et al., 2004) et une hélice amphipathique luminale entre le cinquième et le sixième segment transmembranaire (figure 7) (Nugent et al., 2008). Elle est très conservée et présente, chez l'humain, au moins sept isoformes produits par épissage alternatif. Son orthologue (Btn1) chez la levure Saccharomyces cerevisiae lui est 30% identique et 59% similaire (Pearce and Sherman, 1997). CLN3 compte de nombreux sites de potentielles modifications posttraductionnelles, dont trois pour des N-glycosylation de type hautement mannosylé (high mannose) (Jarvela et al., 1998; Mao et al., 2003; Storch et al., 2007), six de phosphorylation (Nugent et al., 2008), un de N-myristoylation en N-terminal (Kaczmarski et al., 1999) et un motif CaaX de prénylation/farnésylation en C-terminal (Pullarkat and Morris, 1997). CLN3 se retrouve en faible quantité, mais de façon ubiquitaire, et étonnament son expression n'est pas particulièrement marquée dans le système nerveux central (Mole et al., 2011). La présence de deux motifs de triage lysosomial assure sa localisation, indépendante de MPR (Kyttala et al., 2004), dans l'organite ainsi que son passage dans le RE, le Golgi et la voie endolysosomiale (Haskell et al., 2000; Jarvela et al., 1998; Kida et al., 1999; Kyttala et al., 2004). Il s'agit en fait d'un motif dileucine dans une boucle cytosolique (Kyttala et al., 2004; Storch et al., 2004) et d'un motif peu courant en Cterminal, composé d'une méthionine et d'une glycine séparées par neuf résidus (Kyttala et al., 2004). CLN3 se retrouve également à la membrane plasmique dans une proportion d'environ 10% (Mao et al., 2003; Persaud-Sawin et al., 2004; Storch et al., 2007) et aux synapses (synaptosomes) des neurones primaires (Jarvela et al., 1999; Luiro et al., 2001). Les NCL de type CLN3 sont les plus communes avec une prévalence de 0,6 cas par 100 000 naissances au Canada (MacLeod et al., 1976). Du point de vue histologique, la maladie présente un profil en empreintes digitales, parfois accompagné de profils curviligne et rectiligne (Elleder, 1999), avec une accumulation principale de la SCMAS (Elleder et al., 1997). La grande majorité des patients, soit 85%, sont homo ou hétérozygotes pour une même mutation qui consiste en une déplétion de 1 kb provoquant la perte des exons 7 et 8 57 (Cln3Δex7/8) (Munroe et al., 1997). Il en résulte deux transcripts qui, au niveau protéique, se manifestent par la production de deux mutants. Le produit majeur provient d'un changement de cadre de lecture au 153e acide aminé et possède seulement 28 nouveaux résidus en aval de la mutation. Le produit secondaire est tronqué des résidus 154–263 car il est issu de la perte d'un exon, de la création d'un nouveau site d'épissage et du retour au bon cadre de lecture (Kitzmuller et al., 2008). Malgré le phénotype engendré et une forte rétention au réticulum endoplasmique, ces peptides seraient suffisamment biologiquement actifs pour maintenir la morphologie du lysosome (Kitzmuller et al., 2008). D'ailleurs, l'incapacité de CLN3 à atteindre les lysosomes semble directement corrélée avec la sévérité des phénotypes, suggérant un rôle important dans cet organite (Jarvela et al., 1999). Figure 7 - Schéma de la structure de CLN3. CLN3 est une protéine à six domaines transmembranaires comprenant des extrémités cytoplasmiques ainsi qu'une hélice amphipathique entre les cinquième et sixième segments transmembranaires. Figure créée avec Microsoft Office PowerPoint 2007 et inspirée de The transmembrane topology of Batten disease protein CLN3 determined by consensus computational prediction constrained by experimental data. Nugent et al. FEBS Lett 582:1019–1024. À ce jour, aucun homologue de CLN3 n'a été identifié et sa fonction demeure encore inconnue. Il a été proposé que la protéine soit impliquée dans de nombreux processus cellulaires, dont la balance ionique, l'acidification des lysosomes, les modifications de protéolipides, l'apoptose et la signalisation. Brièvement, chez la drosophile, CLN3 interagit avec Notch et Jun N-terminal kinase (JNK) et modulerait leurs voies de signalisation 58 (Tuxworth et al., 2009). L'absence d'activité palmitoyldésaturase chez les cellules déplétées en CLN3 suggère que cette réaction enzymatique pourrait lui être associée (Narayan et al., 2006; Narayan et al., 2008). Une augmentation de l'apoptose a aussi été observée dans ces conditions (Persaud-Sawin and Boustany, 2005; Puranam et al., 1999). CLN3 interagit d'ailleurs avec Calséniline, une protéine à mains EF qui contribue à la mort induite par le calcium (Chang et al., 2007), et protègerait la cellule par la régulation des céramides intracellulaires qui agissent comme second messager dans la cascade apoptotique (Puranam et al., 1999). CLN3 possède un domaine conservé de liaison de galactosylceramide (GalCer lipid raft-binding domain) et participerait à son transport du Golgi vers les radeaux lipidiques influençant la composition membranaire (Rusyn et al., 2008). Dans un autre ordre d'idées, l'absence de CLN3 ou sa mutation entraîne une diminution du pH lysosomial, c'est-à-dire une acidification (Golabek et al., 2000; Holopainen et al., 2001; Padilla-Lopez and Pearce, 2006; Pearce et al., 1999a; Pearce et al., 1999b). CLN3 participe donc au maintien de l'homéostasie protonique et affecte même la maturation dépendante du pH de protéines comme APP et CTSD (Golabek et al., 2000). Le mécanisme qui sous-tend cet effet est encore inconnu, mais pourrait être associé à la régulation de CLN3 par Shwachmann-Bodian-Diamond syndrome protein (SBDS), une protéine impliquée dans la maturation des ribosomes et qui influencerait l'expression de la pompe ATPase vacuolaire (Vitiello et al., 2010). CLN3 agirait aussi comme transporteur puisqu'elle présente une distante, mais significative similarité de séquence avec Equilibrative nucleoside transporter 1 (ENT1), un membre de la famille des Solute carrier family 29 (SLC29), qui équilibre les nucléosides et les nucléobases à la membrane plasmique (Baldwin et al., 2004). D'ailleurs, CLN3 aurait un rôle dans l'osmorégulation rénale, soit le contrôle de l'équilibre en eau et en potassium (Stein et al., 2010). Son expression est aussi plus élevée et sa localisation plus lysosomiale et périphérique en condition d'hyperosmolarité (Getty et al., 2013). Dans le même ordre d'idées, un transport défectueux d'arginines en l'absence de CLN3 ou en présence d'un mutant pathologique tend à démontrer son implication dans l'import de résidus basiques à travers la membrane lysosomiale (Chan et al., 2009; Kim et al., 2003b; Ramirez-Montealegre and Pearce, 2005). Cet effet pourrait simplement être une conséquence de la régulation à la hausse de Btn2 qui module le transporteur Arginine59 lysine-histidine-permease-Can1p, tel que montré chez Saccharomyces cerevisiae (Chattopadhyay and Pearce, 2002). Toutefois, la plupart des évidences pointent vers un rôle dans le transport post-TGN vraisemblablement en lien avec le cytosquelette. En conséquence à une perte de Btn1/CLN3, plusieurs anomalies ont été rapportées relativement à des processus de transport, notamment à l'endocytose (Fossale et al., 2004; Luiro et al., 2004; Uusi-Rauva et al., 2008), au transport axonal du nerf optique (Weimer et al., 2006), à la maturation des vacuoles autophagiques (Cao et al., 2006), à la morphologie golgienne (Codlin and Mole, 2009) et surtout au triage dans le Golgi (Codlin and Mole, 2009). Ce dernier altère la sortie de MPR qui s'accumule au TGN (Metcalf et al., 2008) menant à une maturation aberrante d'enzymes lysosomiales comme des cathepsines (Fossale et al., 2004; Golabek et al., 2000; Metcalf et al., 2008). CLN3 contribue également à la localisation des endosomes et des lysosomes. Plusieurs changements dans la distribution vésiculaire, induits par des mutants de CLN3, ont été observés en fonction des cellules, dont une dispersion causée par Cln3Δex7/8 (Fossale et al., 2004) et un regroupement périnucléaire faisant suite à l'expression de CLN3E295K (Uusi-Rauva et al., 2012). D'ailleurs, CLN3 interagit avec des composantes motrices impliquées à la fois dans le transport antérograde et rétrograde le long des microtubules, soit beta-tubuline, p150Glued du complexe de Dynactine, la chaîne intermédiaire de la Dynéine (DIC) et la sous-unité Kinesin Family Member 3A (KIF3A) du complexe Kinésine-2 (Uusi-Rauva et al., 2012). L'association membranaire de ces composantes est régulée par les protéines Rab ou plus précisément par Rab7 au niveau des endosomes tardifs. Le domaine cytoplasmique N-terminal de CLN3 interagit directement avec Rab7, préférentiellement actif, de façon à influencer son recrutement et possiblement son cycle d'activation (Uusi-Rauva et al., 2012). Il a été montré que CLN3 peut alors former un complexe en liant, avec une de ses boucles cytosoliques, RILP, un effecteur de Rab7 (Uusi-Rauva et al., 2012). Le couple Rab7-RILP est bien connu pour faciliter le recrutement du complexe Dynéine/Dynactine aux endosomes tardifs et pourrait constituer le déterminant moléculaire par lequel CLN3 affecte le positionnement vésiculaire et subséquemment la fonction lysosomiale. Le partenaire d'interaction Btn2, aussi connu sous le nom de Hook1 chez les vertébrés, est aussi impliqué dans la liaison des microtubules 60 (Walenta et al., 2001) et pourrait compétitionner avec CLN3 pour Rab7 (Luiro et al., 2004). Btn2/Hook1 lie également les SNARE (Kama 2007), laissant croire que CLN3 influencerait la fusion. CLN3 est associée à la migration cellulaire, la formation de plaques d'actine (Codlin and Mole, 2009) et la distribution de la Myosine possiblement en lien avec son interaction avec la Myosine-IIB (Getty et al., 2011). CLN3 interagit avec Fodrine et la pompe Na+/K+ ATPase associée (Uusi-Rauva et al., 2008). La première est une protéine associée à la membrane plasmique qui contrôle la distribution de surface de la pompe en la fixant au cytosquelette d'actine. La déplétion de CLN3 cause une perturbation de la distribution de Fodrine ainsi que de l'endocytose de la pompe Na+/K+ ATPase (Uusi-Rauva et al., 2008) mettant en péril le gradient ionique qui fait partie intégrale de la fonction neuronale. La fonction principale de CLN3 reste inconnue et sa multitude de partenaires d'interaction lui confère une potentielle influence directe ou indirecte sur nombre de processus biologiques. 3.3.3 CLN5 Le gène CLN5 a été isolé en 1998 (Savukoski et al., 1998), il est conservé chez les vertébrés uniquement et code une protéine de 407 acides aminés du même nom qui n'a pas d'homologue. À l'origine, les outils informatiques, comme PHDhtm et TMpred, prédisaient que la protéine contiendrait jusqu'à deux segments transmembranaires, soit globalement des résidus 74 à 91 et 352 à 375 (Savukoski et al., 1998). La région hydrophobe 76-91 a ensuite été confirmée comme étant transmembranaire (Vesa et al., 2002) alors que la seconde région pourrait assurer une association membranaire. Aujourd'hui, CLN5 mature est considérée comme étant une protéine soluble (Holmberg et al., 2004; Schmiedt et al., 2010) puisque la portion N-terminale contenant la région transmembranaire est incluse dans un long peptide signal (figure 8). Lors de la synthèse, il y a donc clivage en position 96, par une peptidase signal non identifiée et la région transmembranaire est retenue dans le RE (Schmiedt et al., 2010). Alors que la plupart des espèces possèdent un seul codon initiateur pour CLN5 (Holmberg et al., 2004), chez l'humain, il en existe quatre aux positions 1, 30, 50 et 62 (Savukoski et al., 1998). Les quatre transcrits produits ont des masses moléculaires 61 de 39 à 47 kDa (Isosomppi et al., 2002; Vesa et al., 2002), mais subissent le même clivage de manière à produire une proprotéine identique. Figure 8 - Schéma de la structure de CLN5. (A) CLN5 immature est une protéine transmembranaire de type 2 qui est clivée en position 96 par une peptidase signal suite à son insertion dans la membrane du RE. (B) CLN5 mature est une protéine luminale. Le segment transmembranaire de la forme immature est représentée en rouge, la forme mature correspond à la partie en bleu, les ciseaux représentent le site de clivage par la peptidase signal. Figure créée avec Microsoft Office PowerPoint 2007. Après sa synthèse, CLN5 passe au travers de la voie de synthèse et subit des Nglycosylations hautement mannosylé (high mannose) et de sucres complexes sur jusqu'à huit sites potentiels (Asn179, 192, 227, 252, 304, 320, 330 et 401) (Moharir et al., 2013; Savukoski et al., 1998). Les sucres en position Asn179, 252, 304, 320 et 330 seraient nécessaires à l'export du RE. L'ajout de mannose-6-phosophate au niveau des résidus Asn320, Asn330 et Asn401 (Sleat et al., 2006) a été appuyé par des mesures en spectroscopie de masse, mais seule la modification sur l'Asn401 permettrait de quitter le Golgi (Moharir et al., 2013). CLN5 possède aussi des sites prédits pour d'autres modifications postranductionnelles, dont sept de phosphorylation par des protéine-kinases C, des caséine-kinases II et des tyrosine-kinases ainsi deux sites chevauchants de Nmyristoylation (Savukoski et al., 1998). Ces modifications n'ont pas été confirmées 62 expérimentalement mais expliqueraient la présence, après traitements de déglycosylation, de deux formes de CLN5 (Schmiedt et al., 2010). CLN5 est donc synthétisée comme une préproprotéine d'environ 70 kDa qui est clivée au RE en une proprotéine de 60 kDa avant d'être glycosylée (60–80 kDa) et finalement taillée pour atteindre 50 kDa (Schmiedt et al., 2010). La forme mature résultante est faiblement sécrétée, mais se rend principalement aux lysosomes (Isosomppi et al., 2002; Vesa et al., 2002). Elle emprunte MPR (Kollmann et al., 2005; Sleat et al., 2006) et peut vraisemblablement utiliser une voie alternative comme indiqué par sa présence inaltérée au compartiment acide dans des fibroblastes déficients en MPR (Schmiedt et al., 2010). Le récepteur LIMP-2 serait le principal candidat à cette fonction puisque Sortiline a été montré comme interagissant avec CLN5 sans toutefois influencer son transport (Mamo et al., 2012). Les NCL de type CLN5 ont fait leur apparition en Finlande où l'incidence est de 2,6 cas par million d'habitants (Uvebrant and Hagberg, 1997). En histologie, la maladie partage les mêmes caractéristiques que les types CLN3, 6, 7 et 8, soit l'accumulation majoritaire de la SCMAS (Tyynela et al., 1997) et la présence de profils en empreintes digitales accompagnés de profils curvilignes et rectilignes. La première mutation identifiée, et aussi la plus commune, représente 94% des cas de NCL liés à CLN5 (Mole et al., 2011). Il s'agit d'une délétion de 2 paires de bases dans l'exon 4 (c.1175_1176delAT) produisant un codon de terminaison en position 392 (p.Tyr392∗) (Holmberg et al., 2000) et formant une glycoprotéine de 47-52 kDa (Isosomppi et al., 2002; Vesa et al., 2002). Dans certaines sous-populations côtières de Finlande, jusqu'à 1 personne sur 24 est porteuse de ce défaut génétique (Savukoski et al., 1998), d'où son nom de mutation Finlandaise majeure (FinM) (tableau 2). Cette appellation perdure bien que l'on retrouve maintenant de très rares cas à l'extérieur de ce pays. Parmi les mutations les mieux connues, on compte également la mutation Finlandaise mineure (Finm) (tableau 2) dans l’exon 1 (c.225G>A) qui produit une protéine de 74 résidus et une de 12 kDa (p.Trp75∗) (Holmberg et al., 2000; Xin et al., 2010), la mutation Européenne (EUR) (tableau 2) dans l'exon 4 (c.835G>A) qui provoque un changement de résidu en position 279 (p.Asp279Asn) (Holmberg et al., 2000; Savukoski et al., 1998) et la mutation Suédoise (SWE) (tableau 2) qui entraîne une terminaison prématurée (c.757G>T) et qui 63 code une protéine tronquée de 34 kDa (p.E253∗) ainsi que deux polypeptides de 37 et 40 kDa (Vesa et al., 2002). Ce dernier mutant possède seulement deux des sucres importants pour l'export du RE, mais peut tout de même passer les contrôles de qualité du RE (Vesa et al., 2002). Finalement, quoique peu répandu, le mutant de Terre-Neuve (NFL) (tableau 2) est pertinant à la présente étude et consiste en une mutation dans l'exon 4 (c.1054G>T) qui crée un codon de terminaison (p.Glu352∗) (Moore et al., 2008). Certains mutants seraient retenus au RE cependant l'atteinte du lysosome ne semble pas être en corrélation avec la sévérité du phénotype (Lebrun et al., 2009; Schmiedt et al., 2010). Cela indique que malgré la faible présence de CLN5 au Golgi et au RE (Isosomppi et al., 2002), il pourrait s'agir là de lieux d'interactions importantes. CLN5 est exprimée dans tous les tissus humains (Savukoski et al., 1998), mais surtout au cerveau notamment dans les cellules microgliales, les astrocytes, les oligodendrocytes et les neurones (Heinonen et al., 2000; Holmberg et al., 2004; Schmiedt, 2012). La protéine semble particulièrement importante à ce niveau puisque son absence entraîne un défaut de myélination, une altération du profil lipidique sérique et un ralentissement du transport intracellulaire des sphingolipides. Ces observations tendent à indiquer un rôle dans le métabolisme des lipides (Schmiedt et al., 2012). Tout comme les cellules déficientes en CLN3, 2, 6 et 8 (Lane et al., 1996), les fibroblastes CLN5-/- arborent une augmentation de l'apoptose en lien avec une diminution des niveaux de sphingolipides, comme les céramides (Haddad et al., 2012). Il a été proposé que CLN5 agirait comme activateurs de la Ceramide synthase (CerS), l'enzyme responsable de la synthèse de novo (Haddad et al., 2012). L'absence de CLN5 entraîne également une inflammation avec activation hâtive de la microglie et des astrocytes qui pourrait être à l'origine de la mort neuronale (Schmiedt et al., 2012; von Schantz et al., 2008). En ce qui concerne le transport intracellulaire, l'équipe du Dr. Lefrançois a montré qu'en condition de surexpression CLN5 interagit avec le récepteur Sortiline sans toutefois agir comme cargo (Mamo et al., 2012). La déplétion de CLN5 par ARN interférent entraîne une dégradation de Sortiline, ainsi que de CI-MPR, aux lysosomes (Mamo et al., 2012). En 64 effet, ces derniers ne sont plus recyclés des endosomes vers le TGN ce qui limite la maturation de l'enzyme CTSD (Mamo et al., 2012). CLN5 serait donc nécessaire au fonctionnement du complexe Retromère. Plus particulièrement, la déplétion de CLN5 affecte le recrutement aux endosomes de deux protéines qui amorcent le transport rétrograde et avec lesquelles CLN5 interagit indépendamment de leur activation, soit Rab5 et surtout Rab7 (Mamo et al., 2012). CLN5 serait impliquée dans l’activation de Rab7 puisque la capacité de cette dernière à interagir avec son effecteur RILP est réduite en l'absence de CLN5 (Mamo et al., 2012). Il a été proposé que CLN5 activerait, directement ou par l'intermédiaire d'une protéine transmembranaire, la GEF de Rab7, identifiée comme étant Mon1-Ccz1 chez la levure (Mamo et al., 2012). Suivant la réduction de Rab7 aux endosomes, les sous-unités VPS26 et VPS35 du complexe Rétromère sont également sousreprésentées sur les endosomes (Mamo et al., 2012). Tableau 2 - Principales mutations dans CLN5. Les impacts génétiques (localisations, changements, types) et protéiques (changements et masses moléculaires) des mutations FinM (Finlandaise majeure) et Finm (Finlandaise mineure), EUR (Européenne), SWE (Suédoise) et NFL (Terre-Neuvienne). Basé sur NCL Resource à l'adresse http://www.ucl.ac.uk/ncl/. 3.3.4 Interactosome des CLN De plus en plus d'évidences tendent à démontrer des relations et des effets compensatoires entre les protéines CLN tant en ce qui a trait à leurs fonctions qu'à leurs expressions. Des études in vitro ont montré les premières possibilités d'interactions. Entre autres, CLN5 pourrait lier CLN1 (Lyly et al., 2009), CLN2 via son C-terminal (de CLN5) (Vesa et al., 2002), CLN3 via son N-terminal (de CLN5) (Vesa et al., 2002), CLN6 (Lyly et 65 al., 2009), CLN8 (Lyly et al., 2009). Concernant l'expression, CLN1 et CLN5 partagent des patrons d’expression similaires et, les souris déficientes en CLN5, présentent une augmentation de l’ARNm de CLN1 (Lyly et al., 2009). Il pourrait s'agir là d'un effet compensatoire en lien avec le fait que CLN1 semble influencer positivement le transport de CLN5 (Schmiedt et al., 2010). Il a aussi été proposé que CLN1 et CLN5 formeraient un complexe avec l'ATP synthase (Lyly et al., 2008; Lyly et al., 2009), un récepteur de l’ApoA-I, appuyant davantage un rôle pour CLN5 dans l’homéostasie des lipides (Martinez et al., 2003). Dans le même ordre d'idées, une réduction des protéines CLN2 et CLN5 entraîne une diminution de l’ARNm de CLN3 (Bessa et al., 2006). D'ailleurs, l’activité de CLN2 est plus forte dans les fibroblastes portant une mutation dans CLN5 (Vesa et al., 2002) ou CLN3 (Sleat et al., 1998). Dans ces mêmes cellules, le pH des lysosomes est augmenté laissant croire que les deux protéines pourraient collaborer pour maintenir l’homéostasie du pH (Holopainen et al., 2001; Pearce et al., 1999a). Ces observations suggèrent une relation ou une redondance fonctionnelle entre les CLN2, 3 et 5. CLN5 partagerait aussi le rôle d'activateurs de synthases de ceramides avec CLN8, car CLN8 à la capacité de corriger les défauts apoptotiques engendrés par l'absence de CLN5 (Haddad et al., 2012). De plus amples études seront nécessaires afin de clarifier cet interactosome. La découverte, en 2011, du motif Coordinated Lysosomal Expression And Regulation (CLEAR) pourrait expliquer la coordination d'expression de certaines protéines CLN. Entre autres, CLN1, CLN2, CLN3, CLN5, CLN7, CLN10, CLN11 et CLN13 ont été clairement identifiés comme protéines possédant le motif CLEAR et comme étant vraisemblalement régulées par TFEB (Sardiello et al., 2009). Ce sujet sera plus amplement développé dans la discussion. 3.3.5 Aspects cliniques Les céroïdes-lipofuscinoses neuronales constituent le groupe de maladies neurodégénératives progressives et récessives le plus commun chez les enfants avec une 66 incidence mondiale estimée de 1:100000 (Santavuori, 1988), mais fortement accentuée dans les pays nordiques comme la Finlande à 1:12500 (Rider and Rider, 1988; Williams et al., 1999). Elle est diagnostiquée par la combinaison de plusieurs examens. Les premiers indices de la maladie peuvent être détectés lors d'examens visuels analysant la réponse électrique du nerf optique ou de la rétine (Mole et al., 2011). Aussi, la tomodensitométrie et l'imagerie par résonance magnétique sont utilisées pour révéler des changements structuraux au niveau du cerveau (Williams et al., 2006). Pour parvenir à un diagnostic définitif et précis, les patients sont rapidement soumis à des essais enzymatiques visant à identifier une enzyme défectueuse (Williams et al., 2006). Toutefois, cette technique est valable pour un nombre limité de sous-types de la maladie, n'étant pas tous causés par un problème enzymatique. En microscopie électronique, l'observation de biopsies provenant principalement de la peau, permet également de cibler les NCL grâce aux dépôts cellulaires qui leurs sont caractéristiques (Williams et al., 2006). D'ailleurs, dans les familles affectées par la maladie, l'étude prénatale des inclusions et des ultrastructures sur les prélèvements des villosités choriales révèle la maladie dès le troisième mois de la grossesse (Mole et al., 2011). Finalement, l'analyse de mutations établit un diagnostic incontestable (Williams et al., 2006). À ce sujet, le criblage à haut débit (high-throughput screening) facilite les diagnostiques moléculaires des LSD causées par des mutations dans des protéines autres que des enzymes (Winchester, 2013). La nouvelle approche de séquençage ciblé Lysoplex constitue également une voie prometteuse pour l'identification de mutations pathogéniques dans 891 gènes en lien avec les lysosomes, l'endocytose et l'autophagie (Di Fruscio et al., 2015). Les NCL de même que leur évolution varie considérablement en fonction du soustype de la maladie, de la mutation en cause ainsi que de l'individu. Les premiers signes cliniques de la maladie apparaissent généralement au bout de plusieurs années, mais parfois dès la naissance ou au début de l'âge adulte. Des lésions s'accumulent au niveau des organes entraînant des troubles graves et irréversibles qui provoquent une mort prématurée après des années de symptômes divers. Parmi ces derniers, on compte un retard de développement, de la surdité, de la cécité, des désordres moteurs, des problèmes de posture et des crises d'épilepsie symptomatiques généralisées (Mole et al., 2011). Au quotidien, les 67 parents des enfants atteints doivent gérer de la constipation, des sécrétions orales excessives accompagnées de difficultés à avaler et mastiquer, des troubles du sommeil et des problèmes de communications comme l'emploi de thème récurent de conversation, un manque de fluidité verbale et de la difficulté à initier des phrases (Mole et al., 2011). La santé mentale et émotionnelle des patients est affectée par l'ensemble des symptômes qui entraîne de la frustration, de la démence, de la dépression, de l'anxiété et des hallucinations tant visuelles qu'auditives et sensitives (Mole et al., 2011). Des soins palliatifs sont donc nécessaires pour assurer une certaine qualité de vie aux personnes atteintes. Une routine structurée et variée, un environnement familier et une médication comprenant des antiépileptiques, des antipsychotiques, des laxatifs, des relaxants musculaires, l'installation de tubes nasogastriques ou une gastrotomie, de la physiothérapie et des activités plaisantes aident à soulager plusieurs des symptômes associés aux NCL (Mole et al., 2011). À l'heure actuelle, aucun traitement n'est disponible pour guérir les personnes atteintes de céroïdes-lipofuscinoses neuronales ou autres maladies de surcharge lysosomiale. Par contre, de nombreuses stratégies thérapeutiques sont en développement et montrent des signes prometteurs. Dans plusieurs cas, une thérapie visant le remplacement de l’enzyme défectueuse par infusion intraveineuse (Enzyme remplacement therapy) permet de maintenir des niveaux adéquats (Eng et al., 2001). Lorsque la mutation affecte le repliement plutôt que sa fonctionnalité protéique, il est possible de favoriser le repliement des enzymes thermodynamiquement instables en utilisant des chaperonnes pharmacologiques (Enzyme enhancement therapy) (Desnick, 2004). La thérapie de réduction du substrat (Substrate reduction therapy) constitue une autre approche qui a pour but de réduire la synthèse du substrat de l’enzyme faisant défaut de manière à prévenir son accumulation (Cox, 2005). De façon plus globale, la transplantation de cellules souches de moelle épinière ou de sang de cordon ombilical s’avère efficace pour certaines maladies de surcharge lysosomiale comme pour bien d’autres pathologies (Giugliani et al., 2016). Par contre, l’ensemble de ces thérapies n’affecte pas l’origine génétique des maladies. En ce sens, la thérapie génique, qui consiste à faire pénétrer des gènes dans les cellules pour remplacer un gène défectueux, serait la stratégie expérimentale la plus prometteuse. L’avancement des thérapies repose en grande partie sur l’évolution des connaissances 68 scientifiques relatives à la pathogenèse des maladies. L’identification des gènes impliqués dans les NCL devrait constituer une priorité de même que l’élucidation de la fonction des protéines codées par ces gènes, de leurs substrats et de leurs partenaires d'interaction. Une meilleure compréhension des mécanismes menant aux accumulations cellulaires et à la mort préférentielle des neurones aiderait grandement au développement pharmaceutique. 4 Problématique, hypothèse et objectifs 4.1 Problématique Calnuc est une protéine peu caractérisée dont la fonction au niveau du TGN et des endosomes demeure encore inconnue. Notre groupe a précédemment démontré que Calnuc influence le recyclage du récepteur LRP9 ce qui laisse croire qu’elle pourrait avoir un rôle général dans le transport d'autres récepteurs entre les endosomes et le Golgi. La voie endosomiale est particulièrement importante pour le maintien des lysosomes puisqu’elle est responsable du transit des divers constituants de l’organite. D’ailleurs, des défauts dans les protéines responsables de sa structure et de sa fonction ont été associés à des désordres lysosomiaux, nommés céroïdes-lipofuscinoses neuronales. 4.2 Hypothèse Notre hypothèse de recherche est que Calnuc constituerait un régulateur important du transport rétrograde des endosomes vers le Golgi. De ce fait, une altération de Calnuc, comme une mutation ou une réduction de ses niveaux intracellulaires, causerait un mauvais triage de récepteurs, dont des récepteurs lysosomiaux, menant à des désordres du lysosome, comme les céroïdes-lipofuscinoses neuronales. 69 4.3 Objectifs Article 1 - Impact et mécanisme d'action de Calnuc sur le transport rétrograde des récepteurs lysosomiaux Nos études précédentes ont démontré que Calnuc interagit, colocalise et influence le transport rétrograde de LRP9, un récepteur peu caractérisé qui transite entre le TGN et les endosomes. Nous étudierons donc l'existence d'une fonction plus générale de Calnuc à ce niveau en examinant son impact sur CI-MPR et Sortiline. Nous identifierons ensuite le mécanisme moléculaire responsable de cet effet. Article 2 - Caractérisation topologique de CLN5 et de ses principaux mutants pathologiques Depuis sa découverte, la protéine CLN5 fait l'objet de données contradictoires quant à sa localisation, sa topologie et sa solubilité. Nous caractériserons donc la protéine CLN5 de même que quelques-uns de ses mutants les plus fréquents. Article 3 - Calnuc, un nouveau partenaire d'interaction pour CLN3 et CLN5, est potentiellement impliqué dans les céroïdes-lipofuscinoses neuronales Calnuc affecte le transport rétrograde de récepteurs lysosomiaux de manière semblable à CLN5, une protéine impliquée dans les céroïdes-lipofuscinoses neuronales. Nous étudierons donc le lien entre ces deux protéines de même que l'implication de Calnuc dans la pathologie. 70 ARTICLE 1 Calnuc function in endosomal sorting of lysosomal receptors Auteurs de l’article: Heidi Larkin, Santiago Costantino and Christine Lavoie Statut de l’article: Publié Traffic. 2016 Jan 12. doi: 10.1111/tra.12374. [Epub ahead of print] Avant-propos: J'ai écrit la première version des sections 'Introduction', 'Résultats', 'Matériel et méthodes' ainsi que des descriptions des figures. J'ai réalisé toutes les expériences retrouvées dans l'article ainsi que tous les montages. Christine Lavoie a effectué les corrections du texte et a rédigé la discussion. Elle a également contribué à la prise de photos brutes pour les figures d'immunofluorescence et aux expériences présentées en figures 1C et S1. La quantification de l'intensité des endosomes a été effectuée par Santiago Costantino. 71 RÉSUMÉ – ARTICLE 1 Rôle de Calnuc dans le triage endosomial des récepteurs lysosomiaux Calnuc est une protéine ubiquitaire qui lie le calcium et qui est présente au réseau trans-golgien (TGN) ainsi qu'aux endosomes. Cependant, son rôle précis à ce niveau est peu connu. Notre groupe a précédemment mis en évidence le rôle de Calnuc dans le transport de LRP9, un membre des récepteurs aux lipoprotéines de faible densité (LDL) qui cycle entre le TGN et les endosomes. L'objectif de la présente étude est d'explorer la fonction de Calnuc dans le triage endosomial de deux récepteurs lysosomiaux bien caractérisés, qui cycle aussi entre le TGN et les endosomes, soit le récepteur au mannose-6phosphate (MPR) et Sortiline. Par techniques de biochimie et de microscopie, nous avons montré que Calnuc lie la queue cytoplasmique des deux récepteurs et colocalise avec eux sur les endosomes. La déplétion de Calnuc par ARN interférent cause leur redistribution vers les lysosomes où ils sont dégradés. Ce phénomène résulte d'un défaut dans le mécanisme de recrutement du complexe Rétromère qui agit comme principal mécanisme de transport rétrograde des endosomes vers le TGN. Finalement, nous avons montré que la déplétion de Calnuc altère l'activation et l'association membranaire de Rab7, une petite protéine G nécessaire pour le recrutement des rétromères aux endosomes. Dans l'ensemble, nos données montrent que Calnuc est un important joueur du transport rétrograde des récepteurs lysosomiaux des endosomes vers le TGN via la régulation et le recrutement du complexe Rétromère. 72 Article 1 Calnuc function in endosomal sorting of lysosomal receptors Heidi Larkin1, Santiago Costantino2 and Christine Lavoie1‡ Department of Pharmacology, Faculty of Medicine and Health Sciences, Université de 1 Sherbrooke, Sherbrooke, QC, Canada Centre de recherche de l’Hôpital Maisonneuve-Rosemont, Université de Montréal, 2 Montréal, Canada H1T 2M Running Title: Calnuc modulates lysosomal receptor trafficking Key words: Calnuc, nucleobindin, mannose-6-phosphate receptor, Sortiline, lysosomal receptor, retromers, endosomal sorting, lysosome, Rab7 ‡ Correspondence should be addressed to: Christine Lavoie Department of Pharmacology Faculty of Medicine Université de Sherbrooke 3001—12e Avenue Nord Sherbrooke, QC, Canada J1H 5N4 Telephone: (819) 820-6868, ext. 12732 Fax: (819) 564-5400 Email: [email protected] 73 Early view Synopsis Calnuc is a ubiquitous Ca2+-binding protein which function is poorly characterized. In this study, we demonstrate that Calnuc plays a role in the endosome-to-TGN retrograde transport of the lysosomal receptors CI-MPR and Sortilin through the activation and membrane association of Rab7, a small G protein required for the endosomal recruitment of retromers. Abstract Calnuc is a ubiquitous Ca2+-binding protein present on the trans-Golgi network (TGN) and endosomes. However, the precise role of Calnuc in these organelles is poorly characterized. We previously highlighted the role of Calnuc in the transport of LRP9, a new member of a low density lipoprotein (LDL) receptor subfamily that cycles between the TGN and endosomes. The objective of this study was to explore the role of Calnuc in the endocytic sorting of mannose-6-phosphate receptor (MPR) and Sortilin, two wellcharacterized lysosomal receptors that transit between the TGN and endosomes. Using biochemical and microscopy assays, we showed that Calnuc binds to the cytoplasmic tail of these receptors through an indirect association and colocalizes with them on endosomes. 74 We found that Calnuc depletion (by siRNA) causes the misdelivery to and degradation in lysosomes of CI-MPR and Sortilin due to a defect in the endosomal recruitment of retromers, which are key components of the endosome-to-Golgi retrieval machinery. Indeed, we demonstrated that Calnuc depletion impairs the activation and membrane association of Rab7, a small G protein required for the endosomal recruitment of retromers. Overall, our data indicate a novel role for Calnuc in the endosome-to-TGN retrograde transport of lysosomal receptors through the regulation of Rab7 activity and the recruitment of retromers to endosomes. Introduction Calnuc, also named NUCB1, is a ubiquitous, well-conserved protein; however, its precise physiological and cellular functions are poorly understood. Calnuc is a multifunctional protein comprised of a signal peptide, a putative DNA-binding domain, a leucine zipper and 2 EF-hand motifs, which both possess the ability to bind Ca2+ (Lin, et al., 1998). Calnuc is also a multicompartmental protein, with Golgi luminal, extracellular and cytosolic pools. The predominant Golgi pool of Calnuc is located in the lumen of cisGolgi cisternae, where it acts as a major Ca2+-binding protein and a significant Ca2+ storage pool (Lin, et al., 1998; Lin, et al., 1999). After a long period of retention in the Golgi, Calnuc is secreted via constitutive and constitutive-like pathways (Lavoie, et al., 2002). In addition, a significant cytoplasmic pool of Calnuc (Lin, et al., 1998; Lin, et al., 2000) is either free or associated with the surface of the Golgi, endosomal membranes and secretory granules (Brodeur, et al., 2009; Lin, et al., 2009). The multiple domains and localization of Calnuc potentiates its interactions with various partners, and the growing list of Calnucinteracting partners, such as DNA (Miura, et al., 1992), heterotrimeric Gα proteins (Lin, et al., 1998) or COX (Ballif, et al., 1996), suggests its importance in the regulation of many cellular events. Recently, we have shown that Calnuc interacts with the cytosolic tail of a novel LDLR-related protein known as LRP9 in mice (i.e., LRP10 in humans) (Brodeur, et al., 2009). We have also demonstrated that LRP9 cycles between the trans-Golgi network (TGN) and early endosomes (Boucher, et al., 2008) and that Calnuc is essential for 75 retrieving LRP9 from endosomes to the TGN (Brodeur, et al., 2009). These data highlighted the potential role of Calnuc in endosomal sorting. Similar to LRP9, other receptors such as Sortilin and cationic-dependent and independent mannose-6-phosphate receptor (CD-MPR and CI-MPR) transit between the TGN and endosomes (Braulke and Bonifacino, 2009). These lysosomal receptors are involved in the efficient delivery of soluble digestive enzymes to the lysosome. They bind newly synthesized lysosomal cargos in the Golgi, and are then packaged into clathrincoated vesicles by the adaptors AP1 and GGA (Golgi-localized, γ-ear-containing, ADPribosylation factor-binding proteins), and anterogradely transported to endosomes (Bonifacino and Lippincott-Schwartz, 2003). When the receptor/cargo complex reaches this acidic endosomal environment, the cargo dissociates from the receptor and is transported to lysosomes, while the unoccupied receptor is then recycled back to the Golgi for another round of sorting (Bonifacino and Rojas, 2006). This retrograde transport depends on the retromer complex (Arighi, et al., 2004) composed of two subcomplexes: the vacuolar protein-sorting (Vps) trimer Vps26/Vps29/Vps35, which participates in cargo recognition, and a membrane-targeting complex, which is composed of a heterodimeric sorting nexin (SNX) that consists of an undefined combination of SNX1, SNX2, SNX5 and SNX6 (McGough and Cullen, 2011). The recruitment of the retromer complex is initiated by the sequential action of the small G proteins Rab5 and Rab7 on endosomes. Through the recruitment of class III PI3K (Christoforidis, et al., 1999), Rab5 increases endosomal PI3P levels and consequently increases the membrane recruitment of the SNX subcomplex. The mammalian SNX dimer is necessary but not sufficient for the recruitment of Vps subcomplex (Rojas, et al., 2007). Another Rab5 effector is the Rab7 guanine nucleotide exchange factor (GEF), which recruits and activates Rab7 on endosomal membranes. Active Rab7 recruits the retromer Vps subcomplex on endosome via direct interactions with Vps35/Vps29/Vps26 trimer (Rojas, et al., 2008; Seaman, et al., 2009). The cooperation of Rab5 and Rab7 is thus necessary for retromer recruitment and for the sorting of lysosomal receptors into retrograde tubules that are targeted to the Golgi. Identifying novel players involved in the regulation of lysosomal receptor sorting and trafficking is 76 crucial because a loss of regulation can result in aberrant receptor localization and, subsequently, lysosomal-associated diseases. In the present study, we explored the role of Calnuc in the trafficking of CI-MPR and Sortilin, two well-characterized lysosomal receptors that transit between the TGN and endosomes. We demonstrate the function of Calnuc in the endosomal retrieval of these receptors through modulating Rab7 activity and the retromer recruitment machinery, suggesting a general role in the endosomal sorting of receptors. Materials and Methods Antibodies and reagents Anti-actin monoclonal antibodies (mAbs) were purchased from Sigma-Aldrich (Saint Louis, MO, USA), anti-EEA1 mAbs from BD Transduction Laboratories (Franklin Lakes, NJ, USA), anti-Flag M2 mAbs from Sigma Aldrich (Saint Louis, MO, USA), anti-GFP mAbs from Clontech (Mountain View, CA, USA), anti-HA mAbs from Covance (Emeryville, CA, USA), anti-CI-MPR mAbs from Cedarlane Laboratories (Burlington, ON, Canada), anti-CI-MPR mAbs from Abcam (Cambridge, MA, USA), anti-Myc mAbs from Cell signalling Technology (Danvers, MA, USA), anti-Rab7 mAbs and anti-EGFR pAbs from Cell signalling Technology (Danvers, MA, USA), anti-SNX1 mAbs from BD Transduction Laboratories (Franklin Lakes, NJ, USA), anti-SNX2 mAbs from BD Transduction Laboratories (Franklin Lakes, NJ, USA), anti-Sortilin (Neurotensin receptor 3) mAbs from BD Transduction Laboratories (Franklin Lakes, NJ, USA), anti-TfR mAbs from Invitrogen (Carlsbad, CA, USA). Anti-Calnuc polyclonal antibodies (pAbs) were purchased from Aviva Systems Biology (San Diego, CA, USA), anti-CatD pAbs from Calbiochem (La Jolla, CA, USA), anti-CD8 pAbs from Santa Cruz Biotechnology (Santa Cruz, CA, USA), anti-EEA1 pAbs from Thermo Scientific (Rockford, IL, USA), anti-Flag pAbs from Sigma Aldrich (Saint Louis, MO, USA), anti-GFP pAbs from Molecular Probes (Eugene, OR, USA), anti-hVps26 pAbs from Abcam (Cambridge, MA, USA). 77 DNA constructs Mammalian expression vector pcDNA3.1 containing Calnuc-GFP fusion protein or ΔSP-Calnuc (Calnuc without signal peptide) are described elsewhere (Weiss, et al., 2001). These constructions were used to generate Calnuc-HA, Calnuc--Calnuc-Flag. 3xHA-Rab5 and HA-Rab7 were obtained by subcloning of human Rab5 and Rab7 sequences purchased from Addgene. Cell culture and transfection HeLa cells were purchased from the American Type Culture Collection (Manassas, VA, USA), COS7 cells from Dr Klaus Hahn (University of North Carolina, Chapel Hill, NC, USA) and HEK293T cells were obtained from Dr. Alexandra Newton (University of California, San Diego, CA, USA). HeLa cells stably expressing GFP-rab7 were obtained from Matthew Seaman (University of Cambridge). The cells were grown in Dulbecco’s modified Eagle’s high glucose medium (Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT, USA) and 1% penicillin, Glutamine and streptomycin (Invitrogen, Carlsbad, CA, USA). The stable cell line was also supplemented with 0.5 mg/ml G418 (Invitrogen, Carlsbad, CA, USA). HeLa cells and the HEK cells were transfected with Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturers’ instructions. Coimmunoprecipitation HEK293 cells were plated in 60-mm culture dishes and transfected with the various constructs. After 48 h, the cells were lysed in 50 mM Tris buffer (pH 7.4) containing 100 mM NaCl, 1% NP-40, and protease inhibitors. Cells were incubated for 1 h at 4oC and then centrifuged at 15,800xg for 20 min. The cleared supernatants were incubated with primary antibodies overnight at 4oC and then with protein A-sepharose (GE Healthcare, Piscataway, NJ, USA) or protein G-Sepharose (Zymed, San Francisco, CA, USA) for 1 h. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. Bound immune complexes were analyzed by SDS-PAGE and immunoblotting. 78 Glutathione S-transferase (GST) pull-down assays GST fusion proteins were expressed in E. coli BL21 and were purified on glutathione-Sepharose 4B beads (Pharmacia, Piscataway, NJ, USA) according to the manufacturer’s instructions. Lysate from HeLa cells obtained as described above was incubated overnight at 4oC with 20-100 μg of GST fusion proteins immobilized beads. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. The bound proteins were separated by SDS-PAGE and detected by autoradiography or immunoblotting. Immunoblotting The protein samples were separated on 10% or 12% SDS-PAGE gels and transferred to nitrocellulose membranes (Perkin Elmer, Woodbridge, ON, Canada). The membranes were blocked in Tris-buffered saline (20 mM Tris-HCl, pH 7.5, 150 mM NaCl) containing 0.1% Tween 20 and 5% nonfat dry milk, bovine serum albumin (BSA), or foetal bovine serum (FBS) and incubated with primary antibodies for 1 h at RT and then with horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (Bio-Rad, Richmond, ON, Canada) and enhanced chemiluminescence detection reagent (Pierce Chemical, Rockford, IL, USA). RNA Interference and Rescue Nonspecific (CTL) siRNA scrambled II duplex and Calnuc siRNA were purchased from Dharmacon Research (Chicago, IL, USA). HeLa cells were transfected with a final concentration of 100 nM siRNA duplex using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s instructions. The cells were analyzed 72 h after the transfection. Tagged proteins were transfected with Fugene 12 h before the immunofluorescence experiments. Reversal of phenotype (rescue) was performed as described in (Brodeur, et al., 2009). Briefly, HeLa cells were transfected with cDNA encoding rat Calnuc-GFP, rat ΔSP-Calnuc-GFP or GFP alone using Fugene 8–10 h after the initial transfection with human Calnuc siRNA, and the cells were analyzed after 38–40 h. 79 Cycloheximide chase Mock- and siRNA-treated cells were incubated at 37ºC in complete DMEM containing 25 mM Hepes buffer, pH 7.4, and 40 μg/ml of cycloheximide (Sigma-Aldrich, Saint Louis, MO, USA). At the corresponding time points, the cells were lysed as described above, and analyzed by SDS-PAGE and immunoblotting. Treatment with lysosomal inhibitors Mock- and siRNA-treated cells were incubated for 3h at 37ºC in complete DMEM containing 40 μg/ml of cycloheximide (Sigma-Aldrich, Saint Louis, MO, USA), 1 mg/ml of leupeptin (Roche Diagnostics, Indianapolis, IN, USA), 100 μM of pepstatin and 20 μg/ml of E64 (Sigma-Aldrich, Saint Louis, MO, USA). The cells were collected, lysed, and analyzed by SDS-PAGE and immunoblotting. Immunofluorescence HeLa cells were plated on coverslips. Twelve hours after the transfection, the cells were fixed for 30 min in 3% paraformaldehyde (PFA) 100 mM phosphate buffer, pH 7.4, permeabilized with 0.1% Triton X-100 for 10 min, blocked with 10% goat or fetal bovine serum for 30 min, and incubated with primary antibodies for 1 h at RT, followed by Alexa Fluor-488, 594 or 647-conjugated antibodies (Molecular Probes, Eugene, OR, USA) for 1 h at RT. The specimens were visualized using an inverted confocal laser-scanning microscope (FV1000, Olympus, Tokyo, Japan) equipped with a PlanApo 60x/1.42 oil immersion objective (Olympus, Tokyo, Japan). Olympus Fluoview software version 1.6a was used for image acquisition and analysis. The images were further processed using Adobe Photoshop (Adobe Systems, San Jose, CA, USA). The quantification of the immunofluorescence signal on endosome was previously described (Mamo, et al.). Cell fractionation (S100/P100 centrifugation) HeLa cells were washed in PBS, collected into homogenization buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 5 mM ethylenediaminetetraacetic acid (EDTA) and 200 mM sucrose with protease inhibitors), and centrifuged for 5 min at 500×g. The pellet was homogenized by 60 passages through a 25 G 5/8 needle and centrifuged at 1000×g for 10 80 min to remove nuclei and unbroken cells. The post-nuclear supernatant was collected and centrifuged at 100,000×g for 1 h at 4ºC. The soluble (cytosolic fraction, S100) was collected, and the pellet (membrane fraction, P100) was resuspended using a pestle in an equal volume of buffer. The same volumes of membrane and cytosolic fractions were analyzed by SDS-PAGE and immunoblotting. Statistical analysis Experiments were performed at least in triplicate and results are expressed as means intensity of protein normalised with loading control ± SD. Statistical significance between two groups was assessed using the Student t-test whereas ANOVA was used for comparison with three groups or more. P value < 0.05 was considered significant (*) and p < 0.001 very significant (**). Results Calnuc interacts and colocalizes with the lysosomal sorting receptors Sortilin and CIMPR. Because Calnuc has been previously shown to interact with and alter the endosomal sorting of LRP9 (Brodeur, et al., 2009), we first examined whether Calnuc could interact with Sortilin and CI-MPR, two receptors that follow the same trafficking pathway as LRP9. To verify the specificity of these interactions, we also examined whether Calnuc bound to epidermal growth factor receptor (EGFR), which does not cycle between the TGN and endosomes. We first performed immunoprecipitation experiments using HEK293 cells expressing HA-tagged Calnuc alone or cells expressing GFP-tagged Calnuc or GFP together with Myc-tagged Sortilin or EGFR (Figure 1A). We found a weak but highly reproducible interaction between Calnuc-HA and endogenous CI-MPR as well as between Calnuc-GFP and Myc-Sortilin using anti-CI-MPR or anti-GFP IgG immunoprecipitation, respectively (Figure 1A). However, no interaction was detected between Calnuc-GFP and EGFR using anti-GFP IgG immunoprecipitation (Figure 1A). Furthermore, endogenous Calnuc interacts with endogenous CI-MPR (Figure 1B). Because the cytosolic tail of LRP9 was previously shown to interact with Calnuc (Brodeur, et al., 2009), we then performed 81 Figure 1 - A small proportion of Calnuc interacts with the cytoplasmic tails of CIMPR and Sortilin and colocalizes with CI-MPR and Sortilin on endosomes. (A) Coimmunoprecipitation of Calnuc with full-length CI-MPR and Sortilin but not with 82 EGFR. Lysates from HEK cells transfected with Calnuc-HA alone or with GFP or Calnuc– GFP together with Myc-Sortilin or EGFR were immunoprecipitated with control (lane 1), anti-CI-MPR (lane 2), anti-GFP (lanes 3-6) antibodies and then immunoblotted with antiHA or anti-GFP to detect Calnuc, with anti-Myc to detect Sortilin, with anti-CI-MPR or anti-EGFR. Calnuc-GFP and Calnuc-HA are observed as two bands, the lower of which most likely represents an N-terminal cleavage product. (B) Coimmunoprecipitation of endogenous Calnuc and CI-MPR. Lysates from HeLa cells were immunoprecipitated with control (CTL) or anti-CI-MPR antibodies and then immunoblotted with anti-Calnuc and anti-CI-MPR. (C) Coimmunoprecipitation of Calnuc-GFP with CD8-CI-MPR and CD8Sortilin reporter proteins containing the cytoplasmic domains (C-term) of these receptors. Lysates from HEK cells transfected with CD8-CI-MPR or CD8-Sortilin and GFP or Calnuc-GFP were immunoprecipitated with anti-GFP (lanes 1-4) and then immunoblotted with anti-CD8 to detect CI-MPR C-term and Sortilin C-term or with anti-GFP to detect GFP and Calnuc. (D) Comparison of the intracellular distribution of endogenous Calnuc and CI-MPR. HeLa cells transfected with GFP-tagged Rab5Q79L (to create enlarged endosomes) were fixed, permeabilized, immunostained with anti-Calnuc and anti-CI-MPR and examined by confocal fluorescence microscopy. Calnuc was found mainly in the Golgi cisternae (arrowheads), and a small amount was also detected in Rab5-labeled endosomes (b, e). CI-MPR was localized in Rab5-labeled endosomes (c, d). The merged image (f) shows a partial overlap between Calnuc and CI-MPR in specific membrane microdomains of endosomes. Scale bar, 10 μm. (E) Comparison of the intracellular distribution of the cytoplasmic pool of Calnuc and CI-MPR or Sortilin. COS7 cells were transfected with Calnuc devoid of its signal sequence (ΔSP-Calnuc-GFP), which is only expressed in the cytosol, in the absence (a, b) or presence of Myc-tagged Sortilin (c, d). (a, c) ΔSP-CalnucGFP was mainly detected at the plasma membrane and the cytosolic side of the Golgi cisternae (left insets), and a small amount was found on vesicles corresponding to endosomes (right insets). The merged image (yellow) shows a partial overlap between ΔSP-Calnuc-GFP and CI-MPR (a, right inset) or Myc-Sortlin (c, right inset) on the endosome surface. (b, d) ΔSP-Calnuc-GFP partially colocalized with endogenous CI-MPR and Myc-Sortilin on enlarged early endosomes created by the expression of Rab5Q79L (right insets). Twelve hours after transfection, the cells were fixed, permeabilized and immunostained using anti-GFP (a-d) and anti-CI-MPR (a-b) or anti-Myc antibodies (c-d). The labeled cells were examined by confocal fluorescence microscopy. Scale bar, 10 μm. co-immunoprecipitation experiments using the cytosolic tails of both CI-MPR and Sortilin fused to the reporter protein CD8 (Cluster of differentiation 8), a transmembrane glycoprotein that allows adequate trafficking along the cell (Seaman, 2004) and acts as a tag for detection. HEK293 cells were transfected with GFP-tagged Calnuc or GFP together with CD8-MPR or CD8-Sortilin (Figure 1C), and the proteins were immunoprecipitated using an antibody against GFP. The results showed that Calnuc-GFP, but not GFP alone, interacts with the cytosolic tails of both CI-MPR and Sortilin (Figure 1C). However, using in vitro translation pull-down assays, we were unable to detect a direct interaction between 83 Calnuc and the cytosolic tails of CI-MPR and Sortilin (data not shown), suggesting an indirect interaction between Calnuc and the lysosomal receptors. To determine whether Calnuc colocalizes with these lysosomal receptors, we next compared the intracellular distribution of endogenous Calnuc and CI-MPR in HeLa cells using confocal microscopy. In order to help visualize the colocalization of Calnuc and CIMPR, enlarged endosomes were created by expressing Rab5 GTPase-deficient mutant (Rab5Q79L) that increases homo- and heterotypic fusion, leading to the formation of enlarged early endosomes (Stenmark, et al., 1994). As previously reported (Lin, et al., 1998; Brodeur, et al., 2009), endogenous Calnuc was predominantly localized in the Golgi region (Figure 1D (arrowheads, g)). However, a small amount of Calnuc was also detected in enlarged Rab5-labeled endosomes (Figure 1D(b,e)). Endogenous CI-MPR was also localized in Rab5-labeled endosomes (Figure 1D(c,d)) and partially overlaps with Calnuc in specific membrane microdomains (Figure 1D(f)). However, these images did not indicate whether it is the luminal or cytoplasmic pool of Calnuc that colocalizes with CIMPR. Given that Calnuc interacts with the cytoplasmic tail of MPR, we next examined whether the specific cytosolic pools of Calnuc and CI-MPR colocalize. COS7 cells were transfected with Calnuc-GFP lacking its signal peptide (ΔSP-Calnuc-GFP), which prevents its insertion into the lumen of the Golgi and limits its distribution to the cytoplasm. As previously reported (Brodeur, et al., 2009), ΔSP-Calnuc-GFP accumulated mainly in the cytoplasm and plasma membrane but was also detected at the surface of Golgi (Figure 1E(a), left inset) and of endosomes (Figure 1E(a), right inset). A small proportion of this pool of Calnuc partially colocalized with CI-MPR at the surface of endosomes (Figure 1E(a), right insets). This endosomal colocalization was confirmed in cells co-expressing a Rab5 GTPase-deficient mutant (Rab5Q79L) (Figure 1E(b)). The same confocal analyses were performed in COS7 cells expressing Myc-Sortilin together with ΔSP-Calnuc-GFP in the absence or presence of Rab5Q79L (Figure 1E(c, d)). As observed with CI-MPR, partial colocalization between ΔSP-Calnuc-GFP and Myc-Sortilin was found at the surface of endosomes (Figure 1E(c,d), right insets). Together, these immunofluorescence and biochemical interaction assays suggest that a small proportion of the cytoplasmic pool of Calnuc indirectly interacts with the C-terminal tails of MPR and Sortilin on endosomes. 84 Calnuc is implicated in the trafficking of the lysosomal sorting receptors Sortilin and CI-MPR. To investigate the functional role of Calnuc in CI-MPR and Sortilin trafficking, we examined the steady-state distribution of these lysosomal receptors in Calnuc-depleted cells. HeLa cells were transfected with control or Calnuc siRNA alone or together with Myc-Sortilin and were then analyzed by confocal microscopy. Calnuc was strongly (≥ 90%) depleted at 72 h after transfection with a Calnuc-specific small interfering RNA (siRNA), as observed in immunofluorescence (Figure 2A) and immunoblotting experiments (Figure 2B). In Calnuc knockdown cells, the intensities of the immunofluorescence signals for endogenous CI-MPR and Myc-Sortilin (Figure 2A(d,e)) were reduced compared with control cells (Figure 2A(a,b)). These results were confirmed by western blotting analysis. The steady-state levels of endogenous CI-MPR and MycSortilin were reduced in Calnuc knockdown cells compared with control cells (Figure 2B). In contrast, the level of Transferrin receptor (TfR), a plasma membrane receptor that does not traffic between the TGN and endosomes, was not affected by Calnuc depletion (Figure 2B). The specificity of the effect of Calnuc on the levels of these lysosomal receptors was also validated by a rescue experiment with siRNA-resistant forms of wild-type and cytosolic Calnuc (ΔSP-Calnuc). Rat Calnuc was used as a rescue construct because it differs in many nucleotides from human siRNA target sequences, rendering it resistant to the siRNA. The results showed that reintroducing rat Calnuc-GFP or ΔSP-Calnuc-GFP into Calnuc-depleted cells restored the basal levels of CI-MPR to levels similar to those of control cells (Figure 3A, B). These results suggest that the cytoplasmic pool of Calnuc participates in the intracellular trafficking of lysosomal receptors. To evaluate the impact of Calnuc depletion on CI-MPR and Sortilin stability, we performed cycloheximide chase experiments in which the levels of CI-MPR and Sortilin in control and Calnuc siRNA-treated cells were examined at different times after protein synthesis inhibition (Figure 2C). CI-MPR and Sortilin were degraded more rapidly in the Calnuc-depleted cells than in the control cells (Figure 2C). In mock cells, the half-lives of CI-MPR and Sortilin were greater than 6 h, whereas the half-lives of these receptors were 85 Figure 2 - CI-MPR and Sortilin levels are reduced when Calnuc is depleted due to missorting to lysosomes. HeLa cells were transfected with control or Calnuc siRNA for 3 days and Myc-Sortilin cDNA was transfected 12 h before the experiment. (A) Confocal microscopy analysis of 86 CI-MPR and Myc-Sortilin distribution in control or Calnuc-depleted cells. Cells were fixed with PFA and immunostained with anti-CI-MPR (a, d), anti-Myc (b, e) and antiCalnuc (c, f) and confocal images were acquired using identical instrument settings. CIMPR and Myc-Sortilin localized mainly in the perinuclear region in control cells, and Calnuc knockdown induced a decrease in the intensity level of both receptors. Bars, 10 μm. (B). Western blot analysis of control and Calnuc-depleted cell lysates. Proteins were separated by SDS-PAGE and immunoblotted with anti-CI-MPR, anti-Myc, antiTransferrin receptor (TfR), anti-Early endosome antigen 1 (EEA1) and anti-Calnuc. The steady-state levels of endogenous CI-MPR and Myc-Sortilin, but not TfR, were reduced in Calnuc-depleted cells. (C) The half-lives of CI-MPR and Sortilin are shorter in Calnucdepleted cells than in control cells. Hela cells were incubated with 40 µg/ml of cycloheximide for the time indicated, and total cell lysates were subjected to western blot analysis, as described in (B). (D) Endogenous CI-MPR is degraded in lysosomes in Calnuc-depleted cells. Control and Calnuc-depleted HeLa cells were treated or not with the lysosomal inhibitors leupeptin (1 mg/ml), E64 (20 μg/ml) and pepstatin (100 μM) for 5 h. The cells were lysed, and the proteins were analyzed by western blotting, as described in (B). EEA1 was used as loading control. (E) Bar graph showing the quantification of CI-MPR degradation in cells treated as described in (D) and expressed as a percentage of the CI-MPR present in control cells. The results are shown as the means ± SD (n = 6). There is no significant difference between columns 1 and 3 (P>0.05); **p < 0.01 (compared with control cells). reduced to <4 h in Calnuc siRNA-treated cells (Figure 2C). To determine whether this degradation occurred in lysosomes, we incubated the siRNA-treated cells with the lysosomal inhibitors leupeptin, pepstatin and E64 for 5 h. As shown in Figures 2D and 2E, treatment with lysosomal inhibitors significantly prevented the degradation of endogenous CI-MPR. Forty-one percent of the CI-MPR pool was degraded in the Calnuc knockdown cells in the absence of lysosomal inhibitors, whereas the CI-MPR level was similar to that in the control cells in the presence of lysosomal inhibitors (Figure 2E). Lysosomal inhibitor treatment thus significantly prevented a decrease in CI-MPR levels. Taken together, these observations suggest that depleting Calnuc increases the delivery of CI-MPR and Sortilin to lysosomes. The shorter half-lives of these lysosomal receptors should result in their depletion from the TGN, leading to the missorting of newly synthesized lysosomal cargo proteins, such as Cathepsin D or prosaposin. An immunoblot analysis indicated increased amounts of the precursor and intermediate forms of Cathepsin D in the Calnuc-depleted cells compared to the control cells (Supplementary Fig. S1), which was indicative of impaired transport of the enzyme to lysosomes. Overall, these results are consistent with 87 the role of Calnuc in regulating the endosomal sorting of lysosomal receptors and in the proper delivery of their cargo. Figure 3 - Expression of the cytoplasmic form of Calnuc rescues the lysosomal degradation of CI-MPR. (A) Steady state levels of CI-MPR are increased in Calnuc-depleted cells rescued with siRNA-resistant full-length Calnuc-GFP or cytoplasmic Calnuc-GFP (ΔSP-Calnuc-GFP) unlike cells rescued with GFP. HeLa cells treated with human Calnuc siRNA were transfected with GFP, siRNA-resistant Calnuc-GFP or ΔSP-Calnuc-GFP at 5 h after knockdown. The cells were lysed 3 days later, and the proteins were analyzed by western blotting. Endogenous human Calnuc was below the detection level in the samples treated with the Calnuc-specific siRNA. (B) Bar graph showing the quantification of CI-MPR degradation in cells treated as described in (A) and expressed as a percentage of CI-MPR present in control siRNA cells. Results are shown as means ± SD (n=4). *p<0.05 (compared with control siRNA cells). There is no significant difference between columns 1 and 3 (P>0.05) and columns 1 and 4 (P>0.05). Calnuc is required for the recruitment of retromers to endosomes. Given that Calnuc depletion induces the misdelivery and degradation of CI-MPR and Sortilin to lysosomes, we reasoned that Calnuc is specifically involved in the endosomal sorting step that targets these receptors to retrograde carriers. The key component of the machinery involved in endosomal cargo capture and retrieval is the heteropentameric 88 retromer complex. Accordingly, retromer depletion leads to an effect on CI-MPR and Sortilin stability and localization that is similar to that observed upon Calnuc depletion (Arighi et al. 2004; Rojas et al. 2007). Therefore, we examined whether Calnuc depletion affects the intracellular distribution of retromer subunits in HeLa cells. Two HeLa cell populations were transfected with either control or Calnuc-specific siRNA. After 48 hours, the two populations of cells were mixed and subsequently fixed and stained with antiVps26, anti-SNX2 and anti-Calnuc antibodies (Figure 4A). This allows a direct side-byside comparison of control versus Calnuc-suppressed cells by confocal microscopy. A decrease in the intensity of endogenous hVps26 and SNX2 immunofluorescence staining was observed in Calnuc-depleted cells compared to control cells (Figure 4A). Z-stack imaging was performed on these cells to confirm that the difference in Vps26 and SNX2 staining was observed in all the focal planes of the cells (Supplementary Fig. S2). This observation was evaluated quantitatively using an ad hoc algorithm that was previously developed to specifically detect and quantify the fluorescence intensity of endosomes (Mamo, et al., 2012). This quantification revealed a 35% and 30% decrease in SNX2 and hVps26 signals, respectively, on endosomes in the Calnuc-depleted cells compared to the control cells (Figure 4B). In contrast, no difference was observed in the immunofluorescence staining intensity of the early endosomal marker EEA1 between the control and Calnuc knockdown cells (Figure 4A, B), indicating that the observed effect on retromers was not due to non-specific or general effects of Calnuc depletion on endosomes. Furthermore, the observed decrease in endosomal retromer intensity was not due to degradation because immunoblot analyses did not reveal differences in SNX1, SNX2, hVps26 or hVps35 protein levels between the control and Calnuc knockdown cells (Figure 4C). Altogether, these results suggest a specific role for Calnuc in the trafficking of lysosomal receptors by controlling the recruitment of retromers to endosomal membranes. We next investigated whether the observed effect of Calnuc on retromer recruitment was due to the direct action of Calnuc on the retromer machinery. We first examined whether Calnuc colocalizes with SNX2 and hVps35 retromer components using confocal microscopy. Because retromer subunits are cytosolic proteins, the immunofluorescence experiments were performed using the cytosolic form of Calnuc, 89 Figure 4 - Calnuc depletion alters the recruitment of retromers to endosomes. (A) Knocking down Calnuc affects the intracellular distribution of retromers but not of EEA1. Mixed control and Calnuc siRNA-treated HeLa cells were fixed, permeabilized and triple-labeled with anti-hVps26 or anti-SNX2 together with anti-EEA1 and antiCalnuc antibodies followed by Alexa Fluor-488, 594 and 647-conjugated antibodies. The stained cells were then examined by confocal fluorescence microscopy. Scale bar, 10 μm. 90 (B) Quantification of the relative fluorescence intensities of retromers on endosomes in control and Calnuc-depleted cells. The data represent the relative fluorescence intensities of SNX2, Vps26 and EEA1 for more than 15,000 endosomes per condition. The results are expressed as a percentage of arbitrary units and are shown as the means ± SD (n = 3). * p < 0.05 compared with control siRNA cells. (C) The levels of endogenous retromers are not affected by the depletion of Calnuc. HeLa cells were transfected as described in (A) and lysed. The proteins were separated by SDS-PAGE and detected with specific anti-SNX1, anti-SNX2, anti-hVps26, anti-hVps35 and anti-EEA1 and anti-Calnuc antibodies. (D) Calnuc colocalizes with retromers on endosomes. COS7 cells transfected with ΔSP-Calnuc-GFP (cytoplasmic form of Calnuc) in the absence (a) or presence of Myc-hVps35 (b) were fixed, permeabilized and immunostained with anti-GFP (a, b), antiSNX2 (a) and anti-Myc (b). The stained cells were then examined by confocal fluorescence microscopy. ΔSP-Calnuc-GFP was found at the plasma membrane, the cytosolic side of the Golgi and small vesicular structures in the cytoplasm (insets). Endogenous SNX2 (a) and Myc-hVps35 (b) were found on endosomes distributed throughout the cell. The merged image shows partial overlap in the vesicular distribution of ΔSP-Calnuc-GFP and SNX2 (a) or Myc-hVps35 (b). Scale bar, 10 μm. ΔSP-Calnuc-GFP, transfected either alone (Figure 4D(a)) or with Myc-hVps35 (Figure 4D(b)) into COS7 cells. Both endogenous SNX2 (Figure 4D(a)) and Myc-hVps35 subunits (Figure 4D(b)) were located on punctated structures corresponding to Golgi surrounding vesicles and endosomes. A partial colocalization between ΔSP-Calnuc-GFP and endogenous SNX2 (Figure 4D(a)) as well as with Myc-hVps35 (Figure 4D(b)) was observed on endosomes. However, we were unable to detect an interaction between Calnuc and any subunit of the retromer complex (data not shown). Although these results negate the possibility that Calnuc is an adaptor that recruits retromers to endosomes, they do suggest that Calnuc has an indirect role by possibly acting on the machinery that regulates the recruitment of retromers. Calnuc regulates the localization and activity of Rab7. To identify the molecular mechanism by which Calnuc influences the recruitment of retromers, we next investigated Calnuc’s function with regard to the machinery known to recruit retromers to endosomal membranes. Recently, it was shown that the activation and cooperation of Rab5 and Rab7 are required for the recruitment of Vps retromers to endosomes as well as for lysosomal receptor retrieval (Rojas, et al., 2008; Seaman, et al., 2009). Because we obtained similar effects with Calnuc depletion, we first examined the role of Calnuc in Rab5 and Rab7 localization and activity. 91 Figure 5 - Calnuc partially interacts and colocalizes with Rab7 on endosomes. (A) Comparison of the intracellular distribution of Calnuc with Rab5, Rab7 or Rab9. COS7 cells transfected with ΔSP-Calnuc-Flag (cytoplasmic form of Calnuc) together with HA-tagged Rab5, Rab7 or Rab9 were fixed, permeabilized and immunostained with anti-Flag (a, d, g), and anti-HA (b, e, h). ΔSP-Calnuc-GFP was distributed at the plasma membrane, the cytosolic side of the Golgi and small vesicular structures in the cytoplasm 92 (insets) (a, d, g). Rabs were found on endosomes distributed throughout the cell. The merged images show partial overlap in the vesicular distribution of ΔSP-Calnuc-GFP and Rab5 (c) or Rab7 (f) and also Rab9 (i). Scale bar, 10 μm. (B) Coimmunoprecipitation of Flag-tagged Calnuc with HA-tagged Rab7 but not with HA-tagged Rab5 or HA-tagged Rab9. Lysates from HEK cells transfected with Calnuc-Flag together with HA, 3xHARab5, HA-Rab7 or HA-Rab9 were immunoprecipitated with an anti-Flag antibody and then immunoblotted with anti-HA or anti-Flag to detect Rab proteins or Calnuc, respectively. The 3xHA tag added ≈3-4 kDa to the MW of Rab5 which explain its higher molecular weight. (C) The interaction of Calnuc and Rab7 is independent of the activation status of Rab7. Lysates from HEK cells transfected with Calnuc-Flag together with HA, wild-type HA-Rab7, dominant-active HA-Rab7Q67L or dominant-inactive HARab7T22N were immunoprecipitated and immunoblotted as described in (B). Using confocal microscopy, we examined the presence of Calnuc on Rab5- and Rab7-labeled endosomes in COS7 cells (Figure 5A). Rab9 was also examined because it is involved in late endosome-to-TGN trafficking (Lombardi, et al., 1993) but is not involved in retromer recruitment (Seaman, et al. 2009). ΔSP-Calnuc-Flag was mainly localized on the plasma membrane, Golgi region and cytoplasm (Figure 5A(a,d,g)). A minor pool of ΔSP-Calnuc-Flag was also detected on some punctate structures distributed throughout the cytoplasm that partially overlapped with HA-tagged Rab5 (Figure 5A(c)), HA-tagged Rab7 (Figure 5A(f)) as well as HA-tagged Rab9 (Figure 5A(i)). We next investigated whether Calnuc could form a protein complex with Rab5, Rab7 or Rab9 (Figure 5B). HEK293 cells were transfected with Calnuc-Flag together with HA-Rab5, HA-Rab7 or HA-Rab9. Following immunoprecipitation with anti-Flag antibodies, an interaction between CalnucFlag and HA-Rab7 but not between Calnuc-Flag and HA-Rab5 or HA-Rab9 was observed (Figure 5B), suggesting that Calnuc interacts with Rab7 on endosomes. However, using in vitro translation pull-down assays, we were unable to detect a direct interaction between Calnuc and Rab7 (data not shown). We next examined whether the Calnuc/Rab7 interaction depends on the activation state of Rab7. We tested the interaction of Calnuc-Flag with the dominant-active form of Rab7 (HA-Rab7Q67L) and the dominant-inactive form of Rab7 (HA-Rab7T22N) (Figure 5C). Coimmunoprecipitation experiments showed that wild-type, active and inactive Rab7 bound Calnuc at similar levels (Figure 5C), suggesting that the Rab7 activation state did not influence this interaction. Similarly, to determine whether the interaction of Calnuc and Rab5 is GTP-dependent, we examined the interaction of Calnuc 93 Figure 6 - Rab7 localization and activity is altered in Calnuc-depleted cells. (A) Calnuc knockdown alters the intracellular distribution of Rab7. HeLa cells stably expressing GFP-Rab7 were transfected with control or Calnuc siRNA. After 48 hours, the two populations of cells were mixed and subsequently fixed, permeabilized and triple94 labeled with anti-GFP, anti-EEA1 and anti-Calnuc. The stained cells were then examined by confocal fluorescence microscopy. Scale bar, 10 μm. (B) The fluorescence intensity of endosomal GFP-Rab7 and EEA1 was quantified as described in Figure 4B. The data represent the relative fluorescence intensities of GFP-Rab7 and EEA1 for more than 15,000 endosomes per condition. The results are expressed as a percentage of arbitrary units and are shown as the means ± SD (n = 3). ** p < 0.001 compared with control siRNA cells. (C) Calnuc depletion does not affect the total levels of GFP-Rab7 or EEA1. Lysates from cells treated with control or Calnuc-specific siRNA were analyzed by western blotting with specific anti-GFP, anti-EEA1 and anti-Calnuc antibodies. (D) Depletion of Calnuc increases the amount of Rab7, hVps26 and SNX2 in the cytosolic fraction. HeLa cells treated with control or Calnuc siRNA were fractionated into membrane (P) and cytosolic (S) fractions as described in Material and Methods, followed by immunobloting with antibodies against Rab7, Vps26, SNX2, Calnuc, EEA1, and TfR. (E) Quantification of the amount of Rab7 and Vps26 in the cytosolic and membrane fractions of control and Calnuc-depleted cells. Data (n=3) are presented as percent of total protein (S+P). and dominant-active (Rab5Q79L) and inactive (Rab5S34N) forms of Rab5 (Supplementary Fig. S3). The inability to detect an interaction negates the possibility that Calnuc is a Rab5 effector. These results suggest that Calnuc is not an effector of Rab5 and Rab7 but may be part of the Rab7 activation machinery. To investigate the function of the Calnuc/Rab7 interaction, we used confocal microscopy to compare the intracellular distribution of Rab7 in HeLa cells stably expressing GFP-tagged Rab7 and transfected with control or Calnuc siRNA (Figure 6A). After mixing the two cell populations for a direct side-by-side staining comparison, a decrease in the intensity of GFP-Rab7 immunofluorescence staining on punctate structures was observed in the Calnuc-depleted cells compared to the control cells (Figure 6A). Zstack imaging was performed on these cells to confirm that the difference in Rab7 signal intensity was observed in all focal planes of the cells (Supplementary Fig. S4). Quantification of the endosomal fluorescence intensity revealed a 28% decrease in the Rab7 signal on endosomes in the Calnuc-depleted cells compared to the control cells (Figure 6B), whereas no difference was observed in the endosomal intensity of EEA1 in the same cells (Figure 6A, B). Western blotting using whole-cell lysates revealed that the total amounts of GFP-Rab7 in the control and Calnuc-depleted cells were similar (Figure 6C), suggesting that the change in Rab7 endosomal intensity was not due to the deregulation of 95 Rab7 protein synthesis or degradation. To confirm that the redistribution of Rab7 observed by immunofluorescence resulted from a change in the amount of these proteins that was associated with membranes, HeLa cells treated with control or Calnuc siRNA were fractionated into membrane (pellet, P) and cytosolic (supernatant, S) fractions and analyzed by SDS-PAGE and immunoblotting (Figure 6D). The loss of Calnuc expression resulted in a significant shift of Rab7 to the supernatant fraction, consistent with a redistribution of this protein to the cytoplasm. In agreement and as shown in Figure 4, less membrane-bound and more cytosolic hVps26 and SNX2 was found in the Calnuc-depleted cells (Figure 6D), and this effect was found to be specific because Calnuc depletion did not affect the membrane distribution of other endosomal markers such as EEA1 (Figure 6D). The membrane protein TfR was detected in the pellet fractions, indicating that fractionation did not cause membrane into the supernatant fraction. The data from three separate experiments were quantified and are shown in Figure 6E. In control cells, 64% and 56% of Rab7 and Vps26, respectively, were found in the membrane fraction, whereas, in Calnuc-depleted cells, the membrane fractions of Rab7 and Vps26 were reduced to 45% and 43%, respectively. These results suggest a specific role for Calnuc in the recruitment of Rab7 and consequently of retromers to endosomal membranes. The recruitment of retromers depends on proper Rab5 and Rab7 activation/inactivation in both time and space, which are governed by their cycling between GTP- and GDP-bound forms (Rojas, et al., 2008). Because an interaction was detected between Calnuc and Rab7 (Figure 5B, C) and Calnuc depletion induced the cytoplasmic redistribution of Rab7 (Figure 6), we next investigated whether Calnuc influences Rab7 activation. To test this hypothesis, we used an effector pull-down assay in which the Rab7binding domain of its effector protein RILP (Rab-interacting lysosomal protein) was used to selectively isolate GTP-loaded Rab7 (Romero Rosales, et al., 2009). Recombinant GSTRILP immobilized on glutathione beads was incubated with lysates from control and Calnuc-depleted HeLa cells, and bound (active) Rab7 proteins were analyzed by immunoblotting (Figure 7A). The data from three separate experiments were quantified and indicated that GTP-loaded Rab7 was reduced by 66% in Calnuc-depleted cells compared to control cells (Figure 7B). Taken together, these results suggest that Calnuc regulates the 96 recruitment and activation of Rab7, which is directly involved in the endosomal recruitment of retromers. Figure 7 - Knocking down Calnuc affects the activation of Rab7. (A) Lysates from HeLa cells treated with control or Calnuc-specific siRNA were incubated with GST or GST-RILP, which specifically interacts with the active (GTPbound) form of Rab7. Bound proteins were then analyzed by western blotting with antiRab7. The lower panel is Ponceau stained blot showing GST-tagged proteins. (B) Bar graph showing the quantification of Rab7 associated with GST-RILP in cells treated as described in (A) and expressed as a percentage of the Rab7 present in input lysates. The results are shown as the means ± SD (n = 3). ** p < 0.001 (compared with control cells). Discussion In this study, we demonstrate that Calnuc plays a role in the endosomal sorting of the lysosomal receptors CI-MPR and Sortilin through the modulation of Rab7 activity and the recruitment of retromers to endosomes. Our results show that a small proportion of Calnuc indirectly interacts with the cytoplasmic tails of CI-MPR and Sortilin and partially colocalizes with these receptors on endosomes. Furthermore, Calnuc depletion (by siRNA) causes the misdelivery and degradation of these receptors in lysosomes, leading to a defect in the maturation and secretion of the lysosomal enzyme cathepsin D. These data suggest a role for Calnuc in the endosomal sorting of lysosomal receptors for proper retrieval to the TGN. Using siRNA rescue experiments, we identified the cytoplasmic pool of Calnuc as the functional pool 97 involved in the endosomal sorting of lysosomal receptors. These results are reminiscent of our previous work demonstrating that cytoplasmic Calnuc plays a role in the endosomal retrieval of LRP9, an LDLR subtype that also cycles between the TGN and endosomes (Brodeur, et al., 2009), and suggest that Calnuc is involved in the general mechanism of this endosomal-sorting pathway. In accordance, our results suggest the implication of Calnuc in the recruitment of retromers to endosomes, which is key for the retrograde trafficking of various receptors, including lysosomal sorting receptors (Arighi, et al., 2004; Rojas, et al., 2007; Rojas, et al., 2008). Indeed, as shown by immunofluorescence and membrane fractionation assays, Calnuc depletion resulted in a partial cytosolic redistribution of Vps26 and SNX2. The inability to detect an interaction between Calnuc and retromer subunits negates the possibility that Calnuc is an adaptor that recruits retromers to endosomal membrane. We thus investigated Calnuc’s role on the machinery that recruits retromers onto endosomes, i.e., the small G-proteins Rab5 and Rab7. The activation and cooperation of Rab5 and Rab7 are required for both the recruitment of the SNX and Vps components of retromers to endosomes and for lysosomal receptor retrieval (Rojas, et al., 2008; Seaman, et al., 2009). Our results indicate that Calnuc binds to Rab7 but not to Rab5 or Rab9. Although the interaction between Calnuc and Rab7 is not direct, we observed that Calnuc plays a role in Rab7 endosomal recruitment and activity. The loss of Calnuc led to a reduction in the amount of Rab7-GTP that was in concordance with a decrease in the amount of Rab7 found on endosomal membranes. It has been reported that the depletion of Rab7 or the expression of a constitutively inactive form of Rab7 decreases the association of Vps26 and Vps35 on endosomes, but the association of SNX1 and SNX2 is either unaffected or less dramatically altered than that of the Vps proteins (Rojas, et al., 2008; Seaman, et al., 2009; Vardarajan, et al., 2012). Calnuc depletion induced a partial cytosolic redistribution of the retromer Vps26 and SNX2 subcomplexes, supporting its action on Rab7 and the possibility that Calnuc could affect another effector acting on SNX subcomplex recruitment or stability. The endosomal PI3P levels regulate the membrane recruitment of the SNX subcomplex (Carlton, et al., 2004; Cozier, et al., 2002; Rojas, et al., 2008). Interestingly, calcium/calmodulin have been suggested to impact Vps34/PI3K on early endosomes (Vergne, et al., 2003). Therefore, we examined whether the loss of Calnuc alters the endosomal PI3P levels. Using a FYVE-GFP probe that specifically recognizes 98 PI3P, we did not observe any clear difference between FYVE-GFP labeling in control and Calnuc-depleted cells (data not shown). Furthermore, no difference was observed in the immunofluorescence staining intensity of the early endosomal marker EEA1 (which contains a FYVE domain) between control and Calnuc-knockdown cells (Figure 4A, B and Figure 6A, B). These results suggest that the activity of PI3K is not altered in Calnucdepleted cells, and thus, that Calnuc does not alter SNX2 membrane recruitment by modulating the level of endosomal PI3P. An attractive hypothesis is that the fraction of Calnuc protein that binds to lysosomal receptors at the endosome targets Rab7 activity and retromer recruitment to sites where the receptors are localized. Using various truncation mutants of CI-MPR cytoplasmic tail, we were unable to determine the domain/motif of interaction given that multiple domains interacted with Calnuc (data not shown), preventing the generation of a CI-MPR mutant that cannot bind Calnuc. This could be due to the fact that Calnuc does not bind directly to the cytoplasmic tail of CI-MPR. Indeed, further work is required to elucidate how the association between Calnuc and lysosomal receptors is mediated. Therefore, it is unclear from our results, whether Calnuc actively modulates and recruits the retromer sorting machinery to lysosomal receptors on endosomes. In addition to its role in retromer-dependent endosome-to-TGN retrograde transport, Rab7 has traditionally been involved in transport to late endosomes and lysosomes (Mukhopadhyay et al. 1997; Press et al. 1998; Vonderheit and Helenius 2005; Ceresa and Bahr 2006; Vanlandingham and Ceresa 2009; Zhang et al. 2009). In accordance with Rab7 involvement in protein transport pathways that lead both toward and away from lysosomes, the interference with Rab7 function can lead to the accumulation of CI-MPR in endosomes (Rojas, et al., 2008). In contrast, retromer depletion results in targeting of the CI-MPR to lysosomes (Arighi, et al., 2004; Carlton, et al., 2004; Rojas, et al., 2007; Seaman, 2004), which is caused by a lack of lysosomal receptor removal from maturing endosomes, thus leading to their passive missorting to lysosomes. It is noteworthy that delivery to lysosomes is the fate of the CI-MPR and Sortilin in the absence of Calnuc, although we showed that Rab7 activity was altered. This observation could be explained by the strictly limited action 99 of Calnuc on Rab7 to sites where the retromer is present, thereby allowing Rab7 to regulate fusion of endosomes with lysosomes at other sites within the endocytic pathway or by the exposition of sorting determinants within these lysosomal receptors that would direct its targeting to lysosomes. Whereas there is considerable divergence of sorting nexin proteins between species, the cargo-selective components (CSC) Vps26/Vps29/Vps35 heterotrimer is strikingly conserved throughout eukaryotic evolution (Koumandou, et al., 2011) and is considered to constitute the core functional component of retromer. Indeed, these components were first identified in the budding yeast Saccharomyces cerevisiae and name “class A” vacuolar protein sorting (Vps) proteins and shown as essential to select cargo for endosome-to-Golgi retrieval of the CPY-sorting receptor, Vps10p. The “class B” Vps5p and Vps17p, the yeast SNX1 and SNX2, were identified at the same time and shown to assemble onto the membrane to promote vesicle/tubule formation (Seaman, et al., 1998). Conservation of the yeast retromer complex components in higher eukaryotes suggests an important general role for this complex in endosome protein sorting. However, since their first identification in yeast, a variety of systems have identified multiple cargo proteins that require retromer for their localization, and accessory proteins that function with CSC retromer in endosomal protein sorting. For example, the small GTPase Rab7 associates directly with the CSC complex to mediate its localization to endosomes in yeast and mammals (Balderhaar, et al., 2010; Liu, et al., 2012; Priya, et al., 2015; Rojas, et al., 2008; Seaman, et al., 2009). TBC1D5 regulates the membrane association of retromer through a direct interaction with Vps29 (Harbour, et al., 2010) and via its GAP activity on Rab7 (Seaman, et al., 2009). Recent studies also demonstrated the role of SNX3, a member of SNX family of proteins that do not contain a BAR domain, in retromer recruitment via direct interactions with the CSC subcomplex (Harrison, et al., 2014; Harterink, et al., 2011). However, SNX3 is part of an alternative retromer pathway that mediates endosome sorting of specific cargo proteins such as Wntless and transferrin receptor (Chen, et al., 2013 ; Harterink, et al., 2011 ; Zhang, et al. 2011). Furthermore, unlike Rab7, SNX3 is not conserved across all eukaryotes being absent in plants suggesting that it is dispensable in some eukaryotes (Vardarajan, et al., 2012). Eps15-homology domain 1 (EHD1), as well as EHD-1 interacting protein 100 Rabankyrin-5, have also been shown to associate with the CSC, and have a regulatory role in retromer-mediated endosome-to-Golgi retrieval (Gokool, et al., 2007; Zhang, et al., 2012). In this study, we identify Calnuc as a novel accessory protein that regulates retromer recruitment to endosomes, possibly by impacting Rab7 activity. It is noteworthy that yeast does not seem to have a Calnuc homolog. Therefore, while Calnuc seems to be involved in the recruitment of retromers in mammalian cells, it is dispensable in simpler organisms (e.g. yeast). It suggests that Calnuc serves as a regulatory module in the endosome sorting machinery in higher organisms, rather than an evolutionarily conserved core component. Rab GTPases such as Rab7 cycle between GTP-bound, membrane-associated, and GDP-bound cytosolic states and, in doing so, function to regulate the recruitment of effector proteins, such as retromers, to membranes. Our results show a significant decrease in the active GTP-bound form of Rab7 in Calnuc-depleted cells. However, the manner by which Calnuc alters the amount of active Rab7 requires further clarification. Given that Calnuc does not directly interact with Rab7, these results could be explained by the action of Calnuc on a GDP Exchange Factor (GEF) of Rab7. An alteration of the activity of Rab7GEF would impair the exchange of GDP for GTP, leading to an increase in the amount of inactive Rab7 and its dissociation from membranes. The only GEF identified for Rab7 is Mon1-Ccz1 (Gerondopoulos, et al., 2012; Nordmann, et al., 2010), and it is possible that Calnuc regulates the recruitment of this GEF to localize and activate Rab7. Alternatively, Calnuc could regulate the activity and recruitment of TBC1D5, a Rab7-GAP that negatively regulates retromer recruitment and causes Rab7 to dissociate from membranes (Seaman, et al., 2009). A recent paper has also linked the activity of ceroid-lipofuscinosis neuronal 5 (CLN5) to the regulation of Rab7 endosomal recruitment (Mamo, et al., 2012). CLN5 is mutated in a rare type of lysosomal storage disease named neuronal ceroid lipofuscinosis (NCL), and the loss of CLN5 expression results in reduced levels of active Rab7 and compromised retromer recruitment, leading to degradation of the lysosomal receptors Sortilin and CI-MPR (Mamo, et al., 2012). Because CLN5 depletion phenotypes are reminiscent of that observed upon Calnuc depletion, it will be of interest to test whether Calnuc can interact with and regulate the activity of CLN5. If true, these results would imply a potential role of Calnuc in ceroid lipofuscinosis pathogenesis. 101 Recent studies have shown that the receptor Sortilin regulates the trafficking and processing of amyloid-β (Aβ), a key neurotoxic peptide in Alzheimer’s disease (AD). Sortilin binds to amyloid precursor protein (APP) and thereby promotes the production of a nontoxic soluble product instead of Aβ (Gustafsen, et al., 2013). Moreover, Sortilindeficient mice have increased brain Aβ levels and plaque burden compared with wild-type mice (Carlo, 2013). LRP10 (i.e., LRP9 in mice), another APP receptor that regulates APP trafficking and cleavage into Aβ (Brodeur, et al., 2012), has previously been shown to interact with Calnuc (Brodeur, et al., 2009). As reported for Sortilin and CI-MPR in the present study, the loss of Calnuc leads to the missorting to lysosomes and degradation of LRP10 (Brodeur, et al., 2009). Moreover, the expression of the LRP10 protein is significantly reduced in the brains of AD patients (Brodeur, et al., 2012). Thus, Calnuc expression affects the trafficking and levels of two APP receptors that protect against the amyloidogenic processing of APP and accumulation of Aβ. Interestingly, Calnuc levels are also significantly reduced in the brains of AD patients, and the modulation of Calnuc expression in neuronal cell lines modulates endogenous APP levels (Lin et al. 2007). Our results suggest that the action of Calnuc on Rab7 and retromer recruitment could be involved in the molecular mechanisms by which Calnuc affects APP-sorting receptors and amyloidogenic cleavage. In agreement with this, the loss of retromer function is known to increase the amyloidogenic processing of APP, and studies of tissues from AD patients indicate that retromer expression could be downregulated in neuronal tissue (Small, et al. 2005; Muhammad, et al. 2008; Lane, et al. 2010; Wen, et al. 2011). By altering retromer recruitment, Calnuc is likely to play an important role in AD pathogenesis. In conclusion, the results presented here identify Calnuc as a novel player in the Rab7/retromer-mediated endosomal sorting machinery that determines whether lysosomal receptors are recycled back to the Golgi or degraded in lysosomes. Consequently, an alteration in Calnuc expression or activity may affect the proper sorting of CI-MPR and Sortilin receptors and their cargo and may be potentially involved in the pathological dysfunction that leads to lysosomal storage or AD. 102 Acknowledgments We are grateful to Matthew Seaman (University of Cambridge) for the CD8-MPR and CD8-Sortilin cDNA, the truncation mutants of CI-MPR cytoplasmic tail as well as the HeLa cells stably expressing GFP-Rab7, to Claus M. Petersen (University of Aarhus) for untagged Sortilin cDNA and to Francis S. Lee (Weill Medical College of Cornell University) for Myc-Sortilin cDNA, to Jean-Luc Parent (Université de Sherbrooke) for the HA-Rab9 cDNA, to Aimee Edinger (UC Irvine) for pGEX-RILP construct. We also thank Juan S. Bonifacino (NIH) for Myc-hVps35 cDNA and anti-Vps35 pAbs and Thomas Braulke (University of Hamburg) for the anti-MPR pAbs. We wish to thank Marilène Paquette for her technical assistance. H.L. holds a fellowship from the National Sciences and Engineering Research Council of Canada (NSERC). This work was supported by grants from the Canadian Institutes for Health Research and a Canada Research Chair to C.L.L. The authors declare that they have no conflict of interest. 103 References Arighi CN, Hartnell LM, Aguilar RC, Haft CR, Bonifacino JS. 2004. Role of the mammalian retromer in sorting of the cation-independent mannose 6-phosphate receptor. J Cell Biol 165(1):123-33. Balderhaar HJ, Arlt H, Ostrowicz C, Brocker C, Sundermann F, Brandt R, Babst M, Ungermann C. 2010. The Rab GTPase Ypt7 is linked to retromer-mediated receptor recycling and fusion at the yeast late endosome. J Cell Sci 123(Pt 23):4085-94. Ballif BA, Mincek NV, Barratt JT, Wilson ML, Simmons DL. 1996. Interaction of cyclooxygenases with an apoptosis- and autoimmunity-associated protein. Proc Natl Acad Sci U S A 93(11):5544-9. Bonifacino JS, Lippincott-Schwartz J. 2003. Coat proteins: shaping membrane transport. Nat Rev Mol Cell Biol 4(5):409-14. Bonifacino JS, Rojas R. 2006. Retrograde transport from endosomes to the trans-Golgi network. Nat Rev Mol Cell Biol 7(8):568-79. Boucher R, Larkin H, Brodeur J, Gagnon H, Theriault C, Lavoie C. 2008. Intracellular trafficking of LRP9 is dependent on two acidic cluster/dileucine motifs. Histochem Cell Biol 130(2):315-27. Braulke T, Bonifacino JS. 2009. Sorting of lysosomal proteins. Biochim Biophys Acta 1793(4):605-14. Brodeur J, Larkin H, Boucher R, Theriault C, St-Louis SC, Gagnon H, Lavoie C. 2009. Calnuc binds to LRP9 and affects its endosomal sorting. Traffic 10(8):1098-114. Brodeur J, Theriault C, Lessard-Beaudoin M, Marcil A, Dahan S, Lavoie C. 2012. LDLRrelated protein 10 (LRP10) regulates amyloid precursor protein (APP) trafficking and processing: evidence for a role in Alzheimer's disease. Mol Neurodegener 7:31. Bucci C, Thomsen P, Nicoziani P, McCarthy J, van Deurs B. 2000. Rab7: a key to lysosome biogenesis. Mol Biol Cell 11(2):467-80. Carlo AS. 2013. Sortilin, a novel APOE receptor implicated in Alzheimer disease. Prion 7(5):378-82. Carlton J, Bujny M, Peter BJ, Oorschot VM, Rutherford A, Mellor H, Klumperman J, McMahon HT, Cullen PJ. 2004. Sorting nexin-1 mediates tubular endosome-to-TGN transport through coincidence sensing of high- curvature membranes and 3phosphoinositides. Curr Biol 14(20):1791-800. Chen C, Garcia-Santos D, Ishikawa Y, Seguin A, Li L, Fegan KH, Hildick-Smith GJ, Shah DI, Cooney JD, Chen W and others. 2013. Snx3 regulates recycling of the transferrin receptor and iron assimilation. Cell Metab 17(3):343-52. Christoforidis S, Miaczynska M, Ashman K, Wilm M, Zhao L, Yip SC, Waterfield MD, Backer JM, Zerial M. 1999. Phosphatidylinositol-3-OH kinases are Rab5 effectors. Nat Cell Biol 1(4):249-52. Cozier GE, Carlton J, McGregor AH, Gleeson PA, Teasdale RD, Mellor H, Cullen PJ. 2002. The phox homology (PX) domain-dependent, 3-phosphoinositide-mediated association of sorting nexin-1 with an early sorting endosomal compartment is required for its ability to regulate epidermal growth factor receptor degradation. J Biol Chem 277(50):48730-6. Feng Y, Press B, Wandinger-Ness A. 1995. Rab 7: an important regulator of late endocytic membrane traffic. J Cell Biol 131(6 Pt 1):1435-52. 104 Gerondopoulos A, Langemeyer L, Liang JR, Linford A, Barr FA. 2012. BLOC-3 mutated in Hermansky-Pudlak syndrome is a Rab32/38 guanine nucleotide exchange factor. Curr Biol 22(22):2135-9. Gokool S, Tattersall D, Seaman MN. 2007. EHD1 interacts with retromer to stabilize SNX1 tubules and facilitate endosome-to-Golgi retrieval. Traffic 8(12):1873-86. Gustafsen C, Glerup S, Pallesen LT, Olsen D, Andersen OM, Nykjaer A, Madsen P, Petersen CM. 2013. Sortilin and SorLA display distinct roles in processing and trafficking of amyloid precursor protein. J Neurosci 33(1):64-71. Harbour ME, Breusegem SY, Antrobus R, Freeman C, Reid E, Seaman MN. 2010. The cargo-selective retromer complex is a recruiting hub for protein complexes that regulate endosomal tubule dynamics. J Cell Sci 123(Pt 21):3703-17. Harrison MS, Hung CS, Liu TT, Christiano R, Walther TC, Burd CG. 2014. A mechanism for retromer endosomal coat complex assembly with cargo. Proc Natl Acad Sci U S A 111(1):267-72. Harterink M, Port F, Lorenowicz MJ, McGough IJ, Silhankova M, Betist MC, van Weering JR, van Heesbeen RG, Middelkoop TC, Basler K and others. 2011. A SNX3dependent retromer pathway mediates retrograde transport of the Wnt sorting receptor Wntless and is required for Wnt secretion. Nat Cell Biol 13(8):914-23. Koumandou VL, Klute MJ, Herman EK, Nunez-Miguel R, Dacks JB, Field MC. 2011. Evolutionary reconstruction of the retromer complex and its function in Trypanosoma brucei. J Cell Sci 124(Pt 9):1496-509. Lane RF, Raines SM, Steele JW, Ehrlich ME, Lah JA, Small SA, Tanzi RE, Attie AD, Gandy S. 2010. Diabetes-associated SorCS1 regulates Alzheimer's amyloid-beta metabolism: evidence for involvement of SorL1 and the retromer complex. J Neurosci 30(39):13110-5. Lavoie C, Meerloo T, Lin P, Farquhar MG. 2002. Calnuc, an EF-hand Ca(2+)-binding protein, is stored and processed in the Golgi and secreted by the constitutive-like pathway in AtT20 cells. Mol Endocrinol 16(11):2462-74. Lin P, Fischer T, Lavoie C, Huang H, Farquhar MG. 2009. Calnuc plays a role in dynamic distribution of Galphai but not Gbeta subunits and modulates ACTH secretion in AtT-20 neuroendocrine secretory cells. Mol Neurodegener 4:15. Lin P, Fischer T, Weiss T, Farquhar MG. 2000. Calnuc, an EF-hand Ca(2+) binding protein, specifically interacts with the C-terminal alpha5-helix of G(alpha)i3. Proc Natl Acad Sci U S A 97(2):674-9. Lin P, Le-Niculescu H, Hofmeister R, McCaffery JM, Jin M, Hennemann H, McQuistan T, De Vries L, Farquhar MG. 1998. The mammalian calcium-binding protein, nucleobindin (CALNUC), is a Golgi resident protein. J Cell Biol 141(7):1515-27. Lin P, Yao Y, Hofmeister R, Tsien RY, Farquhar MG. 1999. Overexpression of CALNUC (nucleobindin) increases agonist and thapsigargin releasable Ca2+ storage in the Golgi. J Cell Biol 145(2):279-89. Liu TT, Gomez TS, Sackey BK, Billadeau DD, Burd CG. 2012. Rab GTPase regulation of retromer-mediated cargo export during endosome maturation. Mol Biol Cell 23(13):2505-15. Lombardi D, Soldati T, Riederer MA, Goda Y, Zerial M, Pfeffer SR. 1993. Rab9 functions in transport between late endosomes and the trans Golgi network. Embo J 12(2):67782. 105 Mamo A, Jules F, Dumaresq-Doiron K, Costantino S, Lefrancois S. 2012. The role of ceroid lipofuscinosis neuronal protein 5 (CLN5) in endosomal sorting. Mol Cell Biol 32(10):1855-66. McGough IJ, Cullen PJ. 2011. Recent advances in retromer biology. Traffic 12(8):963-71. Miura K, Titani K, Kurosawa Y, Kanai Y. 1992. Molecular cloning of nucleobindin, a novel DNA-binding protein that contains both a signal peptide and a leucine zipper structure. Biochem Biophys Res Commun 187(1):375-80. Muhammad A, Flores I, Zhang H, Yu R, Staniszewski A, Planel E, Herman M, Ho L, Kreber R, Honig LS and others. 2008. Retromer deficiency observed in Alzheimer's disease causes hippocampal dysfunction, neurodegeneration, and Abeta accumulation. Proc Natl Acad Sci U S A 105(20):7327-32. Ng EL, Gan BQ, Ng F, Tang BL. 2012. Rab GTPases regulating receptor trafficking at the late endosome-lysosome membranes. Cell Biochem Funct 30(6):515-23. Nordmann M, Cabrera M, Perz A, Brocker C, Ostrowicz C, Engelbrecht-Vandre S, Ungermann C. 2010. The Mon1-Ccz1 complex is the GEF of the late endosomal Rab7 homolog Ypt7. Curr Biol 20(18):1654-9. Priya A, Kalaidzidis IV, Kalaidzidis Y, Lambright D, Datta S. 2015. Molecular Insights into Rab7-Mediated Endosomal Recruitment of Core Retromer: Deciphering the Role of Vps26 and Vps35. Traffic. Rojas R, Kametaka S, Haft CR, Bonifacino JS. 2007. Interchangeable but essential functions of SNX1 and SNX2 in the association of retromer with endosomes and the trafficking of mannose 6-phosphate receptors. Mol Cell Biol 27(3):1112-24. Rojas R, van Vlijmen T, Mardones GA, Prabhu Y, Rojas AL, Mohammed S, Heck AJ, Raposo G, van der Sluijs P, Bonifacino JS. 2008. Regulation of retromer recruitment to endosomes by sequential action of Rab5 and Rab7. J Cell Biol 183(3):513-26. Romero Rosales K, Peralta ER, Guenther GG, Wong SY, Edinger AL. 2009. Rab7 activation by growth factor withdrawal contributes to the induction of apoptosis. Mol Biol Cell 20(12):2831-40. Seaman MN. 2004. Cargo-selective endosomal sorting for retrieval to the Golgi requires retromer. J Cell Biol 165(1):111-22. Seaman MN, Harbour ME, Tattersall D, Read E, Bright N. 2009. Membrane recruitment of the cargo-selective retromer subcomplex is catalysed by the small GTPase Rab7 and inhibited by the Rab-GAP TBC1D5. J Cell Sci 122(Pt 14):2371-82. Seaman MN, McCaffery JM, Emr SD. 1998. A membrane coat complex essential for endosome-to-Golgi retrograde transport in yeast. J Cell Biol 142(3):665-81. Small SA, Kent K, Pierce A, Leung C, Kang MS, Okada H, Honig L, Vonsattel JP, Kim TW. 2005. Model-guided microarray implicates the retromer complex in Alzheimer's disease. Ann Neurol 58(6):909-19. Stenmark H, Parton RG, Steele-Mortimer O, Lutcke A, Gruenberg J, Zerial M. 1994. Inhibition of rab5 GTPase activity stimulates membrane fusion in endocytosis. Embo J 13(6):1287-96. Vardarajan BN, Bruesegem SY, Harbour ME, Inzelberg R, Friedland R, St George-Hyslop P, Seaman MN, Farrer LA. 2012. Identification of Alzheimer disease-associated variants in genes that regulate retromer function. Neurobiol Aging 33(9):2231 e152231 e30. 106 Vergne I, Chua J, Deretic V. 2003. Tuberculosis toxin blocking phagosome maturation inhibits a novel Ca2+/calmodulin-PI3K hVPS34 cascade. J Exp Med 198(4):653-9. Wang T, Ming Z, Xiaochun W, Hong W. 2011. Rab7: role of its protein interaction cascades in endo-lysosomal traffic. Cell Signal 23(3):516-21. Weiss TS, Chamberlain CE, Takeda T, Lin P, Hahn KM, Farquhar MG. 2001. Galpha i3 binding to calnuc on Golgi membranes in living cells monitored by fluorescence resonance energy transfer of green fluorescent protein fusion proteins. Proc Natl Acad Sci U S A 98(26):14961-6. Wen L, Tang FL, Hong Y, Luo SW, Wang CL, He W, Shen C, Jung JU, Xiong F, Lee DH and others. 2011. VPS35 haploinsufficiency increases Alzheimer's disease neuropathology. J Cell Biol 195(5):765-79. Zhang J, Reiling C, Reinecke JB, Prislan I, Marky LA, Sorgen PL, Naslavsky N, Caplan S. 2012. Rabankyrin-5 interacts with EHD1 and Vps26 to regulate endocytic trafficking and retromer function. Traffic 13(5):745-57. Zhang P, Wu Y, Belenkaya TY, Lin X. 2011. SNX3 controls Wingless/Wnt secretion through regulating retromer-dependent recycling of Wntless. Cell Res 21(12):167790. 107 Supporting Information Supp. Figure S1 - Depletion of Calnuc causes missorting of the lysosomal enzyme Cathepsin D. Lysates from cells treated with control or Calnuc siRNA were analyzed by western blotting with specific anti-Cathepsin D (CatD), anti-CI-MPR, anti-Calnuc and anti-Actin antibodies. Precursor (p), intermediate (i) and mature (m) forms of Cathepsin D are shown. 108 Supp. Figure S2 - Confocal Z-stack images of cells presented in Figure 4A. Cells were prepared as described in Figure 4A. Series of confocal z-stack images were acquired at 0.12 μm (A) or 0.2 μm (B) intervals. The last panels display merged z stacks and signal intensities. The color bar ranges from low intensity (black) to high intensity (white). Scale bar, 10 μm. 109 Supp. Figure S3 - The interaction of Calnuc and Rab5 is independent of the activation status of Rab5. Lysates from HEK cells transfected with Calnuc-Flag together with HA, wild-type HARab5, dominant-active HA-Rab5Q79L or dominant-inactive HA-Rab5S34N were immunoprecipitated with an anti-Flag antibody and then immunoblotted with anti-HA or anti-Flag to detect Rab5 proteins or Calnuc, respectively. Supp. Figure S4 - Confocal Z-stack images of cells presented in Figure 6A. Cells were prepared as described in Figure 6A. Series of confocal z-stack images were acquired at 0.32 μm intervals. The last panels display merged z stacks and signal intensities. The color bar ranges from low intensity (black) to high intensity (white). Scale bar, 10 μm. 110 ARTICLE 2 Topology and membrane anchoring of the lysosomal storage disease-related protein CLN5 Auteurs de l’article: Heidi Larkin, Maria Gil Ribeiro et Christine Lavoie Statut de l’article: Publié Human mutation, vol. 34, no 12 (déc. 2013), p. 1688-1697 Avant-propos: J'ai écrit la première version des sections 'Matériel et méthodes' ainsi que des descriptions des figures. J'ai réalisé toutes les expériences retrouvées dans l'article ainsi que tous les montages. Christine Lavoie a effectué les corrections du texte et a rédigé l'introduction, les résultats et la discussion. Elle a également effectué l'expérience présentée en figure 2C et les schémas 4A-C. Maria Gil Ribeiro a fourni l'anticorps CLN5-C/32. 111 RÉSUMÉ – ARTICLE 2 Topologie et ancrage membranaire de CLN5, une protéine associée à une maladie du lysosome Certaines variantes infantiles tardives de céroïdes-lipofuscinoses neuronales, une maladie neurodégénérative, sont causées par des mutations dans le gène CLN5. CLN5 code une glycoprotéine lysosomiale dont la structure et la fonction n’ont pas encore été clairement définies. Dans la présente étude, nous utilisons diverses constructions de CLN5 comprenant des étiquettes afin de déterminer la topologie et la solubilité de la protéine. Notre étude indique que CLN5 est synthétisée sous forme d’une glycoprotéine transmembranaire de type II, comprenant un domaine N-terminal cytoplasmique, un segment transmembranaire et un grand domaine C-terminal luminal contenant une hélice amphipathique. Les domaines cytoplasmique et transmembranaire sont rapidement éliminés suivant le clivage du peptide signal de manière à former la protéine CLN5 mature. Cette dernière demeure fortement associée du côté luminal de la membrane par une hélice amphipathique. Les mutants pathologiques de CLN5 dépourvus de cette hélice amphipathique perdent leur association membranaire, sont retenus au réticulum endoplasmique et sont rapidement dégradés par la machinerie du protéasome. Nous avons défini, de façon expérimentale, la topologie de CLN5 et nous avons démontré l’existence d’une hélice amphipathique qui permet l’ancrage de la protéine à la membrane. Notre travail met en lumière les propriétés de base de CLN5 permettant une meilleure compréhension de ses fonctions biologiques et de son implication dans la pathogenèse des céroïdes-lipofuscinoses neuronales. 112 Article 2 Topology and Membrane Anchoring of the Lysosomal Storage Disease-Related Protein CLN5 Heidi Larkin,1 Maria Gil Ribeiro,2,3 and Christine Lavoie1∗ 1 Department of Pharmacology, Faculty of Medicine and Health Sciences, Universite de Sherbrooke, Sherbrooke, Quebec, Canada 2 Research and Development Unit, Department of Genetics, National Health Institute Dr. Ricardo Jorge, Porto, Portugal 3 Faculty of Health Sciences, Fernando Pessoa University, Porto, Portugal Communicated by Elizabeth F. Neufeld ∗Correspondence to: Christine Lavoie, Department of Pharmacology, Faculty of Medicine, Universite de Sherbrooke, 3001-12e Avenue Nord, Sherbrooke, QC J1H 5N4, Canada E-mail: [email protected] Received 28 May 2013 Accepted revised manuscript 6 September 2013. Published online 13 September 2013 in Wiley Online Library (www.wiley.com/humanmutation). DOI: 10.1002/humu.22443 Keywords: neuronal ceroid lipofuscinoses (NCL);CLN5;topology;amphipathic helix;transmembrane domain 113 Early view Elucidation of the topology of human CLN5, a lysosomal protein involved in neuronal ceroid lipofuscinosis pathogenesis. CLN5 is synthesized as a single pass type II transmembrane precursor glycoprotein. The signal peptide (red) is rapidly cleaved, and the resulting mature protein (blue) is tightly associated with the lumen of the membrane through an amphipathic anchor region (AH). Pathological mutants deprived of AH (pGlu253* and pGlu352*) lose their membrane association and are rapidly degraded. Abstract One late infantile variant of the neurodegenerative disease neuronal ceroid lipofuscinosis (NCL) is caused by a mutation in the CLN5 gene. CLN5 encodes a lysosomal glycoprotein whose structure and function have not yet been clearly defined. In the present study, we used epitope-tagged CLN5 to determine the topology and solubility of the CLN5 protein. Our results indicated that CLN5 is synthesized as a type II transmembrane (TM) glycoprotein with a cytoplasmic N-terminus, one TM segment, and a large luminal C-terminal domain containing an amphipathic helix (AH). The cytoplasmic and TM domains were rapidly removed following signal-peptide cleavage, and the resulting mature CLN5 was tightly associated with the lumen of the membrane through the AH. CLN5 pathological mutants deprived of AH lose their membrane association, are 114 retained in the endoplasmic reticulum, and are rapidly degraded by the proteasomal machinery. We experimentally define the topology of CLN5 and demonstrate the existence of an AH that anchors the protein to the membrane. Our work sheds light on the basic properties of CLN5 required to better understand its biological functions and involvement in NCL pathogenesis. Introduction Neuronal ceroid lipofuscinoses (NCLs) are severe neurodegenerative lysosomal storage diseases that manifest mostly in childhood and are characterized by developmental regression, seizures, visual failure, and mental retardation, all of which culminate in premature death [Mole et al., 2005]. NCLs are relatively rare worldwide (1:100,000 live births), but are more common (1:12,500 live births) in certain North American and Northern European populations [Haltia, 2006; Moore et al., 2008; Santavuori, 1988]. These disorders result from mutations in at least 13 ceroid lipofuscinosis neuronal genes (CLN1– CLN14) [Kollmann et al., 2013]. Mutations in the human CLN5 gene (MIM #608102) result in late infantile, juvenile, and adult forms of NCL [Mole et al., 2011]. To date, over 20 disease-causing mutations have been identified in this gene (http://www.ucl.ac.uk/ncl). CLN5 encodes a 407 amino acid (aa) polypeptide with a predicted molecular mass of 46 kDa [Savukoski et al., 1998]. However, CLN5 expressed in cells has been reported to have a higher molecular weight ranging from 50 to 75 kDa due to eight potential Nglycosylation sites [Isosomppi et al., 2002; Schmiedt et al., 2010; Vesa et al., 2002]. CLN5 also contains a signal peptide (SP) that is cleaved in the endoplasmic reticulum (ER) and the mature polypeptide is mainly localized in lysosomes [Holmberg et al., 2004; Isosomppi et al., 2002; Schmiedt et al., 2010; Vesa et al., 2002]. However, the full topology of CLN5 is not known, and its solubility versus membrane-spanning property has been a controversial issue. CLN5 was initially described as having two transmembrane (TM) domains and has been detected in a Triton X-114 (TX-114) TM protein fraction [Bessa et al., 2006; Savukoski et al., 1998; Vesa et al., 2002]. However, it has also been described as a soluble protein and has been detected in culture media [Holmberg et al., 2004; Sleat et al., 115 2006b]. To date, none of the CLN5 disease-causing mutations studied have been shown to have an effect on CLN5 synthesis and maturation, but they are known to affect the lysosomal targeting of CLN5 [Isosomppi et al., 2002; Lebrun et al., 2009; Schmiedt et al., 2010; Vesa et al., 2002]. While CLN5 was identified 15 years ago [Savukoski et al., 1998], its cellular function remains poorly understood. Previous studies have shown that CLN5 interacts with other NCL proteins [Lyly et al., 2009; Vesa et al., 2002], but the purpose of these interactions remains elusive. Moreover, recent studies have shown that CLN5 interacts with the lysosomal-sorting receptor sortilin and affects its endosome-to-TGN (trans-Golgi Network) retrieval by regulating retromer membrane recruitment [Mamo et al., 2012]. This provided the first link between CLN5 and lysosomal function. To better understand the impact of CLN5 mutations and the role and mechanism of the action of CLN5, some of its basic properties, including its topology and solubility, need to be elucidated. By introducing epitope tags on the N- or C-terminus or an internal portion of CLN5, we determined that CLN5 is synthesized as a single pass type II TM protein that is cleaved after its SP to produce a mature luminal form of CLN5 that remains tightly associated with the membrane through an amphipathic anchor region. Materials and Methods Antibodies Anti-Calnexin polyclonal antibody (pAb) was from Millipore (Billerica, MA), antiCalnuc pAb was from Aviva Systems Biology (San Diego, CA), anti-EEA1 pAb was from Thermo Scientific (Rockford, IL), anti-Flag monoclonal antibody (mAb) was from Sigma– Aldrich (Saint Louis, MO), anti-GFP mAb was from Clontech (Mountain View, CA), antiGM130 mAb was from BD Transduction Laboratories (Franklin Lakes, NJ), anti-LAMP1 (H4A3) mAb was from Developmental Studies Hybridoma Bank (University of Iowa, IA), anti-HA mAb and pAb were from Covance (Emeryville, CA), and anti-transferrin receptor (TfR) mAb was from Invitrogen (Carlsbad, CA). CLN5-C/32 rabbit pAb directed against 116 the C-terminus extremity of human CLN5 has been previously described [Schmiedt et al., 2010]. DNA Constructs Human 3xHA-CLN5 in the pReceiver-M06 vector was a generous gift from Dr. Stéphane Lefrançois (Université de Montréal, QC, Canada). Untagged CLN5, CLN5-HA, and SP-Flag-CLN5 were generated by subcloning in the pcDNA3.1 vector. The SP-FlagCLN5 construct was obtained by the insertion of the Flag epitope at amino acid position 97. The pathological mutants EUR (European) (c.835G>A, p.Asp279Asn), the FinM (Finnish major) (c.1175_1176delAT, p.Tyr392*), the SWE (Swedish) (c.757G>T, p.Glu253*), and the NFL (Newfoundland) (c.1054G>T, p.Glu352*) were generated from SP-Flag-CLN5 by QuickChange site-directed mutagenesis as previously described [Schmiedt et al., 2010]. Nucleotide numbering reflects cDNA numbering, with +1 corresponding to the A of the ATG translation initiation codon in the reference sequence (GenBank NM_006493.2) according to journal guidelines (www.hgvs.org/mutnomen). The initiation codon is codon 1. Cell Culture and Transfection HeLa cells were from ATCC (Manassas, VA) and HEK293T cells were from Dr. Alexandra Newton (UCSD, CA). The cells were grown in DMEM high glucose medium (Invitrogen) containing 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT), penicillin, and streptomycin. HEK cells were transfected with Lipofectamine 2000 transfection reagent (Invitrogen), and HeLa cells with Fugene6 transfection reagent (Roche Diagnostic, Indianapolis, IN), both according to the manufacturers’ instructions. In Vitro Translation Assay In vitro translation of human CLN5 pcDNA3.1 was performed using the TNT T7 Quick Coupled Transcription/Translation system (Promega, Madison, WI) in the presence or absence of canine pancreatic microsomes (Promega) according to the manufacturer's protocol. 117 Pulse-Chase Analysis HEK293 cells transiently expressing the various tagged CLN5 proteins were incubated for 30 min in methionine-free DMEM. The cells were pulsed for 5 min with 150 μCi/ml of [35S]methionine/cysteine (>1000 Ci/mmol; EasyTagTM Expression Protein Labeling Mix, Perkin Elmer, Waltham, MA), washed in PBS, and chased for 0, 0.5, 1, 2, or 4 hr in complete culture medium. Cells and medium were collected after each chase. The medium was cleared by centrifugation (15,800g for 30 min). Tagged CLN5 proteins were immunoprecipitated from the cell lysates or medium, separated by SDS-PAGE, and detected by autoradiography. Subcellular Fractionation and Extraction Assays HEK293 cells were washed in cold PBS, delicately scraped into TNES buffer (50 mM tris, pH 7.4, 150 mM NaCl, 5 mM EDTA, 250 mM sucrose, and protease inhibitors), and centrifuged for 3 min at 500g. The pellet was homogenized in TNES buffer by 30 passages through a 25 G 5/8 needle. Nuclei and unbroken cells were removed by centrifugation at 1,000g for 10 min. The postnuclear supernatant (PNS) was collected and was centrifuged at 100,000g for 1 hr at 4°C. The soluble fraction (S100) was collected, and the pellet or membrane fraction (P100) was resuspended using a pestle in an equal volume of TNES buffer. Approximately 200 μg of P100 was then incubated with 500 μL of 1 M NaCl, or 150 mM Na2CO3 (pH 11.5), or 5 M urea for 30 min on ice. The solution was centrifuged at 100,000g for 1 hr at 4°C. The soluble fraction (SP100) was precipitated with 10% trichloroacetic acid (TCA) overnight at −80°C. The precipitate (PP100) was directly resuspended in loading buffer. The same volumes of S100 and P100 and the entire SP100 and PP100 fractions were analyzed by SDS-PAGE and immunoblotting. TX-114 Fractionation TX-114 fractionation was performed on the P100 membrane fraction using MemPER Eukaryotic Membrane Protein Extraction Reagent kits from Thermo Scientific (Waltham, MA) using the manufacturer's protocol [Qoronfleh et al., 2003]. The aqueous phase was precipitated with 10% TCA overnight at −80°C, whereas the detergent phase was precipitated using 1 ml of ethanol/ether (4:1) overnight at −20°C. The fractions were 118 centrifuged at 15,800g for 20 min and the pellets were resuspended in SDS-PAGE-loading buffer. Treatment with Glycosidase and Proteinase K Deglycosylation was carried out according to the manufacturer's instructions. Briefly, 50 μg of cell lysate were incubated with 1 unit of PNGase F (NEB, Ipswich, MA) overnight at 37°C. For the topology assays, 50 μg of P100 membrane fraction were incubated with 100 μg/ml of proteinase K (Invitrogen) in the absence or presence of 0.5% TX-100. The protease reactions were stopped by adding 1 mM phenylmethylsulfonyl fluoride (Sigma– Aldrich) and incubating for 10 min on ice. Immunoblotting The protein samples were separated on SDS-PAGE gels and were transferred to nitrocellulose membranes. The membranes were blocked in tris-buffered saline containing 0.1% Tween 20 and 5% nonfat dry milk and were incubated with primary antibodies for 1 hr at room temperature (RT) followed by horseradish peroxidase-conjugated goat antirabbit or anti-mouse immunoglobulin G (Bio-Rad, Richmond, ON) for 1 hr at RT and then with enhanced chemiluminescence (ECL) detection reagent (Pierce Chemical, Rockford, IL). Immunofluorescence HeLa cells were plated on coverslips. Twelve hours after transfection, the cells were fixed and permeabilized in cold methanol for 10 min. The cells were blocked with 10% goat or bovine fetal serum (FBS) for 30 min and were incubated with primary antibodies for 1 hr at RT followed by Alexa Fluor-conjugated secondary antibodies (Molecular Probes, Eugene, OR). The specimens were visualized using an inverted confocal laserscanning microscope FV1000 (Olympus, Tokyo, Japan) equipped with a PlanApo 60x/1.42 oil immersion objective. Olympus Fluoview software version 1.6a was used to acquire and analyze the images, which were further processed using Adobe Photoshop (Adobe Systems, San Jose, CA). 119 Results Tagging the CLN5 Protein Does not Affect Its Maturation, Secretion, or Localization To clarify the topology of CLN5, three different epitope-tagged human CLN5 constructs were generated. The first had three HA tags at the N-terminus of the first initiation methionine of CLN5 (3xHA-CLN5), the second had an HA tag at the C-terminus (CLN5-HA), and the third had a Flag tag after the putative SP cleavage site (SP-FlagCLN5) (Fig. 1A). To validate the reliability of these tools, we first confirmed that the epitope-tagged forms of CLN5 were expressed and processed appropriately. To that end, wild-type (WT; untagged) and epitope-tagged CLN5 were transiently expressed in HeLa cells, and their molecular weights (MW) were analyzed by Western blotting. Since CLN5 contains an SP cleavage site and eight potential N-glycosylation sites (Fig. 1A), the size of the protein detected depends on the fragment recognized by the antibody and the level of glycosylation. In Figure 1B, proteins were detected using either antibodies directed against the Flag or HA tags or a specific CLN5 peptide antibody (α-C/32) directed against the Cterminal region of CLN5 (aa 393–407) that has already been used to detect the mature form of overexpressed untagged CLN5 [Schmiedt et al., 2010]. In addition, portions of cell lysates were treated with PNGase F to remove all N-linked oligosaccharide side chains from the CLN5 proteins. As previously reported [Schmiedt et al., 2010], untagged CLN5 is detected as a 60-kDa polypeptide in the absence of PNGase F and a 35-kDa polypeptide in the presence of PNGase F, which correspond to the glycosylated and unglycosylated cleaved (mature) forms of CLN5 (96–407 aa fragment), respectively. CLN5-HA and SPFlag-CLN5 detected with antibodies directed against their tags had the same expression pattern as untagged CLN5 in the absence and presence of PNGase F, suggesting that the insertion of these tags did not affect the cleavage or glycosylation of CLN5. N-terminustagged 3xHA-CLN5 was detected as three bands of ∼73, 50, and 15 kDa corresponding to the predicted MW of the glycosylated and unglycosylated full-length (uncleaved or precursor) forms of CLN5 and the SP fragment (aa 1–96), respectively. This is in agreement with the results of Vesa et al. (2002), who detected a 75-kDa glycosylated polypeptide and a 47-kDa unglycosylated polypeptide using antibodies raised against aa 1– 75 of CLN5. Using the α-C/32 antibody, we also detected the mature 60 kDa fragment in 120 Figure 1 - Epitope-tagged CLN5 are appropriately expressed, processed, and secreted. A: Schematic representation of the different CLN5 cDNA constructs. Putative TM domains are represented by hatched boxes, the SP cleavage site as a dotted line, and the Nglycosylation sites by stars. HA or Flag epitope tags (represented by black boxes) were inserted at the N- or C-terminus as well as after the putative SP (aa 96) of CLN5. B: Analysis of expression patterns of untagged and epitope-tagged CLN5 by Western blotting. Lysates from HeLa cells transiently transfected with the untagged or tagged CLN5 were 121 incubated in the absence or presence of PNGase F. The proteins were separated by SDSPAGE and were immunoblotted with anti-HA, anti-Flag, or anti-CLN5 C/32 antibody. Schematic representations of the corresponding proteins are indicated on the left or right side of the bands. The 3xHA tag added ∼3–4 kDa to the MW of CLN5, whereas the HA and Flag tags added ∼1 kDa. C: Analysis of the maturation and secretion of untagged and tagged-CLN5. HEK293 cells transiently expressing CLN5-HA, SP-Flag-CLN5, or untagged CLN5 were pulsed for 5 min with 35S-methionine and were chased for the indicated periods of time. Tagged and untagged CLN5 proteins were immunoprecipitated from the cell lysates, and culture media was collected after each chase period using antiHA, anti-Flag, or anti-C/32 antibodies. The proteins were separated by SDS-PAGE and analyzed by autoradiography. *, Nonspecific band; black arrowhead, precursor polypeptide; white arrowhead, mature polypeptide; boxed lane, longer exposure of the 0 min chase of SP-Flag-CLN5 to show the presence of precursor and mature polypeptides. the 3xHA-CLN5 cell lysate (data not shown), confirming that the appropriate cleavage and glycosylation had occurred. These results indicated that the different epitope-tagged CLN5 proteins were correctly expressed, cleaved, and glycosylated. To confirm that the epitope-tagged forms of CLN5 were appropriately localized, the intracellular distributions of 3xHA-CLN5, CLN5-HA, SP-Flag-CLN5, and untagged CLN5 expressed in HeLa cells were next analyzed by confocal microscopy (Supp. Fig. S1). Schmiedt et al. (2010) reported that the precursor form or N-terminal fragment of GFPCLN5 is retained in the ER, whereas the mature form of CLN5 exits the ER and is mainly targeted to lysosomes. We observed that the intracellular distribution of CLN5-HA and SPFlag-CLN5 was very similar to the untagged CLN5 stained with anti-C/32 antibody and partially colocalized with the lysosomal marker LAMP1, which is in agreement with the results of Schmiedt et al. (2010). These three CLN5 proteins were also detected in a juxtanuclear compartment corresponding to TGN since they partially colocalized with TGN46 (data not shown). On the other hand, 3xHA-CLN5 colocalized exclusively with the ER marker Calnexin. These results indicated that the different epitope-tagged CLN5 were appropriately distributed in the cells. WT CLN5 has also previously been reported to be secreted [Isosomppi et al., 2002]. To confirm that the maturation and secretion of epitope-tagged CLN5 were not altered, we next compared the kinetics of synthesis and secretion of SP-Flag-CLN5 and CLN5-HA to 122 that of untagged CLN5 using a pulse-chase assay. HEK293 cells expressing SP-FlagCLN5, CLN5-HA, or untagged CLN5 were labeled for 5 min with [35S]methionine, chased for up to 2 hr, immunoprecipitated from the cell lysate and the medium, separated by SDSPAGE, and detected by autoradiography (Fig. 1C). Similar kinetics was obtained with tagged and untagged CLN5 proteins. After a 0-min chase, untagged and tagged CLN5 were detected as a ∼70-kDa glycosylated precursor and a ∼60-kDa processed/mature form of CLN5 in the cell lysates (Fig. 1C; intracellular, black and white arrowheads, respectively). The weaker signal obtained with SP-Flag-CLN5 may be due to less efficient detection of the internal epitope tag. With longer chases, the 70-kDa precursor of tagged and untagged CLN5 was rapidly processed, giving rise to the 60-kDa mature form, which was still detected in the cell lysate after a 2-hr chase (Fig. 1C, intracellular). The fact that the 60-kDa mature form of CLN5 was observed after a 5-min pulse indicated that the N-glycosylation and SP cleavage of CLN5 are cotranslational events, as previously suggested by Schmiedt et al. (2010). Moreover, beginning with a 1-hr chase, small amounts of CLN5-HA, SPFlag-CLN5, and untagged CLN5 were detected in the extracellular medium (Fig. 1C, extracellular), indicating that mature CLN5 is partially secreted. This was confirmed with a chase at 20°C, which is known to arrest protein transport in the TGN [Matlin and Simons, 1983], that completely blocked the secretion of tagged and untagged CLN5 (Supp. Fig. S2). In summary, the presence of tags did not markedly alter the kinetics of maturation and secretion of CLN5. The small difference observed may be due to slight intracellular retention or slower degradation of the tagged CLN5 proteins compared with their WT counterpart or to differences in antibody affinity or native protein epitope recognition. Overall, these results confirmed that tagging CLN5 does not alter its intracellular maturation, secretion, or localization since the tagged CLN5 behaved mostly as untagged CLN5. Therefore, these tagged CLN5 proteins represent suitable and valuable tools for addressing key issues regarding the basic properties of CLN5. 123 Determination of the Topology of CLN5 A bioinformatic analysis of the CLN5 sequence using online SACS and ExPASy TM prediction tools identified up to two hydrophobic regions predicted to function as membrane-spanning regions (Fig. 1A). One TM domain was unanimously predicted at residues ranging from 73 to 93, whereas a second potential TM domain (aa 353–373) was only predicted by a few tools. Two models can thus be proposed based on these predictions and on the presence of an SP cleavage site at position 96 (Fig. 2A). In one model, CLN5 encodes two TM regions with the N- and C-termini in the cytoplasm and an intraluminal loop (Fig. 2A, Model 1). In the second model, CLN5 contains one TM, a cytoplasmic Nterminus, and an intraluminal C-terminus (Fig. 2A, Model 2). To determine the topological orientation of the amino- and carboxyl-termini of CLN5, we performed a protease protection assay using the previously characterized N-terminally, C-terminally, and internally tagged CLN5 proteins (Fig. 1). Since the protease cannot cross the membrane, regions or tags in the lumen of intact compartments would be protected from digestion, whereas the cytoplasmic regions or tags would be digested. Crude membrane compartment fractions isolated from HeLa cells expressing the different tagged CLN5 were incubated with proteinase K in the presence or absence of TX-100 and were analyzed by Western blotting (Fig. 2B). In the absence of TX-100, the N-terminus HA tags were digested by the protease, resulting in the complete loss of immunoreactivity, whereas the C-terminus HAtag and the Flag-tag inserted after the SP were not digested. However, when TX-100 was added to the proteinase K digestion mixture, these two tags were also digested (Fig. 2B). These results indicated that the N-terminus tail of CLN5 is located in the cytoplasm, whereas its C-terminus tail is located in the lumen, as depicted in Model 2 (Fig. 2A). The topology of CLN5 was confirmed using a cell-free protein expression system containing canine pancreatic microsomes. This in vitro system, which functionally and topologically mimics the ER, accurately reproduces the translation, membrane insertion and translocation, SP cleavage, and N-linked glycosylation of membrane and secreted proteins that occurs in vivo. Protease protection assays can thus be used with this system to determine the topology of membrane proteins. Untagged CLN5 was translated in vitro in the absence or presence of microsomal membranes and was analyzed by SDS-PAGE and 124 Figure 2 - Topology of CLN5. A: Schematic representation of two models representing a possible topology of CLN5. Model 1 depicts two TM regions with cytoplasmic N- and C-termini and an intraluminal loop. Model 2 depicts one TM region, a cytoplasmic N-terminus, and an intraluminal Cterminus. B: Membrane fractions isolated from HeLa cells transiently expressing the tagged CLN5 proteins were incubated in the absence or presence of proteinase K, with or without 125 TX-100. The proteins were separated by SDS-PAGE and were immunoblotted with antiHA or anti-Flag antibody. Endogenous luminal calnuc and cytoplasmic EEA1 were used as controls. C: Untagged CLN5 cDNA subcloned into the pcDNA3.1 vector was translated in vitro with [35S]methionine using reticulocytes in the absence or presence of microsomes. The translated proteins were incubated in the absence or presence of PNGase F or proteinase K, with or without TX-100. The proteins were separated by SDS-PAGE and were analyzed by autoradiography. Schematic representations of the corresponding bands are indicated on the right side of the gel. autoradiography. In the absence of microsomes, a major 48-kDa band was observed (Fig. 2C, lane I, band 1), which corresponded to the MW of the unglycosylated precursor form of CLN5. The lower MW bands most likely corresponded to CLN5 translated from other potential AUG codons (aa 30, 50, and 62), as previously reported [Isosomppi et al., 2002; Vesa et al., 2002], or to degradation products. In the presence of microsomes, additional bands with apparent MW of 70–72 kDa and 60 kDa were observed (Fig. 2C, lane II, bands 2 and 3, respectively). These bands corresponded to the expected MW of the glycosylated precursor and the mature forms (in which the SP has been cleaved) of CLN5, respectively. Indeed, the PNGase F treatment induced the deglycosylation and a ∼25-kDa downshift of the precursor and mature forms of CLN5, resulting in 48 and 34 kDa bands (Fig. 2C, lane III, bands 1 and 4, respectively). These results indicated that CLN5 is glycosylated and cleaved in the presence of microsomes. When the microsomes were treated with proteinase K, the 60-kDa mature form of CLN5 was protected (Fig. 2C, lane IV, band 3), whereas the 70–72-kDa precursor was shifted to an apparent MW of 65 kDa (Fig. 2C lane IV, band 5). This 8–9 kDa shift corresponded to the digestion of the cytoplasmic N-terminus of CLN5 (76 amino acids). This is in agreement with topology Model 2 (Fig. 2A). Furthermore, the unprocessed precursors and the degradation products observed in lane I were also degraded in the presence of proteinase K, indicating that they are located on the cytoplasmic side of the microsomes (Fig. 2C, lane IV). When TX-100 was added to the proteinase K treatment, the bands observed in lane IV became sensitive to digestion (Fig. 2C, lane V), showing that these proteins are located in the lumen. Taken together, these results indicated that the Nterminus of CLN5 is located in the cytoplasm, that its first hydrophobic region is a TM domain, that the second hydrophobic region does not cross the membrane, and that the protein C-terminus is located in the lumen of the microsomes. CLN5 is thus synthesized as a type II TM precursor protein that is rapidly glycosylated and cleaved at the SP site 126 located after the TM domain, with the resulting mature form of CLN5 being entirely located in the lumen of the intracellular compartments. CLN5 is Tightly Bound to the Membrane CLN5 has been previously reported to be a soluble protein [Holmberg et al., 2004; Isosomppi et al., 2002]. However, even though the mature form of CLN5 is devoid of a TM domain, we observed that only a small amount is secreted into the medium (Fig. 1C), suggesting that it is directly or indirectly attached to the membrane. We then investigated whether CLN5 was associated with the membrane using subcellular fractionation assays on HEK293 cells expressing the various tagged CLN5 proteins. Cell homogenates (PNS) were centrifuged at high speed, and the resulting pelleted membrane fractions (P) and supernatants containing soluble proteins (S) were analyzed by Western blotting using antibodies directed against the tagged CLN5 proteins. 3xHA-CLN5 was mainly found in the membrane fraction, as expected, since this precursor form contains a TM domain. However, the CLN5-HA and SP-Flag-CLN5 proteins, which do not have a TM region, were also mainly found in the membrane fraction (Fig. 3A). To determine the biochemical nature of the association of CLN5 with the membrane, we conducted a series of extraction experiments to discriminate between peripheral and integral membrane associations. The isolated membrane fraction described in Figure 3A was incubated with 1 M NaCl or 150 mM Na2CO3 (pH 11.5), two agents known to release loosely attached peripheral membrane proteins, or with 5 M urea, one chaotropic agent that can extract strongly attached peripheral membrane proteins (Fig. 3B). The membranes were centrifuged, and the pellets (P) and supernatants (S) were analyzed for the partitioning of tagged CLN5 proteins. As shown in Figure 3B, both the precursor and the mature forms of CLN5 remained associated with the membrane under all the extraction conditions. These results strongly suggested that mature CLN5 is tightly associated with membrane and behaves as an integral membrane protein. To further analyze the association of mature CLN5 with the membrane, TX-114 partitioning was performed to segregate soluble and membrane-bound proteins into the 127 Figure 3 - Mature CLN5 is tightly associated with the membranes. A: Membrane fractionation of tagged CLN5 proteins. HEK293 cells transiently expressing 3×HA-CLN5, CLN5-HA, or SP-Flag-CLN5 were lysed in TNES buffer and were centrifuged at low speed. The PNS was centrifuged at 100,000g to separate the membrane fraction (P) and the soluble proteins (S). The same volumes of S and P were separated by SDS-PAGE and were immunoblotted using anti-HA or anti-Flag antibody. B: Membrane subfractionation of tagged CLN5 proteins. Membranes (P) isolated as described in (A) were resuspended in TNES buffer containing 1 M NaCl, 150 mM sodium carbonate (pH 11.5), or 5 M urea, and were incubated for 30 min at 4°C. The membranes were sedimented at 100,000g. The supernatant (S) and pellet (P) fractions were separated by SDS-PAGE and were immunoblotted using anti-HA or anti-Flag antibody. C: TX-114 partitioning of tagged CLN5 proteins. Membranes (P) isolated as described in (A) were solubilized and were extracted using TX-114 membrane protein extraction reagent Mem-PER protein extraction kits. The supernatant (aqueous [AQ]) and pellet (detergent [Det]) fractions were separated by SDS-PAGE and were immunoblotted using anti-HA or anti-Flag antibody. Endogenous cytosolic protein EEA1, peripheral protein GM130, and transmembrane protein TfR were used as fractionation controls. aqueous phase and integral membrane proteins into the detergent phase [Bordier, 1981]. The membrane pellet fraction described in Figure 3A was dissolved in TX-114 buffer, incubated, and centrifuged. The two phases were then analyzed by Western blotting (Fig. 3C). The 3xHA-CLN5 precursor form containing a TM domain remained mainly membrane bound, whereas CLN5-HA and SP-Flag-CLN5 were distributed almost evenly between the aqueous and detergent fractions. These results indicated that the mature form of CLN5 is tightly bound to the membrane but that it does not behave like a typical integral membrane protein. 128 CLN5 is Anchored to the Membrane Through an Amphipathic Helix The aa 353–373 region of human CLN5, which was predicted to be a potential TM domain, contains a pattern of hydrophobic and nonpolar residues that is well conserved among CLN5 orthologues (Fig. 4A). We showed that this region is not a TM region. A helical wheel projection using the amino acid interfacial hydrophobicity scale described by Wimley and White (1996) revealed that the aa 353–392 fragment of CLN5 can adopt an amphipathic helix (AH) conformation, with two hydrophobic patches at each end (Fig. 4B). As depicted in the helical net diagram (Fig. 4C), these two patches lie on one face of the helix and can directly adsorb at the interface of the lipidic membrane. To test the hypothesis that the predicted AH mediates the association with the membrane, several deletion mutants of the mature form of CLN5 with or without the AH were generated and fused to the Cterminus of GFP (Fig. 4D). The membrane association of these GFP-tagged proteins expressed in HEK293 cells was examined by subcellular fractionation and urea extraction followed by Western blotting (Fig. 4E). All the GFP fusion proteins containing the AH domain (GFP-CLN5 353–392, GFP-CLN5 353–407, and GFP-CLN5 97–407) were mainly present in the membrane-enriched fraction, even after urea extraction, indicating that the AH is sufficient for membrane association. On the other hand, the GFP-fusion protein without the AH (GFP-CLN5 97–352) was found exclusively in the soluble fraction after the urea treatment, as was GFP alone. These results suggested that the AH is a key factor in membrane anchoring of the mature form of CLN5. Effect of Genetic Pathological Mutations on the Membrane Anchoring and Secretion of CLN5 To improve our understanding of the impact of NCL pathological mutations on the function of CLN5, we analyzed the membrane association of SP-Flag-CLN5 carrying four variant late infantile CLN5 disease-causing mutations (Fig. 5A): the EUR p.Asp279Asn mutation [Savukoski et al., 1998], the FinM p.Tyr392* mutation [Savukoski et al., 1998], the SWE p.Glu253* mutation [Holmberg et al., 2000], and the NFL p.Glu352* mutation [Moore et al., 2008]. The SWE and NFL mutations both resulted in truncated proteins without the AH and the C-terminal region, whereas the FinM mutant retained the AH but not the C-terminal region. The EUR mutant was included because it retained the AH and 129 Figure 4 - An AH in CLN5 mediates tight membrane association. A: Sequence comparison of the potential AH from different species. The numbers refer to amino acid positions in human CLN5. *, Fully conserved residues., similar residues. B: Helical wheel projection of CLN5 aa 353–392. The amino acids are color-coded based on the hydrophobic character of the residue: green for hydrophobic (W, F, Y, L, I, M, C), white for neutral (A, T, S, V, G), and red for hydrophilic and charged (E, D, K, R, H, Q, P, N). Two hydrophobic patches can be seen on the two sides of the helix. C: Helical net 130 projection of CLN5 aa 353–392. In these representations, the cylindrical helix region is cut longitudinally and flattened into the plane of the page. The amino acids are color-coded as described in (B). Two hydrophobic patches can also be seen in this diagram. D: Schematic representation of the different GFP-tagged CLN5 cDNA constructs. GFP-CLN5 97–407, the mature form of CLN5; GFP-CLN5 97–352, which does not contain the putative AH; GFP-CLN5 353–407, which contains the AH and the C-terminus of CLN5; and GFP-CLN5 353–392, which contains only the putative AH. E: Membrane association analysis of the different GFP-tagged CLN5 proteins. HEK293 cells transiently expressing the GFP or the GFP-CLN5 proteins described in (D) were subfractionated and treated with 5 M urea as described in Figure 3. The PNS, supernatant (S), and pellet (P) fractions were separated by SDS-PAGE and were immunoblotted using anti-GFP antibody. Cytosolic EEA1, peripheral GM130, and TM protein TfR were used as fractionation controls. Representative blots are shown. the C-terminus. The membrane associations of these pathological CLN5 mutants expressed in HEK293 cells were examined by subcellular fractionation and urea extraction followed by Western blotting (Fig. 5B). The EUR mutation did not affect the membrane association since it behaves like WT SP-Flag-CLN5. The FinM mutant protein was more sensitive to urea since it was partially found in the soluble fraction following urea extraction. In contrast, the SWE and NFL mutant proteins were found exclusively in the soluble fraction following urea extraction. These results suggested that CLN5 pathological mutants deprived of their AH are not tightly associated with the membrane. Pulse-chase assays were then performed to determine whether the mutations altered the secretion of CLN5. HEK293 cells expressing WT SP-Flag-CLN5 or carrying the EUR, FinM, SWE, or NFL mutations were pulsed for 5 min and were chased for different periods of time, after which CLN5 was immunoprecipitated from the medium and cell extracts (Fig. 5C). The EUR and FinM mutations did not affect the kinetics of maturation and secretion of CLN5 since these mutants behave like WT SP-Flag-CLN5. A deglycosylation analysis confirmed that the glycosylation of the EUR and FinM mutant proteins are not altered (Supp. Fig. S3). However, the SWE and NFL mutations affected the maturation and stability of CLN5 (Fig. 5C; intracellular). After a 0-min chase, bands were detected at the predicted MW of the SWE and NFL glycosylated precursors (37–39 and 60 kDa bands, respectively; Fig. 5C, black arrowheads) and glycosylated mature forms (∼29 and 50 kDa bands, respectively; Fig. 5C, white arrowheads), suggesting cotranslational glycosylation 131 Figure 5 - Membrane association and secretion of Flag-tagged CLN5 carrying the NCL pathological mutations. A: Schematic representation of the relative localization and amino acid changes of four pathological CLN5 mutations causing variant late infantile CLN5 disease: EUR (c.835G>A, p.Asp279Asn), FinM (c.1175_1176delAT, p.Tyr392*) SWE (c.757G>T, p.Glu253*), and NFL (c.1054G>T, p.Glu352*) mutations. Numbering is based on the 132 cDNA sequence (GenBank: NM_006493.2). The A of ATG of the initiator methionine codon is denoted as nucleotide 1. TM, transmembrane region; black box, Flag tag. B: Membrane association analysis of Flag-tagged CLN5 mutant proteins. HEK293 cells transiently expressing the WT SP-Flag-CLN5 or carrying the EUR, FinM, SWE, or NFL pathological mutation were subfractionated and treated with 5 M urea as described in Figure 3. The supernatant (S) and pellet (P) fractions were separated by SDS-PAGE and were immunoblotted using anti-Flag antibody. Cytosolic EEA1, peripheral GM130, and TM protein TfR were used as fractionation controls. Representative blots are shown. C: Analysis of the maturation and secretion of Flag-tagged CLN5 mutant proteins. HEK293 cells transiently expressing the WT SP-Flag-CLN5, or the EUR, FinM, SWE, or NFL mutants were pulsed for 5 min with 35S-methionine and were chased for the indicated periods of time. The Flag-tagged CLN5 proteins were immunoprecipitated from the cell lysates, and the culture media collected after each chase period. The proteins were separated by SDS-PAGE and were analyzed by autoradiography. The films were exposed for 48 hr. *, Nonspecific band; black arrowhead, precursor polypeptide; white arrowhead, mature polypeptide. and SP cleavage. Surprisingly, the size and/or amounts of the mature forms of these mutants were clearly modified following the different chase periods. After a 2-hr chase, the amount of mature NFL protein was significantly reduced, whereas the MW of SWE decreased to ∼25 kDa, which was probably caused by partial deglycosylation. Indeed, the PNGase F treatment indicated that this 25-kDa band is a glycosylated form of SP-cleaved SWE (Supp. Fig. S3). After a 4-hr chase, both mutants were barely detectable in the cell lysate. Since the levels of secreted SWE and NFL mutant proteins were no higher than that of WT CLN5 (Fig. 5C; extracellular), we hypothesized that this was due to the degradation of the mutated CLN5 proteins by the proteasomal machinery. This was confirmed by the restoration of SWE and NFL protein levels when the cells were chased for 4 hr in the presence of proteasomal enzyme inhibitor MG132 (Supp. Fig. S4). Moreover, the amount of secreted NFL protein was lower than that for the other mutants (Fig. 5C; extracellular), suggesting that the high level of degradation reduces NFL protein secretion. Lastly, the lower MW bands detected in the lysates following the different chase periods of the SWE and NFL mutants most likely correspond to differential glycosylation of the SP-cleaved proteins and to degradation products. Taken together, these findings indicated that the EUR, FinM, SWE, and NFL mutations do not affect the processing (cleavage of SP) or glycosylation of CLN5, whereas the SWE and NFL mutations affect the membrane association and stability of CLN5. However, all the mutations induced the relocalization of 133 CLN5 to the ER (Supp. Fig. S5). This is the first time that the NFL mutation has been characterized. Discussion Mutations in the CLN5 gene result in NCL, a severe neurodegenerative disorder that mainly affects children. The basic properties of CLN5 such as its structure and solubility need to be clarified to better understand the role of this protein in the pathogenesis of NCL. Previous analyses of the amino acid sequence of CLN5 using various prediction programs resulted in different representations of the structure of CLN5. The original model predicted two TM domains and cytoplasmic-oriented N- and C-termini [Savukoski et al., 1998]. In contrast, current bioinformatic tools predict one TM domain, a cytoplasmic N-terminus, and a luminal C-terminus. However, some programs have also predicted that CLN5 is a soluble protein [Holmberg et al., 2004]. The present study is the first to experimentally define the topology of CLN5. Using proteinase K protection assays on various epitopetagged CLN5 proteins expressed in cells and on untagged CLN5 translated in vitro, we showed that the N-terminus of CLN5 is located in the cytoplasm, whereas its C-terminus is located in the lumen of organelles. We thus ruled out the presence of a second TM domain and demonstrated that CLN5 is synthesized as a type II single pass TM protein. This precursor form is rapidly cleaved at the SP site located after the TM domain, with the resulting mature form of CLN5 being entirely located in the lumen of intracellular compartments. Mature CLN5 has mostly been reported to be a soluble protein based on the fact that it is secreted, that it contains mannose-6-phosphate residues, and that mouse CLN5 can be extracted with TX-114 [Holmberg et al., 2004; Isosomppi et al., 2002; Sleat et al., 2006a; Vesa et al., 2002]. In the present study, we showed that the mature form of CLN5, which is devoid of a TM domain, is not a soluble protein but is tightly associated with the membrane through an AH. Mature CLN5 was resistant to high salt, alkaline sodium carbonate, and urea membrane extraction whereas TX-114 induced a partial partitioning into the aqueous fraction, indicating that CLN5 does not behave like a typical peripheral protein or integral 134 membrane protein. This can be explained by the presence of a predicted AH in the aa 353– 392 region composed of two in-plane hydrophobic patches that would directly adsorb at the interface of the lipidic membrane. We used GFP fusion experiments to show that the predicted AH of CLN5 is sufficient to tightly hook GFP to the membrane. This is the first report of the presence of an AH in CLN5 involved in its membrane association. The presence of this AH may explain the slow trafficking of CLN5 and its localization to different intracellular compartments. AH is a key structure for the membrane anchoring and retention of other luminal proteins that lack a TM domain, including the viral Erns glycoprotein [Burrack et al., 2012; Fetzer et al., 2005], torsin A [Vander Heyden et al., 2011], and the cyclooxygenase isozymes COX-1 and COX-2 [Simmons et al., 2004]. Furthermore, while CLN3, another NCL protein, has been modeled with one luminal AH and six TM helices [Nugent et al., 2008], the specific role of AH in the function of CLN3 is unknown. AH appears to mediate more than just membrane anchoring and has been variously described as an autoinhibitory protein activity domain, a protein interaction domain, or a membrane curvature sensor/modulator that can respond to and modulate the physical properties of membranes [Boucrot et al., 2012; Cornell and Taneva, 2006; Shih et al., 2011; Stern et al., 2013]. As such, the involvement of the AH of CLN5 in functions other than membrane anchoring needs to be investigated. The mature form of overexpressed CLN5 has been detected in the extracellular medium in several studies [Isosomppi et al., 2002; Kollmann et al., 2005; Vesa et al., 2002], including the present work. We found that a small amount (∼20%) of mature CLN5 is secreted into the medium after a 2–4-hr chase. However, CLN5 is secreted considerably more slowly (>1 hr) than most constitutively secreted proteins, which is consistent with our results indicating that CLN5 is membrane anchored. This raises a question concerning the functional significance of extracellular CLN5. The secretion of CLN5 may be necessary for its turnover, or CLN5 may have unknown extracellular functions. In addition, overexpression may alter the retention of mature CLN5. The secretion of endogenous CLN5 will thus have to be validated when the appropriate antibodies become available. 135 We noted that the predicted AH domain in CLN5 is highly conserved among species, suggesting that it has major structural constraints with respect to function and localization. Since many known NCL pathological missense mutations are located in or immediately adjacent to the predicted AH domain [Mole et al., 2011], we analyzed the impact of these mutations on CLN5 membrane anchoring and secretion. The membrane association of two AH-deficient CLN5 mutants (p.Glu253* [SWE] and p.Glu352* [NFL]) was dramatically impaired, whereas mutants truncated immediately after the AH (p.Tyr392* [FinM]) and mutations that did not alter the AH (p.Asp279Asn [EUR]) were not or only slightly affected. The partial membrane dissociation of FinM might be because the C-terminal region plays a partial role in membrane association, or because its absence renders the AH unstable. Surprisingly, the absence of the membrane anchoring AH domain in the p.Glu253* and p.Glu352* mutants did not lead to increased secretion of CLN5 but, rather, resulted in their recognition and degradation by the ER quality control system, suggesting that AH could play a role in protein stability in addition to membrane anchoring. Schmiedt et al. (2010) also reported that the p.Glu253* and p.Tyr392* mutations resulted in unstable proteins. However, in our hands, the stability of the p.Tyr392* mutant was not affected. This discrepancy might be due to the different analytical methods used. As previously reported, none of the CLN5 disease mutations alter the SP cleavage or glycosylation of CLN5, but they all result in the mislocalization of CLN5 in the ER, suggesting that the mutated proteins are misfolded and are retained in the ER [Lyly et al., 2009; Schmiedt et al., 2010]. However, based on the pulse-chase assay, the mutant CLN5 polypeptides (except NFL that is rapidly degraded) were secreted at the same level as WT CLN5, suggesting that they are sufficiently correctly folded to pass the quality control of the intracellular sorting system in the Golgi but are not sorted to lysosomes. These results are in line with the observation that CLN5 mutants contain complex sugars, suggesting that they reach the Golgi but are transported back to the ER [Schmiedt et al., 2010]. Two other mutations in the AH (p.Leu358Alafs*4 and p.Trp379Cys) have been reported to lead to the redistribution of CLN5 to the ER [Lebrun et al., 2009]. To better understand the role of the AH in CLN5 function, it would be interesting to determine whether these mutations, as well as others identified in the AH [Mole et al., 2011], alter the membrane anchoring, secretion, and stability of CLN5. 136 Summary and Conclusion Figure 6 - Schematic model for human CLN5. A: CLN5 is synthesized as a single pass type II TM precursor glycoprotein. The Nterminus faces the cytoplasm, and the C-terminus is located in the lumen of the intracellular compartment. B: The SP is rapidly cleaved, and the resulting mature protein is tightly associated with the lumen of the membrane through an amphipathic anchor region. The TM domain (aa 76–91) and anchor region (aa 353–392) of human CLN5 are depicted as a dark rectangle. The predicted SP cleavage site is indicated (aa 96). Nucleotide numbering is based on GenBank reference sequence NM_006493.2. The figure is not to scale. We propose a new topological model for CLN5 in which its structure contains a cytoplasmic N-terminus and one TM segment, which are cotranslationally cleaved in the ER by a signal peptidase, and a large luminal C-terminal domain with a previously unidentified AH, which is involved in the membrane anchoring of the mature form of CLN5 (Fig. 6). This model provides new molecular insights into the results of previous studies. It suggests that the reported interactions of mature CLN5 with CLN1, CLN2, CLN3, CLN6, and CLN8 involve the luminal domains of these proteins [Lyly et al., 2009; Vesa et al., 2002]. It also indicates that the reported role of CLN5 in the recruitment of retromers on the cytoplasmic side of the membrane is indirect and is mediated by a TM interactive partner [Mamo et al., 2012]. Furthermore, the lysosomal transport of CLN5 would not involve a direct interaction with cytosolic trafficking components but would rather involve a luminal sorting motif or a domain of interaction with a lysosomal sorting protein. CLN5 lysosomal targeting has been reported to be independent of the mannose-6137 phosphate receptor and sortilin [Mamo et al., 2012; Schmiedt et al., 2010]. Other possible candidates are LIMP-II, the sorting receptor for glucocerebrosidase [Reczek et al., 2007], and palmitoyl protein thiosterase 1 (PPT1/CLN1), which facilitates the lysosomal transport of mutated CLN5 [Lyly et al., 2009]. More work is required to elucidate the sorting and trafficking mechanisms for CLN5. In conclusion, the elucidation of the proper topology of CLN5 represents a crucial step toward acquiring a better understanding of the mechanism of action of CLN5 and clarifying its role in the pathogenesis of NCL. Acknowledgments We wish to thank Dr. Stéphane Lefrançois (Université de Montréal, QC, Canada) for the human 3xHA-CLN5 cDNA and helpful discussions and Dr. Pierre Lavigne (Université de Sherbrooke, QC) for his valuable advice regarding the analysis of the amphipathic helix. We also wish to thank Caroline Thériault and Marilène Paquette for their technical assistance. H.L. holds a fellowship from the National Sciences and Engineering Research Council of Canada (NSERC). C.L. holds a Canada Research Chair in Cellular Pharmacology. 138 References Bessa C, Teixeira CA, Mangas M, Dias A, Sa Miranda MC, Guimaraes A, Ferreira JC, Canas N, Cabral P, Ribeiro MG. 2006. Two novel CLN5 mutations in a Portuguese patient with vLINCL: insights into molecular mechanisms of CLN5 deficiency. Mol Genet Metab 89:245–253. Bordier C. 1981. Phase separation of integral membrane proteins in Triton X-114 solution. J Biol Chem 256:1604–1607. Boucrot E, Pick A, Camdere G, Liska N, Evergren E, McMahon HT, Kozlov MM. 2012. Membrane fission is promoted by insertion of amphipathic helices and is restricted by crescent BAR domains. Cell 149:124–136. Burrack S, Aberle D, Burck J, Ulrich AS, Meyers G. 2012. A new type of intracellular retention signal identified in a pestivirus structural glycoprotein. FASEB J 26:3292– 3305. Cornell RB, Taneva SG. 2006. Amphipathic helices as mediators of the membrane interaction of amphitropic proteins, and as modulators of bilayer physical properties. Curr Protein Pept Sci 7:539–552. Fetzer C, Tews BA, Meyers G. 2005. The carboxy-terminal sequence of the pestivirus glycoprotein E(rns) represents an unusual type of membrane anchor. J Virol 79:11901–11913. Haltia M. 2006. The neuronal ceroid-lipofuscinoses: from past to present. Biochim Biophys Acta 1762:850–856. Holmberg V, Jalanko A, Isosomppi J, Fabritius AL, Peltonen L, Kopra O. 2004. The mouse ortholog of the neuronal ceroid lipofuscinosis CLN5 gene encodes a soluble lysosomal glycoprotein expressed in the developing brain. Neurobiol Dis 16:29–40. Holmberg V, Lauronen L, Autti T, Santavuori P, Savukoski M, Uvebrant P, Hofman I, Peltonen L, Jarvela I. 2000. Phenotype-genotype correlation in eight patients with Finnish variant late infantile NCL (CLN5). Neurology 55:579–581. Isosomppi J, Vesa J, Jalanko A, Peltonen L. 2002. Lysosomal localization of the neuronal ceroid lipofuscinosis CLN5 protein. Hum Mol Genet 11:885–891. Kollmann K, Mutenda KE, Balleininger M, Eckermann E, von Figura K, Schmidt B, Lubke T. 2005. Identification of novel lysosomal matrix proteins by proteome analysis. Proteomics 5:3966–3978. Kollmann K, Uusi-Rauva K, Scifo E, Tyynela J, Jalanko A, Braulke T. 2013. Cell biology and function of neuronal ceroid lipofuscinosis-related proteins. Biochim Biophys Acta 1832:1866–1881. Lebrun AH, Storch S, Ruschendorf F, Schmiedt ML, Kyttala A, Mole SE, Kitzmuller C, Saar K, Mewasingh LD, Boda V, Kohlschütter A, Ullrich K, Braulke T, Schulz A. 2009. Retention of lysosomal protein CLN5 in the endoplasmic reticulum causes neuronal ceroid lipofuscinosis in Asian sibship. Hum Mutat 30:E651–E661. Lyly A, von Schantz C, Heine C, Schmiedt ML, Sipila T, Jalanko A, Kyttala A. 2009. Novel interactions of CLN5 support molecular networking between Neuronal Ceroid Lipofuscinosis proteins. BMC Cell Biol 10:83. Mamo A, Jules F, Dumaresq-Doiron K, Costantino S, Lefrancois S. 2012. The role of ceroid lipofuscinosis neuronal protein 5 (CLN5) in endosomal sorting. Mol Cell Biol 32:1855–1866. 139 Matlin KS, Simons K. 1983. Reduced temperature prevents transfer of a membrane glycoprotein to the cell surface but does not prevent terminal glycosylation. Cell 34:233–243. Mole SE, Williams RE, Goebel HH. 2005. Correlations between genotype, ultrastructural morphology and clinical phenotype in the neuronal ceroid lipofuscinoses. Neurogenetics 6:107–126. Mole SE, Williams RE, Goebel HH. 2011. The neuronal ceroid lipofuscinoses (Batten disease). In: Mole SE, Williams RE, Goebel HH, editors. New York: Oxford University Press. p 444. Moore SJ, Buckley DJ, MacMillan A, Marshall HD, Steele L, Ray PN, Nawaz Z, Baskin B, Frecker M, Carr SM, Ives E, Parfrey PS. 2008. The clinical and genetic epidemiology of neuronal ceroid lipofuscinosis in Newfoundland. Clin Genet 74:213–222. Nugent T, Mole SE, Jones DT. 2008. The transmembrane topology of Batten disease protein CLN3 determined by consensus computational prediction constrained by experimental data. FEBS Lett 582:1019–1024. Qoronfleh MW, Benton B, Ignacio R, Kaboord B. 2003. Selective enrichment of membrane proteins by partition phase separation for proteomic studies. J Biomed Biotechnol 2003:249–255. Reczek D, Schwake M, Schroder J, Hughes H, Blanz J, Jin X, Brondyk W, Van Patten S, Edmunds T, Saftig P. 2007. LIMP-2 is a receptor for lysosomal mannose-6phosphate-independent targeting of beta-glucocerebrosidase. Cell 131:770–783. Santavuori P. 1988. Neuronal ceroid-lipofuscinoses in childhood. Brain Dev 10:80–83. Savukoski M, Klockars T, Holmberg V, Santavuori P, Lander ES, Peltonen L. 1998. CLN5, a novel gene encoding a putative transmembrane protein mutated in Finnish variant late infantile neuronal ceroid lipofuscinosis. Nat Genet 19:286–288. Schmiedt ML, Bessa C, Heine C, Ribeiro MG, Jalanko A, Kyttala A. 2010. The neuronal ceroid lipofuscinosis protein CLN5: new insights into cellular maturation, transport, and consequences of mutations. Hum Mutat 31:356–365. Shih YL, Huang KF, Lai HM, Liao JH, Lee CS, Chang CM, Mak HM, Hsieh CW, Lin CC. 2011. The N-terminal amphipathic helix of the topological specificity factor MinE is associated with shaping membrane curvature. PLoS One 6:e21425. Simmons DL, Botting RM, Hla T. 2004. Cyclooxygenase isozymes: the biology of prostaglandin synthesis and inhibition. Pharmacol Rev 56:387–437. Sleat DE, Wang Y, Sohar I, Lackland H, Li Y, Li H, Zheng H, Lobel P. 2006a. Identification and validation of mannose 6-phosphate glycoproteins in human plasma reveal a wide range of lysosomal and non-lysosomal proteins. Mol Cell Proteomics 5:1942–1956. Sleat DE, Zheng H, Qian M, Lobel P. 2006b. Identification of sites of mannose 6phosphorylation on lysosomal proteins. Mol Cell Proteomics 5:686–701. Stern O, Hung YF, Valdau O, Yaffe Y, Harris E, Hoffmann S, Willbold D, Sklan EH. 2013. An N-terminal amphipathic helix in dengue virus nonstructural protein 4A mediates oligomerization and is essential for replication. J Virol 87:4080–4085. Vander Heyden AB, Naismith TV, Snapp EL, Hanson PI. 2011. Static retention of the lumenal monotopic membrane protein torsinA in the endoplasmic reticulum. EMBO J 30:3217–3231. 140 Vesa J, Chin MH, Oelgeschlager K, Isosomppi J, DellAngelica EC, Jalanko A, Peltonen L. 2002. Neuronal ceroid lipofuscinoses are connected at molecular level: interaction of CLN5 protein with CLN2 and CLN3. Mol Biol Cell 13:2410–2420. Wimley WC, White SH. 1996. Experimentally determined hydrophobicity scale for proteins at membrane interfaces. Nat Struct Biol 3:842–848. 141 Supporting Information Supp. Figure S1 - Steady state localization of untagged and tagged human CLN5. HeLa cells transiently transfected with 3xHA-CLN5 (A–C), CLN5-HA (D-F),SP-FLAGCLN5 (G-I), or untagged CLN5 (J-L) together with LAMP1-GFP (D-I) were fixed with methanol and immunostained using anti-HA (A, D), anti-FLAG (G), anti-c/32 (J), antiGFP (E, H), anti-Calnexin (B), or anti-LAMP1 (K) antibody. The stained cells were examined by confocal fluorescence microscopy. The merged images (C, F, I, K) show a partial overlap between CLN5-HA, SP-FLAG-CLN5, and untagged CLN5 and the lysosomal marker LAMPI while 3HA-CLN5 colocalized with the ER marker Calnexin. Scale bar, 10 μm. 142 Supp. Figure S2 - The secretion of CLN5 is inhibited at 20°C. HEK293 cells transiently expressing the untagged and tagged CLN5 proteins were pulsed for 5 min with 35S-methionine and were chased for 1 h at 37°C or 20°C. Untagged and tagged CLN5 proteins were immunoprecipitated from the cell lysates, and culture media collected after each chase period using anti-C/32 and anti-HA/-Flag antibodies, respectively. The proteins were separated by SDS-PAGE and were analyzed by autoradiography. Supp. Figure S3 - CLN5 pathological mutants are N-glycosylated. Lysates from HEK293 cells transiently expressing the WT SP-Flag-CLN5 or the EUR (p.Asp279Asn), FinM (p.Tyr392*), SWE (p.Glu253*), or NFL (p.Glu352*) mutations were incubated in the absence or presence of PNGaseF. The proteins were separated by SDS-PAGE and were immunoblotted using anti-Flag antibody. 143 Supp. Figure S4 - Stabilization of the SWE and NFL mutants by proteasome inhibitors. HEK293 cells transiently expressing the SWE (p.Glu253*) or NFL (p.Gly352*) mutants were pulsed for 5 min with 35S-methionine. The cells were then chased for the indicated periods of time in the absence or presence of the lysosomal enzyme inhibitors leupeptin (Leu, 1 mg/mL), pepstatin (Pep, 100 μM), and E64 (10 μg/mL), or the proteasomal enzyme inhibitor MG132 (10 μM). The Flag-tagged CLN5 proteins were immunoprecipitated from the cell lysates collected after each chase period. The proteins were separated by SDS-PAGE and were analyzed by autoradiography. The films were exposed for 48 h. 144 Supp. Figure S5 - Steady state localization of CLN5 pathological mutants in HeLa cells. SP-Flag-CLN5 constructs carrying the p.Asp279Asn (EUR, A-C), p.Tyr392* (FinM, D– F), p.Glu253* (SWE, G-I), or p.Glu352* (NFL, J-L) mutation were transiently transfected into HeLa cells. The localizations of the CLN5 mutant proteins were studied by immunofluorescent labeling followed by confocal microscopy using anti-Flag antibody (A, D, G, J). Calnexin was used to label the ER (B, E, H, K). Colocalization is indicated in yellow. Scale bar 10 μm. 145 ARTICLE 3 Rôle de Calnuc dans les céroïdes-lipofuscinoses neuronales Auteurs de l’article: Heidi Larkin, Susan Cotman, Sara Mole et Christine Lavoie Statut de l’article: En cours de rédaction Avant-propos: J'ai écrit la première version de l'article et j'ai réalisé toutes les expériences ainsi que tous les montages. Christine Lavoie a effectué les corrections du texte et a effectué la prise des photos pour les expériences d'immunofluorescence. Sara Mole a fourni les fibroblastes issus de patients atteints de céroïdes-lipofuscinoses neuronales et Susan Cotman les lignées cellulaires stables de cervelet. 146 RÉSUMÉ – ARTICLE 3 Calnuc, un nouveau partenaire d'interaction pour CLN3 et CLN5, est potentiellement impliqué dans les céroïdes-lipofuscinoses neuronales Les céroïdes-lipofuscinoses neuronales (NCL), aussi connues sous le nom de maladies de Batten, forment un groupe de maladies neurodégénératives caractérisées par une accumulation de lipopigments autofluorescents, nommés lipofuscines, dans les lysosomes. Jusqu'à présent, 14 gènes (désignés CLN1-14) ont été associés à la pathologie. Récemment, CLN3 et CLN5 ont été montrées comme étant impliquées dans le transport intracellulaire par une interaction et une régulation de Rab7, une petite protéine G nécessaire pour le recrutement endosomial du complexe Rétromère. Nous avons récemment démontré le rôle de Calnuc, une protéine ubiquitaire qui lie le calcium, dans le transport rétrograde de récepteurs lysosomiaux des endosomes vers le réseau trans-golgien, en raison de son importance pour l'association membranaire et l'activité de Rab7. Dans la présente étude, nous examinons les interactions fonctionnelles entre Calnuc, CLN3 et CLN5, en plus de la potentielle implication de Calnuc dans les NCL. À l'aide d'analyses de biochimie et de microscopie, nous avons montré que Calnuc interagit avec CLN3 et CLN5 et colocalise avec ces protéines sur les endosomes. Nous avons découvert des niveaux réduits de CLN3 et CLN5 dans les cellules déplétées en Calnuc et, inversement, des niveaux réduits de Calnuc dans les cellules déplétées en CLN3 et CLN5. Ces résultats suggèrent que Calnuc, CLN3 et CLN5 peuvent former un complexe stabilisé par leurs interactions intramoléculaires. En microscopie confocale et électronique, nous avons observé que la déplétion de Calnuc induit l’apparition de caractéristiques morphologiques cellulaires typiques des NCL, soit un engorgement des lysosomes, une accumulation de matériel autofluorescent et de la sous-unité C de l’ATP synthase mitochondriale, ainsi qu'une l’induction accrue de l’autophagie. Dans les lignées de cellules de patients atteints de NCL, le profil d'expression de Calnuc indique des niveaux protéiques diminués de Calnuc dans les cas confirmés de type CLN3 et CLN5, de même que dans tous les cas de type orphelin, c'est-à-dire qui sont génétiquement indéfinis mais présentent des similarités phénotypiques avec NCL. Dans l'ensemble, ces données suggèrent que la fonction de Calnuc est altérée dans les NCL et que Calnuc pourrait être impliquée dans la pathogenèse de la maladie. 147 Article 3 Evidence for a role of Calnuc, a novel interacting partner for CLN3 and CLN5, in neuronal ceroid lipofuscinosis Heidi Larkin1 and Christine Lavoie1‡ 1 Department of Pharmacology, Faculty of Medicine and Health Sciences, Université de Sherbrooke, Sherbrooke, QC, Canada ‡ Correspondence to: Christine Lavoie, Department of Pharmacology, Faculty of Medicine, Universite de Sherbrooke, 3001-12e Avenue Nord, Sherbrooke, QC J1H 5N4, Canada E-mail: [email protected] Key words: Calnuc, nucleobindin, CLN5, CLN3, lipofuscinosis, lysosomal receptor, endosomal sorting, lysosome 148 Abstract The neuronal ceroid lipofuscinoses (NCLs), also known as Batten disease, are a group of neurodegenerative diseases characterized by accumulation of autofluorescent lipopigments appointed lipofuscins, in lysosomes. So far, 14 genes (designated CLN1-14) have been associated with the pathology. Recently, CLN3 and CLN5 have been involved in the proper trafficking of lysosomal receptors through the interaction and regulation of Rab7, a small G protein required for the endosomal recruitment of retromers. We previously highlighted the role of Calnuc, an ubiquitous calcium-binding protein, in the retrograde transport of lysosomal receptors from endosomes to trans-Golgi network (TGN) via the activation and membrane association of Rab7 (Larkin et al. 2016). In the present study, we examined the functional interactions between Calnuc, CLN3 and CLN5 and the potential implication of Calnuc in NCL pathogenesis. Using biochemical and microscopy assays, we showed that Calnuc interacts with CLN3 and CLN5 and colocalizes with them on endosomes. We found reduced levels of CLN3 and CLN5 in Calnuc-depleted cells and vice-versa, a lower level of Calnuc in CLN3- and CLN5-depleted cells, suggesting that Calnuc, CLN3 and CLN5 could form a protein complex stabilized by their intramolecular interactions. Using confocal and electron microscopy, we observed that the depletion of Calnuc induces cellular morphological features reminiscent of NCL such as lysosome enlargement, accumulation of autofluorescent material and of subunit C of mitochondrial ATP synthase (SCMAS), as well as an increased of autophagy induction. Expression profiling of Calnuc in fibroblasts cell lines from NCL patients indicated a lower proteic level of Calnuc in CLN3 and CLN5 genetically confirmed cases of NCL as well as in orphan cases of NCL which remain undefined genetically but display phenotypes similar to CLN gene deficiencies. Overall, these data suggest that Calnuc function is altered in human NCL and that Calnuc could be a novel protein involved in NCL pathogenesis. Introduction Neuronal Ceroid Lipofuscinoses (NCLs) are severe neurodegenerative lysosomal storage diseases that manifest in childhood and are characterized by developmental 149 regression, seizures, visual failure and mental retardation, culminating in premature death (Mole et al. 2005; Jalanko and Braulke 2009). NCLs are relatively rare, with an incidence of 1 in 100,000 live births worldwide, but appear to be more common (1:12,500 live births) in northern Europe (Finland and Sweden) and Newfoundland (Santavuori 1988; Palmer et al. 1989). At the cellular level, an accumulation of autofluorescent ceroid lipopigments (i.e., lipofuscin) and specific storage material ultrastructures are morphological indications of NCL (Goebel and Wisniewski 2004; Mole et al. 2005; Cotman and Staropoli 2012). These disorders result from mutations in at least 14 Ceroid-Lipofuscinosis Neuronal genes (CLN1-CLN14) (Jalanko and Braulke 2009) and are classified clinically based on either the mutation of the CLN gene or the age of onset of disease symptoms (Jalanko and Braulke 2009). Furthermore, there are still several groups of NCL patients that remain undefined genetically. Infantile NCL (INCL; 0-2 years old) is associated with mutations in CLN1, late infantile NCL (LINCL; 2-4 years old) is linked to mutations in CLN2, 5, 6, 7 and 8, and juvenile NCL (JNCL or Batten disease; 5-10 years old) is related to mutations in CLN3. The precise function and regulation of most of these CLN proteins are unknown. However, recent studies have implicated CLN3 and CLN5 in the trafficking of lysosomal receptors. CLN3, a multipass transmembrane (TM) protein localized mainly on endosomes and lysosomes, has been reported to be involved in both post-Golgi anterograde and endosometo-TGN retrograde trafficking pathways (Cotman and Staropoli 2012). This role of CLN3 is via its interaction with various proteins involved in different trafficking steps. CLN3 binds to Rab7 and with various retro- and anterograde microtubular motor complexes responsible for the movement of late endosomes and lysosomes (Uusi-Rauva et al. 2012). It regulates Sed5, a Golgi SNARE involved in endosome-to-Golgi retrieval (Kama et al. 2011), as well as Btn2p, a retromer accessory protein that facilitates specific protein retrieval from the endosome to the Golgi (Kama et al. 2007). Recently, the activity of CLN5 has also been linked to Rab7 activity and recruitment to endosomes (Mamo et al. 2012) providing the first link to lysosomal function. Indeed, CLN5 depletion results in reduced levels of active Rab7 and retromer membrane recruitment, key elements of the endosmal retrograde sorting machinery, leading to the misdelivery and degradation of Sortilin and CI-MPR to lysosome (Mamo et al. 2012). CLN5, is a luminal protein associated with the endolysosomal membrane (Larkin et al. 2013). This indicates that the reported effect of CLN5 on the 150 lysosomal receptor trafficking is indirect and is mediated by a TM interactive partner. Interestingly, CLN5 has been shown to directly interact with CLN3 (Vesa et al. 2002; Lyly et al. 2009), suggesting that CLN3 and CLN5 cooperate on the endosome-to-Golgi retrograde machinery for lysosomal receptors retrieval. Taken together, these studies imply that a functional deficiency in CLN3 or CLN5 would impair the retrograde trafficking of lysosomal receptors and the proper delivery of their cargos, which underlies the lysosomal dysfunction observed in NCL disorders. Calnuc (nucleobindin or NUCB1) is a ubiquitous and well-conserved protein suggesting that it has important biological functions (Wendel, Sommarin et al. 1995; Miura et al. 1996; Lin et al. 1998; Kawano et al. 2000). Indeed, the modulation of Calnuc expression is associated with pathologies, such as cancer, lupus and Alzheimer’s, but its precise cellular functions remained poorly understood. Calnuc is predominantly localized in the cis-Golgi lumen where it is retained for a long period and constitutes the major Ca2+ storage pool via 2 EF-hand motifs (Lin et al. 1998; Lin et al. 1999). It is then secreted via constitutive and constitutive-like pathways to the extracellular space where it has been involved in bone matrix maturation (Lavoie et al. 2002). Calnuc is also found free in cytosol or associated with the surface of the Golgi, endosomal membranes and secretory granules (Brodeur et al. 2009; Lin et al. 2009). In addition to being a multicompartmental protein, Calnuc is also a multifunctional protein comprised of a signal peptide, a putative DNA-binding domain, a leucine zipper and 2 EF-hand motifs (Lin et al. 1998). These characteristics potentiates its interactions with a growing list of partners, such as DNA (Miura et al. 1992), heterotrimeric Gα proteins (Lin et al. 1998), COX (Ballif et al. 1996), Necdin (Taniguchi et al. 2000) and LRP9 (Brodeur et al. 2009). Known as LRP10 in humans, this latter is a LDL receptor cycling between the trans-Golgi network (TGN) and early endosomes (Boucher et al. 2008). We have shown that cytoplasmic pool of Calnuc plays an essential role in LRP9 retrieving from the endosomes to the TGN (Brodeur et al. 2009). Similarly, Calnuc affects other receptors transiting between the TGN and endosomes, such as lysosomal receptors Sortilin and CI-MPR (Larkin et al. 2016), suggesting that Calnuc plays a general role in endosomal receptor sorting. We demonstrated that Calnuc interacts indirectly and colocalizes with Rab7 on endosome, that 151 Calnuc depletion impairs the activation and membrane association of Rab7 that lead to a defect in the endosomal recruitment of retromer and the misdelivery and degradation of MPR and Sortilin in lysosomes (Larkin et al. 2016). Thus Calnuc is novel player involved in the endosome-to-TGN retrograde transport of lysosomal receptors. Given that Calnuc affects the same molecular components of the endosomal sorting machinery as CLN3 and CLN5, and have similar effect on lysosomal receptor retrieval, this study aim at investigating the functional interactions/cooperation between Calnuc, CLN3 and CLN5. We observed a physical association between Calnuc, CLN3 and CLN5 and that Calnuc depletion induces cellular morphological features reminiscent of NCL, suggesting that Calnuc is new protein involved in NCLs. Materials and methods Antibodies and reagents Anti-Flag M2 monoclonal antibodies (mAbs) were purchased from Sigma Aldrich (Saint Louis, MO, USA), anti-HA mAbs from Covance (Emeryville, CA, USA),, anti-GFP mAbs from Clontech (Mountain View, CA, USA), anti-Rab7 mAbs from Cell signalling Technology (Danvers, MA, USA), anti-HA mAbs from Covance (Emeryville, CA, USA), anti-LAMP 2 (H4B4) m Abs from Developmental Studies Hybridoma Bank at the University of Iowa (Iowa, IA, USA). Anti-Calnuc polyclonal antibodies (pAbs) were purchased from Aviva Systems Biology (San Diego, CA, USA), anti-EEA1 pAbs from Thermo Scientific (Rockford, IL, USA), anti-Flag pAbs from Sigma Aldrich (Saint Louis, MO, USA), anti-GFP pAbs from Molecular Probes (Eugene, OR, USA) and anti-LC3 pAbs from Novus Biologicals (Littleton, CO, USA). Anti-Calnuc pAbs against mouse was a gift from Marilyn Farqhuar (University of California San Diego) and anti-SCMAS pAbs from Susan Cotman (Harvard Medical School). DNA constructs Mammalian expression vector pcDNA3.1 containing Calnuc-GFP fusion protein or ΔSP-Calnuc (Calnuc without signal peptide) or ΔEF-Calnuc (Calnuc without calcium 152 binding domains) are described elsewhere (Weiss et al. 2001). GST-RILP was a gift from Aimee L. Edinger (UC Irvine) and LC3-GFP from Sheela Ramanathan (Université de Sherbrooke). Human CLN1 and CLN3 in pCMV6-XL5 vector as well as 3xHA-CLN5 in the pReceiver-M06 vector were generous gifts from Stéphane Lefrançois (Université de Montréal, QC, Canada). CLN1-HA, CLN3-HA and CLN5-HA were generated by subcloning in the pcDNA3.1 vector. The SP-Flag-CLN5 construct was obtained by the insertion of the Flag epitope at amino acid position 97. Cell culture and transfection HeLa cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and HEK293T cells were obtained from Alexandra Newton (University of California, San Diego, CA, USA). The cells were grown at 37°C, with 5% CO2 atmosphere control in Dulbecco’s modified Eagle’s high glucose medium (DMEM) (Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT, USA) and 1% penicillin, Glutamine and streptomycin (PSG) (Invitrogen, Carlsbad, CA, USA). Cerebellar cells stably expressing CLN3+/+ or CLN3Δex7/8/Δex7/8 were obtained from Susan Cotman (Harvard Medical School) (Fossale 2004, BMC Neurosci. 5, 57). Cb cells maintained between 30 and 90% confluency and were grown at 33°C, with 5% CO2 atmosphere control, in DMEM with 10% heat-inactivated FBS, 24 mM KCl, 1% PSG, and 200 μg/ml G418 (Invitrogen, Carlsbad, CA, USA). Fibroblasts cells from patients were obtained from Sara Mole (University College London) and controls fibroblasts were purchase at Coriell Biorepository. Those cells were culture at 37°C, with 5% CO2 atmosphere control in MEM or DMEM with 10% FBS and 1% PSG accordingly to instructions. HeLa cells, HEK cells, and Cb cells were transfected with Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturers’ instructions. Coimmunoprecipitation HEK293 cells were plated in 60-mm culture dishes and transfected with the various constructs. After 48 h, the cells were lysed in 50 mM Tris buffer (pH 7.4) containing 100 mM NaCl, 1% NP-40, and protease inhibitors. Cells were incubated for 1 h at 4oC and then 153 centrifuged at 15,800xg for 20 min. The cleared supernatants were incubated with primary antibodies overnight at 4oC and then with protein A-sepharose (GE Healthcare, Piscataway, NJ, USA) or protein G-Sepharose (Zymed, San Francisco, CA, USA) for 1 h. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. Experiments containing multiple transmembrane protein CLN3 were instead incubated 1h at TP in Laemmli sample buffer. Bound immune complexes were analyzed by SDS-PAGE and immunoblotting. Glutathione S-transferase (GST) pull-down assays GST fusion proteins were expressed in E. coli BL21 and were purified on glutathione-Sepharose 4B beads (Pharmacia, Piscataway, NJ, USA) according to the manufacturer’s instructions. Lysate from HeLa cells obtained as described above, was incubated overnight at 4oC with 20 μg of GST fusion proteins immobilized beads. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. The bound proteins were separated by SDS-PAGE and detected by autoradiography or immunoblotting. Immunoblotting The protein samples were separated on 10% or 12% SDS-PAGE gels and transferred to nitrocellulose membranes (Perkin Elmer, Woodbridge, ON, Canada). The membranes were blocked in Tris-buffered saline (20 mM Tris-HCl, pH 7.5, 150 mM NaCl) containing 0.1% Tween 20 and 5% nonfat dry milk, bovine serum albumin (BSA), or foetal bovine serum (FBS) and incubated with primary antibodies for 1 h at RT or O/N at 4°C and then with horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (Bio-Rad, Hercules, CA, USA) and enhanced chemiluminescence detection reagent (Pierce Chemical, Rockford, IL, USA). RNA Interference and Rescue Nonspecific (CTL), Calnuc, CLN1, CLN3 and CLN5 siRNA scrambled II duplex were purchased from Dharmacon Research (Chicago, IL, USA). HeLa cells were transfected with a final concentration of 100 nM siRNA duplex using Lipofectamine 2000 154 reagent (Invitrogen) according to the manufacturer’s instructions. The cells were analyzed 72 h after the transfection. Tagged proteins were transfected with Fugene 12 h before the immunofluorescence experiments. Reversal of phenotype (rescue) was performed as described in (Brodeur et al. 2009). Briefly, HeLa cells were transfected with cDNA encoding rat Calnuc-GFP, rat ΔSP-Calnuc-GFP, rat ΔEF-Calnuc-GFP or GFP alone using Fugene 8–10 h after the initial transfection with human Calnuc siRNA, and the cells were analyzed after 38–40 h. Immunofluorescence Cb cells were plated on coverslips. Twelve hours after the transfection, the cells were fixed for 30 min in 3% paraformaldehyde (PFA) 100 mM phosphate buffer, pH 7.4, permeabilized with 0.1% Triton X-100 for 10 min, blocked with 10% goat or fetal bovine serum for 30 min, and incubated with primary antibodies for 1 h at RT, followed by Alexa Fluor-488, 594 or 647-conjugated antibodies (Molecular Probes, Eugene, OR, USA) for 1 h at RT. The specimens were visualized using an inverted confocal laser-scanning microscope (FV1000, Olympus, Tokyo, Japan) equipped with a PlanApo 60x/1.42 oil immersion objective (Olympus, Tokyo, Japan). Olympus Fluoview software version 1.6a was used for image acquisition and analysis. The images were further processed using Adobe Photoshop (Adobe Systems, San Jose, CA, USA). The quantification of the immunofluorescence signal on endosome was previously described (Mamo et al. 2012). Electronic microscopy Cells were rinsed with PBS, prefixed for 15 min with a 1:1 mixture of culture medium (Dulbecco’s modified Eagle’s medium) and freshly prepared 2.8% glutaraldehyde in cacodylate buffer (0.1 M cacodylate and 7.5% sucrose) and then fixed for 30 min with 2.8% glutaraldehyde at room temperature. After two rinses, specimens were postfixed for 1 h with 2% osmium tetroxide in cacodylate buffer. Cells were then dehydrated using graded ethanol concentrations (40, 70, 90, 95, and 100%, three times each) and coated two times for 3 h with a thin layer of Araldite 502 resin (for ethanol substitution). Finally, the resin was allowed to polymerize at 60°C for 48 h. The specimens were detached from the plastic vessels, inverted in embedding molds, immersed in Araldite 502, and polymerized at 60°C 155 for 48 h. Ultramicrotome- prepared thin sections were contrasted with lead citrate and uranyl acetate and then observed on a Jeol 100 CX transmission electron microscope. All reagents were purchased from Electron Microscopy Sciences (Cedarlane, Hornby, ON, Canada). RNA analysis RNA analysis were performed by the Plateforme RNomique de l’Université de Sherbrooke. Total RNA extractions were performed on cell pellets using TRIzol (Invitrogen, Carlsbad, CA, USA) with chloroform, following the manufacturer’s protocol. The aqueous layer was recovered, mixed with one volume of 70% ethanol and applied directly to a RNeasy Mini Kit column (Qiagen, Venlo, Netherlands). DNAse treatment on the column and total RNA recovery were performed as per the manufacturer’s protocol. RNA quality and presence of contaminating genomic DNA was verified as previously described (Brosseau et al. 2010). RNA integrity was assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Reverse transcription was performed on 750 ng total RNA with Transcriptor reverse transcriptase, random hexamers, dNTPs (Roche Diagnostics, Basel, Switzerland), and 10 units of RNAseOUT (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol in a total volume of 20 µl. All forward and reverse primers were individually resuspended to 20–100 μM stock solution in Tris-EDTA buffer (IDT) and diluted as a primer pair to 1 μM in RNase DNase-free water (IDT). Quantitative PCR (qPCR) reactions were performed in 10 µl in 384 well plates on a CFX-384 thermocycler (Bio-Rad, Hercules, CA, USA) with 5 μl of 2X iTaq Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA), 10 ng (3 µl) cDNA, and 200 nM final (2 µl) primer pair solutions. The following cycling conditions were used: 3 min at 95°C; 50 cycles: 15 sec at 95°C, 30 sec at 60°C, 30 sec at 72°C. Relative expression levels were calculated using the qBASE framework (Hellemans et al. 2007) and the housekeeping genes Gapdh, Lamp1 and Clk2 for human cDNA. Primer design and validation was evaluated as described elsewhere (Brosseau et al. 2010). In every qPCR run, a no-template control was performed for each primer pair and these were consistently negative. All primer sequences are available in Supplemental Table 1. 156 Statistical analysis Experiments were performed at least in triplicate and results are expressed as means intensity of protein normalized with loading control ± SD. Statistical significance between two groups was assessed using the Student t-test whereas ANOVA was used for comparison with three groups or more. P value < 0.05 was considered significant (*) and p < 0.001 very significant (**). Results Calnuc interacts and colocalizes with CLN3 and CLN5. Because we previously showed that Calnuc operates on the same molecular components of the endosomal retrieval machinery (Rab7 and retromer) as CLN5, we first investigated whether Calnuc could interact with CLN5. We first performed immunoprecipitation experiments using HEK293 cells expressing GFP-tagged Calnuc or GFP together with HA-tagged CLN5, and the proteins were immunoprecipitated using an antibody against GFP (Figure 1A). We found an interaction between CLN5-HA and Calnuc-GFP, but not with GFP alone (Figure 1A). As the cytosolic pool of Calnuc has been involved in the recruitment of retromers acting on the activity of Rab7 and that CLN5 is a luminal protein, the interaction between those two proteins must be indirect, probably through a transmembrane protein. CLN3 is a six multipass transmembrane protein and a CLN5 and Rab7 interacting partner (Vesa et al. 2002; Lyly et al. 2009; Uusi-Rauva et al. 2012). We thus tested if Calnuc can interact with CLN3. In HEK293 cells expressing GFPtagged Calnuc or GFP together with HA-tagged CLN3, we found an interaction between CLN3-HA and Calnuc-GFP whereas no unspecific interaction was detect with GFP alone (Figure 1B). Given that CLN3 interacts directly with CLN5 (Vesa et al. 2002; Lyly et al. 2009), we also determined whether Calnuc can form a complex with both CLN5 and CLN3. HEK293 cells expressing Calnuc-GFP, CLN3-HA and SP-Flag-CLN5 were immunonoprecipitated using antibody against Flag or Myc (as negative control). No interaction was detected with anti-Myc antibody, but CLN3 and Calnuc co157 immunoprecipitated with CLN5 when anti-Flag antibody was used (Figure 1C). These results suggest that Calnuc form a protein complex with both CLN3 and CLN5. Figure 1 - Calnuc form a protein complex with CLN3 and CLN5 on endosomes. Coimmunoprecipitation of Calnuc with (A) CLN5 and (B) CLN3. Lysates from HEK cells transfected with GFP or Calnuc–GFP together with CLN3-HA or CLN5-HA were 158 immunoprecipitated with anti-GFP antibodies and then immunoblotted with anti-GFP or anti-HA to detect Calnuc and CLN3 or CLN5 respectively. Calnuc-GFP is observed as two bands, the lower of which most likely represents an N-terminal cleavage product. (C) Coimmunoprecipitation of Calnuc with both CLN3 and CLN5. Lysates from HEK cells transfected with Calnuc-GFP, CLN3-HA and SP-Flag-CLN5 were immunoprecipitated with anti-Myc of anti-Flag and then immunoblotted with anti-GFP to detect Calnuc, antiHA to detect CLN3 or with anti-HA to detect CLN5. (D) Comparison of the intracellular distribution of Calnuc, CLN3 and CLN5 in HeLa cells. (a, d, e) Calnuc-GFP, (b, d, f) CLN3-HA, (c, e, f) SP-Flag-CLN5 are found on enlarged early endosomes created by the expression of Rab5QL. The merged images (right insets) show a partial overlap (yellow) between (d) Calnuc-GFP and CLN3-HA, (e) Calnuc-GFP and SP-Flag-CLN5 or (f) CLN3HA and SP-Flag-CLN5. (Central pannel) The three proteins partially colocalize together on specific membrane microdomains of enlarged early endosomes. HeLa cells were transfected with Calnuc-GFP, CLN3-HA, SP-Flag-CLN5 and Rab5QL.Twelve hours after transfection, the cells were fixed, permeabilized and immunostained using anti-GFP, antiHA and anti-Flag antibodies. The labeled cells were examined by confocal fluorescence microscopy. Scale bar, 10 μm. To validate the presence of this protein complex in intact cells, we next compared the intracellular distribution of Calnuc-GFP, CLN3-HA and SP-Flag-CLN5 in HeLa cells using confocal microscopy. In order to help visualize the colocalization of Calnuc, CLN3 and CLN5, enlarged endosomes were created by expressing Rab5 GTPase-deficient mutant (Rab5Q79L) that increases homo- and heterotypic fusion, leading to the formation of enlarged early endosomes (Stenmark et al. 1994). As previously reported Calnuc-GFP predominantly localized in the Golgi region (Figure 1D, central panel), but a small amount is also detected on cytoplasmic face of Rab5-enlarged endosomes (Figure 1Da). CLN3-HA (Figure 1Db) and SP-Flag-CLN5 (Figure 1Dc) were mainly localized in Rab5-enlarged endosomes and are already known to colocalize (Figure 1Df). Both proteins also partially overlap with Calnuc in specific membrane microdomains (Figure 1Dd and 1De, respectively). When the staining of Calnuc-GFP, CLN3-HA and SP-Flag-CLN5 were superposed, we observe partial colocalization and close proximity of these protein in early endosomes microdomains (Figure 1D, central panel inset). Together, these immunofluorescence and biochemical interaction assays suggest that Calnuc interacts with CLN3 and CLN5, on microdomains of early endosomes, and that Calnuc would act on this complex to regulate endosomal sorting of lysosomal receptors. 159 Calnuc depletion alters the levels of CLN3 and CLN5. Because various NCL-causing mutations in CLN3 and CLN5 alter their intracellular distribution and level of expression (Jarvela et al. 1999; Isosomppi et al. 2002; Vesa et al. 2002; Bessa et al. 2006; Lebrun et al. 2009; Schmiedt et al. 2010) and because CLN3 and CLN5 depletion alter the trafficking of lysosomal receptors (Metcalf et al. 2008; Mamo et al. 2012), we next examined whether Calnuc knockdown alters the levels of CLN3 and CLN5. HeLa cells were transfected with control or Calnuc siRNA alone or together with CLN3-HA or CLN5-HA and were then analyzed by western blotting. CLN1-HA was also used as a control because it is a cargo molecule of MPR but does not affect its trafficking (Lefrancois et al. 2003; Jalanko and Braulke 2009). Endogenous Calnuc was strongly (≥ 90%) depleted at 72 h after transfection with specific small interfering RNA (siRNA), as observed in immunoblotting experiments (Figure 2A). The steady-state levels of CLN3-HA and CLN5-HA, but not CLN1-HA, were reduced in Calnuc knockdown cells compared with control cells (Figure 2A). Conversely, we next analyzed the effect of CLN3 and CLN5 depletion on Calnuc protein level and distribution. When CLN3 is knockdown in HeLa cells, we observe a reduction in endogenous Calnuc and CLN5-HA. Levels of Calnuc and CLN3-HA also decrease in CLN5 depleted cells. The specificity of the effect of Calnuc on the levels of CLN5 was also validated by a rescue experiment with siRNA-resistant forms of wild-type, Calnuc without calcium binding domains (ΔEF-Calnuc) and cytosolic Calnuc (ΔSP-Calnuc). Rat Calnuc was used as a rescue construct because it differs in many nucleotides from human siRNA target sequences, rendering it resistant to the siRNA. The results showed that reintroducing rat Calnuc-GFP, ΔEF-Calnuc-GFP or ΔSP-Calnuc-GFP into Calnuc-depleted cells restored the basal levels of CLN5-HA to levels similar to those of control cells (Figure 2B). These results suggest that the cytoplasmic pool of Calnuc, but not the EF-hands, participates in the stabilization of CLN5. We next examined whether the level of Calnuc is altered in CLN3 pathological culture cell model. We used two immortalized mouse cerebellar precursor cells lines established from wild-type mice (CbCln3+/+) (WT1 and WT2, Fig2C) or from a mouse model which contains a genetically mutation of CLN3 found in the majority of human patients (CbCln3Δex7/8/Δex7/8) (Fossale et al. 2004; Cao et al. 2006; Cao et al. 2011) (MT1 and MT2, Fig. 2C). This CLN3 mutation 160 Figure 2 - Calnuc knockdown reduces CLN3 and CLN5 levels and vice-versa. (A) Western blot analysis showing reduced levels of CLN3 and CLN5 in Calnuc-depleted cells, reduced levels of Calnuc and CLN5 in CLN3-depleted and reduced levels of Calnuc and CLN3 in CLN5 depleted-cells. Used as a negative control, CLN1-HA level was not altered upon the depletion of Calnuc. HeLa cells were transfected with control, Calnuc, CLN1, CLN3, or CLN5 siRNA for 3 days and CLN1-HA, CLN3-HA or CLN5-HA cDNA was transfected 12 h before the experiment. Proteins were separated by SDS-PAGE and immunoblotted with anti-HA, anti-Calnuc, and with anti-EEA1 as loading controls. (B) Steady state levels of CLN5 are irecovered in Calnuc-depleted cells rescued with siRNAresistant full-length Calnuc-GFP or cytoplasmic Calnuc-GFP (ΔSP-Calnuc-GFP) or calcium-binding deficient Calnuc-GFP (ΔEF-Calnuc-GFP), in contrast to cells rescued with GFP. HeLa cells treated with human Calnuc siRNA were transfected with GFP, siRNAresistant Calnuc-GFP or ΔSP-Calnuc-GFP or ΔEF-Calnuc-GFP at 5 h after knockdown. 161 The cells were lysed 3 days later, and the proteins were analyzed by western blotting. (C) Levels of Calnuc in cerebellar culture cells (Cb) from CLN3+/+ or CbCln3Δex7/8/Δex7/8 (MT) mice. Calnuc expression is reduced in MT compared to WT cells. (D) Lysates from WT and MT cerebellar cells were incubated with GST or GST-RILP, which specifically interacts with the active (GTP-bound) form of Rab7. Bound proteins were then analyzed by western blotting with anti-Rab7. is a 1.03 kb deletion causing exons 7 and 8 deletion and resulting in a protein harboring a frame shift at amino acid 153, leading to a truncated and unstable CLN3 protein. Western blot analysis of these cell lines indicated a lower level of endogenous Calnuc in CbCln3Δex7/8/Δex7/8 cells compared to CbCln3+/+ cells (Figure 2C). These results suggest that Calnuc, CLN3 and CLN5 could form a protein complex stabilized by their intramolecular interactions and that the downregulation of one member of this complex induces the downregulation of the other ones. These results strengthen our hypothesis that Calnuc, CLN3 and CLN5 are part of a functional protein complex. Given that Calnuc levels has been previously shown to alter Rab7 activity (Larkin et al. 2016) and that Calnuc is reduced in CbCln3Δex7/8/Δex7/8 cells, we next compared the level of active Rab7 (Rab7-GTP) in CbCln3Δex7/8/Δex7/8 (MT1 and MT2, Fig 2D) to CbCln3+/+ cells (WT1 and WT2, Fig 2D). We used an effector pull-down assay in which the Rab7binding domain of its effector protein RILP (Rab-interacting lysosomal protein) was used to selectively isolate GTP-loaded Rab7 (Romero Rosales et al. 2009). Recombinant GSTRILP immobilized on glutathione beads was incubated with lysates from CbCln3 +/+ and CbCln3Δex7/8/Δex7/8 cells, and bound (active) Rab7 proteins were analyzed by immunoblotting (Figure 2D). GTP-loaded Rab7 was reduced CbCln3Δex7/8/Δex7/8 compared to CbCln3+/+ cells (Figure 2D). Taken together, these results indicate that the level of Calnuc and active Rab7 are reduced in CbCln3Δex7/8/Δex7/8 cells, suggesting that this is the mechanism by which this genetically mutation of CLN3 found in the majority of human patients alters the lysosomal receptor trafficking. 162 Calnuc-knockdown cells exhibit the characteristic lysosomal features of NCL. Calnuc depletion causes the misdelivery and degradation of lysosomal receptors CIMPR and Sortilin in lysosomes (Larkin et al. 2016). A consequence of the defective transport of lysosomal sorting receptors is the missorting of lysosomal enzymes and consequently, the accumulation of undegraded substrates in lysosomes. In NCL, ceroid lipofuscins accumulate in lysosomes (Goebel and Wisniewski 2004; Jalanko and Braulke 2009). The accumulation of this material induces the enlargement of lysosomes as well as specific ultrastructural changes. Based on our observations that Calnuc interacts with and alters the levels of CLN3 and CLN5, two proteins involved in NCL, we investigated whether Calnuc depletion generate phenotype typical to lysosomal disease associated to CLN3 and CLN5 mutations. The morphology of the lysosomes were first be investigated by confocal microscopy in HeLa cells transfected with control or Calnuc siRNA. Using the lysosomal marker LAMP2, bright and large lysosomes were observed in Calnuc-depleted cells compares to control cells (Figure 3A), suggesting an enlargement of the lysosomes. In concordance, immunoblot analysis indicated that LAMP2 protein levels were increased in Calnuc-depleted cells compared to control (data not shown). To further characterize the lysosomal compartments, we next used the greater resolving power of transmission electron microscopy (EM) to examine the fine ultrastructural details of lysosomes in Calnucdepleted cells. We observed storage-like material that resembled curvilinear as well as autophagic vacuoles, which are commonly find in CLN3 and CLN5 NCL subtype (Figure 3B). Autophagy is emerging as a major pathway involved in NCL (Koike et al. 2005; Cao et al. 2006; Thelen et al. 2012). Autophagic vacuoles accumulate in different forms of NCL due to a defect in autophagic clearance through the lysosome (Koike et al. 2005; Cao et al. 2006; Thelen al. 2012). We therefore used LC3, considered one of the most specific markers for autophagosomes (Mizushima 2004) to examine autophagy defects in Calnuc depleted cells. Under normal conditions, LC3 is reported to be cytosolic; however, after autophagy induction, LC3 is lipidated and recruited onto autophagosome (Tanida et al. 2008). Control- and Calnuc-depleted HeLa cells transfected with LC3-GFP were treated 163 Figure 3 - Calnuc depletion alters lysosomes morphology, induces curvilinear-like pattern and activates autophagy. (A) Comparison of the lysosomes sizes using LAMP2 marker in control versus Calnuc depleted cells. HeLa cells were immunostained using anti-Calnuc and anti-LAMP2 antibodies. LAMP2 immunostaining showed enhanced signal intensity with less perinuclear clustering in Calnuc delpleted cells. (B) TEM analysis of inclusions in Calnuc 164 PC12 depleted cells is shown. Curvilinear pattern and large electron-dense inclusion are present in Calnuc depleted cells. 10,000 × magnification. (C) Level of chloroquine induced autophagy showed by LC3 in control and Calnuc knockdown cells. Calnuc depleted cells show increased level of LC3 compared to control cells. HeLa cells were transfected with control or Calnuc-specific siRNA for 3 days, fixed, permeabilized and immunostained. Confocal microscopy images were acquired using identical instrument settings. For LC3 experiment cells were also transfected with LC3-GFP cDNA using Fugene HD and were treated with 50 μM Chloroquine to induces autophagic vacuole formation 12 h before the experiment. Scale bar, 10 μm. with choloroquine over night to induce autophagy and analyze by confocal microscopy. In control siRNA treated HeLa, LC3-GFP was mainly detected in the cytoplasm and on small puncta that do not colocalize with the lysosomal marker LAMPII (Figure 3Ca-c). In contrast, Calnuc siRNA treated cells displayed large LC3-positive puncta (Figure 3Cd), that partially colocalize with LAMP2 (Figure 3Cf). These results indicate an accumulation of LC3-positive autophagosomes in Calnuc-depleted cells, in concordance with the presence of dense inclusions bodies associated with autophagy observed by electronic microscopy (Figure 3B). The pathological hallmark of NCL is the accumulation of autofluorescent ceroid lipofuscin deposits that are enriched in saposins or subunit C of the mitochondrial ATP synthase complex (Palmer et al. 1989; Palmer et al. 1992), in the lysosomes of neurons and other cell types. These lipofuscin deposits usually appear as small, punctate intracellular structures that are strongly fluorescent under any excitation ranging from 360 nm to 647 nm (Fossale et al. 2004; Lojewski et al. 2013). Using confocal microscopy with excitation wavelength of 408 nm, we therefore examined control- and Calnuc-depleted HeLa cells for the presence of autofluorescent lipofuscin deposits. No autofluorencent material was observed in the cytoplasm of control cells (Figure 4Aa). However, autofluorescent puncta that colocalize with lysosomal marker LAMP2 was detected in Calnuc-depleted cells (Figure 4Af). We next examined the cells for the presence of the pore-forming subunit C of the mitochondrial ATP synthase complex (SCMAS), considered as a specific markers for lipofuscin deposits. Control and Calnuc siRNA treated HeLa cells were immunostained with antibodies against endogenous LAMP2 and SCMAS and analyzed by confocal microscopy (Figure 4B). An accumulation of SCMAS puncta that partially colocalize with 165 LAMP2 was observed in Calnuc-depleted cells (Fig. 4Bd-f), consistent with the abnormal accumulation of lipofuscins occurring in a lysosomal compartment. Figure 4 - Calnuc depletion causes an accumulation of autofluorescent material and of SCMAS in lysosomes. 166 Autofluorescence and SCMAS immunostaining is shown in HeLa cells transfected with control or Calnuc siRNA for 3 days, fixed, permeabilized and immunostained. (A) Calnuc knockdown cells present autofluorescent material which was not observed in control cells and which strongly, though not perfectly, overlaps with lysosomal marker LAMP2. Cells were immunostained using anti-LAMP2 and autofluorescence was detected at 408 nm excitation wavelength. (B) Control cells exhibited limited SCMAS immunostaining whereas Calnuc depleted cells contained numerous SCMAS puncta wich colocalise with lysosomal marker LAMP2. Cells were immunostained using anti-LAMP2 and antiSCMAS. The labeled cells were examined by confocal fluorescence microscopy and images were acquired using identical instrument settings. Scale bar, 10 μm. Calnuc expression levels are reduced in NCL fibroblasts. Based on our results suggesting that the depletion of Calnuc induces morphological features reminiscent of NCL, and that Calnuc interacts and regulates CLN proteins involved in this disease, we hypothesized that Calnuc can be involved in ceroid lipofuscinosis pathogenesis. This hypothesis was tested by analyzing the expression level of Calnuc expression in fibroblast cell lines from NCL patients (Figure 5). Expression profiling of Calnuc was performed by comparing Calnuc protein and mRNA levels in fibroblasts from four apparently healthy individual acting as controls (CTL), in one fibroblast from patients suffering from Niemann Pick disease (NPD), a related lysosomal storage disease, four genetically confirmed cases of NCL, including one from CLN1 patient, one from CLN3 patient (homozygous for 1.02-kb deletion, the most common mutation), two from CLN5 patients (p.Leu358AlafsX4 and p.Thr176AsnfsX11 mutants), as well as five orphan cases of NCL (Orphan) which remain undefined genetically, i.e., not involving CLN mutation (Mole et al. 2011; Kousi et al. 2012). Immunoblot analysis indicated that all patients suffering from NCL, both genetically confirmed and orphan, expressed lower level of Calnuc protein compared to the healthy or NPD controls (Figure 5A). Quantification analysis of Calnuc levels normalized to the loading control early endosome antigen 1 (EEA1) confirmed that Calnuc expression was approximately 50% lower in fibroblasts from CLN1, CLN3, CLN5 and orphans patients compared to control fibroblasts (Figure 5A). Interestingly, no decrease in Calnuc protein levels was observed in NPD fibroblasts suggesting a Calnuc independent mechanism for the pathogenesis of this 167 Figure 5 - Calnuc levels in healthy and NCL fibroblasts. Analysis of Calnuc protein and mRNA levels in control (CTL) and Neuronal ceroid lipofuscinosis (NCL) fibroblasts. (A) Calnuc protein expression in fibroblasts of healthy (CTL), Niemann pick disease type C1 (NPD) and NCL patients was compared by Western blotting. Representative data showing lysates subjected to SDS-PAGE and immunoblotted with antibodies directed against Calnuc and EEA1. (B) Densitometric analysis of Western blots, such as those shown in (A), normalized to the signal of EEA1. Results are expressed as means ± SD (n ≥ 3) compared with the mean of healthy patients. (C) Calnuc, CLN1, CLN3 and CLN5 mRNA levels in fibroblasts of CTL, NPD and NCL patients were compared by qRT-PCR. Total mRNA was reverse transcribed, and the levels of Calnuc, CLN1, CLN3 and CLN5 cDNA were analyzed by qPCR with iTaq Universal SYBR Green 168 Supermix and were expressed relative to the housekeeping genes Gapdh, Lamp1 and Clk2 using the qBASE framework. Results are expressed as means ± SD (n = 5 samples, in duplicate). The difference between CTL and NCL was not significant. other lysosomal storage disease. To determine whether the decrease in Calnuc protein levels was due to a change in gene expression, total quantitative RT-PCR assays were next performed on the same fibroblasts cells lines (Figure 5B). Results were normalized using the housekeeping genes Gapdh, Lamp1 and Clk2 and expressed relatively to the first arbitrary establish control. No significant difference was observed between Calnuc mRNA levels in all the NCL cells, NPD cells and control cells. However, as expected, we noted that CLN1 fibroblasts shows highly reduced levels of CLN1 mRNA as well as CLN3 mRNA in CLN3 fibroblasts and CLN5 mRNA in one of the CLN5 fibroblast cells line. In summary, our results indicated that Calnuc protein levels, but not mRNA levels, are lower in fibroblasts from NCL patients. Discussion We previously showed that the cytosolic pool of Calnuc plays an essential role in endosomes-to-TGN transport of the lysosomal receptors MPR and Sortilin. We reported that, through an indirect interaction, Calnuc regulates membrane association and activation of Rab7, a small G protein required for the endosomal recruitment of retromers and crucial for the retrieval of lysosomal receptors to the TGN (Larkin et al. 2016). In the present study, we extended our previous findings and provide further mechanistic insight into how Calnuc regulates Rab7 activity. We demonstrated that Calnuc interacts with and stabilizes CLN3 and CLN5, two proteins involved in the proper trafficking of lysosomal receptors through the interaction with and regulation of Rab7. CLN3 and CLN5 are mutated/altered in NCL, a rare lysosomal storage disease. Interestingly, we observed that the depletion of Calnuc induced cellular morphological features reminiscent of NCL and that Calnuc expression is reduced in fibroblasts from NCL patients. Overall, this study suggests that Calnuc function is altered in human NCL and that Calnuc could be a novel protein involved in NCL pathogenesis. 169 Our results demonstrate that Calnuc forms a complex with CLN3 and CLN5 on endosomes. CLN3 is known to interact directly with CLN5 and our immunoprecipitation results showed that Calnuc can interact with both CLN3 and CLN5, suggesting that they are in the same protein complex. Using confocal microscopy, all three proteins were found together on early endosomes microdomains. Furthermore, we showed that knockdown of one of the protein of the complex (Calnuc, CLN3 or CLN5) induces the downregulation of the two other partners. Similarly, we observed a reduction of Calnuc level in mouse cerebellar precursor cell lines established from a mouse model containing a genetically mutation of CLN3 (CbCln3Δex7/8/Δex7/8) (Fossale et al. 2004; Cao et al. 2006; Cao et al. 2011) in which CLN3 protein level was reported as reduced (Fossale et al. 2004). This codepletion phenomenon is often observed in protein complexes (e.g., retromers, nucleoporins) (Boehmer et al. 2003; Arighi et al. 2004) and strengthens our hypothesis that Calnuc, CLN3 and CLN5 are part of a functional protein complex. In concordance, CLN5 could be rescued by the expression of the cytoplasmic form of Calnuc (ΔSP-Calnuc-GFP), suggesting the implication of the transmembrane CLN3. Based on these findings, we propose a model (Figure 6) in which cytosolic Calnuc interacts on endolysosomal membranes with the 6TM protein CLN3, which can in turn associated with the highly glycosylated luminal protein CLN5. Once assembled, this complex would modulate the recruitment and activation of Rab7, leading to lysosomal receptor recycling to the TGN. Indeed, the loss of Calnuc and CLN5 led to a reduction of Rab7-GTP, a decrease in the recruitment of retromers on endosomal membrane and the missorting of the lysosomal receptors MPR and Sortilin (Mamo et al. 2012; Larkin et al. 2016). CLN3 has been reported to bind directly Rab7, preferentially Rab7-GTP, and various retro and anterograde microtubular motor complexes responsible for the movement of late endosomes and lysosomes (Uusi-Rauva et al. 2012). However, CLN3 has never been shown to regulate Rab7 activity. We found that the level of active Rab7 (Rab7-GTP) was reduced in CbCln3Δex7/8/Δex7/8 cells, suggesting that an alteration of CLN3 levels also modulates Rab7 activity. Thus, it is tempting to speculate the complex Calnuc-CLN3-CLN5 is involved in the regulation of Rab7 activity. But how is this complex activated in order to promote the sorting of lysosomal receptors? Based on previous finding by Mamo et al showing that CLN5 interact with Sortilin (Mamo et al. 2012), we propose that when the cargo-loaded 170 lysosomal sorting receptor arrives at the endosome, the change in pH causes a dissociation of the cargo from the receptor that subsequently leads to interaction with CLN5. This would provide a signal to initiate their retrograde transport to the TGN. Through direct interaction between CLN5 and CLN3, and possibly conformational modifications of CLN3, this signal would cross the membrane to reach Calnuc. On the cytosolic side, Calnuc would next participate in the activation of Rab7 and, subsequently, in the recruitment of the retromer retrograde machinery. However, the manner by which Calnuc alters the amount of active Rab7 requires further clarification. This model is in accordance with the favored binding of active GTP-bound form of Rab7 with the N-terminal domain of CLN3 and on the reported shorter residence of Rab7 on endolysosomal membranes devoided of CLN3 protein. This unbalanced GTP/GDP cycle of Rab7 would be caused by the CLN5-CLN3 signal unable to reach Calnuc. Figure 6 - Proposed schematic model for Calnuc-CLN3-CLN5 complex influencing receptors retrograde transport. CLN5 is tightly associated with the lumen of the endosomal membrane. After cargo dissociation from the lysosomal sorting receptor, CLN5 binds to it and subsequently to CLN3, a transmembrane protein. This CLN5-CLN3 interaction would acts as a signal to initiate the action of Calnuc on the cytosolic side of the endosome, the activation of Rab7 leading to the recruitment of retromer and the retrograde transport of the lysosomal receptor to the TGN. The figure is not to scale. 171 One question that remains is how the pathological mutations identified in CLN5 and CLN3 affect the sorting machinery leading to lysosomal disorders. Some CLN5 mutants such as FinM, SWE and EUR maintain their physical abilities to interact with CLN3 (Lyly et al. 2009). However, in some cases, they are unstable and rapidly degraded (SWE and NFL) or mislocated to the endoplasmic reticulum (EUR, FinM, SWE, and NFL) (Schmiedt et al. 2010; Larkin et al. 2013; Moharir et al. 2013) Thus, it is possible that interactions with lysosomal receptors may be compromised in some CLN5 mutants. Similarly, mutations in CLN3 can lead to their incapacity to interact with Calnuc or Rab7 proteins. For instance, CbCln3Δex7/8/Δex7/8 mutation results in a truncated protein that is retained in the ER where it cannot interact with endolysosomal proteins. To further understand the underlying mechanism by which Calnuc-CLN3-CLN5 complex regulates lysosome receptor trafficking and its potential implication in NCL pathogenesis, the specific regions of interaction of these proteins need to be characterized as well as how NCL-associated mutations in CLN5 and CLN3 affect the complex. Based on our observations that Calnuc form a complex with CLN3 and CLN5, two proteins involved in NCL, we next investigated whether Calnuc depletion generate phenotype typical to this lysosomal storage disease. Using confocal and electron microscopy, we showed that loss of Calnuc induces cellular morphological features reminiscent of NCL, such as lysosome enlargement, accumulation of autofluorescent material and of subunit C of mitochondrial ATP synthase (SCMAS), as well as the presence of storage-like material that resembled curvilinear deposits and an increase of autophagy vacuoles. To this day, these structure and content of storage material are still used as diagnostic clues to help define the lysosomal defective function in NCL and suggest that Calnuc deficiency affect the endosomal/lysosomal system and could cause lysosomal storage disease. The cellular changes observed upon Calnuc depletion are also highly consistent with what has been described in animal models and in patients with mutations in CLN3 and CLN5 genes, suggesting the Calnuc, CLN3 and CLN5 cooperate in a shared pathway or complex. Indeed, CLN3 and CLN5 mutations are typically associated with fingerprint, curvilinear or rectilinear lysosomal ultrastructures (Santavuori 1988; Mole et al. 2005; Pineda-Trujillo et al. 2005), combined with SCMAS as the main component of 172 the storage material (Palmer et al. 1992; Tyynela et al. 1997; Tyynela et al. 1997). In contrast, CLN1 mutation is associated with the accumulation of sphingolipid-activator proteins A and D storage material and granular osmiophilic deposits (Carpenter et al. 1973; Das, Becerra et al. 1998; Wisniewski et al. 1998; Santavuori et al. 2000; Gupta et al. 2001; Van Diggelen et al. 2001; Ramadan et al. 2007), suggesting its implication in a different complex/pathway than CLN3 and CLN5. Disrupted autophagic clearance through to the lysosome is thought to be particularly damaging for vulnerable neuron cells and is rightly associated with neurodegenerative disorders like NCL and Alzheimer's disease. Interestingly, recent evidences lead toward lysosomal calcium signalling as a major regulator of lysosome biogenesis and autophagy through calcineurin-mediated induction of TFEB (Medina et al. 2015). As Calnuc possesses the ability to bind calcium, it may respond to such local calcium increase to act in this pathway. The observation that Calnuc forms a complex with NCL proteins and that loss of Calnuc induces cellular morphological features reminiscent of NCL suggests a potential role of Calnuc in ceroid lipofuscinosis pathogenesis. Interestingly, the analysis of Calnuc expression profile in human fibroblasts from genetically-confirmed cases of NCL and orphan NCL (NCL phenotypes but undefined genes) indicated a lower level of Calnuc protein in CLN3, CLN5 as well as all the orphan NCL fibroblasts tested. These results are in accordance with our data showing a decrease of Calnuc protein level in CLN3 and CLN5 siRNA depleted Hela cells. It is noteworthy that Calnuc is not affected in fibroblasts from patients suffering from Niemann-Pick disease (NPD) patients, a related lysosomal storage disease, suggesting the specific involvement of Calnuc in NCL associated lysosomal storage disease. On the other hand, the reduction of Calnuc protein level in fibroblasts from CLN1 patient may indicate its implication with the Calnuc-CLN3-CLN5 complex. CLN1 is an enzyme which removes fatty acids from cysteine residues in lipid-modified proteins (Camp et al. 1994). CLN1 colocalizes and interacts with CLN5 (Lyly et al. 2009) and has been proposed to facilitate its lysosomal transport (Schmiedt et al.). Therefore, CLN1 activity may impact the complex formation and Calnuc stability. Although these preliminary data suggest a link between Calnuc and NCL, the interpretation is limited by the low amount of samples analyzed due to the limited availability of the samples. 173 Further analysis of Calnuc expression profile using qRT-PCR analysis ruled out the possibility that the decrease in Calnuc protein level was caused by an alteration of its mRNA levels. No significant change in Calnuc mRNA was detected in all the analyzed fibroblasts from LSD patients. However, mutations in Calnuc could be causative of its defective function without necessarily influencing its level of expression. Indeed, various mutations that affect the CLN genes lead to hasten mRNA degradation, result in truncated proteins that are nonfunctional or less stable, alter interaction with partners or cause their mislocalization (Kousi et al. 2012). For example, in fibrobasts from CLN5 (p.Leu358AlafsX4) patient, the mRNA level of CLN5 was not altered whereas in fibrobasts from CLN5 (p.Thr176AsnfsX11) patient, in which a STOP codon is inserted at aa 176 due to a frameshift, the mRNA level is reduced, probably due to the instability of the transcript. Our data also confirmed previous work reporting a lower level of CLN3 mRNA in CLN3 mutant fibroblasts. CLN1 mRNA was previously reported to be upregulated in a mouse model of CLN5 FinM mutation (Lyly et al. 2009). No such transcriptional behavior related to a mutation was observed. However, we noted that the mRNA levels of Calnuc, CLN3 and CLN5 increase together in some orphan cell lines without being reflected as an increase in Calnuc protein levels. Given that genetically undefined forms of NCL have been reported, it would be interesting to determine whether Calnuc is mutated in the orphan cases of NCL, potentially identifying Calnuc as a new gene involved in NCLs. On the other hand, the direct involvement of Calnuc in genetically-confirmed cases of CLN would redefine the paradigm of NCL classification: it would not solely be based on the CLN gene mutated but also on the expression of disease-modifier genes. These genes, such as Calnuc, could alter the expression of CLN proteins and aggravate the progression of the disease. These modifier genes could explain the high phenotype variability observed in CLN patients with an identical CLN genotype. In summary, herein, we have showed the first evidence for a role of Calnuc in neuronal ceroid lipofuscinosis through novel interaction with CLN3 and CLN5. The opportunity to identify new players in the disease provides new insights into the understanding of the cellular mechanism and physiopathology. Above all, it offers the possibility to develop new diagnostic tools and best chance of developing treatments. 174 Acknowledgments We are grateful to Susan Cotman (Harvard Medical School) and Sara Mole (University College London) for cells and antibodies link to NCL pathology. H.L. holds a fellowship from the National Sciences and Engineering Research Council of Canada (NSERC). This work was supported by grants from the Canadian Institutes for Health Research and a Canada Research Chair to C.L.L. The authors declare that they have no conflict of interest. 175 References Arighi, C. N., L. M. Hartnell, et al. (2004). "Role of the mammalian retromer in sorting of the cation-independent mannose 6-phosphate receptor." J Cell Biol 165(1): 123-133. Ballif, B. A., N. V. Mincek, et al. (1996). "Interaction of cyclooxygenases with an apoptosis- and autoimmunity-associated protein." Proc Natl Acad Sci U S A 93(11): 5544-5549. Bessa, C., C. A. Teixeira, et al. (2006). "Two novel CLN5 mutations in a Portuguese patient with vLINCL: insights into molecular mechanisms of CLN5 deficiency." Mol Genet Metab 89(3): 245-253. Boehmer, T., J. Enninga, et al. (2003). "Depletion of a single nucleoporin, Nup107, prevents the assembly of a subset of nucleoporins into the nuclear pore complex." Proc Natl Acad Sci U S A 100(3): 981-985. Boucher, R., H. Larkin, et al. (2008). "Intracellular trafficking of LRP9 is dependent on two acidic cluster/dileucine motifs." Histochem Cell Biol 130(2): 315-327. Brodeur, J., H. Larkin, et al. (2009). "Calnuc binds to LRP9 and affects its endosomal sorting." Traffic 10(8): 1098-1114. Brosseau, J. P., J. F. Lucier, et al. (2010). "High-throughput quantification of splicing isoforms." Rna 16(2): 442-449. Camp, L. A., L. A. Verkruyse, et al. (1994). "Molecular cloning and expression of palmitoyl-protein thioesterase." J Biol Chem 269(37): 23212-23219. Cao, Y., J. A. Espinola, et al. (2006). "Autophagy is disrupted in a knock-in mouse model of juvenile neuronal ceroid lipofuscinosis." J Biol Chem 281(29): 20483-20493. Cao, Y., J. F. Staropoli, et al. (2011). "Distinct early molecular responses to mutations causing vLINCL and JNCL presage ATP synthase subunit C accumulation in cerebellar cells." PLoS One 6(2): e17118. Carpenter, S., G. Karpati, et al. (1973). "A type of juvenile cerebromacular degeneration characterized by granular osmiophilic deposits." J Neurol Sci 18(1): 67-87. Cotman, S. L. and J. F. Staropoli (2012). "The juvenile Batten disease protein, CLN3, and its role in regulating anterograde and retrograde post-Golgi trafficking." Clin Lipidol 7(1): 79-91. Das, A. K., C. H. Becerra, et al. (1998). "Molecular genetics of palmitoyl-protein thioesterase deficiency in the U.S." J Clin Invest 102(2): 361-370. Fossale, E., P. Wolf, et al. (2004). "Membrane trafficking and mitochondrial abnormalities precede subunit C deposition in a cerebellar cell model of juvenile neuronal ceroid lipofuscinosis." BMC Neurosci 5: 57. Goebel, H. H. and K. E. Wisniewski (2004). "Current state of clinical and morphological features in human NCL." Brain Pathol 14(1): 61-69. Gupta, P., A. A. Soyombo, et al. (2001). "Disruption of PPT1 or PPT2 causes neuronal ceroid lipofuscinosis in knockout mice." Proc Natl Acad Sci U S A 98(24): 1356613571. Hellemans, J., G. Mortier, et al. (2007). "qBase relative quantification framework and software for management and automated analysis of real-time quantitative PCR data." Genome Biol 8(2): R19. 176 Isosomppi, J., J. Vesa, et al. (2002). "Lysosomal localization of the neuronal ceroid lipofuscinosis CLN5 protein." Hum Mol Genet 11(8): 885-891. Jalanko, A. and T. Braulke (2009). "Neuronal ceroid lipofuscinoses." Biochim Biophys Acta 1793(4): 697-709. Jarvela, I., M. Lehtovirta, et al. (1999). "Defective intracellular transport of CLN3 is the molecular basis of Batten disease (JNCL)." Hum Mol Genet 8(6): 1091-1098. Kama, R., V. Kanneganti, et al. (2011). "The yeast Batten disease orthologue Btn1 controls endosome-Golgi retrograde transport via SNARE assembly." J Cell Biol 195(2): 203-215. Kama, R., M. Robinson, et al. (2007). "Btn2, a Hook1 ortholog and potential Batten disease-related protein, mediates late endosome-Golgi protein sorting in yeast." Mol Cell Biol 27(2): 605-621. Kawano, J., T. Kotani, et al. (2000). "CALNUC (nucleobindin) is localized in the Golgi apparatus in insect cells." Eur J Cell Biol 79(3): 208-217. Koike, M., M. Shibata, et al. (2005). "Participation of autophagy in storage of lysosomes in neurons from mouse models of neuronal ceroid-lipofuscinoses (Batten disease)." Am J Pathol 167(6): 1713-1728. Kousi, M., A. E. Lehesjoki, et al. (2012). "Update of the mutation spectrum and clinical correlations of over 360 mutations in eight genes that underlie the neuronal ceroid lipofuscinoses." Hum Mutat 33(1): 42-63. Larkin, H., Costantino, S and Lavoie, C (2016). "Calnuc function in endosomal sorting of lysosomal receptors." Traffic. 2016 Jan 12. doi:10.1111/ tra.12374. [Epub ahead of print] Larkin, H., M. G. Ribeiro, et al. (2013). "Topology and membrane anchoring of the lysosomal storage disease-related protein CLN5." Hum Mutat 34(12): 1688-1697. Larkin, H., C. Thériault, et al. (2013). "Topology and membrane anchoring of the lysosomal storage disease-related membrane protein CLN5." Human mutation Submitted. Lavoie, C., T. Meerloo, et al. (2002). "Calnuc, an EF-hand Ca(2+)-binding protein, is stored and processed in the Golgi and secreted by the constitutive-like pathway in AtT20 cells." Mol Endocrinol 16(11): 2462-2474. Lebrun, A. H., S. Storch, et al. (2009). "Retention of lysosomal protein CLN5 in the endoplasmic reticulum causes neuronal ceroid lipofuscinosis in Asian sibship." Hum Mutat 30(5): E651-661. Lefrancois, S., J. Zeng, et al. (2003). "The lysosomal trafficking of sphingolipid activator proteins (SAPs) is mediated by sortilin." EMBO J 22(24): 6430-6437. Lin, P., T. Fischer, et al. (2009). "Calnuc plays a role in dynamic distribution of Galphai but not Gbeta subunits and modulates ACTH secretion in AtT-20 neuroendocrine secretory cells." Mol Neurodegener 4: 15. Lin, P., H. Le-Niculescu, et al. (1998). "The mammalian calcium-binding protein, nucleobindin (CALNUC), is a Golgi resident protein." J Cell Biol 141(7): 15151527. Lin, P., Y. Yao, et al. (1999). "Overexpression of CALNUC (nucleobindin) increases agonist and thapsigargin releasable Ca2+ storage in the Golgi." J Cell Biol 145(2): 279-289. 177 Lojewski, X., J. F. Staropoli, et al. (2013). "Human iPSC models of neuronal ceroid lipofuscinosis capture distinct effects of TPP1 and CLN3 mutations on the endocytic pathway." Hum Mol Genet. Lyly, A., C. von Schantz, et al. (2009). "Novel interactions of CLN5 support molecular networking between Neuronal Ceroid Lipofuscinosis proteins." BMC Cell Biol 10: 83. Mamo, A., F. Jules, et al. (2012). "The role of ceroid lipofuscinosis neuronal protein 5 (CLN5) in endosomal sorting." Mol Cell Biol 32(10): 1855-1866. Medina, D. L., S. Di Paola, et al. (2015). "Lysosomal calcium signalling regulates autophagy through calcineurin and TFEB." Nat Cell Biol 17(3): 288-299. Metcalf, D. J., A. A. Calvi, et al. (2008). "Loss of the Batten disease gene CLN3 prevents exit from the TGN of the mannose 6-phosphate receptor." Traffic 9(11): 1905-1914. Miura, K., M. Hirai, et al. (1996). "Organization of the human gene for nucleobindin (NUC) and its chromosomal assignment to 19q13.2-q13.4." Genomics 34(2): 181186. Miura, K., K. Titani, et al. (1992). "Molecular cloning of nucleobindin, a novel DNAbinding protein that contains both a signal peptide and a leucine zipper structure." Biochem Biophys Res Commun 187(1): 375-380. Mizushima, N. (2004). "Methods for monitoring autophagy." Int J Biochem Cell Biol 36(12): 2491-2502. Mole, S. E., R. E. Williams, et al. (2005). "Correlations between genotype, ultrastructural morphology and clinical phenotype in the neuronal ceroid lipofuscinoses." Neurogenetics 6(3): 107-126. Mole, S. E., R. E. Williams, et al. (2011). The neuronal Ceroid Lipofuscinoses (Batten Disease), Oxford University Press. Palmer, D. N., I. M. Fearnley, et al. (1989). "Lysosomal storage of the DCCD reactive proteolipid subunit of mitochondrial ATP synthase in human and ovine ceroid lipofuscinoses." Adv Exp Med Biol 266: 211-222; discussion 223. Palmer, D. N., I. M. Fearnley, et al. (1992). "Mitochondrial ATP synthase subunit C storage in the ceroid-lipofuscinoses (Batten disease)." Am J Med Genet 42(4): 561-567. Pineda-Trujillo, N., W. Cornejo, et al. (2005). "A CLN5 mutation causing an atypical neuronal ceroid lipofuscinosis of juvenile onset." Neurology 64(4): 740-742. Ramadan, H., A. S. Al-Din, et al. (2007). "Adult neuronal ceroid lipofuscinosis caused by deficiency in palmitoyl protein thioesterase 1." Neurology 68(5): 387-388. Romero Rosales, K., E. R. Peralta, et al. (2009). "Rab7 activation by growth factor withdrawal contributes to the induction of apoptosis." Mol Biol Cell 20(12): 28312840. Santavuori, P. (1988). "Neuronal ceroid-lipofuscinoses in childhood." Brain Dev 10(2): 8083. Santavuori, P., L. Lauronen, et al. (2000). "Neuronal ceroid lipofuscinoses in childhood." Neurol Sci 21(3 Suppl): S35-41. Schmiedt, M. L., C. Bessa, et al. "The neuronal ceroid lipofuscinosis protein CLN5: new insights into cellular maturation, transport, and consequences of mutations." Hum Mutat 31(3): 356-365. 178 Schmiedt, M. L., C. Bessa, et al. (2010). "The neuronal ceroid lipofuscinosis protein CLN5: new insights into cellular maturation, transport, and consequences of mutations." Hum Mutat 31(3): 356-365. Stenmark, H., R. G. Parton, et al. (1994). "Inhibition of rab5 GTPase activity stimulates membrane fusion in endocytosis." Embo J 13(6): 1287-1296. Tanida, I., T. Ueno, et al. (2008). "LC3 and Autophagy." Methods Mol Biol 445: 77-88. Taniguchi, N., H. Taniura, et al. (2000). "The postmitotic growth suppressor necdin interacts with a calcium-binding protein (NEFA) in neuronal cytoplasm." J Biol Chem 275(41): 31674-31681. Thelen, M., M. Damme, et al. (2012). "Correction: Disruption of the Autophagy-Lysosome Pathway Is Involved in Neuropathology of the nclf Mouse Model of Neuronal Ceroid Lipofuscinosis." PLoS One 7(5). Tyynela, J., J. Suopanki, et al. (1997). "Sphingolipid activator proteins (SAPs) in neuronal ceroid lipofuscinoses (NCL)." Neuropediatrics 28(1): 49-52. Tyynela, J., J. Suopanki, et al. (1997). "Variant late infantile neuronal ceroid-lipofuscinosis: pathology and biochemistry." J Neuropathol Exp Neurol 56(4): 369-375. Uusi-Rauva, K., A. Kyttala, et al. (2012). "Neuronal ceroid lipofuscinosis protein CLN3 interacts with motor proteins and modifies location of late endosomal compartments." Cell Mol Life Sci. Van Diggelen, O. P., J. L. Keulemans, et al. (2001). "Pre- and postnatal enzyme analysis for infantile, late infantile and adult neuronal ceroid lipofuscinosis (CLN1 and CLN2)." Eur J Paediatr Neurol 5 Suppl A: 189-192. Vesa, J., M. H. Chin, et al. (2002). "Neuronal ceroid lipofuscinoses are connected at molecular level: interaction of CLN5 protein with CLN2 and CLN3." Mol Biol Cell 13(7): 2410-2420. Weiss, T. S., C. E. Chamberlain, et al. (2001). "Galpha i3 binding to calnuc on Golgi membranes in living cells monitored by fluorescence resonance energy transfer of green fluorescent protein fusion proteins." Proc Natl Acad Sci U S A 98(26): 14961-14966. Wendel, M., Y. Sommarin, et al. (1995). "Isolation, characterization, and primary structure of a calcium-binding 63-kDa bone protein." J Biol Chem 270(11): 6125-6133. Wimley, W. C. and S. H. White (1996). "Experimentally determined hydrophobicity scale for proteins at membrane interfaces." Nat Struct Biol 3(10): 842-848. Wisniewski, K. E., F. Connell, et al. (1998). "Palmitoyl-protein thioesterase deficiency in a novel granular variant of LINCL." Pediatr Neurol 18(2): 119-123. 179 Supplementary figures Table S1 - Primer sequences for qRT-PCR 180 DISCUSSION Mise en contexte - Calnuc lie LRP9 et influence son triage endosomial La présente thèse a pour but de déterminer la fonction de la protéine Calnuc dans le transport entre le TGN et les endosomes. L'étude fait suite à un premier article dans lequel notre groupe a montré que Calnuc joue un rôle dans le transport rétrograde de LRP9, un membre peu caractérisé de la famille des récepteurs au LDL, qui cycle entre le TGN et les endosomes (Sardiello et al., 2009). Ce dernier est la première protéine transmembranaire à être identifiée comme partenaire d'interaction direct pour Calnuc. En effet, la partie Nterminale de Calnuc interagit avec une région riche en arginine de la queue cytosolique de LRP9. Ceci va de pair avec la colocalisation partielle de LRP9 avec la forme cytosolique de Calnuc au niveau de la surface du TGN et des endosomes précoces. Toutefois, la région riche en arginines semble insuffisante pour permettre l'interaction indiquant que d'autres déterminants structuraux pourraient être impliqués. Pour déterminer l'impact de Calnuc sur la distribution subcellulaire et les niveaux de LRP9, nous avons procédé à la déplétion de Calnuc par ARN interférent. Il en résulte une redistribution de LRP9 vers les lysosomes où il est dégradé. Ce phénomène est démontré par le traitement avec des inhibiteurs d'enzymes lysosomiales qui préviennent la diminution des niveaux des récepteurs. L'expression du mutant cytosolique de Calnuc contrecarre aussi l'effet ce qui indique que c'est bien cette fraction qui est en cause dans le triage de LRP9. Cette première étude tend donc à démontrer que Calnuc pourrait jouer un rôle dans le recyclage de LRP9 vers le TGN de manière à prévenir sa dégradation. Toutefois, de plus amples travaux étaient nécessaires afin de déterminer le mécanisme précis par lequel Calnuc régule le transport rétrograde de LRP9 et potentiellement d'autres récepteurs. En effet, Calnuc interagit également avec LRP3, qui possède une région riche en arginines, et influence son transport de la même manière que pour LRP9. À l'opposé, aucune interaction et aucun effet n'est observé avec LRP1 qui ne contient pas de région riche en arginines. Ces données tendent à démontrer que Calnuc aurait un effet sur certains récepteurs spécifiques plutôt qu'un effet général. 181 Article 1 - Rôle de Calnuc dans le triage endosomial des récepteurs lysosomiaux Le premier article qui fait l'objet du présent ouvrage visait donc à déterminer si Calnuc joue un rôle dans le transport rétrograde d'autres récepteurs et à mettre en lumière le mécanisme moléculaire qui sous-tend ce rôle. L'étude s'est orientée vers CI-MPR et Sortiline, deux récepteurs bien caractérisés qui transitent entre le TGN et les endosomes et qui permettent l'acheminement d'enzymes solubles vers le lysosome. Nous avons montré que Calnuc interagit avec la queue cytosolique de Sortiline et de CI-MPR et colocalise avec ces deux récepteurs à la surface des endosomes. À l'image de l'effet observé sur LRP9 et LRP3, la déplétion de Calnuc cause une redistribution de CI-MPR et Sortiline vers les lysosomes où ils sont dégradés entraînant une diminution des niveaux intracellulaires des récepteurs. Cet effet peut être prévenu par un traitement avec des inhibiteurs d'enzymes lysosomiales et peut être renversé par la surexpression du mutant cytosolique de Calnuc indiquant, encore une fois, que cette fraction est en cause. De surcroît, la diminution de CIMPR, associées à l'absence de Calnuc, limite le transport de son cargo CTSD à la sortie du TGN. La maturation intracellulaire de CTSD est alors limitée et il s'en suit une augmentation de la sécrétion de la forme mature dans le milieu extracellulaire. Ces résultats indiquent que Calnuc est impliquée dans le transport rétrograde des récepteurs lysosomiaux CI-MPR et Sortiline et, indirectement, dans la livraison subséquente de leurs cargos dans la voie endosomiale. Nos études montrent que Calnuc interagit avec de nombreux récepteurs qui cyclent entre le Golgi et les endosomes, dont CI-MPR, CD-MPR (figure 9E), Sortiline, LRP3 et LRP9, de façon à prévenir leur acheminement aux lysosomes. Des résultats préliminaires obtenus par notre groupe suggèrent que la fonction de Calnuc pourrait s'étendre à des récepteurs internalisés à la membrane plasmique, dont les récepteurs de type tyrosine-kinase Tropomyosin receptor kinase (Trk)A (figure 9D), TrkB, TrkC, de même que la protéase à sérine Furine (figure 9A), aussi connue sur le nom de PACE (Paired basic amino acid cleaving enzyme). Par contre, aucune interaction n'a été décelée avec LRP1, Neurotrophin receptor (p75/NTR) (figure 9B) et le Epidermal growth factor receptor (EGFR) (figure 9C). Initialement, un segment amphiphile riche en arginines de la queue cytosolique de LRP9 a été identifié comme étant important, mais non essentiel pour l'interaction avec Calnuc, laissant croire que d'autres déterminants structuraux pourraient 182 être en cause. En effet, une brève étude structurale a montré que ce segment n'est pas un élément commun des queues cytoplasmiques de l'ensemble des récepteurs interagissant avec Calnuc. De plus, contrairement à LRP9, certains récepteurs, dont CI-MPR et Sortiline, interagissent indirectement avec Calnuc. Une protéine intermédiaire pourrait alors être impliquée complexifiant l'identification d'un motif d'interaction. Nos données préliminaires indiquent également que la déplétion de Calnuc entraîne une diminution des niveaux de Furine, CD-MPR et EGFR en plus de CI-MPR (figure 10). De plus amples études seront nécessaires afin de clarifier comment Calnuc affecte ces récepteurs. Figure 9 - Calnuc interagit avec Furine, TrkA et CD-MPR, mais pas avec p75/NTR ou EGFR. Des cellules HEK ont été transfectées avec GFP ou Calnuc-GFP avec (A) Furine-Flag, (B) p75/NTR-HA, (C) EGFR ou avec HA ou Calnuc-HA avec (D) TrkA et (E) CD-MPR-GFP. Les cellules ont été lysées et les protéines immunoprécipitées et détectées avec les anticorps indiqués. La principale machinerie impliquée dans le recyclage des récepteurs vers le TGN est le complexe nommé Rétromère. La déplétion des sous-unités Vps26 (Brodeur et al., 2009), Vps35 (Lin et al., 2007) ou SNX1 (Brodeur et al., 2012) de cet hétéropentamère produit des effets sur divers récepteurs qui sont similaires à ceux observés à la suite de la déplétion de Calnuc. Nous avons alors émis l'hypothèse que Calnuc serait impliquée dans le triage 183 endosomiale médié par les rétromères. La déplétion de Calnuc a montré, par l'exemple de Vps26 et SNX2, que la protéine est nécessaire au recrutement du complexe à la surface des endosomes sans perturber les niveaux intracellulaires des sous-unités testées, soit Vps26, Vps35, SNX1 et SNX2. Toutefois, aucune interaction n'a été détectée entre Calnuc et les différentes sous-unités du complexe Rétromère indiquant que Calnuc n'agit pas comme adaptateur pour leur recrutement membranaire (données non montrées). Figure 10 - La déplétion de Calnuc influence les niveaux de CI-MPR, Furine, CDMPR et EGFR. Les niveaux de CI-MPR, Furine, CD-MPR et EGFR sont réduits dans les cellules déplétées en Calnuc. Les cellules HeLa ont été transfectées avec un siRNA contrôle ou dirigé contre Calnuc pendant 3 jours. Les cellules ont aussi été transfectées avec Furine-Flag, CD-MPRGFP ou EGFR 12h avant l'expérience. Les cellules ont été lysées et les protéines séparées sur gel SDS-PAGE avant d'être détectées avec les anticorps anti-CI-MPR, anti-Flag, antiGFP, anti-EGFR, anti-Calnuc et anti-actine. L'actine sert de contrôle de quantité. L'initiation et la régulation de l'attachement du complexe Rétromère à la surface cytosolique de l'endosome dépendent de l'action séquentielle de Rab5 et de Rab7. Nous avons alors examiné ces deux petites protéines G en tant que potentiels partenaires d'interaction pour Calnuc. Nos études de microscopie confocale révèlent que Calnuc cytosolique colocalise partiellement avec Rab5, Rab7 et Rab9 au niveau des endosomes. Toutefois, en condition de surexpression, Calnuc interagit seulement avec Rab7 et de façon indépendante de son état d'activation (lié au GDP ou au GTP). Afin de déterminer l'impact 184 Figure 11 - Rôles potentiels de Calnuc dans le cycle d'activation des Rab. (A) Calnuc pourrait influencer positivement une GEF de manière à favoriser l'activation de Rab7. Calnuc pourrait également influencer négativement ou (B) séquestrer une GAP de façon à réduire l'inactivation de Rab7. + signifie influencer positivement et - signifie influencer négativement. Figure modifiée de RAB11-mediated trafficking in host–pathogen interactions. Guichard et al. Nat Rev Microbiol. 2014 Sep;12(9):624-34 avec la permission de Nature Publishing Group ( Macmillan Publishers Ltd). de Calnuc sur la localisation et l'activité de Rab7 nous avons procédé à la déplétion de Calnuc. Dans des cellules transfectées de façon stable, il s'en suit une diminution d'environ 28% du signal correspondant à Rab7-GFP alors que les niveaux totaux de la protéine restent inchangés. Ces résultats sont appuyés par une expérience de fractionnement cellulaire qui montre un déplacement de Rab7, et subséquemment de Vps26 et SNX2, de la fraction membranaire vers la fraction cytosolique. Ce défaut de recrutement est vraisemblablement causé par une réduction de l'activation de Rab7. En effet, la quantité de Rab7-GTP liant l'effecteur RILP est diminuée d'environ 66% en absence de Calnuc lors de réactions de liaison par affinité. Calnuc serait donc nécessaire à l'activation de Rab7, qui permet son recrutement aux endosomes et conséquemment l'assemblage du complexe Rétromère ainsi que le recyclage de nombreux récepteurs vers le TGN. Toutefois, Calnuc ne peut pas agir comme activatrice de Rab7, c'est-à-dire comme GEF, étant donné l'absence d'interaction directe entre les deux protéines. L'action de Calnuc pourrait alors s'expliquer de deux façons (figure 11) (Guichard et al., 2014). Calnuc pourrait influencer positivement une GEF de Rab7, comme Mon1-Ccz1 (Small et al., 2005). En l'absence de Calnuc, l'échange par la GEF du GDP pour le GTP serait moindre limitant l'activation de Rab7. À l'opposé, Calnuc pourrait séquestrer ou réduire l'activité d'une GAP de Rab7, comme 185 TBC1D5 (Seaman et al., 2009). La déplétion de Calnuc favoriserait alors l'inactivation précoce et la dissociation membranaire de Rab7. Calnuc n'est pas la première protéine liant le calcium à influencer le transport de récepteurs. Calmodulin, une autre protéine à mains EF liant le calcium, interagit de manière dépendante au calcium avec la queue cytoplasmique d'EGFR (Li and Villalobo, 2002) afin de réguler son transport (Llado et al., 2008), son activation et sa signalisation (Tebar et al., 2002a; Tebar et al., 2002b). D'ailleurs, malgré une absence d'interaction, des données préliminaires de notre groupe montrent que Calnuc jouerait un rôle dans le triage intracellulaire d'EGFR en modulant la protéine GGA3. EGFR est un récepteur à activité tyrosine kinase qui, suivant la liaison de son ligand, homodimérise pour permettre son autophosphorylation et sa mono-ubiquitinylation au niveau de sa queue cytoplasmique. Cette phosphorylation permet l'activation de plusieurs voies de signalisation contrôlant, entre autres, la croissance cellulaire, la différenciation, l'apoptose et l'expression génique. Une fois le récepteur internalisé, l'ubiquitine assure son acheminement vers les endosomes tardifs puis les lysosomes. GGA3 reconnaît et concentre les cargos ubiquitilylés (Staub and Rotin, 2006). Il a été montré que Calnuc interagit avec GGA3 de manière accrue en l'absence de calcium. De plus, à l'image des résultats obtenus dans la présente étude, la déplétion de Calnuc entraîne la redistribution d'EGFR vers les lysosomes où il est dégradé. Le mécanisme d'action proposé pour cet effet de Calnuc diverge cependant de celui montré pour CI-MPR et Sortiline. Au niveau du cytosol, Calnuc séquestrerait GGA3 pour l'empêcher d’interagir avec EGFR. Une fois le récepteur activité, la relâche calcique intracellulaire engendrée libèrerait GGA3 qui participerait alors à l’acheminement d'EGFR aux endosomes tardifs. Éventuellement, l'abaissement du niveau de calcium, permettrait à GGA3 d'être capturée de nouveau par Calnuc empêchant le récepteur de quitter les endosomes précoces. Calnuc régulerait donc la distribution intracellulaire de GGA3 et, en son absence, ce dernier serait constamment recruté aux endosomes pour acheminer des récepteurs aux lysosomes. Il est à noter que ce mécanisme influençant EGFR accorde une grande importance à l'impact du calcium, alors que, jusqu'à présent, l'ion ne semble pas affecter le mécanisme influençant CI-MPR et Sortiline. En ce sens, après la déplétion de Calnuc endogène, l'expression d'un mutant de Calnuc ne pouvant pas lier le calcium (ΔEF186 Calnuc) permet de restaurer les niveaux de CI-MPR au même titre que la forme sauvage (résultat non montré). Article 2 - Topologie et ancrage membranaire de CLN5, une protéine associée à une maladie du lysosome À l'image de Calnuc, CLN5 a récemment été identifiée comme étant un régulateur du recrutement endosomial de Rab7. Il a été démontré que l'absence de CLN5 entraîne une diminution de l'activation de Rab7 qui limite le recrutement du complexe Rétromère aux endosomes menant à un défaut de recyclage des récepteurs CI-MPR et Sortiline qui sont alors dégradés aux lysosomes (Mamo et al., 2012). N'ayant jusqu'à présent identifié aucun partenaire d'interaction direct soutenant le rôle de Calnuc dans le transport rétrograde, nous avons voulu savoir si Calnuc pourrait agir par l'intermédiaire de CLN5. Cette dernière est principalement connue pour son implication dans les NCL quoique son rôle exact dans la pathogenèse et l'impact mécanistique de ses mutants demeurent nébuleux. De plus, la topologie de CLN5 fait l'objet d'études contradictoires, la protéine ayant été décrite comme possédant un ou deux domaines transmembranaires ou encore comme étant soluble et sécrétée. La simple localisation de CLN5 et de ses différents mutants pathologiques a été compliquée par l'usage d'anticorps variés et d'étiquettes moléculaires localisées à différents endroits sur la protéine. Dans le but de mieux comprendre la topologie, l'impact des mutations et le mécanisme d'action de CLN5, nous avons décidé d'étudier d'abord ses propriétés de bases, à savoir son clivage, sa glycosylation et sa localisation intracellulaire. Nous avons élaboré différentes constructions présentant des étiquettes localisées en Nterminal, C-terminal et après le site prédit de clivage du SP. Nous avons ainsi confirmé l'expression d'un précurseur hautement glycosylé de 73 kDa qui subit un clivage, vraisemblablement en position 96, de manière à produire un fragment de 15 kDa correspondant au peptide signal ainsi qu'une protéine mature glycosylée de 55-60 kDa (Larkin et al., 2013). Le précurseur se retrouve coincé au RE alors que la protéine mature, détectée via un épitope en C-terminal ou un anticorps dirigé contre cette extrémité, se retrouve au niveau du Golgi, des lysosomes et est partiellement sécrétée en condition de surexpression (Larkin et al., 2013). Nous avons ensuite voulu vérifier l'existence des deux 187 segments transmembranaires prédits, soit les aa 77-93 et 353-373. Par essai de protection contre la protéinase K in vitro et in cellulo, nous avons établi que le N-terminal du précurseur de CLN5 se trouve dans le cytoplasme alors que le C-terminal est dans la lumière des organites confirmant la présence d'un seul segment transmembranaire et une topologie de type II (Larkin et al., 2013). Afin de permettre l'accessibilité des enzymes luminales au site de clivage du SP, le segment transmembranaire doit correspondre à la région hydrophobe en N-terminal, c'est-à-dire la région la plus souvent prédite. La protéine mature résultante est donc luminale indiquant que ses partenaires d'interaction directs doivent se retrouver en totalité ou en partie dans lumière des organites. Ceci implique également que l'action de CLN5 sur la machinerie cytosolique de transport rétrograde se fait par un intermédiaire transmembranaire qui pourrait être son partenaire d'interaction CLN3. Malgré l'absence de segment transmembranaire dans la forme mature de CLN5, cette dernière demeure fortement liée à la membrane lors d'expériences de fractionnement cellulaire. Nous avons alors émis l'hypothèse que les aa 353-373, contenant de nombreux résidus hydrophobes et non polaires, pourraient permettre l'ancrage de la protéine. L'analyse de la séquence a révélé un haut niveau de conservation de la région 353-392 au travers des espèces ainsi qu'une possible organisation en hélice amphipathique présentant deux surfaces hydrophobes (Larkin et al., 2013). Les hélices amphipathiques sont généralement localisées aux extrémités N-terminale ou C-terminale des protéines (Cornell and Taneva, 2006), comme c'est le cas pour CLN5. Elles permettent un ancrage de faible affinité, contrôlé et réversible, à des lipides généralement anioniques (phosphatidylinositol et phosphatidylsérine), sans la nécessité d'un motif de sélectivité lipidique (Cornell and Taneva, 2006). Par fractionnements cellulaires de segments protéiques en fusion avec GFP, nous avons été en mesure de démontrer que les aa 353–392 suffisent en effet à l'association membranaire (Larkin et al., 2013). Le rôle de l'hélice amphipathique de CLN5 pourrait s'étendre au-delà de l'attachement membranaire. En effet, cette structure peut aussi servir de domaine d'autoinhibition pour prévenir l'activation de protéines à l'état soluble, comme Arf, ou pour cacher des sites d'interaction, comme c'est le cas chez la Vinculine (Cornell and Taneva, 2006). Elle permet également de détecter et de moduler les propriétés physiques de 188 bicouches lipidiques. Les protéines à domaines N-BAR, telles que l'Amphiphysine, qui répondent et induisent la courbure membranaire, en sont de bons exemples (Jao et al., 2010). De plus amples études fonctionnelles et structurelles seront nécessaires afin de déterminer le rôle exact de CLN5 et de son hélice amphipathique. Il est toutefois intéressant de noter que, selon les prédictions informatiques, CLN3 possède une hélice semblable entre son cinquième et sixième segment transmembranaire (Nugent et al., 2008). Il est alors tentant de spéculer sur l'implication de cette structure dans l'interaction entre CLN3 et CLN5 et leur rôle dans le transport intracellulaire. Pour déterminer l'impact des mutations pathologiques sur la maturation, la sécrétion et la localisation de CLN5 nous avons étudié les mutants communs EUR, FinM, SWE et NFL. Ces différentes mutations n'affectent pas la maturation ni la glycosylation générale de CLN5, mais entraînent leur délocalisation vers le RE (Larkin et al., 2013). De plus, les mutants pathologiques dépourvus de leur hélice amphipathique, soit SWE et NFL, perdent leur association membranaire et subissent une dégradation par le protéasome, plutôt qu'une sécrétion accrue, témoignant de leur instabilité (Larkin et al., 2013). Cette deuxième étude a permis de clarifier la localisation et la structure de CLN5 ainsi que de quelques mutants pathologiques. Elle a également démontré que la présence d'une étiquette en C-terminal ne nuit pas la maturation, la localisation ou encore la sécrétion de CLN5 et permet la détection de la forme mature. Au moment où paraissait notre article sur la topologie de CLN5, un autre groupe publiait une étude des différents sites de glycosylation de CLN5. La mutation en glutamines des huit sites de séquence consensus N-X-T/S, soit Asn179, 192, 227, 252, 304, 320, 330 et 401, a permis de confirmer leur glycosylation in vivo. Il existe des différences fonctionnelles dans ces sites affectant le repliement, le transport ou la fonction de CLN5. Les modifications en position 179, 252, 304, 320 et, plus modestement, 330 sont nécessaires à la sortie du RE indiquant leur importance dans le repliement protéique alors que la glycosylation Asn401 est essentielle pour le transport post-Golgi vers les lysosomes (Moharir et al., 2013). Les glycosylations en Asn192 et Asn227 n'ont pas d'impact sur la localisation, mais le caractère pathologique de la mutation N192S indique un rôle 189 fonctionnel pour cette modification (Moharir et al., 2013). À l'image de ce que nous avons observé, les mutants pathologiques EUR (p.D279N), qui introduit un site de glycosylation, et FinM (p.Y392*), qui perd le potentiel site de glycosylation en position 401, sont retenues au RE (Moharir et al., 2013; Schmiedt et al., 2010). Par contre, contrairement à nos résultats, ces deux mutants, de même que ceux retenus au RE, ne sont pas sécrétés (Moharir et al., 2013). Cette divergence peut s'expliquer par la sensibilité de détection accrue de notre système de précipitation radioactif par rapport à la capacité de détection par anticorps sur des lysats cellulaires. Il n'en demeure pas moins que la sécrétion est faible et peut être attribuée à la saturation de la machinerie cellulaire en condition de surexpression. Les auteurs ont également remarqué que le mutant FinM est moins stable que le mutant EUR ce qui est en accord avec la déstabilisation de l'hélice amphipathique et la sensibilité accrue à l'extraction membranaire rapportée par notre groupe. Bref, la délocalisation de plusieurs mutants pathologiques, dont p.D279N et p.Y392*, mais aussi p.R112P, p.R112H, p.E253* et p.W379C, pourrait expliquer leur manque de fonctionnalité (Isosomppi et al., 2002; Lebrun et al., 2009; Moharir et al., 2013; Schmiedt et al., 2010). Récemment, la mise au point d'un nouvel anticorps par la compagnie Abcam (ab170899) a rendu possible la détection efficace de CLN5 endogène dans de nombreux tissus chez l'humain comme chez la souris. Ce nouvel outil reconnaît une région entre les aa 150 et 250 de CLN5. Il a permis de montrer l'existence de deux formes persistantes de CLN5 dans la plupart des types cellulaires, soit une proforme de 56 kDa et une forme mature dominante de 52 kDa (Silva et al., 2015). Cette différence de 4 kDa est attribuable à la perte, suivant l'action d'une protéase à cystéine, d'environ 10-15 acides aminés en Cterminal (Silva et al., 2015). Dans notre étude de la topologique de CLN5, l'utilisation de l'étiquette Flag après le peptide signal nous a d'ailleurs permis d'observer ce doublet pour la forme WT, mais aussi pour le mutant EUR (Article 2 - figure 5) (Larkin et al., 2013). Par contre, l'anticorps CLN5-C/32 ne permet pas cette détection puisqu'il est dirigé contre la séquence terminale clivée (EEIPLPIRNKTLSGL). L'utilisation d'une étiquette intramoléculaire ou d'anticorps reconnaissant la partie centrale de la protéine serait donc à favoriser pour assurer l'étude de CLN5 mature. L'importance du clivage post-traductionnel 190 en aval de l'hélice amphipathique reste à être déterminée mais pourrait être nécessaire pour l'ancrage et pour l'activité de CLN5. La détection endogène de la protéine CLN5 a également permis de clarifier sa sécrétion. En effet, aucune des formes endogènes de la protéine n'a été détectée dans le milieu extracellulaire laissant croire que la sécrétion préalablement observée en système de surexpression serait un artefact causé par la congestion et l'insuffisance de la machinerie cellulaire (Silva et al., 2015). Ceci va de pair avec le modèle structural proposé par notre équipe décrivant CLN5 comme étant une protéine soluble ancrée aux membranes luminales par une hélice amphipathique afin d'assurer sa rétention cellulaire (Larkin et al., 2013). Bref, la compréhension grandissante de la maturation de CLN5 et la faisabilité des études endogènes assurent graduellement un consensus dans les études tout en expliquant les nombreuses discordances antérieures. Article 3 - Rôle de Calnuc dans les céroïdes-lipofuscinoses neuronales Après avoir clarifié la topologie de CLN5, nous avons voulu vérifer l'existence d'un lien fonctionnel entre Calnuc et CLN5. En effet, ces deux protéines influencent le recrutement de Rab7 au niveau des endosomes et, conséquemment, affectent le transport de récepteurs qui transitent entre le TGN et les endosomes. Alors que notre modèle dépeint CLN5 mature comme étant une protéine luminale, notre étude du rôle de Calnuc a montré l'implication de la fraction cytosolique dans le routage cellulaire. Ces données laissent croire que Calnuc et CLN5 partagent une mécanistique commune qui implique un partenaire d'interaction transmembranaire. CLN3, une protéine à six domaines transmembranaires impliquée dans les lipofuscinoses et interagissant avec CLN5, s'est avérée être la candidate de choix pour constituer cet intermédiaire. CLN3 serait impliquée dans le transport intracellulaire puisqu'elle interagit directement avec Rab7 de façon à influencer son recrutement (Uusi-Rauva et al., 2012). De plus, des études montrent que son absence mène à une distribution anormale de protéines de la voie endolysosomiale de même qu'à des défauts morphologiques au niveau des organites de cette voie (Fossale et al., 2004; Lojewski et al., 2013; Metcalf et al., 2008). Tout comme l'absence de Calnuc et de 191 CLN5, l'expression du mutant pathologique CLN3Δex7/8/Δex7/8 (MT), dépourvu des exons 7 et 8 et arborant un changement de cadre de lecture au 153e acide aminé, entraîne une forte diminution des niveaux de Rab7 actif. Dans un premier temps, nous avons donc montré que Calnuc, CLN3 et CLN5 interagissent ensemble et colocalisent aux endosomes précoces. Ces résultats ont mené à l'élaboration d'un modèle selon lequel Calnuc cytosolique, CLN3 transmembranaire et CLN5 luminale forment un complexe important pour le transport intracellulaire (figure 12). La dynamique de ce modèle et le rôle exact de chaque partenaire demeurent pour le moment inconnus. À l'avenir, l'identification des régions d'interaction protéique impliquées permettra de mieux comprendre la mécanistique et l'impact des diverses mutations pathologiques. Figure 12 - Modèle du complexe Calnuc-CLN3-CLN5 influençant le recrutement de Rab7 et du complexe Rétromère. Calnuc cytosolique, la protéine à six segments transmembranaires CLN3 et la protéine luminale ancrée CLN5 forment un complexe. Ce dernier permet le recyclage de récepteurs en influençant l'activation de Rab7 et conséquemment le recrutement du complexe Rétromère à la surface des endosomes. Figure créée avec Microsoft Office PowerPoint 2007. L'implication du calcium dans le complexe Calnuc-CLN3-CLN5 est également une avenue intéressante. Calnuc est principalement connue pour sa capacité à lier le calcium via ses mains EF et à former une large réserve mobilisable de l'ion au niveau du Golgi. De son côté, CLN3 interagit, de façon sensible au calcium, avec Calséniline, une protéine à mains 192 EF qui réprime la transcription (Chang et al., 2007). Ceci suggère la possibilité d'autres interactions influencées par le calcium, avec des protéines comme Calnuc. Aussi, l'étude de cellules exprimant des mutants pathologiques de CLN3 a permis d'associer la protéine à l'homéostasie calcique du réticulum, des mitochondries et des lysosomes (Chandrachud et al., 2015; Chang et al., 2007). Étant connu pour interagir avec des canaux ioniques (UusiRauva et al., 2008), CLN3 pourrait agir comme régulateur du trafic ou de l'activité de ces derniers ou constituer un nouveau transporteur ionique. D'ailleurs, le calcium lysosomial prend de plus en plus d'importance dans les phénomènes intracellulaires permettant, entre autres, d'activer des protéines en lien avec la biogenèse des lysosomes et l'induction de l'autophagie (Medina et al., 2015). Ainsi, la relâche de ce calcium pourrait permettre à Calnuc de lier l'ion dans le cytosol et, possiblement, de moduler son activité. Quoi qu'il en soit, au niveau des lipofuscinoses, les ions commencent à être considérés en raison de leur importance pour la transmission synaptique au cerveau, mais surtout de leur neurotoxicité. Une accumulation et une dérégulation du transport de métaux (zinc, cuivre, manganèse, fer, cobalt), précédant la neurodégénération, ont d'ailleurs été rapportées chez des moutons de type CLN6 (Grubman et al., 2014a; Kanninen et al., 2013) et des souris de type CLN1, CLN3 et CLN5 (Grubman et al., 2014b). La deuxième partie de l'étude consistait à déterminer si Calnuc, à l'image de ses deux partenaires, CLN3 et CLN5, pourrait être impliquée dans les maladies du lysosome. Nous avons vérifié l'impact de la déplétion de Calnuc sur l'apparition de phénotypes cellulaires couramment retrouvés dans les lipofuscinoses. Par microscopie, nous avons observé la formation d'ultrastructures de type curviligne, un grossissement des lysosomes, une accumulation de matériel autofluorescent et de la sous-unité C de l'ATP synthase mitochondriale (SCMAS) ainsi qu'une augmentation de l'autophagie. Ces évidences tendent à démontrer le rôle de Calnuc dans les lipofuscinoses. Par contre, cette partie de l'étude gagnerait en validité par l'usage d'un modèle cellulaire neurologique. C'est pourquoi nous envisageons de reprendre ces expériences de microscopie en utilisant des cellules de cervelet de souris. En ce qui a trait à la caractérisation du phénotype, des souris déficientes en Calnuc, actuellement offertes chez Taconic Biosciences, seraient également un atout pour la poursuite du projet. 193 Dans l'espoir d'associer une modulation de Calnuc à des cas de lipofuscinoses, nous avons performé une analyse des niveaux de protéines et d'ARNm dans des lignées de fibroblastes. Outre les lignées cellulaires contrôles, nous avons obtenu des lignées provenant de patients atteints de NCL de types CLN1, CLN3 et CLN5, de même que de type orphelin. En effet, à ce jour, malgré l'identification de 14 gènes porteurs de mutations pathologiques, il existe encore des patients dont la cause génétique de leur maladie demeure inconnue. Nous avons observé une réduction d'au moins 40% des niveaux protéiques de Calnuc chez toutes les lignées de fibroblastes issues de patients atteints de NCL, ce qui suggère une implication de cette dernière dans la maladie. Malgré l'absence de CLN1 dans notre modèle, la lignée mutante pour cette protéine présente une forte diminution de Calnuc. CLN1 ou PPT1 (Palmitoyl-protein thioesterase 1) est une enzyme lysosomiale qui élimine les groupements lipidiques des protéines et qui a été montrée comme pouvant interagir avec CLN5 (Lyly et al., 2009) de manière à influencer son transport vers les lysosomes (Schmiedt et al., 2010). Il se peut que, par ce moyen, CLN1 influence indirectement Calnuc. Aussi, il est intéressant de noter que Calnuc n'est pas affectée dans la lignée contrôle Niemann-Pick Disease de type C1 (NPD), qui constitue une maladie du lysosome en lien avec un défaut d'élimination des lipides intracellulaires et mécanistiquement distincte des NCL. De son côté, l'analyse de l'ARNm de Calnuc n'a pas révélé de variations significatives dans les différents échantillons. Par contre, Calnuc pourrait présenter des mutants pathologiques stables dans des lipofuscinoses de type orphelin. La recherche de ces mutations, de même que la confirmation des résultats présentés ici, sont limitées par la faible disponibilité et diversité du matériel biologique issu de patients. À l'heure actuelle, selon le site 'GeneCards: The Human Gene Database' une centaine de mutations influençant la séquence de la protéine ont été rapportées, mais aucune n'est associée à une pathologie. La fonction de Calnuc au niveau des lysosomes pourrait prochainement être appuyée par la récente découverte du motif CLEAR (Sardiello et al., 2009). Il s'agit d'une séquence palindromique de 10 paires de bases (G/C/A-T/C/G-C-A-C/G/T-G/C-T/C-G-A/G-C/T/G ou GTCACGTGAC) située à proximité du site d'initiation de la transcription et permettant de coordonner l'expression et la régulation de gènes en lien avec la fonction lysosomiale. Elle 194 est reconnue par le facteur de transcription EB (TFEB) qui agit comme responsable de la clairance cellulaire via la régulation de 291 gènes, dont plusieurs en lien avec la biogenèse des lysosomes, la voie de l’autophagosome et les hydrolases lysosomiales (Sardiello et al., 2009). TFEB constituerait une importante cible thérapeutique puisque sa surexpression permet d'augmenter la genèse de lysosomes et d'augmenter la dégradation de molécules complexes (Sardiello et al., 2009). Entre autres, le motif CLEAR est retrouvé au niveau des gènes codant LAMP1, MPR, CLN1 (PTT1, CLN2 (TPP1), CLN3, CLN5, MFSD8 (CLN7), CTSD (CLN10), GRN (CLN11), CTSF (CLN13) (Palmieri et al., 2011; Sardiello et al., 2009). De son côté, Calnuc a été identifiée parmi les 25 gènes dont l'expression varie le plus en fonction de Psap, CTSA ou CTSD et dans les 90 premiers variant pour CLN2 (Sardiello et al., 2009). Cette coordination d'expression suggère que Calnuc serait aussi régulée par TFEB. Une rapide analyse de la séquence entourant les promoteurs de Calnuc révèle en effet la présence d'une séquence GTCAAGAGGC à 101 nucléotides en amont du site CCAAT. Il serait intéressant de vérifier si l'expression de Calnuc est modulée par TFEB. Les lipofuscinoses ne sont pas les premières maladies neurodégénératives dans lesquelles Calnuc aurait un rôle à jouer. De plus en plus d'évidences soutiennent son implication dans la maladie d'Alzheimer. Entre autres, elle permet de maintenir les niveaux intracellulaires de Sortiline et de LRP10 (Brodeur et al., 2009), deux récepteurs d'APP qui cycle entre le TGN et les endosomes, et protègent contre le clivage amyloïdogénique et l'accumulation du peptide neurotoxique Aβ. Cela va de pair avec une réduction de Calnuc (Lin et al., 2007) et de LRP10 (Brodeur et al., 2012) dans les cerveaux de patients atteints de la maladie d'Alzheimer. Ces mêmes tissus présentent une réduction des rétromères Vps26 et Vps35 (Small et al., 2005). Il se pourrait donc que l'effet de Calnuc sur l'activation de Rab7 et le recrutement du complexe Rétromère soit également en cause. 195 CONCLUSION La présente étude a permis de mettre en lumière un nouveau rôle pour la protéine Calnuc dans le transport rétrograde de récepteurs entre les endosomes et le TGN. Nos données démontrent que Calnuc influence positivement l'activité de la petite protéine G Rab7 qui agit comme régulateur spatio-temporel lors de multiples processus intracellulaires, dont le transport rétrograde. En fait, Calnuc formerait un complexe avec les protéines CLN3 et CLN5 qui ont également cette fonction. La mécanistique exacte de la formation de ce complexe reste inconnue et l'absence d'interactions directes entre les partenaires laisse supposer l'implication d'autres protéines et la complexification du modèle proposé ici. Calnuc partage donc une fonction commune avec des protéines impliquées dans les céroïdes-lipofuscinoses neuronales suggérant son implication dans cette maladie neurodégénérative. La déplétion de Calnuc entraîne, en effet, l'apparition de phénotypes cellulaires typiques. La forte modulation de son expression dans toutes les lignées de fibroblastes de patients atteints de NCL à notre disposition soutient cette hypothèse malgré l'absence de variation au niveau de l'ARNm. Il se peut donc qu'il existe des mutants pathologiques stables de Calnuc responsables de formes orphelines de la maladie. La recherche de ces mutations et l'étude de la fonction exacte de Calnuc dans la pathogenèse seraient alors primordiales puisque la protéine pourrait constituer un nouvel outil diagnostique et une nouvelle cible thérapeutique pour le traitement de lipofuscinoses. 196 REMERCIEMENTS Pour débuter, je tiens à remercier les Pr Alexey Pshezhetsky, Xavier Roucou et Michel Grandbois pour la révision de ma thèse et pour les commentaires constructifs que vous y avez apportés. Ma reconnaissance va surtout à ma directrice de recherche Pre Christine Lavoie qui m'a accueilli dans son laboratoire, d'abord comme stagiaire, puis durant sept ans. Merci de m'avoir guidé au travers mon doctorat et d'avoir su me pousser en temps et lieu. Tu es un modèle de travail acharné et de persévérance digne des scientifiques les plus passionnés. La qualité, l'étendue et la rigueur de la formation que tu m'as offerte me sera un atout pour ma carrière. J'ai également trouvé auprès de toi une amie qui m'a aidé à passer au travers les épreuves des études aux cycles supérieurs, mais aussi de la vie. En cette difficile période pour le financement de la recherche, je crois qu'il est important de souligner le rôle des organismes subventionnaires. Mon projet a été supporté par les IRSC et j'ai eu la chance d'obtenir une des rares bourses d’études supérieures du CRSNG, ce qui a grandement facilité mon parcours. Aussi, à trois occasions, j'ai participé à des congrès à l'international, entre autres, grâce à la Bourse IPS-RECPUS pour congrès international et à la Bourse de voyage d’appui communautaire des Instituts (IRSC). Ces fonds aident à établir des contacts qui peuvent faire toute la différence dans le cheminement d'un étudiant. Finalement, merci à l'Université de Sherbrooke pour les bourses de stage, de maîtrise et de doctorat qui constituent un encouragement et un bonus salariale non négligeable. Je suis heureuse d'avoir eu le privilège de participer à un projet de collaboration avec Marc-André Bonin, Hassan Trabulsi et, dernièrement, Étienne Marousseau du laboratoire de Pr Éric Marsault. Merci les gars d'avoir mis un peu de chimie dans mes études. Grâce à vous, et à Leonid Volkov (responsable de la Plateforme de biophotonique du Centre de recherche clinique Étienne-Le Bel), j'ai entre autres appris les rudiments du Fluorescenceactivated cell sorting (FACS). Également, un merci à Pr Pierre Lavigne pour les conseils en matière de structure protéique qui m'ont été fort utiles pour l'article concernant CLN5. 197 Je voudrais aussi remercier les chercheurs, employés et étudiants des laboratoires voisins que j'ai côtoyés tout au long de ces années et avec qui j'ai partagé anecdotes, conseils expérimentaux et matériel. Il m'a fait plaisir de vous connaître. Merci aux membres présents et passés de mon laboratoire, soit Hugo Gagnon, Julie Brodeur avec qui j'ai travaillé pour l'article à propos de LRP9, Stéphanie Rosciglione, Catherine Duclos, Sébastien Grastilleur, Erwan Lanchec, Xuezhi Li, Étienne Marousseau et Marilène Paquette. Parmi mes collègues certains sont devenus des amis : le très critique Élie Simard, Yannick Miron, le gastronome, Yannik Réginbal-Dumas, qui m'a montré l'importance de l'implication sociale et scolaire, Thomas Söllradl, avec qui j'ai siroté moultes bières en pratiquant mon anglais et mon allemand, Marc-Olivier Frégeau, mon collègue depuis le baccalauréat, le toujours souriant Alexandre Desroches, Martin Houde l'encyclopédie et Jean-Sébastien Maltais, mon 'buddy' de salle de culture. En 2006, alors que je faisais mes premières armes dans le département, j'ai eu le privilège de connaître des étudiants finissants qui sont devenus mes modèles. Je parle ici de Pascale Lepage et Marc Lussier, 'mon grand-frère académique'. Au cours de ces années, deux personnes sont devenues des amis indispensables avec qui je garderai sans doute contact toute ma vie. Stéphanie Rosciglione, alias Darling, ma collègue doctorante, mon amie et ma confidente. En raison de notre cheminement parallèle, nous avons partagé nos épreuves, nos réussites, nos peines, nos anecdotes, nos découragements et nos joies. J'espère que nos chemins se recroiseront et au plaisir de se faire de nouveau calciner au soleil avec une bouteille de Martini Rosso. Je me souviens encore du jour où, après qu'il m'ait offert d'utiliser son incubateur, j'ai dit à Dave Boucher qu'il était devenu 'mon nouveau meilleur ami'. À ce moment-là, je ne savais pas à quel point c'était vrai. Dave, tu as toujours été l'oreille attentive, le gars qui connaît tous les petits trucs de lab, le super motivé, l'hyperactif, l'homme capable de tout faire en même temps ; un futur grand chercheur quoi! Merci pour ces soirées à discuter au laboratoire, pour le sofa et pour ce merveilleux voyage en Australie. Le laboratoire fut une grosse portion de ma vie pendant les dernières années et plusieurs personnes ont veillé à y mettre de l'équilibre. Merci donc à mes amis Luc 198 Paquette, Alexei Nordel-Markovits, Marie-Sol Poirier et Marc-André Daviault. Un merci aussi à Catherine Bourdon et aux autres personnes du Domaine Équi-Sphère pour m'avoir fait sortir du monde universitaire ainsi qu'à Tiptoe et Tao pour la zoothérapie de fin de doctorat. Pour terminer, je remercie chaleureusement mes parents, Pierre et Francine, de m'avoir soutenue moralement et financièrement pendant mon long parcours scolaire. Je suis reconnaissante envers ma fratrie, c'est-à-dire mon frère David qui m'a hébergé dans son duplex, ma studieuse soeur Annie-Pier ainsi que mon frère Samuel et sa conjointe Nadine. Finalement, les mots ne suffisent guère pour exprimer ma gratitude envers mon conjoint Marc Therrien pour ces soirées à m'attendre et ces fins de semaine à m'accompagner au travail, mais surtout pour son amour, sa compréhension et ses encouragements. Merci à vous tous de m'avoir montré ce qui est le plus important dans la vie. 199 LISTE DES PUBLICATIONS Adam, P. J., Boyd, R., Tyson, K. L., Fletcher, G. C., Stamps, A., Hudson, L., Poyser, H. R., Redpath, N., Griffiths, M., Steers, G., Harris, A. L., Patel, S., Berry, J., Loader, J. A., Townsend, R. R., Daviet, L., Legrain, P., Parekh, R., and Terrett, J. A. (2003): Comprehensive proteomic analysis of breast cancer cell membranes reveals unique proteins with potential roles in clinical cancer. J Biol Chem 278, 6482-9. Afshar, N., Black, B. E., and Paschal, B. M. (2005): Retrotranslocation of the chaperone calreticulin from the endoplasmic reticulum lumen to the cytosol. Mol Cell Biol 25, 8844-53. Alory, C., and Balch, W. E. (2001): Organization of the Rab-GDI/CHM superfamily: the functional basis for choroideremia disease. Traffic 2, 532-43. Alvarez, C., Garcia-Mata, R., Brandon, E., and Sztul, E. (2003): COPI recruitment is modulated by a Rab1b-dependent mechanism. Mol Biol Cell 14, 2116-27. Antonin, W., Fasshauer, D., Becker, S., Jahn, R., and Schneider, T. R. (2002): Crystal structure of the endosomal SNARE complex reveals common structural principles of all SNAREs. Nat Struct Biol 9, 107-11. Aradhyam, G. K., Balivada, L. M., Kanuru, M., Vadivel, P., and Vidhya, B. S. (2010): Calnuc: Emerging roles in calcium signaling and human diseases. IUBMB Life 62, 436-46. Baldwin, S. A., Beal, P. R., Yao, S. Y., King, A. E., Cass, C. E., and Young, J. D. (2004): The equilibrative nucleoside transporter family, SLC29. Pflugers Arch 447, 735-43. Ballabio, A., and Gieselmann, V. (2009): Lysosomal disorders: from storage to cellular damage. Biochim Biophys Acta 1793, 684-96. Ballif, B. A., Mincek, N. V., Barratt, J. T., Wilson, M. L., and Simmons, D. L. (1996): Interaction of cyclooxygenases with an apoptosis- and autoimmunity-associated protein. Proc Natl Acad Sci U S A 93, 5544-9. Bandyopadhyay, U., and Cuervo, A. M. (2008): Entering the lysosome through a transient gate by chaperone-mediated autophagy. Autophagy 4, 1101-3. Bandyopadhyay, U., Kaushik, S., Varticovski, L., and Cuervo, A. M. (2008): The chaperone-mediated autophagy receptor organizes in dynamic protein complexes at the lysosomal membrane. Mol Cell Biol 28, 5747-63. Barnikol-Watanabe, S., Gross, N. A., Gotz, H., Henkel, T., Karabinos, A., Kratzin, H., Barnikol, H. U., and Hilschmann, N. (1994): Human protein NEFA, a novel DNA binding/EF-hand/leucine zipper protein. Molecular cloning and sequence analysis of the cDNA, isolation and characterization of the protein. Biol Chem Hoppe Seyler 375, 497-512. Batten, F. (1903): Cerebral degeneration with symmetrical changes in the maculae in two members of a family. Trans Ophthalmol Soc UK 23, 386-390. Batten, F. (1914): Family cerebral degeneration with macular change (so-called juvenile form of family amaurotic idiocy). Q J Med 7, 444-454. Belin, D., Bost, S., Vassalli, J. D., and Strub, K. (1996): A two-step recognition of signal sequences determines the translocation efficiency of proteins. Embo J 15, 468-78. 200 Bellettato, C. M., and Scarpa, M. (2010): Pathophysiology of neuropathic lysosomal storage disorders. J Inherit Metab Dis 33, 347-62. Bento, C. F., Puri, C., Moreau, K., and Rubinsztein, D. C. (2013): The role of membranetrafficking small GTPases in the regulation of autophagy. J Cell Sci 126, 1059-69. Berggard, T., Miron, S., Onnerfjord, P., Thulin, E., Akerfeldt, K. S., Enghild, J. J., Akke, M., and Linse, S. (2002): Calbindin D28k exhibits properties characteristic of a Ca2+ sensor. J Biol Chem 277, 16662-72. Bessa, C., Teixeira, C. A. F., Mangas, M., Dias, A., Sá Miranda, M. C., Guimarães, A., Ferreira, J. C., Canas, N., Cabral, P., and Ribeiro, M. G. (2006): Two novel CLN5 mutations in a Portuguese patient with vLINCL: Insights into molecular mechanisms of CLN5 deficiency. Molecular Genetics and Metabolism 89, 245-253. Bonifacino, J. S. (2004): The GGA proteins: adaptors on the move. Nat Rev Mol Cell Biol 5, 23-32. Bonifacino, J. S., and Traub, L. M. (2003): Signals for sorting of transmembrane proteins to endosomes and lysosomes. Annu Rev Biochem 72, 395-447. Boucher, R., Larkin, H., Brodeur, J., Gagnon, H., Theriault, C., and Lavoie, C. (2008): Intracellular trafficking of LRP9 is dependent on two acidic cluster/dileucine motifs. Histochem Cell Biol 130, 315-27. Brailoiu, G. C., Dun, S. L., Brailoiu, E., Inan, S., Yang, J., Chang, J. K., and Dun, N. J. (2007): Nesfatin-1: distribution and interaction with a G protein-coupled receptor in the rat brain. Endocrinology 148, 5088-94. Brasaemle, D. L. (2007): Thematic review series: adipocyte biology. The perilipin family of structural lipid droplet proteins: stabilization of lipid droplets and control of lipolysis. J Lipid Res 48, 2547-59. Braulke, T., and Bonifacino, J. S. (2009): Sorting of lysosomal proteins. Biochim Biophys Acta 1793, 605-14. Braulke, T., Gartung, C., Hasilik, A., and von Figura, K. (1987): Is movement of mannose 6-phosphate-specific receptor triggered by binding of lysosomal enzymes? J Cell Biol 104, 1735-42. Brean, A. (2004): [An account of a strange instance of disease--Stengel-Batten-SpielmayerVogt disease]. Tidsskr Nor Laegeforen 124, 970-1. Breuer, P., Korner, C., Boker, C., Herzog, A., Pohlmann, R., and Braulke, T. (1997): Serine phosphorylation site of the 46-kDa mannose 6-phosphate receptor is required for transport to the plasma membrane in Madin-Darby canine kidney and mouse fibroblast cells. Mol Biol Cell 8, 567-76. Brocker, C., Engelbrecht-Vandre, S., and Ungermann, C. (2010): Multisubunit tethering complexes and their role in membrane fusion. Curr Biol 20, R943-52. Brodeur, J., Larkin, H., Boucher, R., Theriault, C., St-Louis, S. C., Gagnon, H., and Lavoie, C. (2009): Calnuc binds to LRP9 and affects its endosomal sorting. Traffic 10, 1098-114. Brodeur, J., Theriault, C., Lessard-Beaudoin, M., Marcil, A., Dahan, S., and Lavoie, C. (2012): LDLR-related protein 10 (LRP10) regulates amyloid precursor protein (APP) trafficking and processing: evidence for a role in Alzheimer's disease. Mol Neurodegener 7, 31. 201 Brown, J., Esnouf, R. M., Jones, M. A., Linnell, J., Harlos, K., Hassan, A. B., and Jones, E. Y. (2002): Structure of a functional IGF2R fragment determined from the anomalous scattering of sulfur. Embo J 21, 1054-62. Brunk, U. T., and Terman, A. (2002a): Lipofuscin: mechanisms of age-related accumulation and influence on cell function. Free Radic Biol Med 33, 611-9. Brunk, U. T., and Terman, A. (2002b): The mitochondrial-lysosomal axis theory of aging: accumulation of damaged mitochondria as a result of imperfect autophagocytosis. Eur J Biochem 269, 1996-2002. Bucci, C., Parton, R. G., Mather, I. H., Stunnenberg, H., Simons, K., Hoflack, B., and Zerial, M. (1992): The small GTPase rab5 functions as a regulatory factor in the early endocytic pathway. Cell 70, 715-28. Bugarcic, A., Zhe, Y., Kerr, M. C., Griffin, J., Collins, B. M., and Teasdale, R. D. (2011): Vps26A and Vps26B subunits define distinct retromer complexes. Traffic 12, 175973. Bulankina, A. V., Deggerich, A., Wenzel, D., Mutenda, K., Wittmann, J. G., Rudolph, M. G., Burger, K. N., and Honing, S. (2009): TIP47 functions in the biogenesis of lipid droplets. J Cell Biol 185, 641-55. Burd, C. G. (2011): Physiology and pathology of endosome-to-Golgi retrograde sorting. Traffic 12, 948-55. Campellone, K. G., and Welch, M. D. (2010): A nucleator arms race: cellular control of actin assembly. Nat Rev Mol Cell Biol 11, 237-51. Cang, C., Zhou, Y., Navarro, B., Seo, Y. J., Aranda, K., Shi, L., Battaglia-Hsu, S., Nissim, I., Clapham, D. E., and Ren, D. (2013): mTOR regulates lysosomal ATP-sensitive two-pore Na(+) channels to adapt to metabolic state. Cell 152, 778-90. Canuel, M., Libin, Y., and Morales, C. R. (2009): The interactomics of sortilin: an ancient lysosomal receptor evolving new functions. Histol Histopathol 24, 481-92. Cao, X., and Sudhof, T. C. (2001): A transcriptionally [correction of transcriptively] active complex of APP with Fe65 and histone acetyltransferase Tip60. Science 293, 11520. Cao, Y., Espinola, J. A., Fossale, E., Massey, A. C., Cuervo, A. M., MacDonald, M. E., and Cotman, S. L. (2006): Autophagy is disrupted in a knock-in mouse model of juvenile neuronal ceroid lipofuscinosis. J Biol Chem 281, 20483-93. Carlo, A. S., Gustafsen, C., Mastrobuoni, G., Nielsen, M. S., Burgert, T., Hartl, D., Rohe, M., Nykjaer, A., Herz, J., Heeren, J., Kempa, S., Petersen, C. M., and Willnow, T. E. (2013): The pro-neurotrophin receptor sortilin is a major neuronal apolipoprotein E receptor for catabolism of amyloid-beta peptide in the brain. J Neurosci 33, 35870. Carlton, J., Bujny, M., Peter, B. J., Oorschot, V. M., Rutherford, A., Mellor, H., Klumperman, J., McMahon, H. T., and Cullen, P. J. (2004): Sorting nexin-1 mediates tubular endosome-to-TGN transport through coincidence sensing of highcurvature membranes and 3-phosphoinositides. Curr Biol 14, 1791-800. Carroll, K. S., Hanna, J., Simon, I., Krise, J., Barbero, P., and Pfeffer, S. R. (2001): Role of Rab9 GTPase in facilitating receptor recruitment by TIP47. Science 292, 1373-6. Chan, C. H., Ramirez-Montealegre, D., and Pearce, D. A. (2009): Altered arginine metabolism in the central nervous system (CNS) of the Cln3-/- mouse model of juvenile Batten disease. Neuropathol Appl Neurobiol 35, 189-207. 202 Chandrachud, U., Walker, M. W., Simas, A. M., Heetveld, S., Petcherski, A., Klein, M., Oh, H., Wolf, P., Zhao, W. N., Norton, S., Haggarty, S. J., Lloyd-Evans, E., and Cotman, S. L. (2015): Unbiased Cell-based Screening in a Neuronal Cell Model of Batten Disease Highlights an Interaction between Ca2+ Homeostasis, Autophagy, and CLN3 Protein Function. J Biol Chem 290, 14361-80. Chang, J. W., Choi, H., Kim, H. J., Jo, D. G., Jeon, Y. J., Noh, J. Y., Park, W. J., and Jung, Y. K. (2007): Neuronal vulnerability of CLN3 deletion to calcium-induced cytotoxicity is mediated by calsenilin. Hum Mol Genet 16, 317-26. Chattopadhyay, S., and Pearce, D. A. (2002): Interaction with Btn2p is required for localization of Rsglp: Btn2p-mediated changes in arginine uptake in Saccharomyces cerevisiae. Eukaryot Cell 1, 606-12. Chavrier, P., Parton, R. G., Hauri, H. P., Simons, K., and Zerial, M. (1990): Localization of low molecular weight GTP binding proteins to exocytic and endocytic compartments. Cell 62, 317-29. Chen, H. J., Remmler, J., Delaney, J. C., Messner, D. J., and Lobel, P. (1993): Mutational analysis of the cation-independent mannose 6-phosphate/insulin-like growth factor II receptor. A consensus casein kinase II site followed by 2 leucines near the carboxyl terminus is important for intracellular targeting of lysosomal enzymes. J Biol Chem 268, 22338-46. Chen, H. J., Yuan, J., and Lobel, P. (1997): Systematic mutational analysis of the cationindependent mannose 6-phosphate/insulin-like growth factor II receptor cytoplasmic domain. An acidic cluster containing a key aspartate is important for function in lysosomal enzyme sorting. J Biol Chem 272, 7003-12. Chen, Y., Lin, P., Qiu, S., Peng, X. X., Looi, K., Farquhar, M. G., and Zhang, J. Y. (2007): Autoantibodies to Ca2+ binding protein Calnuc is a potential marker in colon cancer detection. Int J Oncol 30, 1137-44. Cheng, X., Shen, D., Samie, M., and Xu, H. (2010): Mucolipins: Intracellular TRPML1-3 channels. FEBS Lett 584, 2013-21. Chia, P. Z., Gasnereau, I., Lieu, Z. Z., and Gleeson, P. A. (2011): Rab9-dependent retrograde transport and endosomal sorting of the endopeptidase furin. J Cell Sci 124, 2401-13. Chia, P. Z., Gunn, P., and Gleeson, P. A. (2013): Cargo trafficking between endosomes and the trans-Golgi network. Histochem Cell Biol 140, 307-15. Choy, R. W., Cheng, Z., and Schekman, R. (2012): Amyloid precursor protein (APP) traffics from the cell surface via endosomes for amyloid beta (Abeta) production in the trans-Golgi network. Proc Natl Acad Sci U S A 109, E2077-82. Christoforidis, S., McBride, H. M., Burgoyne, R. D., and Zerial, M. (1999a): The Rab5 effector EEA1 is a core component of endosome docking. Nature 397, 621-5. Christoforidis, S., Miaczynska, M., Ashman, K., Wilm, M., Zhao, L., Yip, S. C., Waterfield, M. D., Backer, J. M., and Zerial, M. (1999b): Phosphatidylinositol-3OH kinases are Rab5 effectors. Nat Cell Biol 1, 249-52. Codlin, S., and Mole, S. E. (2009): S. pombe btn1, the orthologue of the Batten disease gene CLN3, is required for vacuole protein sorting of Cpy1p and Golgi exit of Vps10p. J Cell Sci 122, 1163-73. 203 Collins, B. M., Skinner, C. F., Watson, P. J., Seaman, M. N., and Owen, D. J. (2005a): Vps29 has a phosphoesterase fold that acts as a protein interaction scaffold for retromer assembly. Nat Struct Mol Biol 12, 594-602. Collins, K. M., Thorngren, N. L., Fratti, R. A., and Wickner, W. T. (2005b): Sec17p and HOPS, in distinct SNARE complexes, mediate SNARE complex disruption or assembly for fusion. Embo J 24, 1775-86. Cooper, G. M., and Hausman, R. E. (2006): The Cell: A Molecular Approach. Sinauer Associates, Inc. Cornell, R. B., and Taneva, S. G. (2006): Amphipathic helices as mediators of the membrane interaction of amphitropic proteins, and as modulators of bilayer physical properties. Curr Protein Pept Sci 7, 539-52. Cox, T. M. (2005): Substrate reduction therapy for lysosomal storage diseases. Acta Paediatr Suppl 94, 69-75; discussion 57. Cozier, G. E., Carlton, J., McGregor, A. H., Gleeson, P. A., Teasdale, R. D., Mellor, H., and Cullen, P. J. (2002): The phox homology (PX) domain-dependent, 3phosphoinositide-mediated association of sorting nexin-1 with an early sorting endosomal compartment is required for its ability to regulate epidermal growth factor receptor degradation. J Biol Chem 277, 48730-6. Cuajungco, M. P., and Lees, G. J. (1997a): Zinc and Alzheimer's disease: is there a direct link? Brain Res Brain Res Rev 23, 219-36. Cuajungco, M. P., and Lees, G. J. (1997b): Zinc metabolism in the brain: relevance to human neurodegenerative disorders. Neurobiol Dis 4, 137-69. Cullen, P. J. (2011): Phosphoinositides and the regulation of tubular-based endosomal sorting. Biochem Soc Trans 39, 839-50. Cullen, P. J., and Korswagen, H. C. (2011): Sorting nexins provide diversity for retromerdependent trafficking events. Nat Cell Biol 14, 29-37. da Silva, A. C., and Reinach, F. C. (1991): Calcium binding induces conformational changes in muscle regulatory proteins. Trends Biochem Sci 16, 53-7. Dahms, N. M., and Hancock, M. K. (2002): P-type lectins. Biochim Biophys Acta 1572, 317-40. Damen, E., Krieger, E., Nielsen, J. E., Eygensteyn, J., and van Leeuwen, J. E. (2006): The human Vps29 retromer component is a metallo-phosphoesterase for a cationindependent mannose 6-phosphate receptor substrate peptide. Biochem J 398, 399409. Daumke, O., Lundmark, R., Vallis, Y., Martens, S., Butler, P. J., and McMahon, H. T. (2007): Architectural and mechanistic insights into an EHD ATPase involved in membrane remodelling. Nature 449, 923-7. de Alba, E., and Tjandra, N. (2004): Structural studies on the Ca2+-binding domain of human nucleobindin (calnuc). Biochemistry 43, 10039-49. De Duve, C., Pressman, B. C., Gianetto, R., Wattiaux, R., and Appelmans, F. (1955): Tissue fractionation studies. 6. Intracellular distribution patterns of enzymes in ratliver tissue. Biochem J 60, 604-17. De Luca, A., Progida, C., Spinosa, M. R., Alifano, P., and Bucci, C. (2008): Characterization of the Rab7K157N mutant protein associated with Charcot-MarieTooth type 2B. Biochem Biophys Res Commun 372, 283-7. 204 Derivery, E., and Gautreau, A. (2010): Evolutionary conservation of the WASH complex, an actin polymerization machine involved in endosomal fission. Commun Integr Biol 3, 227-30. Derivery, E., Sousa, C., Gautier, J. J., Lombard, B., Loew, D., and Gautreau, A. (2009): The Arp2/3 activator WASH controls the fission of endosomes through a large multiprotein complex. Dev Cell 17, 712-23. Desnick, R. J. (2004): Enzyme replacement and enhancement therapies for lysosomal diseases. J Inherit Metab Dis 27, 385-410. Di Fruscio, G., Schulz, A., De Cegli, R., Savarese, M., Mutarelli, M., Parenti, G., Banfi, S., Braulke, T., Nigro, V., and Ballabio, A. (2015): Lysoplex: An efficient toolkit to detect DNA sequence variations in the autophagy-lysosomal pathway. Autophagy 11, 928-38. Diaz, E., and Pfeffer, S. R. (1998): TIP47: a cargo selection device for mannose 6phosphate receptor trafficking. Cell 93, 433-43. Dingwall, C., and Laskey, R. A. (1991): Nuclear targeting sequences--a consensus? Trends Biochem Sci 16, 478-81. Dittmer, F., Ulbrich, E. J., Hafner, A., Schmahl, W., Meister, T., Pohlmann, R., and von Figura, K. (1999): Alternative mechanisms for trafficking of lysosomal enzymes in mannose 6-phosphate receptor-deficient mice are cell type-specific. J Cell Sci 112 ( Pt 10), 1591-7. Doray, B., Ghosh, P., Griffith, J., Geuze, H. J., and Kornfeld, S. (2002): Cooperation of GGAs and AP-1 in packaging MPRs at the trans-Golgi network. Science 297, 17003. Doray, B., Knisely, J. M., Wartman, L., Bu, G., and Kornfeld, S. (2008): Identification of acidic dileucine signals in LRP9 that interact with both GGAs and AP-1/AP-2. Traffic 9, 1551-62. Ducharme, N. A., and Bickel, P. E. (2008): Lipid droplets in lipogenesis and lipolysis. Endocrinology 149, 942-9. Duncan, J. R., and Kornfeld, S. (1988): Intracellular movement of two mannose 6phosphate receptors: return to the Golgi apparatus. J Cell Biol 106, 617-28. Eaton, S. (2008): Retromer retrieves wntless. Dev Cell 14, 4-6. Elleder, L., Gobel (1999): Definitions of the ultrastructural patterns found in NCL. Elleder, M., Sokolova, J., and Hrebicek, M. (1997): Follow-up study of subunit c of mitochondrial ATP synthase (SCMAS) in Batten disease and in unrelated lysosomal disorders. Acta Neuropathol 93, 379-90. Emmerzaal, T. L., and Kozicz, T. (2013): Nesfatin-1; implication in stress and stressassociated anxiety and depression. Curr Pharm Des 19, 6941-8. Eng, C. M., Banikazemi, M., Gordon, R. E., Goldman, M., Phelps, R., Kim, L., Gass, A., Winston, J., Dikman, S., Fallon, J. T., Brodie, S., Stacy, C. B., Mehta, D., Parsons, R., Norton, K., O'Callaghan, M., and Desnick, R. J. (2001): A phase 1/2 clinical trial of enzyme replacement in fabry disease: pharmacokinetic, substrate clearance, and safety studies. Am J Hum Genet 68, 711-22. Espinosa, E. J., Calero, M., Sridevi, K., and Pfeffer, S. R. (2009): RhoBTB3: a Rho GTPase-family ATPase required for endosome to Golgi transport. Cell 137, 938-48. 205 Evans, J. H., Spencer, D. M., Zweifach, A., and Leslie, C. C. (2001): Intracellular calcium signals regulating cytosolic phospholipase A2 translocation to internal membranes. J Biol Chem 276, 30150-60. Ezaki, J., Takeda-Ezaki, M., Koike, M., Ohsawa, Y., Taka, H., Mineki, R., Murayama, K., Uchiyama, Y., Ueno, T., and Kominami, E. (2003): Characterization of Cln3p, the gene product responsible for juvenile neuronal ceroid lipofuscinosis, as a lysosomal integral membrane glycoprotein. J Neurochem 87, 1296-308. Fasshauer, D., Sutton, R. B., Brunger, A. T., and Jahn, R. (1998): Conserved structural features of the synaptic fusion complex: SNARE proteins reclassified as Q- and RSNAREs. Proc Natl Acad Sci U S A 95, 15781-6. Finan, G. M., Okada, H., and Kim, T. W. (2011): BACE1 retrograde trafficking is uniquely regulated by the cytoplasmic domain of sortilin. J Biol Chem 286, 12602-16. Fjorback, A. W., Seaman, M., Gustafsen, C., Mehmedbasic, A., Gokool, S., Wu, C., Militz, D., Schmidt, V., Madsen, P., Nyengaard, J. R., Willnow, T. E., Christensen, E. I., Mobley, W. B., Nykjaer, A., and Andersen, O. M. (2012): Retromer binds the FANSHY sorting motif in SorLA to regulate amyloid precursor protein sorting and processing. J Neurosci 32, 1467-80. Ford, M. G., Mills, I. G., Peter, B. J., Vallis, Y., Praefcke, G. J., Evans, P. R., and McMahon, H. T. (2002): Curvature of clathrin-coated pits driven by epsin. Nature 419, 361-6. Fossale, E., Wolf, P., Espinola, J. A., Lubicz-Nawrocka, T., Teed, A. M., Gao, H., Rigamonti, D., Cattaneo, E., MacDonald, M. E., and Cotman, S. L. (2004): Membrane trafficking and mitochondrial abnormalities precede subunit c deposition in a cerebellar cell model of juvenile neuronal ceroid lipofuscinosis. BMC Neurosci 5, 57. Fukuda, M. (1990): [Lysosomal membrane glycoproteins; structure, biosynthesis, and trafficking to lysosome]. Seikagaku 62, 1225-40. Ganley, I. G., Carroll, K., Bittova, L., and Pfeffer, S. (2004): Rab9 GTPase regulates late endosome size and requires effector interaction for its stability. Mol Biol Cell 15, 5420-30. Garcia-Galiano, D., Navarro, V. M., Roa, J., Ruiz-Pino, F., Sanchez-Garrido, M. A., Pineda, R., Castellano, J. M., Romero, M., Aguilar, E., Gaytan, F., Dieguez, C., Pinilla, L., and Tena-Sempere, M. (2010): The anorexigenic neuropeptide, nesfatin1, is indispensable for normal puberty onset in the female rat. J Neurosci 30, 778392. Garcia-Galiano, D., Pineda, R., Ilhan, T., Castellano, J. M., Ruiz-Pino, F., SanchezGarrido, M. A., Vazquez, M. J., Sangiao-Alvarellos, S., Romero-Ruiz, A., Pinilla, L., Dieguez, C., Gaytan, F., and Tena-Sempere, M. (2012): Cellular distribution, regulated expression, and functional role of the anorexigenic peptide, NUCB2/nesfatin-1, in the testis. Endocrinology 153, 1959-71. Garcia-Galiano, D., and Tena-Sempere, M. (2013): Emerging roles of NUCB2/nesfatin-1 in the metabolic control of reproduction. Curr Pharm Des 19, 6966-72. Garcia-Marcos, M., Kietrsunthorn, P. S., Wang, H., Ghosh, P., and Farquhar, M. G. (2011): G Protein binding sites on Calnuc (nucleobindin 1) and NUCB2 (nucleobindin 2) define a new class of G(alpha)i-regulatory motifs. J Biol Chem 286, 28138-49. 206 Garmroudi, F., Devi, G., Slentz, D. H., Schaffer, B. S., and MacDonald, R. G. (1996): Truncated forms of the insulin-like growth factor II (IGF-II)/mannose 6-phosphate receptor encompassing the IGF-II binding site: characterization of a point mutation that abolishes IGF-II binding. Mol Endocrinol 10, 642-51. Getty, A., Kovacs, A. D., Lengyel-Nelson, T., Cardillo, A., Hof, C., Chan, C. H., and Pearce, D. A. (2013): Osmotic stress changes the expression and subcellular localization of the Batten disease protein CLN3. PLoS One 8, e66203. Getty, A. L., Benedict, J. W., and Pearce, D. A. (2011): A novel interaction of CLN3 with nonmuscle myosin-IIB and defects in cell motility of Cln3(-/-) cells. Exp Cell Res 317, 51-69. Geuze, H. J., Stoorvogel, W., Strous, G. J., Slot, J. W., Bleekemolen, J. E., and Mellman, I. (1988): Sorting of mannose 6-phosphate receptors and lysosomal membrane proteins in endocytic vesicles. J Cell Biol 107, 2491-501. Ghai, R., Bugarcic, A., Liu, H., Norwood, S. J., Skeldal, S., Coulson, E. J., Li, S. S., Teasdale, R. D., and Collins, B. M. (2013): Structural basis for endosomal trafficking of diverse transmembrane cargos by PX-FERM proteins. Proc Natl Acad Sci U S A 110, E643-52. Gifford, J. L., Walsh, M. P., and Vogel, H. J. (2007): Structures and metal-ion-binding properties of the Ca2+-binding helix-loop-helix EF-hand motifs. Biochem J 405, 199-221. Gilchrist, A., Au, C. E., Hiding, J., Bell, A. W., Fernandez-Rodriguez, J., Lesimple, S., Nagaya, H., Roy, L., Gosline, S. J., Hallett, M., Paiement, J., Kearney, R. E., Nilsson, T., and Bergeron, J. J. (2006): Quantitative proteomics analysis of the secretory pathway. Cell 127, 1265-81. Giugliani, R., Federhen, A., Vairo, F., Vanzella, C., Pasqualim, G., da Silva, L. M., Giugliani, L., de Boer, A. P., de Souza, C. F., Matte, U., and Baldo, G. (2016): Emerging drugs for the treatment of mucopolysaccharidoses. Expert Opin Emerg Drugs, 1-18. Goda, Y., and Pfeffer, S. R. (1988): Selective recycling of the mannose 6-phosphate/IGF-II receptor to the trans Golgi network in vitro. Cell 55, 309-20. Goebel, M., Stengel, A., Wang, L., and Tache, Y. (2011): Central nesfatin-1 reduces the nocturnal food intake in mice by reducing meal size and increasing inter-meal intervals. Peptides 32, 36-43. Gokool, S., Tattersall, D., and Seaman, M. N. (2007): EHD1 interacts with retromer to stabilize SNX1 tubules and facilitate endosome-to-Golgi retrieval. Traffic 8, 187386. Golabek, A. A., Kida, E., Walus, M., Kaczmarski, W., Michalewski, M., and Wisniewski, K. E. (2000): CLN3 protein regulates lysosomal pH and alters intracellular processing of Alzheimer's amyloid-beta protein precursor and cathepsin D in human cells. Mol Genet Metab 70, 203-13. Gomez, T. S., and Billadeau, D. D. (2009): A FAM21-containing WASH complex regulates retromer-dependent sorting. Dev Cell 17, 699-711. Gonatas, N. K., Steiber, A., Kim, S. U., Graham, D. I., and Avrameas, S. (1975): Internalization of neuronal plasma membrane ricin receptors into the Golgi apparatus. Exp Cell Res 94, 426-31. 207 Grabe, M., and Oster, G. (2001): Regulation of organelle acidity. J Gen Physiol 117, 32944. Grohovaz, F., Bossi, M., Pezzati, R., Meldolesi, J., and Tarelli, F. T. (1996): High resolution ultrastructural mapping of total calcium: electron spectroscopic imaging/electron energy loss spectroscopy analysis of a physically/chemically processed nerve-muscle preparation. Proc Natl Acad Sci U S A 93, 4799-803. Grubman, A., Lidgerwood, G. E., Duncan, C., Bica, L., Tan, J. L., Parker, S. J., Caragounis, A., Meyerowitz, J., Volitakis, I., Moujalled, D., Liddell, J. R., Hickey, J. L., Horne, M., Longmuir, S., Koistinaho, J., Donnelly, P. S., Crouch, P. J., Tammen, I., White, A. R., and Kanninen, K. M. (2014a): Deregulation of subcellular biometal homeostasis through loss of the metal transporter, Zip7, in a childhood neurodegenerative disorder. Acta Neuropathol Commun 2, 25. Grubman, A., Pollari, E., Duncan, C., Caragounis, A., Blom, T., Volitakis, I., Wong, A., Cooper, J., Crouch, P. J., Koistinaho, J., Jalanko, A., White, A. R., and Kanninen, K. M. (2014b): Deregulation of biometal homeostasis: the missing link for neuronal ceroid lipofuscinoses? Metallomics 6, 932-43. Gu, X. (1998): Early metazoan divergence was about 830 million years ago. J Mol Evol 47, 369-71. Guichard, A., Nizet, V., and Bier, E. (2014): RAB11-mediated trafficking in host-pathogen interactions. Nat Rev Microbiol 12, 624-34. Gustafsen, C., Glerup, S., Pallesen, L. T., Olsen, D., Andersen, O. M., Nykjaer, A., Madsen, P., and Petersen, C. M. (2013): Sortilin and SorLA display distinct roles in processing and trafficking of amyloid precursor protein. J Neurosci 33, 64-71. Haddad, S. E., Khoury, M., Daoud, M., Kantar, R., Harati, H., Mousallem, T., Alzate, O., Meyer, B., and Boustany, R. M. (2012): CLN5 and CLN8 protein association with ceramide synthase: biochemical and proteomic approaches. Electrophoresis 33, 3798-809. Hales, C. M., Vaerman, J. P., and Goldenring, J. R. (2002): Rab11 family interacting protein 2 associates with Myosin Vb and regulates plasma membrane recycling. J Biol Chem 277, 50415-21. Haltia, M. (2003): The neuronal ceroid-lipofuscinoses. J Neuropathol Exp Neurol 62, 1-13. Hancock, M. K., Haskins, D. J., Sun, G., and Dahms, N. M. (2002): Identification of residues essential for carbohydrate recognition by the insulin-like growth factor II/mannose 6-phosphate receptor. J Biol Chem 277, 11255-64. Hanna, J., Carroll, K., and Pfeffer, S. R. (2002): Identification of residues in TIP47 essential for Rab9 binding. Proc Natl Acad Sci U S A 99, 7450-4. Hannover, A. (1843): Mikrokopiske undersögelser af nervesystemet. Det kongelige Danske Videnskabernes Selskabs Naturvidenskabelige og Mathematiske Afhandlinger 10, 1112. Harbour, M. E., Breusegem, S. Y., Antrobus, R., Freeman, C., Reid, E., and Seaman, M. N. (2010): The cargo-selective retromer complex is a recruiting hub for protein complexes that regulate endosomal tubule dynamics. J Cell Sci 123, 3703-17. Harterink, M., Port, F., Lorenowicz, M. J., McGough, I. J., Silhankova, M., Betist, M. C., van Weering, J. R., van Heesbeen, R. G., Middelkoop, T. C., Basler, K., Cullen, P. J., and Korswagen, H. C. (2011): A SNX3-dependent retromer pathway mediates 208 retrograde transport of the Wnt sorting receptor Wntless and is required for Wnt secretion. Nat Cell Biol 13, 914-23. Haskell, R. E., Carr, C. J., Pearce, D. A., Bennett, M. J., and Davidson, B. L. (2000): Batten disease: evaluation of CLN3 mutations on protein localization and function. Hum Mol Genet 9, 735-44. Haze, K., Yoshida, H., Yanagi, H., Yura, T., and Mori, K. (1999): Mammalian transcription factor ATF6 is synthesized as a transmembrane protein and activated by proteolysis in response to endoplasmic reticulum stress. Mol Biol Cell 10, 3787-99. Heard, J. M., Bruyere, J., Roy, E., Bigou, S., Ausseil, J., and Vitry, S. (2010): Storage problems in lysosomal diseases. Biochem Soc Trans 38, 1442-7. Heinonen, O., Salonen, T., Jalanko, A., Peltonen, L., and Copp, A. (2000): CLN-1 and CLN-5, genes for infantile and variant late infantile neuronal ceroid lipofuscinoses, are expressed in the embryonic human brain. J Comp Neurol 426, 406-12. Hierro, A., Rojas, A. L., Rojas, R., Murthy, N., Effantin, G., Kajava, A. V., Steven, A. C., Bonifacino, J. S., and Hurley, J. H. (2007): Functional architecture of the retromer cargo-recognition complex. Nature 449, 1063-7. Hirst, J., Barlow, L. D., Francisco, G. C., Sahlender, D. A., Seaman, M. N., Dacks, J. B., and Robinson, M. S. (2011): The fifth adaptor protein complex. PLoS Biol 9, e1001170. Holmberg, V., Jalanko, A., Isosomppi, J., Fabritius, A. L., Peltonen, L., and Kopra, O. (2004): The mouse ortholog of the neuronal ceroid lipofuscinosis CLN5 gene encodes a soluble lysosomal glycoprotein expressed in the developing brain. Neurobiol Dis 16, 29-40. Holmberg, V., Lauronen, L., Autti, T., Santavuori, P., Savukoski, M., Uvebrant, P., Hofman, I., Peltonen, L., and Jarvela, I. (2000): Phenotype-genotype correlation in eight patients with Finnish variant late infantile NCL (CLN5). Neurology 55, 57981. Holopainen, J. M., Saarikoski, J., Kinnunen, P. K., and Jarvela, I. (2001): Elevated lysosomal pH in neuronal ceroid lipofuscinoses (NCLs). Eur J Biochem 268, 58516. Hong, Z., Yang, Y., Zhang, C., Niu, Y., Li, K., Zhao, X., and Liu, J. J. (2009): The retromer component SNX6 interacts with dynactin p150(Glued) and mediates endosome-to-TGN transport. Cell Res 19, 1334-49. Horiuchi, H., Lippe, R., McBride, H. M., Rubino, M., Woodman, P., Stenmark, H., Rybin, V., Wilm, M., Ashman, K., Mann, M., and Zerial, M. (1997): A novel Rab5 GDP/GTP exchange factor complexed to Rabaptin-5 links nucleotide exchange to effector recruitment and function. Cell 90, 1149-59. Hueck, W. (1912): Pigmentstudien. Beitr. Pathol. Anat., 68–232. Hunker, C. M., Galvis, A., Kruk, I., Giambini, H., Veisaga, M. L., and Barbieri, M. A. (2006): Rab5-activating protein 6, a novel endosomal protein with a role in endocytosis. Biochem Biophys Res Commun 340, 967-75. Huotari, J., and Helenius, A. (2011): Endosome maturation. Embo J 30, 3481-500. Ikura, M. (1996): Calcium binding and conformational response in EF-hand proteins. Trends Biochem Sci 21, 14-7. 209 Islam, A., Adamik, B., Hawari, F. I., Ma, G., Rouhani, F. N., Zhang, J., and Levine, S. J. (2006): Extracellular TNFR1 release requires the calcium-dependent formation of a nucleobindin 2-ARTS-1 complex. J Biol Chem 281, 6860-73. Isosomppi, J., Vesa, J., Jalanko, A., and Peltonen, L. (2002): Lysosomal localization of the neuronal ceroid lipofuscinosis CLN5 protein. Hum Mol Genet 11, 885-91. Jahn, R., and Scheller, R. H. (2006): SNAREs--engines for membrane fusion. Nat Rev Mol Cell Biol 7, 631-43. Janes, R. W., Munroe, P. B., Mitchison, H. M., Gardiner, R. M., Mole, S. E., and Wallace, B. A. (1996): A model for Batten disease protein CLN3: functional implications from homology and mutations. FEBS Lett 399, 75-7. Jao, C. C., Hegde, B. G., Gallop, J. L., Hegde, P. B., McMahon, H. T., Haworth, I. S., and Langen, R. (2010): Roles of amphipathic helices and the bin/amphiphysin/rvs (BAR) domain of endophilin in membrane curvature generation. J Biol Chem 285, 20164-70. Jarvela, I., Lehtovirta, M., Tikkanen, R., Kyttala, A., and Jalanko, A. (1999): Defective intracellular transport of CLN3 is the molecular basis of Batten disease (JNCL). Hum Mol Genet 8, 1091-8. Jarvela, I., Sainio, M., Rantamaki, T., Olkkonen, V. M., Carpen, O., Peltonen, L., and Jalanko, A. (1998): Biosynthesis and intracellular targeting of the CLN3 protein defective in Batten disease. Hum Mol Genet 7, 85-90. Jay, P., Rougeulle, C., Massacrier, A., Moncla, A., Mattei, M. G., Malzac, P., Roeckel, N., Taviaux, S., Lefranc, J. L., Cau, P., Berta, P., Lalande, M., and Muscatelli, F. (1997): The human necdin gene, NDN, is maternally imprinted and located in the Prader-Willi syndrome chromosomal region. Nat Genet 17, 357-61. Jia, D., Gomez, T. S., Billadeau, D. D., and Rosen, M. K. (2012): Multiple repeat elements within the FAM21 tail link the WASH actin regulatory complex to the retromer. Mol Biol Cell 23, 2352-61. Johannes, L., and Wunder, C. (2011): Retrograde transport: two (or more) roads diverged in an endosomal tree? Traffic 12, 956-62. Johansson, M., Rocha, N., Zwart, W., Jordens, I., Janssen, L., Kuijl, C., Olkkonen, V. M., and Neefjes, J. (2007): Activation of endosomal dynein motors by stepwise assembly of Rab7-RILP-p150Glued, ORP1L, and the receptor betalll spectrin. J Cell Biol 176, 459-71. Johnson, K. F., and Kornfeld, S. (1992a): The cytoplasmic tail of the mannose 6phosphate/insulin-like growth factor-II receptor has two signals for lysosomal enzyme sorting in the Golgi. J Cell Biol 119, 249-57. Johnson, K. F., and Kornfeld, S. (1992b): A His-Leu-Leu sequence near the carboxyl terminus of the cytoplasmic domain of the cation-dependent mannose 6-phosphate receptor is necessary for the lysosomal enzyme sorting function. J Biol Chem 267, 17110-5. Jolly, R. D., Douglas, B. V., Davey, P. M., and Roiri, J. E. (1995): Lipofuscin in bovine muscle and brain: a model for studying age pigment. Gerontology 41 Suppl 2, 28395. Jung, T., Bader, N., and Grune, T. (2007): Lipofuscin: formation, distribution, and metabolic consequences. Ann N Y Acad Sci 1119, 97-111. 210 Kaczmarski, W., Wisniewski, K. E., Golabek, A., Kaczmarski, A., Kida, E., and Michalewski, M. (1999): Studies of membrane association of CLN3 protein. Mol Genet Metab 66, 261-4. Kalnina, Z., Silina, K., Bruvere, R., Gabruseva, N., Stengrevics, A., Barnikol-Watanabe, S., Leja, M., and Line, A. (2009): Molecular characterisation and expression analysis of SEREX-defined antigen NUCB2 in gastric epithelium, gastritis and gastric cancer. Eur J Histochem 53, 7-18. Kanai, Y., Isonishi, S., Katagiri, T., Koizumi, T., Miura, K., and Kurosawa, Y. (1990): Early induction of anti-double-stranded DNA antibodies in lupus-prone MRL mice inoculated with Ly-24+ cells cloned from lymph node cells of an MRL/Mp-lpr/lpr mouse: possible effect of putative cytokines produced by cloned Ly-24+ cells. Immunol Lett 24, 49-55. Kanai, Y., Katagiri, T., Mori, S., and Kubota, T. (1986): An established MRL/Mp-lpr/lpr cell line with null cell properties produces a B cell differentiation factor(s) that promotes anti-single-stranded DNA antibody production in MRL spleen cell culture. Int Arch Allergy Appl Immunol 81, 92-4. Kanai, Y., Kyuwa, S., Miura, K., and Kurosawa, Y. (1995a): Induction and natural occurrence of serum nucleosomal DNA in autoimmune MRL/lpr/lpr mice: its relation to apoptosis in the thymus. Immunol Lett 46, 207-14. Kanai, Y., Takeda, O., Miura, K., Amagai, M., Kaneko, T., Kubota, T., Tanuma, S., and Kurosawa, Y. (1995b): Induction of autoantibodies in normal mice by injection of nucleobindin and natural occurrence of antibodies against nucleobindin in autoimmune MRL/lpr/lpr mice. Immunol Lett 45, 35-42. Kanai, Y., Takeda, O., Miura, K., and Kurosawa, Y. (1993): Novel autoimmune phenomena induced in vivo by a new DNA binding protein Nuc: a study on MRL/n mice. Immunol Lett 39, 83-9. Kanai, Y., and Tanuma, S. (1992): Purification of a novel B cell growth and differentiation factor associated with lupus syndrome. Immunol Lett 32, 43-8. Kanninen, K. M., Grubman, A., Meyerowitz, J., Duncan, C., Tan, J. L., Parker, S. J., Crouch, P. J., Paterson, B. M., Hickey, J. L., Donnelly, P. S., Volitakis, I., Tammen, I., Palmer, D. N., and White, A. R. (2013): Increased zinc and manganese in parallel with neurodegeneration, synaptic protein changes and activation of Akt/GSK3 signaling in ovine CLN6 neuronal ceroid lipofuscinosis. PLoS One 8, e58644. Kanuru, M., Raman, R., and Aradhyam, G. K. (2013): Serine protease activity of calnuc: regulation by Zn2+ and G proteins. J Biol Chem. Kanuru, M., Samuel, J. J., Balivada, L. M., and Aradhyam, G. K. (2009): Ion-binding properties of Calnuc, Ca2+ versus Mg2+--Calnuc adopts additional and unusual Ca2+-binding sites upon interaction with G-protein. Febs J 276, 2529-46. Kapoor, N., Gupta, R., Menon, S. T., Folta-Stogniew, E., Raleigh, D. P., and Sakmar, T. P. (2010): Nucleobindin 1 is a calcium-regulated guanine nucleotide dissociation inhibitor of G{alpha}i1. J Biol Chem 285, 31647-60. Karabinos, A., Bhattacharya, D., Morys-Wortmann, C., Kroll, K., Hirschfeld, G., Kratzin, H. D., Barnikol-Watanabe, S., and Hilschmann, N. (1996): The divergent domains of the NEFA and nucleobindin proteins are derived from an EF-hand ancestor. Mol Biol Evol 13, 990-8. 211 Kerr, M. C., Bennetts, J. S., Simpson, F., Thomas, E. C., Flegg, C., Gleeson, P. A., Wicking, C., and Teasdale, R. D. (2005): A novel mammalian retromer component, Vps26B. Traffic 6, 991-1001. Kida, E., Kaczmarski, W., Golabek, A. A., Kaczmarski, A., Michalewski, M., and Wisniewski, K. E. (1999): Analysis of intracellular distribution and trafficking of the CLN3 protein in fusion with the green fluorescent protein in vitro. Mol Genet Metab 66, 265-71. Kim, H. S., Kim, E. M., Lee, J. P., Park, C. H., Kim, S., Seo, J. H., Chang, K. A., Yu, E., Jeong, S. J., Chong, Y. H., and Suh, Y. H. (2003a): C-terminal fragments of amyloid precursor protein exert neurotoxicity by inducing glycogen synthase kinase-3beta expression. Faseb J 17, 1951-3. Kim, Y., Ramirez-Montealegre, D., and Pearce, D. A. (2003b): A role in vacuolar arginine transport for yeast Btn1p and for human CLN3, the protein defective in Batten disease. Proc Natl Acad Sci U S A 100, 15458-62. Kitzmuller, C., Haines, R. L., Codlin, S., Cutler, D. F., and Mole, S. E. (2008): A function retained by the common mutant CLN3 protein is responsible for the late onset of juvenile neuronal ceroid lipofuscinosis. Hum Mol Genet 17, 303-12. Kjolby, M., Andersen, O. M., Breiderhoff, T., Fjorback, A. W., Pedersen, K. M., Madsen, P., Jansen, P., Heeren, J., Willnow, T. E., and Nykjaer, A. (2010): Sort1, encoded by the cardiovascular risk locus 1p13.3, is a regulator of hepatic lipoprotein export. Cell Metab 12, 213-23. Koivu, T., Laitinen, S., Riento, K., and Olkkonen, V. M. (1997): Sequence of a human cDNA encoding Cab45, a Ca2+-binding protein with six EF-hand motifs. DNA Seq 7, 217-20. Kollmann, K., Mutenda, K. E., Balleininger, M., Eckermann, E., von Figura, K., Schmidt, B., and Lubke, T. (2005): Identification of novel lysosomal matrix proteins by proteome analysis. Proteomics 5, 3966-78. Konczol, K., Bodnar, I., Zelena, D., Pinter, O., Papp, R. S., Palkovits, M., Nagy, G. M., and Toth, Z. E. (2010): Nesfatin-1/NUCB2 may participate in the activation of the hypothalamic-pituitary-adrenal axis in rats. Neurochem Int 57, 189-97. Koneff, H. (1886): Beiträge zur Kenntniss der Nervenzellen den peripheren Ganglien. Mitt. Naturforsch. Gesellsch 44, 13–14. Kornfeld, S., and Mellman, I. (1989): The biogenesis of lysosomes. Annu Rev Cell Biol 5, 483-525. Koumandou, V. L., Klute, M. J., Herman, E. K., Nunez-Miguel, R., Dacks, J. B., and Field, M. C. (2011): Evolutionary reconstruction of the retromer complex and its function in Trypanosoma brucei. J Cell Sci 124, 1496-509. Kretsinger, R. H. (1987): Calcium coordination and the calmodulin fold: divergent versus convergent evolution. Cold Spring Harb Symp Quant Biol 52, 499-510. Kroll, K. A., Otte, S., Hirschfeld, G., Barnikol-Watanabe, S., Gotz, H., Sternbach, H., Kratzin, H. D., Barnikol, H. U., and Hilschmann, N. (1999): Heterologous overexpression of human NEFA and studies on the two EF-hand calcium-binding sites. Biochem Biophys Res Commun 260, 1-8. Kubota, T., Miyauchi, M., Miura, K., Hirokawa, G., Awaya, A., Miyasaka, N., Kurosawa, Y., Kanai, Y., and Maruyama, K. (1998): Upregulation of nucleobindin expression in human-activated lymphocytes and non-Hodgkin's lymphoma. Pathol Int 48, 22-8. 212 Kurten, R. C., Eddington, A. D., Chowdhury, P., Smith, R. D., Davidson, A. D., and Shank, B. B. (2001): Self-assembly and binding of a sorting nexin to sorting endosomes. J Cell Sci 114, 1743-56. Kutay, U., Hartmann, E., and Rapoport, T. A. (1993): A class of membrane proteins with a C-terminal anchor. Trends Cell Biol 3, 72-5. Kyttala, A., Ihrke, G., Vesa, J., Schell, M. J., and Luzio, J. P. (2004): Two motifs target Batten disease protein CLN3 to lysosomes in transfected nonneuronal and neuronal cells. Mol Biol Cell 15, 1313-23. LaFerla, F. M. (2002): Calcium dyshomeostasis and intracellular signalling in Alzheimer's disease. Nat Rev Neurosci 3, 862-72. Lambright, D. G., Sondek, J., Bohm, A., Skiba, N. P., Hamm, H. E., and Sigler, P. B. (1996): The 2.0 A crystal structure of a heterotrimeric G protein. Nature 379, 311-9. Lane, S. C., Jolly, R. D., Schmechel, D. E., Alroy, J., and Boustany, R. M. (1996): Apoptosis as the mechanism of neurodegeneration in Batten's disease. J Neurochem 67, 677-83. Laplante, M., and Sabatini, D. M. (2012): mTOR signaling in growth control and disease. Cell 149, 274-93. Larkin, H., Ribeiro, M. G., and Lavoie, C. (2013): Topology and membrane anchoring of the lysosomal storage disease-related protein CLN5. Hum Mutat 34, 1688-97. Lauffer, B. E., Melero, C., Temkin, P., Lei, C., Hong, W., Kortemme, T., and von Zastrow, M. (2010): SNX27 mediates PDZ-directed sorting from endosomes to the plasma membrane. J Cell Biol 190, 565-74. Laurent-Matha, V., Derocq, D., Prebois, C., Katunuma, N., and Liaudet-Coopman, E. (2006): Processing of human cathepsin D is independent of its catalytic function and auto-activation: involvement of cathepsins L and B. J Biochem 139, 363-71. Lavoie, C., Meerloo, T., Lin, P., and Farquhar, M. G. (2002): Calnuc, an EF-hand Ca(2+)binding protein, is stored and processed in the Golgi and secreted by the constitutive-like pathway in AtT20 cells. Mol Endocrinol 16, 2462-74. Lebrun, A. H., Storch, S., Ruschendorf, F., Schmiedt, M. L., Kyttala, A., Mole, S. E., Kitzmuller, C., Saar, K., Mewasingh, L. D., Boda, V., Kohlschutter, A., Ullrich, K., Braulke, T., and Schulz, A. (2009): Retention of lysosomal protein CLN5 in the endoplasmic reticulum causes neuronal ceroid lipofuscinosis in Asian sibship. Hum Mutat 30, E651-61. Leclerc, P., Biarc, J., St-Onge, M., Gilbert, C., Dussault, A. A., Laflamme, C., and Pouliot, M. (2008): Nucleobindin co-localizes and associates with cyclooxygenase (COX)-2 in human neutrophils. PLoS One 3, e2229. Lee, A. Y., Gulnik, S. V., and Erickson, J. W. (1998): Conformational switching in an aspartic proteinase. Nat Struct Biol 5, 866-71. Lefrancois, S., Zeng, J., Hassan, A. J., Canuel, M., and Morales, C. R. (2003): The lysosomal trafficking of sphingolipid activator proteins (SAPs) is mediated by sortilin. Embo J 22, 6430-7. Leung, K. F., Baron, R., and Seabra, M. C. (2006): Thematic review series: lipid posttranslational modifications. geranylgeranylation of Rab GTPases. J Lipid Res 47, 467-75. Li, H., and Villalobo, A. (2002): Evidence for the direct interaction between calmodulin and the human epidermal growth factor receptor. Biochem J 362, 499-505. 213 Li, Q. C., Wang, H. Y., Chen, X., Guan, H. Z., and Jiang, Z. Y. (2010): Fasting plasma levels of nesfatin-1 in patients with type 1 and type 2 diabetes mellitus and the nutrient-related fluctuation of nesfatin-1 level in normal humans. Regul Pept 159, 72-7. Lin, P., Fischer, T., Lavoie, C., Huang, H., and Farquhar, M. G. (2009): Calnuc plays a role in dynamic distribution of Galphai but not Gbeta subunits and modulates ACTH secretion in AtT-20 neuroendocrine secretory cells. Mol Neurodegener 4, 15. Lin, P., Fischer, T., Weiss, T., and Farquhar, M. G. (2000): Calnuc, an EF-hand Ca(2+) binding protein, specifically interacts with the C-terminal alpha5-helix of G(alpha)i3. Proc Natl Acad Sci U S A 97, 674-9. Lin, P., Le-Niculescu, H., Hofmeister, R., McCaffery, J. M., Jin, M., Hennemann, H., McQuistan, T., De Vries, L., and Farquhar, M. G. (1998): The mammalian calciumbinding protein, nucleobindin (CALNUC), is a Golgi resident protein. J Cell Biol 141, 1515-27. Lin, P., Li, F., Zhang, Y. W., Huang, H., Tong, G., Farquhar, M. G., and Xu, H. (2007): Calnuc binds to Alzheimer's beta-amyloid precursor protein and affects its biogenesis. J Neurochem 100, 1505-14. Lin, P., Yao, Y., Hofmeister, R., Tsien, R. Y., and Farquhar, M. G. (1999): Overexpression of CALNUC (nucleobindin) increases agonist and thapsigargin releasable Ca2+ storage in the Golgi. J Cell Biol 145, 279-89. Lippe, R., Miaczynska, M., Rybin, V., Runge, A., and Zerial, M. (2001): Functional synergy between Rab5 effector Rabaptin-5 and exchange factor Rabex-5 when physically associated in a complex. Mol Biol Cell 12, 2219-28. Liu, T. T., Gomez, T. S., Sackey, B. K., Billadeau, D. D., and Burd, C. G. (2012): Rab GTPase regulation of retromer-mediated cargo export during endosome maturation. Mol Biol Cell 23, 2505-15. Llado, A., Timpson, P., Vila de Muga, S., Moreto, J., Pol, A., Grewal, T., Daly, R. J., Enrich, C., and Tebar, F. (2008): Protein kinase Cdelta and calmodulin regulate epidermal growth factor receptor recycling from early endosomes through Arp2/3 complex and cortactin. Mol Biol Cell 19, 17-29. Lobel, P., Dahms, N. M., and Kornfeld, S. (1988): Cloning and sequence analysis of the cation-independent mannose 6-phosphate receptor. J Biol Chem 263, 2563-70. Lodhi, I. J., Chiang, S. H., Chang, L., Vollenweider, D., Watson, R. T., Inoue, M., Pessin, J. E., and Saltiel, A. R. (2007): Gapex-5, a Rab31 guanine nucleotide exchange factor that regulates Glut4 trafficking in adipocytes. Cell Metab 5, 59-72. Lojewski, X., Staropoli, J. F., Biswas-Legrand, S., Simas, A. M., Haliw, L., Selig, M. K., Coppel, S. H., Goss, K. A., Petcherski, A., Chandrachud, U., Sheridan, S. D., Lucente, D., Sims, K. B., Gusella, J. F., Sondhi, D., Crystal, R. G., Reinhardt, P., Sterneckert, J., Scholer, H., Haggarty, S. J., Storch, A., Hermann, A., and Cotman, S. L. (2013): Human iPSC models of neuronal ceroid lipofuscinosis capture distinct effects of TPP1 and CLN3 mutations on the endocytic pathway. Hum Mol Genet 23, 2005-22. Lombardi, D., Soldati, T., Riederer, M. A., Goda, Y., Zerial, M., and Pfeffer, S. R. (1993): Rab9 functions in transport between late endosomes and the trans Golgi network. Embo J 12, 677-82. 214 Luiro, K., Kopra, O., Lehtovirta, M., and Jalanko, A. (2001): CLN3 protein is targeted to neuronal synapses but excluded from synaptic vesicles: new clues to Batten disease. Hum Mol Genet 10, 2123-31. Luiro, K., Yliannala, K., Ahtiainen, L., Maunu, H., Jarvela, I., Kyttala, A., and Jalanko, A. (2004): Interconnections of CLN3, Hook1 and Rab proteins link Batten disease to defects in the endocytic pathway. Hum Mol Genet 13, 3017-27. Lyly, A., Marjavaara, S. K., Kyttala, A., Uusi-Rauva, K., Luiro, K., Kopra, O., Martinez, L. O., Tanhuanpaa, K., Kalkkinen, N., Suomalainen, A., Jauhiainen, M., and Jalanko, A. (2008): Deficiency of the INCL protein Ppt1 results in changes in ectopic F1ATP synthase and altered cholesterol metabolism. Hum Mol Genet 17, 1406-17. Lyly, A., von Schantz, C., Heine, C., Schmiedt, M. L., Sipila, T., Jalanko, A., and Kyttala, A. (2009): Novel interactions of CLN5 support molecular networking between Neuronal Ceroid Lipofuscinosis proteins. BMC Cell Biol 10, 83. MacLeod, P. M., Dolman, C. L., Chang, E., Applegarth, D. A., and Bryant, B. (1976): The neuronal ceroid lipofuscinoses in British Columbia: a clinical epidemiologic and ultrastructural study. Birth Defects Orig Artic Ser 12, 289-96. Maejima, Y., Sedbazar, U., Suyama, S., Kohno, D., Onaka, T., Takano, E., Yoshida, N., Koike, M., Uchiyama, Y., Fujiwara, K., Yashiro, T., Horvath, T. L., Dietrich, M. O., Tanaka, S., Dezaki, K., Oh, I. S., Hashimoto, K., Shimizu, H., Nakata, M., Mori, M., and Yada, T. (2009): Nesfatin-1-regulated oxytocinergic signaling in the paraventricular nucleus causes anorexia through a leptin-independent melanocortin pathway. Cell Metab 10, 355-65. Mamo, A., Jules, F., Dumaresq-Doiron, K., Costantino, S., and Lefrancois, S. (2012): The role of Ceroid Lipofuscinosis Neuronal protein-5 (CLN5) in endosomal sorting. Mol Cell Biol. Mao, Q., Foster, B. J., Xia, H., and Davidson, B. L. (2003): Membrane topology of CLN3, the protein underlying Batten disease. FEBS Lett 541, 40-6. Marcusson, E. G., Horazdovsky, B. F., Cereghino, J. L., Gharakhanian, E., and Emr, S. D. (1994): The sorting receptor for yeast vacuolar carboxypeptidase Y is encoded by the VPS10 gene. Cell 77, 579-86. Mari, M., Bujny, M. V., Zeuschner, D., Geerts, W. J., Griffith, J., Petersen, C. M., Cullen, P. J., Klumperman, J., and Geuze, H. J. (2008): SNX1 defines an early endosomal recycling exit for sortilin and mannose 6-phosphate receptors. Traffic 9, 380-93. Markmann, S., Thelen, M., Cornils, K., Schweizer, M., Brocke-Ahmadinejad, N., Willnow, T., Heeren, J., Gieselmann, V., Braulke, T., and Kollmann, K. (2015): Lrp1/LDL Receptor Play Critical Roles in Mannose 6-Phosphate-Independent Lysosomal Enzyme Targeting. Traffic 16, 743-59. Marron-Terada, P. G., Hancock, M. K., Haskins, D. J., and Dahms, N. M. (2000): Recognition of Dictyostelium discoideum lysosomal enzymes is conferred by the amino-terminal carbohydrate binding site of the insulin-like growth factor II/mannose 6-phosphate receptor. Biochemistry 39, 2243-53. Martinez, L. O., Jacquet, S., Esteve, J. P., Rolland, C., Cabezon, E., Champagne, E., Pineau, T., Georgeaud, V., Walker, J. E., Terce, F., Collet, X., Perret, B., and Barbaras, R. (2003): Ectopic beta-chain of ATP synthase is an apolipoprotein A-I receptor in hepatic HDL endocytosis. Nature 421, 75-9. 215 Maruyama, K., Usami, M., Aizawa, T., and Yoshikawa, K. (1991): A novel brain-specific mRNA encoding nuclear protein (necdin) expressed in neurally differentiated embryonal carcinoma cells. Biochem Biophys Res Commun 178, 291-6. Marz, K. E., Lauer, J. M., and Hanson, P. I. (2003): Defining the SNARE complex binding surface of alpha-SNAP: implications for SNARE complex disassembly. J Biol Chem 278, 27000-8. Matsui, Y., Kikuchi, A., Araki, S., Hata, Y., Kondo, J., Teranishi, Y., and Takai, Y. (1990): Molecular cloning and characterization of a novel type of regulatory protein (GDI) for smg p25A, a ras p21-like GTP-binding protein. Mol Cell Biol 10, 4116-22. Mattson, M. P., and Chan, S. L. (2003): Neuronal and glial calcium signaling in Alzheimer's disease. Cell Calcium 34, 385-97. McBride, H. M., Rybin, V., Murphy, C., Giner, A., Teasdale, R., and Zerial, M. (1999): Oligomeric complexes link Rab5 effectors with NSF and drive membrane fusion via interactions between EEA1 and syntaxin 13. Cell 98, 377-86. McCormick, P. J., Dumaresq-Doiron, K., Pluviose, A. S., Pichette, V., Tosato, G., and Lefrancois, S. (2008): Palmitoylation controls recycling in lysosomal sorting and trafficking. Traffic 9, 1984-97. McGough, I. J., and Cullen, P. J. (2011): Recent advances in retromer biology. Traffic 12, 963-71. McNew, J. A., Parlati, F., Fukuda, R., Johnston, R. J., Paz, K., Paumet, F., Sollner, T. H., and Rothman, J. E. (2000): Compartmental specificity of cellular membrane fusion encoded in SNARE proteins. Nature 407, 153-9. Medina, D. L., Di Paola, S., Peluso, I., Armani, A., De Stefani, D., Venditti, R., Montefusco, S., Scotto-Rosato, A., Prezioso, C., Forrester, A., Settembre, C., Wang, W., Gao, Q., Xu, H., Sandri, M., Rizzuto, R., De Matteis, M. A., and Ballabio, A. (2015): Lysosomal calcium signalling regulates autophagy through calcineurin and TFEB. Nat Cell Biol 17, 288-99. Meikle, P. J., Hopwood, J. J., Clague, A. E., and Carey, W. F. (1999): Prevalence of lysosomal storage disorders. Jama 281, 249-54. Meikle, P. J., Ranieri, E., Simonsen, H., Rozaklis, T., Ramsay, S. L., Whitfield, P. D., Fuller, M., Christensen, E., Skovby, F., and Hopwood, J. J. (2004): Newborn screening for lysosomal storage disorders: clinical evaluation of a two-tier strategy. Pediatrics 114, 909-16. Merali, Z., Cayer, C., Kent, P., and Anisman, H. (2008): Nesfatin-1 increases anxiety- and fear-related behaviors in the rat. Psychopharmacology (Berl) 201, 115-23. Meresse, S., Ludwig, T., Frank, R., and Hoflack, B. (1990): Phosphorylation of the cytoplasmic domain of the bovine cation-independent mannose 6-phosphate receptor. Serines 2421 and 2492 are the targets of a casein kinase II associated to the Golgi-derived HAI adaptor complex. J Biol Chem 265, 18833-42. Mermelstein, C. S., Rebello, M. I., Amaral, L. M., and Costa, M. L. (2003): Changes in cell shape, cytoskeletal proteins and adhesion sites of cultured cells after extracellular Ca2+ chelation. Braz J Med Biol Res 36, 1111-6. Metcalf, D. J., Calvi, A. A., Seaman, M., Mitchison, H. M., and Cutler, D. F. (2008): Loss of the Batten disease gene CLN3 prevents exit from the TGN of the mannose 6phosphate receptor. Traffic 9, 1905-14. 216 Michalak, M., Burns, K., Andrin, C., Mesaeli, N., Jass, G. H., Busaan, J. L., and Opas, M. (1996): Endoplasmic reticulum form of calreticulin modulates glucocorticoidsensitive gene expression. J Biol Chem 271, 29436-45. Mitchison, H. M., Lim, M. J., and Cooper, J. D. (2004): Selectivity and types of cell death in the neuronal ceroid lipofuscinoses. Brain Pathol 14, 86-96. Miura, K., Hirai, M., Kanai, Y., and Kurosawa, Y. (1996): Organization of the human gene for nucleobindin (NUC) and its chromosomal assignment to 19q13.2-q13.4. Genomics 34, 181-6. Miura, K., Kurosawa, Y., and Kanai, Y. (1994): Calcium-binding activity of nucleobindin mediated by an EF hand moiety. Biochem Biophys Res Commun 199, 1388-93. Miura, K., Titani, K., Kurosawa, Y., and Kanai, Y. (1992): Molecular cloning of nucleobindin, a novel DNA-binding protein that contains both a signal peptide and a leucine zipper structure. Biochem Biophys Res Commun 187, 375-80. Mixon, M. B., Lee, E., Coleman, D. E., Berghuis, A. M., Gilman, A. G., and Sprang, S. R. (1995): Tertiary and quaternary structural changes in Gi alpha 1 induced by GTP hydrolysis. Science 270, 954-60. Mochizuki, N., Hibi, M., Kanai, Y., and Insel, P. A. (1995): Interaction of the protein nucleobindin with G alpha i2, as revealed by the yeast two-hybrid system. FEBS Lett 373, 155-8. Moharir, A., Peck, S. H., Budden, T., and Lee, S. Y. (2013): The role of N-glycosylation in folding, trafficking, and functionality of lysosomal protein CLN5. PLoS One 8, e74299. Mole, S., Williams, R., and Goebel, H. (2011): The neuronal ceroid lipofuscinoses (Batten disease). Oxford University Press. Moore, S. J., Buckley, D. J., MacMillan, A., Marshall, H. D., Steele, L., Ray, P. N., Nawaz, Z., Baskin, B., Frecker, M., Carr, S. M., Ives, E., and Parfrey, P. S. (2008): The clinical and genetic epidemiology of neuronal ceroid lipofuscinosis in Newfoundland. Clin Genet 74, 213-22. Morel-Huaux, V. M., Pypaert, M., Wouters, S., Tartakoff, A. M., Jurgan, U., Gevaert, K., and Courtoy, P. J. (2002): The calcium-binding protein p54/NEFA is a novel luminal resident of medial Golgi cisternae that traffics independently of mannosidase II. Eur J Cell Biol 81, 87-100. Morris, N. J., Ross, S. A., Lane, W. S., Moestrup, S. K., Petersen, C. M., Keller, S. R., and Lienhard, G. E. (1998): Sortilin is the major 110-kDa protein in GLUT4 vesicles from adipocytes. J Biol Chem 273, 3582-7. Muhammad, A., Flores, I., Zhang, H., Yu, R., Staniszewski, A., Planel, E., Herman, M., Ho, L., Kreber, R., Honig, L. S., Ganetzky, B., Duff, K., Arancio, O., and Small, S. A. (2008): Retromer deficiency observed in Alzheimer's disease causes hippocampal dysfunction, neurodegeneration, and Abeta accumulation. Proc Natl Acad Sci U S A 105, 7327-32. Mullins, C., and Bonifacino, J. S. (2001): The molecular machinery for lysosome biogenesis. Bioessays 23, 333-43. Munck Petersen, C., Nielsen, M. S., Jacobsen, C., Tauris, J., Jacobsen, L., Gliemann, J., Moestrup, S. K., and Madsen, P. (1999): Propeptide cleavage conditions sortilin/neurotensin receptor-3 for ligand binding. Embo J 18, 595-604. 217 Munroe, P. B., Mitchison, H. M., O'Rawe, A. M., Anderson, J. W., Boustany, R. M., Lerner, T. J., Taschner, P. E., de Vos, N., Breuning, M. H., Gardiner, R. M., and Mole, S. E. (1997): Spectrum of mutations in the Batten disease gene, CLN3. Am J Hum Genet 61, 310-6. Nakada-Tsukui, K., Saito-Nakano, Y., Ali, V., and Nozaki, T. (2005): A retromerlike complex is a novel Rab7 effector that is involved in the transport of the virulence factor cysteine protease in the enteric protozoan parasite Entamoeba histolytica. Mol Biol Cell 16, 5294-303. Nakayama, S., Moncrief, N. D., and Kretsinger, R. H. (1992): Evolution of EF-hand calcium-modulated proteins. II. Domains of several subfamilies have diverse evolutionary histories. J Mol Evol 34, 416-48. Narayan, S. B., Rakheja, D., Tan, L., Pastor, J. V., and Bennett, M. J. (2006): CLN3P, the Batten's disease protein, is a novel palmitoyl-protein Delta-9 desaturase. Ann Neurol 60, 570-7. Narayan, S. B., Tan, L., and Bennett, M. J. (2008): Intermediate levels of neuronal palmitoyl-protein Delta-9 desaturase in heterozygotes for murine Batten disease. Mol Genet Metab 93, 89-91. Nash, P. D., Opas, M., and Michalak, M. (1994): Calreticulin: not just another calciumbinding protein. Mol Cell Biochem 135, 71-8. Nesselhut, J., Jurgan, U., Onken, E., Gotz, H., Barnikol, H. U., Hirschfeld, G., BarnikolWatanabe, S., and Hilschmann, N. (2001): Golgi retention of human protein NEFA is mediated by its N-terminal Leu/Ile-rich region. FEBS Lett 509, 469-75. Ni, X., and Morales, C. R. (2006): The lysosomal trafficking of acid sphingomyelinase is mediated by sortilin and mannose 6-phosphate receptor. Traffic 7, 889-902. Nielsen, M. S., Gustafsen, C., Madsen, P., Nyengaard, J. R., Hermey, G., Bakke, O., Mari, M., Schu, P., Pohlmann, R., Dennes, A., and Petersen, C. M. (2007a): Sorting by the cytoplasmic domain of the amyloid precursor protein binding receptor SorLA. Mol Cell Biol 27, 6842-51. Nielsen, M. S., Jacobsen, C., Olivecrona, G., Gliemann, J., and Petersen, C. M. (1999): Sortilin/neurotensin receptor-3 binds and mediates degradation of lipoprotein lipase. J Biol Chem 274, 8832-6. Nielsen, M. S., Madsen, P., Christensen, E. I., Nykjaer, A., Gliemann, J., Kasper, D., Pohlmann, R., and Petersen, C. M. (2001): The sortilin cytoplasmic tail conveys Golgi-endosome transport and binds the VHS domain of the GGA2 sorting protein. Embo J 20, 2180-90. Nielsen, R., Courtoy, P. J., Jacobsen, C., Dom, G., Lima, W. R., Jadot, M., Willnow, T. E., Devuyst, O., and Christensen, E. I. (2007b): Endocytosis provides a major alternative pathway for lysosomal biogenesis in kidney proximal tubular cells. Proc Natl Acad Sci U S A 104, 5407-12. Niinobe, M., Koyama, K., and Yoshikawa, K. (2000): Cellular and subcellular localization of necdin in fetal and adult mouse brain. Dev Neurosci 22, 310-9. Niphakis, M. J., Lum, K. M., Cognetta, A. B., 3rd, Correia, B. E., Ichu, T. A., Olucha, J., Brown, S. J., Kundu, S., Piscitelli, F., Rosen, H., and Cravatt, B. F. (2015): A Global Map of Lipid-Binding Proteins and Their Ligandability in Cells. Cell 161, 1668-80. 218 Norwood, S. J., Shaw, D. J., Cowieson, N. P., Owen, D. J., Teasdale, R. D., and Collins, B. M. (2010): Assembly and solution structure of the core retromer protein complex. Traffic 12, 56-71. Nothwehr, S. F., Ha, S. A., and Bruinsma, P. (2000): Sorting of yeast membrane proteins into an endosome-to-Golgi pathway involves direct interaction of their cytosolic domains with Vps35p. J Cell Biol 151, 297-310. Nugent, T., Mole, S. E., and Jones, D. T. (2008): The transmembrane topology of Batten disease protein CLN3 determined by consensus computational prediction constrained by experimental data. FEBS Lett 582, 1019-24. Nykjaer, A., Lee, R., Teng, K. K., Jansen, P., Madsen, P., Nielsen, M. S., Jacobsen, C., Kliemannel, M., Schwarz, E., Willnow, T. E., Hempstead, B. L., and Petersen, C. M. (2004): Sortilin is essential for proNGF-induced neuronal cell death. Nature 427, 843-8. Oh, I. S., Shimizu, H., Satoh, T., Okada, S., Adachi, S., Inoue, K., Eguchi, H., Yamamoto, M., Imaki, T., Hashimoto, K., Tsuchiya, T., Monden, T., Horiguchi, K., Yamada, M., and Mori, M. (2006): Identification of nesfatin-1 as a satiety molecule in the hypothalamus. Nature 443, 709-12. Olsnes, S., and Pihl, A. (1972): Ricin - a potent inhibitor of protein synthesis. FEBS Lett 20, 327-329. Olson, L. J., Castonguay, A. C., Lasanajak, Y., Peterson, F. C., Cummings, R. D., Smith, D. F., and Dahms, N. M. (2015): Identification of a fourth mannose 6-phosphate binding site in the cation-independent mannose 6-phosphate receptor. Glycobiology. Orsel, J. G., Sincock, P. M., Krise, J. P., and Pfeffer, S. R. (2000): Recognition of the 300kDa mannose 6-phosphate receptor cytoplasmic domain by 47-kDa tail-interacting protein. Proc Natl Acad Sci U S A 97, 9047-51. Osei-Hyiaman, N., Encinas (2011): Fasting co-supresses Nesfatin-1 and GPR12 in Mouse Hypothlamic Appetite Center: Implications for Energy Metabolism. Endocrine Reviews 32, 2-300. Padilla-Lopez, S., and Pearce, D. A. (2006): Saccharomyces cerevisiae lacking Btn1p modulate vacuolar ATPase activity to regulate pH imbalance in the vacuole. J Biol Chem 281, 10273-80. Palmer, D. N., Fearnley, I. M., Medd, S. M., Walker, J. E., Martinus, R. D., Bayliss, S. L., Hall, N. A., Lake, B. D., Wolfe, L. S., and Jolly, R. D. (1989): Lysosomal storage of the DCCD reactive proteolipid subunit of mitochondrial ATP synthase in human and ovine ceroid lipofuscinoses. Adv Exp Med Biol 266, 211-22; discussion 223. Palmieri, M., Impey, S., Kang, H., di Ronza, A., Pelz, C., Sardiello, M., and Ballabio, A. (2011): Characterization of the CLEAR network reveals an integrated control of cellular clearance pathways. Hum Mol Genet 20, 3852-66. Pan, X., Eathiraj, S., Munson, M., and Lambright, D. G. (2006): TBC-domain GAPs for Rab GTPases accelerate GTP hydrolysis by a dual-finger mechanism. Nature 442, 303-6. Pankiv, S., Alemu, E. A., Brech, A., Bruun, J. A., Lamark, T., Overvatn, A., Bjorkoy, G., and Johansen, T. (2010): FYCO1 is a Rab7 effector that binds to LC3 and PI3P to mediate microtubule plus end-directed vesicle transport. J Cell Biol 188, 253-69. Pearce, D. A., Ferea, T., Nosel, S. A., Das, B., and Sherman, F. (1999a): Action of BTN1, the yeast orthologue of the gene mutated in Batten disease. Nat Genet 22, 55-8. 219 Pearce, D. A., Nosel, S. A., and Sherman, F. (1999b): Studies of pH regulation by Btn1p, the yeast homolog of human Cln3p. Mol Genet Metab 66, 320-3. Pearce, D. A., and Sherman, F. (1997): BTN1, a yeast gene corresponding to the human gene responsible for Batten's disease, is not essential for viability, mitochondrial function, or degradation of mitochondrial ATP synthase. Yeast 13, 691-7. Pereira-Leal, J. B., and Seabra, M. C. (2001): Evolution of the Rab family of small GTPbinding proteins. J Mol Biol 313, 889-901. Persaud-Sawin, D. A., and Boustany, R. M. (2005): Cell death pathways in juvenile Batten disease. Apoptosis 10, 973-85. Persaud-Sawin, D. A., McNamara, J. O., 2nd, Rylova, S., Vandongen, A., and Boustany, R. M. (2004): A galactosylceramide binding domain is involved in trafficking of CLN3 from Golgi to rafts via recycling endosomes. Pediatr Res 56, 449-63. Peter, B. J., Kent, H. M., Mills, I. G., Vallis, Y., Butler, P. J., Evans, P. R., and McMahon, H. T. (2004): BAR domains as sensors of membrane curvature: the amphiphysin BAR structure. Science 303, 495-9. Peter, F., Nuoffer, C., Pind, S. N., and Balch, W. E. (1994): Guanine nucleotide dissociation inhibitor is essential for Rab1 function in budding from the endoplasmic reticulum and transport through the Golgi stack. J Cell Biol 126, 1393406. Petersen, C. M., Nielsen, M. S., Nykjaer, A., Jacobsen, L., Tommerup, N., Rasmussen, H. H., Roigaard, H., Gliemann, J., Madsen, P., and Moestrup, S. K. (1997): Molecular identification of a novel candidate sorting receptor purified from human brain by receptor-associated protein affinity chromatography. J Biol Chem 272, 3599-605. Petersson, U., Somogyi, E., Reinholt, F. P., Karlsson, T., Sugars, R. V., and Wendel, M. (2004): Nucleobindin is produced by bone cells and secreted into the osteoid, with a potential role as a modulator of matrix maturation. Bone 34, 949-60. Platt, F. M., Boland, B., and van der Spoel, A. C. (2012): The cell biology of disease: lysosomal storage disorders: the cellular impact of lysosomal dysfunction. J Cell Biol 199, 723-34. Plemel, R. L., Lobingier, B. T., Brett, C. L., Angers, C. G., Nickerson, D. P., Paulsel, A., Sprague, D., and Merz, A. J. (2011): Subunit organization and Rab interactions of Vps-C protein complexes that control endolysosomal membrane traffic. Mol Biol Cell 22, 1353-63. Pocha, S. M., Wassmer, T., Niehage, C., Hoflack, B., and Knust, E. (2011): Retromer controls epithelial cell polarity by trafficking the apical determinant Crumbs. Curr Biol 21, 1111-7. Porta, E. A. (1991): Advances in age pigment research. Arch Gerontol Geriatr 12, 303-20. Porta, E. A. (2002): Pigments in aging: an overview. Ann N Y Acad Sci 959, 57-65. Poteryaev, D., Datta, S., Ackema, K., Zerial, M., and Spang, A. (2010): Identification of the switch in early-to-late endosome transition. Cell 141, 497-508. Praefcke, G. J., and McMahon, H. T. (2004): The dynamin superfamily: universal membrane tubulation and fission molecules? Nat Rev Mol Cell Biol 5, 133-47. Prasad, K., Barouch, W., Greene, L., and Eisenberg, E. (1993): A protein cofactor is required for uncoating of clathrin baskets by uncoating ATPase. J Biol Chem 268, 23758-61. 220 Pullarkat, R. K., and Morris, G. N. (1997): Farnesylation of Batten disease CLN3 protein. Neuropediatrics 28, 42-4. Puranam, K. L., Guo, W. X., Qian, W. H., Nikbakht, K., and Boustany, R. M. (1999): CLN3 defines a novel antiapoptotic pathway operative in neurodegeneration and mediated by ceramide. Mol Genet Metab 66, 294-308. Puthenveedu, M. A., Lauffer, B., Temkin, P., Vistein, R., Carlton, P., Thorn, K., Taunton, J., Weiner, O. D., Parton, R. G., and von Zastrow, M. (2010): Sequence-dependent sorting of recycling proteins by actin-stabilized endosomal microdomains. Cell 143, 761-73. Qian, M., Sleat, D. E., Zheng, H., Moore, D., and Lobel, P. (2008): Proteomics analysis of serum from mutant mice reveals lysosomal proteins selectively transported by each of the two mannose 6-phosphate receptors. Mol Cell Proteomics 7, 58-70. Quistgaard, E. M., Madsen, P., Groftehauge, M. K., Nissen, P., Petersen, C. M., and Thirup, S. S. (2009): Ligands bind to Sortilin in the tunnel of a ten-bladed betapropeller domain. Nat Struct Mol Biol 16, 96-8. Ramanjaneya, M., Chen, J., Brown, J. E., Tripathi, G., Hallschmid, M., Patel, S., Kern, W., Hillhouse, E. W., Lehnert, H., Tan, B. K., and Randeva, H. S. (2010): Identification of nesfatin-1 in human and murine adipose tissue: a novel depot-specific adipokine with increased levels in obesity. Endocrinology 151, 3169-80. Ramesh, N., Mohan, H., and Unniappan, S. (2015): Nucleobindin-1 encodes a nesfatin-1like peptide that stimulates insulin secretion. Gen Comp Endocrinol. Ramirez-Montealegre, D., and Pearce, D. A. (2005): Defective lysosomal arginine transport in juvenile Batten disease. Hum Mol Genet 14, 3759-73. Rawlings, N. D., and Barrett, A. J. (1995): Families of aspartic peptidases, and those of unknown catalytic mechanism. Methods Enzymol 248, 105-20. Ren, J., Wen, L., Gao, X., Jin, C., Xue, Y., and Yao, X. (2009): DOG 1.0: illustrator of protein domain structures. Cell Res 19, 271-3. Rider, J. A., and Rider, D. L. (1988): Batten disease: past, present, and future. Am J Med Genet Suppl 5, 21-6. Riederer, M. A., Soldati, T., Shapiro, A. D., Lin, J., and Pfeffer, S. R. (1994): Lysosome biogenesis requires Rab9 function and receptor recycling from endosomes to the trans-Golgi network. J Cell Biol 125, 573-82. Risselada, H. J., Kutzner, C., and Grubmuller, H. (2011): Caught in the act: visualization of SNARE-mediated fusion events in molecular detail. Chembiochem 12, 1049-55. Roberts, D. L., Weix, D. J., Dahms, N. M., and Kim, J. J. (1998): Molecular basis of lysosomal enzyme recognition: three-dimensional structure of the cation-dependent mannose 6-phosphate receptor. Cell 93, 639-48. Rohrer, J., and Kornfeld, R. (2001): Lysosomal hydrolase mannose 6-phosphate uncovering enzyme resides in the trans-Golgi network. Mol Biol Cell 12, 1623-31. Rojas, A. M., Fuentes, G., Rausell, A., and Valencia, A. (2012): The Ras protein superfamily: evolutionary tree and role of conserved amino acids. J Cell Biol 196, 189-201. Rojas, R., Kametaka, S., Haft, C. R., and Bonifacino, J. S. (2007): Interchangeable but essential functions of SNX1 and SNX2 in the association of retromer with endosomes and the trafficking of mannose 6-phosphate receptors. Mol Cell Biol 27, 1112-24. 221 Rojas, R., van Vlijmen, T., Mardones, G. A., Prabhu, Y., Rojas, A. L., Mohammed, S., Heck, A. J., Raposo, G., van der Sluijs, P., and Bonifacino, J. S. (2008): Regulation of retromer recruitment to endosomes by sequential action of Rab5 and Rab7. J Cell Biol 183, 513-26. Rosenthal, J. A., Chen, H., Slepnev, V. I., Pellegrini, L., Salcini, A. E., Di Fiore, P. P., and De Camilli, P. (1999): The epsins define a family of proteins that interact with components of the clathrin coat and contain a new protein module. J Biol Chem 274, 33959-65. Rosorius, O., Mieskes, G., Issinger, O. G., Korner, C., Schmidt, B., von Figura, K., and Braulke, T. (1993): Characterization of phosphorylation sites in the cytoplasmic domain of the 300 kDa mannose-6-phosphate receptor. Biochem J 292 ( Pt 3), 8338. Rotty, J. D., Wu, C., and Bear, J. E. (2013): New insights into the regulation and cellular functions of the ARP2/3 complex. Nat Rev Mol Cell Biol 14, 7-12. Roux, A., Cuvelier, D., Nassoy, P., Prost, J., Bassereau, P., and Goud, B. (2005): Role of curvature and phase transition in lipid sorting and fission of membrane tubules. Embo J 24, 1537-45. Rusyn, E., Mousallem, T., Persaud-Sawin, D. A., Miller, S., and Boustany, R. M. (2008): CLN3p impacts galactosylceramide transport, raft morphology, and lipid content. Pediatr Res 63, 625-31. Saadipour, K., Yang, M., Lim, Y., Georgiou, K., Sun, Y., Keating, D., Liu, J., Wang, Y. R., Gai, W. P., Zhong, J. H., Wang, Y. J., and Zhou, X. F. (2013): Amyloid beta(1)()(4)(2) (Abeta(4)(2)) up-regulates the expression of sortilin via the p75(NTR)/RhoA signaling pathway. J Neurochem 127, 152-62. Saftig, P., and Klumperman, J. (2009): Lysosome biogenesis and lysosomal membrane proteins: trafficking meets function. Nat Rev Mol Cell Biol 10, 623-35. Sagne, C., and Gasnier, B. (2008): Molecular physiology and pathophysiology of lysosomal membrane transporters. J Inherit Metab Dis 31, 258-66. Sancak, Y., Bar-Peled, L., Zoncu, R., Markhard, A. L., Nada, S., and Sabatini, D. M. (2010): Ragulator-Rag complex targets mTORC1 to the lysosomal surface and is necessary for its activation by amino acids. Cell 141, 290-303. Sandvig, K., Garred, O., Prydz, K., Kozlov, J. V., Hansen, S. H., and van Deurs, B. (1992): Retrograde transport of endocytosed Shiga toxin to the endoplasmic reticulum. Nature 358, 510-2. Santavuori, P. (1988): Neuronal ceroid-lipofuscinoses in childhood. Brain Dev 10, 80-3. Sardiello, M., Palmieri, M., di Ronza, A., Medina, D. L., Valenza, M., Gennarino, V. A., Di Malta, C., Donaudy, F., Embrione, V., Polishchuk, R. S., Banfi, S., Parenti, G., Cattaneo, E., and Ballabio, A. (2009): A gene network regulating lysosomal biogenesis and function. Science 325, 473-7. Sarret, P., Krzywkowski, P., Segal, L., Nielsen, M. S., Petersen, C. M., Mazella, J., Stroh, T., and Beaudet, A. (2003): Distribution of NTS3 receptor/sortilin mRNA and protein in the rat central nervous system. J Comp Neurol 461, 483-505. Savukoski, M., Klockars, T., Holmberg, V., Santavuori, P., Lander, E. S., and Peltonen, L. (1998): CLN5, a novel gene encoding a putative transmembrane protein mutated in Finnish variant late infantile neuronal ceroid lipofuscinosis. Nat Genet 19, 286-8. 222 Scherer, P. E., Lederkremer, G. Z., Williams, S., Fogliano, M., Baldini, G., and Lodish, H. F. (1996): Cab45, a novel (Ca2+)-binding protein localized to the Golgi lumen. J Cell Biol 133, 257-68. Schlossman, D. M., Schmid, S. L., Braell, W. A., and Rothman, J. E. (1984): An enzyme that removes clathrin coats: purification of an uncoating ATPase. J Cell Biol 99, 723-33. Schmidt, B., Kiecke-Siemsen, C., Waheed, A., Braulke, T., and von Figura, K. (1995): Localization of the insulin-like growth factor II binding site to amino acids 15081566 in repeat 11 of the mannose 6-phosphate/insulin-like growth factor II receptor. J Biol Chem 270, 14975-82. Schmiedt, M. L. (2012): The CLN5 disease - protein maturation, trafficking and pathology: Medical Faculty, University of Helsinki. Schmiedt, M. L., Bessa, C., Heine, C., Ribeiro, M. G., Jalanko, A., and Kyttala, A. (2010): The neuronal ceroid lipofuscinosis protein CLN5: new insights into cellular maturation, transport, and consequences of mutations. Hum Mutat 31, 356-65. Schmiedt, M. L., Blom, T., Kopra, O., Wong, A., von Schantz-Fant, C., Ikonen, E., Kuronen, M., Jauhiainen, M., Cooper, J. D., and Jalanko, A. (2012): Cln5deficiency in mice leads to microglial activation, defective myelination and changes in lipid metabolism. Neurobiol Dis 46, 19-29. Schu, P. V., Takegawa, K., Fry, M. J., Stack, J. H., Waterfield, M. D., and Emr, S. D. (1993): Phosphatidylinositol 3-kinase encoded by yeast VPS34 gene essential for protein sorting. Science 260, 88-91. Schultz, M. L., Tecedor, L., Chang, M., and Davidson, B. L. (2011): Clarifying lysosomal storage diseases. Trends Neurosci. Schweizer, A., Kornfeld, S., and Rohrer, J. (1996): Cysteine34 of the cytoplasmic tail of the cation-dependent mannose 6-phosphate receptor is reversibly palmitoylated and required for normal trafficking and lysosomal enzyme sorting. J Cell Biol 132, 57784. Schweizer, A., Kornfeld, S., and Rohrer, J. (1997): Proper sorting of the cation-dependent mannose 6-phosphate receptor in endosomes depends on a pair of aromatic amino acids in its cytoplasmic tail. Proc Natl Acad Sci U S A 94, 14471-6. Seabra, M. C., Goldstein, J. L., Sudhof, T. C., and Brown, M. S. (1992): Rab geranylgeranyl transferase. A multisubunit enzyme that prenylates GTP-binding proteins terminating in Cys-X-Cys or Cys-Cys. J Biol Chem 267, 14497-503. Seaman, M. N. (2004): Cargo-selective endosomal sorting for retrieval to the Golgi requires retromer. J Cell Biol 165, 111-22. Seaman, M. N. (2007): Identification of a novel conserved sorting motif required for retromer-mediated endosome-to-TGN retrieval. J Cell Sci 120, 2378-89. Seaman, M. N. (2012): The retromer complex - endosomal protein recycling and beyond. J Cell Sci 125, 4693-702. Seaman, M. N., Harbour, M. E., Tattersall, D., Read, E., and Bright, N. (2009): Membrane recruitment of the cargo-selective retromer subcomplex is catalysed by the small GTPase Rab7 and inhibited by the Rab-GAP TBC1D5. J Cell Sci 122, 2371-82. Seaman, M. N., Marcusson, E. G., Cereghino, J. L., and Emr, S. D. (1997): Endosome to Golgi retrieval of the vacuolar protein sorting receptor, Vps10p, requires the function of the VPS29, VPS30, and VPS35 gene products. J Cell Biol 137, 79-92. 223 Seaman, M. N., McCaffery, J. M., and Emr, S. D. (1998): A membrane coat complex essential for endosome-to-Golgi retrograde transport in yeast. J Cell Biol 142, 66581. Seidah, N. G., and Chretien, M. (1999): Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides. Brain Res 848, 45-62. Semerdjieva, S., Shortt, B., Maxwell, E., Singh, S., Fonarev, P., Hansen, J., Schiavo, G., Grant, B. D., and Smythe, E. (2008): Coordinated regulation of AP2 uncoating from clathrin-coated vesicles by rab5 and hRME-6. J Cell Biol 183, 499-511. Sharpe, H. J., Stevens, T. J., and Munro, S. (2010): A comprehensive comparison of transmembrane domains reveals organelle-specific properties. Cell 142, 158-69. Shimizu, H., Oh, I. S., Okada, S., and Mori, M. (2009): Nesfatin-1: an overview and future clinical application. Endocr J 56, 537-43. Silva, B., Adams, J., and Lee, S. Y. (2015): Proteolytic processing of the neuronal ceroid lipofuscinosis related lysosomal protein CLN5. Exp Cell Res. Sincock, P. M., Ganley, I. G., Krise, J. P., Diederichs, S., Sivars, U., O'Connor, B., Ding, L., and Pfeffer, S. R. (2003): Self-assembly is important for TIP47 function in mannose 6-phosphate receptor transport. Traffic 4, 18-25. Sivars, U., Aivazian, D., and Pfeffer, S. R. (2003): Yip3 catalyses the dissociation of endosomal Rab-GDI complexes. Nature 425, 856-9. Skanland, S. S., Walchli, S., Brech, A., and Sandvig, K. (2009): SNX4 in complex with clathrin and dynein: implications for endosome movement. PLoS One 4, e5935. Sleat, D. E., Sohar, I., Pullarkat, P. S., Lobel, P., and Pullarkat, R. K. (1998): Specific alterations in levels of mannose 6-phosphorylated glycoproteins in different neuronal ceroid lipofuscinoses. Biochem J 334 ( Pt 3), 547-51. Sleat, D. E., Zheng, H., Qian, M., and Lobel, P. (2006): Identification of sites of mannose 6-phosphorylation on lysosomal proteins. Mol Cell Proteomics 5, 686-701. Small, S. A., Kent, K., Pierce, A., Leung, C., Kang, M. S., Okada, H., Honig, L., Vonsattel, J. P., and Kim, T. W. (2005): Model-guided microarray implicates the retromer complex in Alzheimer's disease. Ann Neurol 58, 909-19. Sollner, T., Bennett, M. K., Whiteheart, S. W., Scheller, R. H., and Rothman, J. E. (1993): A protein assembly-disassembly pathway in vitro that may correspond to sequential steps of synaptic vesicle docking, activation, and fusion. Cell 75, 409-18. Somogyi, E., Petersson, U., Hultenby, K., and Wendel, M. (2003): Calreticulin--an endoplasmic reticulum protein with calcium-binding activity is also found in the extracellular matrix. Matrix Biol 22, 179-91. Somogyi, E., Petersson, U., Sugars, R. V., Hultenby, K., and Wendel, M. (2004): Nucleobindin--a Ca2+-binding protein present in the cells and mineralized tissues of the tooth. Calcif Tissue Int 74, 366-76. Spielmeyer, W. (1905a): über familiäre amaurotische Idioten. Neurol Cbl 24, 620-621. Spielmeyer, W. (1905b): Weitere Mitteilungen über eine besondere Form von familiärer amaurotischer Idiotie. Neuro Cbl 24, 1131-1132. Sprang, S. R. (1997): G protein mechanisms: insights from structural analysis. Annu Rev Biochem 66, 639-78. St-Arnaud, R., Prud'homme, J., Leung-Hagesteijn, C., and Dedhar, S. (1995): Constitutive expression of calreticulin in osteoblasts inhibits mineralization. J Cell Biol 131, 1351-9. 224 Staub, O., and Rotin, D. (2006): Role of ubiquitylation in cellular membrane transport. Physiol Rev 86, 669-707. Stein, C. S., Yancey, P. H., Martins, I., Sigmund, R. D., Stokes, J. B., and Davidson, B. L. (2010): Osmoregulation of ceroid neuronal lipofuscinosis type 3 in the renal medulla. Am J Physiol Cell Physiol 298, C1388-400. Stein, M. P., Feng, Y., Cooper, K. L., Welford, A. M., and Wandinger-Ness, A. (2003): Human VPS34 and p150 are Rab7 interacting partners. Traffic 4, 754-71. Steinberg, F., Gallon, M., Winfield, M., Thomas, E. C., Bell, A. J., Heesom, K. J., Tavare, J. M., and Cullen, P. J. (2013): A global analysis of SNX27-retromer assembly and cargo specificity reveals a function in glucose and metal ion transport. Nat Cell Biol 15, 461-71. Stengel, A., Goebel, M., and Tache, Y. (2010): Nesfatin-1: a novel inhibitory regulator of food intake and body weight. Obes Rev 12, 261-71. Stengel, A., Goebel, M., Yakubov, I., Wang, L., Witcher, D., Coskun, T., Tache, Y., Sachs, G., and Lambrecht, N. W. (2009): Identification and characterization of nesfatin-1 immunoreactivity in endocrine cell types of the rat gastric oxyntic mucosa. Endocrinology 150, 232-8. Stengel, O. (1826a): Beretning om et maerkeligt Sygdomstilfaelde hos fire Sodskende i Naerheden af Roraas. Eyr Medicinsk Tidskrift (Christiana) 1, 347-352. Stenmark, H. (2009): Rab GTPases as coordinators of vesicle traffic. Nat Rev Mol Cell Biol 10, 513-25. Storch, S., Pohl, S., and Braulke, T. (2004): A dileucine motif and a cluster of acidic amino acids in the second cytoplasmic domain of the batten disease-related CLN3 protein are required for efficient lysosomal targeting. J Biol Chem 279, 53625-34. Storch, S., Pohl, S., Quitsch, A., Falley, K., and Braulke, T. (2007): C-terminal prenylation of the CLN3 membrane glycoprotein is required for efficient endosomal sorting to lysosomes. Traffic 8, 431-44. Stow, J. L., de Almeida, J. B., Narula, N., Holtzman, E. J., Ercolani, L., and Ausiello, D. A. (1991): A heterotrimeric G protein, G alpha i-3, on Golgi membranes regulates the secretion of a heparan sulfate proteoglycan in LLC-PK1 epithelial cells. J Cell Biol 114, 1113-24. Su, Y., Zhang, J., Tang, Y., Bi, F., and Liu, J. N. (2010): The novel function of nesfatin-1: anti-hyperglycemia. Biochem Biophys Res Commun 391, 1039-42. Sugiyama, T., Kumagai, H., Morikawa, Y., Wada, Y., Sugiyama, A., Yasuda, K., Yokoi, N., Tamura, S., Kojima, T., Nosaka, T., Senba, E., Kimura, S., Kadowaki, T., Kodama, T., and Kitamura, T. (2000): A novel low-density lipoprotein receptorrelated protein mediating cellular uptake of apolipoprotein E-enriched beta-VLDL in vitro. Biochemistry 39, 15817-25. Sunahara, R. K., Tesmer, J. J., Gilman, A. G., and Sprang, S. R. (1997): Crystal structure of the adenylyl cyclase activator Gsalpha. Science 278, 1943-7. Sutton, R. B., Fasshauer, D., Jahn, R., and Brunger, A. T. (1998): Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4 A resolution. Nature 395, 347-53. Swarbrick, J. D., Shaw, D. J., Chhabra, S., Ghai, R., Valkov, E., Norwood, S. J., Seaman, M. N., and Collins, B. M. (2011): VPS29 is not an active metallo-phosphatase but is 225 a rigid scaffold required for retromer interaction with accessory proteins. PLoS One 6, e20420. Tabuchi, M., Yanatori, I., Kawai, Y., and Kishi, F. (2010): Retromer-mediated direct sorting is required for proper endosomal recycling of the mammalian iron transporter DMT1. J Cell Sci 123, 756-66. Taniguchi, N., Taniura, H., Niinobe, M., Takayama, C., Tominaga-Yoshino, K., Ogura, A., and Yoshikawa, K. (2000): The postmitotic growth suppressor necdin interacts with a calcium-binding protein (NEFA) in neuronal cytoplasm. J Biol Chem 275, 3167481. Taniura, H., Matsumoto, K., and Yoshikawa, K. (1999): Physical and functional interactions of neuronal growth suppressor necdin with p53. J Biol Chem 274, 16242-8. Taylor, R. S., Jones, S. M., Dahl, R. H., Nordeen, M. H., and Howell, K. E. (1997): Characterization of the Golgi complex cleared of proteins in transit and examination of calcium uptake activities. Mol Biol Cell 8, 1911-31. Tebar, F., Llado, A., and Enrich, C. (2002a): Role of calmodulin in the modulation of the MAPK signalling pathway and the transactivation of epidermal growth factor receptor mediated by PKC. FEBS Lett 517, 206-10. Tebar, F., Villalonga, P., Sorkina, T., Agell, N., Sorkin, A., and Enrich, C. (2002b): Calmodulin regulates intracellular trafficking of epidermal growth factor receptor and the MAPK signaling pathway. Mol Biol Cell 13, 2057-68. Temkin, P., Lauffer, B., Jager, S., Cimermancic, P., Krogan, N. J., and von Zastrow, M. (2011): SNX27 mediates retromer tubule entry and endosome-to-plasma membrane trafficking of signalling receptors. Nat Cell Biol 13, 715-21. Teng, H. K., Teng, K. K., Lee, R., Wright, S., Tevar, S., Almeida, R. D., Kermani, P., Torkin, R., Chen, Z. Y., Lee, F. S., Kraemer, R. T., Nykjaer, A., and Hempstead, B. L. (2005): ProBDNF induces neuronal apoptosis via activation of a receptor complex of p75NTR and sortilin. J Neurosci 25, 5455-63. Terman, A., and Brunk, U. T. (1998): Lipofuscin: mechanisms of formation and increase with age. Apmis 106, 265-76. The International Batten Disease Consortium (1995): Isolation of a novel gene underlying Batten disease, CLN3. The International Batten Disease Consortium. Cell 82, 94957. Tsukumo, Y., Tomida, A., Kitahara, O., Nakamura, Y., Asada, S., Mori, K., and Tsuruo, T. (2007): Nucleobindin 1 controls the unfolded protein response by inhibiting ATF6 activation. J Biol Chem 282, 29264-72. Tsukumo, Y., Tsukahara, S., Saito, S., Tsuruo, T., and Tomida, A. (2009): A novel endoplasmic reticulum export signal: proline at the +2-position from the signal peptide cleavage site. J Biol Chem 284, 27500-10. Tuxworth, R. I., Vivancos, V., O'Hare, M. B., and Tear, G. (2009): Interactions between the juvenile Batten disease gene, CLN3, and the Notch and JNK signalling pathways. Hum Mol Genet 18, 667-78. Tyynela, J., Palmer, D. N., Baumann, M., and Haltia, M. (1993): Storage of saposins A and D in infantile neuronal ceroid-lipofuscinosis. FEBS Lett 330, 8-12. 226 Tyynela, J., Suopanki, J., Santavuori, P., Baumann, M., and Haltia, M. (1997): Variant late infantile neuronal ceroid-lipofuscinosis: pathology and biochemistry. J Neuropathol Exp Neurol 56, 369-75. Ullrich, O., Horiuchi, H., Bucci, C., and Zerial, M. (1994): Membrane association of Rab5 mediated by GDP-dissociation inhibitor and accompanied by GDP/GTP exchange. Nature 368, 157-60. Uusi-Rauva, K., Kyttala, A., van der Kant, R., Vesa, J., Tanhuanpaa, K., Neefjes, J., Olkkonen, V. M., and Jalanko, A. (2012): Neuronal ceroid lipofuscinosis protein CLN3 interacts with motor proteins and modifies location of late endosomal compartments. Cell Mol Life Sci. Uusi-Rauva, K., Luiro, K., Tanhuanpaa, K., Kopra, O., Martin-Vasallo, P., Kyttala, A., and Jalanko, A. (2008): Novel interactions of CLN3 protein link Batten disease to dysregulation of fodrin-Na+, K+ ATPase complex. Exp Cell Res 314, 2895-905. Uvebrant, P., and Hagberg, B. (1997): Neuronal ceroid lipofuscinoses in Scandinavia. Epidemiology and clinical pictures. Neuropediatrics 28, 6-8. Valencia, C. A., Cotten, S. W., Duan, J., and Liu, R. (2008): Modulation of nucleobindin-1 and nucleobindin-2 by caspases. FEBS Lett 582, 286-90. van Weering, J. R., Verkade, P., and Cullen, P. J. (2012): SNX-BAR-mediated endosome tubulation is co-ordinated with endosome maturation. Traffic 13, 94-107. Vesa, J., Chin, M. H., Oelgeschlager, K., Isosomppi, J., DellAngelica, E. C., Jalanko, A., and Peltonen, L. (2002): Neuronal ceroid lipofuscinoses are connected at molecular level: interaction of CLN5 protein with CLN2 and CLN3. Mol Biol Cell 13, 241020. Vesa, J., Hellsten, E., Verkruyse, L. A., Camp, L. A., Rapola, J., Santavuori, P., Hofmann, S. L., and Peltonen, L. (1995): Mutations in the palmitoyl protein thioesterase gene causing infantile neuronal ceroid lipofuscinosis. Nature 376, 584-7. Vitiello, S. P., Benedict, J. W., Padilla-Lopez, S., and Pearce, D. A. (2010): Interaction between Sdo1p and Btn1p in the Saccharomyces cerevisiae model for Batten disease. Hum Mol Genet 19, 931-42. Vitner, E. B., Platt, F. M., and Futerman, A. H. (2010): Common and uncommon pathogenic cascades in lysosomal storage diseases. J Biol Chem 285, 20423-7. von Blume, J., Alleaume, A. M., Kienzle, C., Carreras-Sureda, A., Valverde, M., and Malhotra, V. (2012): Cab45 is required for Ca(2+)-dependent secretory cargo sorting at the trans-Golgi network. J Cell Biol 199, 1057-66. von Rotz, R. C., Kohli, B. M., Bosset, J., Meier, M., Suzuki, T., Nitsch, R. M., and Konietzko, U. (2004): The APP intracellular domain forms nuclear multiprotein complexes and regulates the transcription of its own precursor. J Cell Sci 117, 443548. von Schantz, C., Saharinen, J., Kopra, O., Cooper, J. D., Gentile, M., Hovatta, I., Peltonen, L., and Jalanko, A. (2008): Brain gene expression profiles of Cln1 and Cln5 deficient mice unravels common molecular pathways underlying neuronal degeneration in NCL diseases. BMC Genomics 9, 146. Wagner, U. G., Petersen, E. I., Schwab, H., and Kratky, C. (2002): EstB from Burkholderia gladioli: a novel esterase with a beta-lactamase fold reveals steric factors to discriminate between esterolytic and beta-lactam cleaving activity. Protein Sci 11, 467-78. 227 Walenta, J. H., Didier, A. J., Liu, X., and Kramer, H. (2001): The Golgi-associated hook3 protein is a member of a novel family of microtubule-binding proteins. J Cell Biol 152, 923-34. Wang, S. N., Miyauchi, M., Koshikawa, N., Maruyama, K., Kubota, T., Miura, K., Kurosawa, Y., Awaya, A., and Kanai, Y. (1994): Antigen expression associated with lymph node metastasis in gastric adenocarcinomas. Pathol Int 44, 844-9. Wassmer, T., Attar, N., Bujny, M. V., Oakley, J., Traer, C. J., and Cullen, P. J. (2007): A loss-of-function screen reveals SNX5 and SNX6 as potential components of the mammalian retromer. J Cell Sci 120, 45-54. Wassmer, T., Attar, N., Harterink, M., van Weering, J. R., Traer, C. J., Oakley, J., Goud, B., Stephens, D. J., Verkade, P., Korswagen, H. C., and Cullen, P. J. (2009): The retromer coat complex coordinates endosomal sorting and dynein-mediated transport, with carrier recognition by the trans-Golgi network. Dev Cell 17, 110-22. Weimbs, T., Low, S. H., Chapin, S. J., Mostov, K. E., Bucher, P., and Hofmann, K. (1997): A conserved domain is present in different families of vesicular fusion proteins: a new superfamily. Proc Natl Acad Sci U S A 94, 3046-51. Weimer, J. M., Custer, A. W., Benedict, J. W., Alexander, N. A., Kingsley, E., Federoff, H. J., Cooper, J. D., and Pearce, D. A. (2006): Visual deficits in a mouse model of Batten disease are the result of optic nerve degeneration and loss of dorsal lateral geniculate thalamic neurons. Neurobiol Dis 22, 284-93. Weiss, T. S., Chamberlain, C. E., Takeda, T., Lin, P., Hahn, K. M., and Farquhar, M. G. (2001): Galpha i3 binding to calnuc on Golgi membranes in living cells monitored by fluorescence resonance energy transfer of green fluorescent protein fusion proteins. Proc Natl Acad Sci U S A 98, 14961-6. Wen, L., Tang, F. L., Hong, Y., Luo, S. W., Wang, C. L., He, W., Shen, C., Jung, J. U., Xiong, F., Lee, D. H., Zhang, Q. G., Brann, D., Kim, T. W., Yan, R., Mei, L., and Xiong, W. C. (2011): VPS35 haploinsufficiency increases Alzheimer's disease neuropathology. J Cell Biol 195, 765-79. Wendel, M., Sommarin, Y., Bergman, T., and Heinegard, D. (1995): Isolation, characterization, and primary structure of a calcium-binding 63-kDa bone protein. J Biol Chem 270, 6125-33. Whitley, P., Grahn, E., Kutay, U., Rapoport, T. A., and von Heijne, G. (1996): A 12residue-long polyleucine tail is sufficient to anchor synaptobrevin to the endoplasmic reticulum membrane. J Biol Chem 271, 7583-6. Williams, P., Tulke, S., Ilegems, E., Berggren, P. O., and Broberger, C. (2014): Expression of nucleobindin 1 (NUCB1) in pancreatic islets and other endocrine tissues. Cell Tissue Res 358, 331-42. Williams, R. E., Aberg, L., Autti, T., Goebel, H. H., Kohlschutter, A., and Lonnqvist, T. (2006): Diagnosis of the neuronal ceroid lipofuscinoses: an update. Biochim Biophys Acta 1762, 865-72. Williams, R. E., Gardiner, R. M., and Goebel, H. H. (1999): The European Concerted Action NCL Clinical Case Registry. Mol Genet Metab 66, 407-8. Wilson, B. S., Komuro, M., and Farquhar, M. G. (1994): Cellular variations in heterotrimeric G protein localization and expression in rat pituitary. Endocrinology 134, 233-44. 228 Winchester, B. (2013): Lysosomal diseases: diagnostic update. J Inherit Metab Dis 37, 599-608. Wolins, N. E., Rubin, B., and Brasaemle, D. L. (2001): TIP47 associates with lipid droplets. J Biol Chem 276, 5101-8. Xin, W., Mullen, T. E., Kiely, R., Min, J., Feng, X., Cao, Y., O'Malley, L., Shen, Y., ChuShore, C., Mole, S. E., Goebel, H. H., and Sims, K. (2010): CLN5 mutations are frequent in juvenile and late-onset non-Finnish patients with NCL. Neurology 74, 565-71. Yamashiro, D. J., Fluss, S. R., and Maxfield, F. R. (1983): Acidification of endocytic vesicles by an ATP-dependent proton pump. J Cell Biol 97, 929-34. Yamashiro, D. J., and Maxfield, F. R. (1987): Acidification of morphologically distinct endosomes in mutant and wild-type Chinese hamster ovary cells. J Cell Biol 105, 2723-33. Yang, Y., and Cook, D. G. (2004): Presenilin-1 deficiency impairs glutamate-evoked intracellular calcium responses in neurons. Neuroscience 124, 501-5. Ye, J., Rawson, R. B., Komuro, R., Chen, X., Dave, U. P., Prywes, R., Brown, M. S., and Goldstein, J. L. (2000): ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs. Mol Cell 6, 1355-64. Yoshida, N., Maejima, Y., Sedbazar, U., Ando, A., Kurita, H., Damdindorj, B., Takano, E., Gantulga, D., Iwasaki, Y., Kurashina, T., Onaka, T., Dezaki, K., Nakata, M., Mori, M., and Yada, T. (2010): Stressor-responsive central nesfatin-1 activates corticotropin-releasing hormone, noradrenaline and serotonin neurons and evokes hypothalamic-pituitary-adrenal axis. Aging (Albany NY) 2, 775-84. Yosten, G. L., and Samson, W. K. (2009): Nesfatin-1 exerts cardiovascular actions in brain: possible interaction with the central melanocortin system. Am J Physiol Regul Integr Comp Physiol 297, R330-6. Yosten, G. L., and Samson, W. K. (2010): The anorexigenic and hypertensive effects of nesfatin-1 are reversed by pretreatment with an oxytocin receptor antagonist. Am J Physiol Regul Integr Comp Physiol 298, R1642-7. Yosten, G. L., and Samson, W. K. (2013): Cardiovascular and antidipsogenic effects of nesfatin-1. Curr Pharm Des 19, 6973-5. Yosten, G. L., and Samson, W. K. (2014): Neural circuitry underlying the central hypertensive action of nesfatin-1: melanocortins, corticotropin-releasing hormone, and oxytocin. Am J Physiol Regul Integr Comp Physiol 306, R722-7. Zelazny, E., Santambrogio, M., Pourcher, M., Chambrier, P., Berne-Dedieu, A., FobisLoisy, I., Miege, C., Jaillais, Y., and Gaude, T. (2013): Mechanisms governing the endosomal membrane recruitment of the core retromer in Arabidopsis. J Biol Chem 288, 8815-25. Zeman, W., and Dyken, P. (1969): Neuronal ceroid-lipofuscinosis (Batten's disease): relationship to amaurotic family idiocy? Pediatrics 44, 570-83. Zimprich, A., Benet-Pages, A., Struhal, W., Graf, E., Eck, S. H., Offman, M. N., Haubenberger, D., Spielberger, S., Schulte, E. C., Lichtner, P., Rossle, S. C., Klopp, N., Wolf, E., Seppi, K., Pirker, W., Presslauer, S., Mollenhauer, B., Katzenschlager, R., Foki, T., Hotzy, C., Reinthaler, E., Harutyunyan, A., Kralovics, R., Peters, A., Zimprich, F., Brucke, T., Poewe, W., Auff, E., Trenkwalder, C., Rost, B., Ransmayr, G., Winkelmann, J., Meitinger, T., and Strom, T. M. (2011): A mutation 229 in VPS35, encoding a subunit of the retromer complex, causes late-onset Parkinson disease. Am J Hum Genet 89, 168-75. Zoncu, R., Bar-Peled, L., Efeyan, A., Wang, S., Sancak, Y., and Sabatini, D. M. (2011): mTORC1 senses lysosomal amino acids through an inside-out mechanism that requires the vacuolar H(+)-ATPase. Science 334, 678-83. 230 ANNEXE 1 Intracellular trafficking of LRP9 is dependent on two acidic cluster/dileucine motifs Boucher, R., Larkin, H., Brodeur, J., Gagnon, H., Thériault, C., and Lavoie, C. Histochem Cell Biol 130(2):315-27 (2008) Travaux de stage IF=2,32 231 Intracellular trafficking of LRP9 is dependent on two acidic cluster/dileucine motifs Boucher, R., Larkin, H., Brodeur, J., Gagnon, H., Thériault, C., and Lavoie, C. Abstract LDL receptor-related protein 9 (LRP9) is a distant member of the low-density lipoprotein receptor (LDLR) superfamily. To date, there are no reports on the cellular distribution of LRP9 or the signals responsible for its localization. Here, we investigated the intracellular localization and trafficking of LRP9. Using confocal microscopy, we demonstrated that LRP9 was not present at the plasma membrane but co-localized with various markers of the trans-Golgi network (TGN) and endosomes. This co-localization was dependent on the presence of two acidic cluster/dileucine (DXXLL) motifs in the cytoplasmic tail of LRP9, which interact with GGA proteins, clathrin adaptors involved in transport between the TGN and endosomes. LRP9 is the first example of a transmembrane protein with an internal GGA-binding sequence in addition to the usual C-terminal motif. An inactivating mutation (LL → AA) in both DXXLL motifs, which completely inhibited the interaction of LRP9 with GGA proteins, led to an intracellular redistribution of LRP9 from the TGN to early endosomes and the cell surface, indicating that the two DXXLL motifs are essential sorting determinants of LRP9. In conclusion, our results suggest that LRP9 cycles between the TGN, endosomes and the plasma membrane through a GGA dependent-trafficking mechanism. Keywords LDLR related protein 9 (LRP9), GGA (Golgi-localized, γ ear-containing, ADP ribosylation factor binding protein), trans-Golgi network, Endosome, Acidic cluster/dileucine (DXXLL) motif, LDL receptor Introduction Introduction The low-density lipoprotein receptor (LDLR) family comprises a large number of members that share common structural features as the LDLRA repeats, LDLRB repeats, epidermal growth factor precursor-type repeats and a cytoplasmic region which includes internalization motifs. The variable numbers and combinations of these elements correspond with the huge variety of ligands and functions of these receptors. A new member of this growing family, termed LDLR-related protein 9 (LRP9), was recently identified (Sugiyama et al. 2000). LRP9 is part of a new subfamily of the LDLR superfamily that includes two other receptors, LRP3 and LRP12 (previously named st7) (Battle et al. 2003). Like all LDLR family members, this new subgroup contains LDLRA domain repeats in their extracellular domains (Hussain et al. 1999; Strickland et al. 2002) that function as ligand binding sites (Yamamoto et al. 1984). However, several unique 232 structural features distinguish this new LRP subfamily from prototypical members of the LDLR superfamily (Battle et al. 2003): (1) the absence of EGF-like repeats and YWTDcontaining domains in their extracellular domain; (2) the presence of extracellular CUB domains that are found in a diverse array of functionally unrelated extracellular proteins (Bork and Beckmann 1993) and that are involved in ligand binding (Christensen and Birn 2002); and (3) a large cytoplasmic tail that contains a proline-rich domain. The important role of many members of the LDLR subfamily like LDL, LRP, megalin, VDL and ApoER receptors in endocytosis and lipid metabolism is well known (Willnow 1999). However, several other biological functions are becoming more and more evident. Members of the LDLR family are involved in the regulation of cell surface protease activity, control of cellular entry of bacterial toxins and viruses, the transport and activation of steroid hormones, the regulation of Ca2+ homeostasis and also in cellular signaling events (Sugiyama et al. 2000; Herz 2001; Strickland et al. 2002; Battle et al. 2003; Kikuchi et al. 2007; May et al. 2007) Little is known about the function of LRP9. It has been reported to be ubiquitously expressed and to be involved in the internalization of apolipoprotein E (ApoE)-enriched βVLDL (Sugiyama et al. 2000). Additionally, putative signals for endocytosis and intracellular trafficking (YXXϕ, dileucine, acidic cluster/dileucine [DXXLL] motifs) are present in the cytoplasmic tail of LRP9 (Sugiyama et al. 2000). Moreover, several putative signal transduction motifs (e.g., PDZ-binding domain, proline-rich region) as well as potential phosphorylation sites have also been identified in the cytoplasmic tail of LRP9 (Battle et al. 2003). LRP9 thus has the molecular characteristics of a trafficking and signalling member of the LDLR family. GGA proteins (Golgi localized, γ ear-containing, ADP ribosylation factor-binding proteins) are monomeric adaptor proteins that mediate the transport of cargo proteins between the trans-Golgi Network (TGN) and endosomes (reviewed by Bonifacino 2004; Ghosh and Kornfeld 2004). They have a modular structure consisting of an amino-terminal VHS (Vps27, Hrs, STAM) domain, a GAT (GGA and TOM1) domain, a hinge region and a C-terminal γ-adaptin “ear” (GAE) domain. The VHS domain binds to acidic cluster/dileucine (DXXLL (where X is any amino acid)) sorting signals that are present in the cytoplamic tail of various transmembrane proteins such as mannose-6-phosphate receptor (MPR) (Puertollano et al. 2001; Zhu et al. 2001), sortilin (Nielsen et al. 2001), βsecretase (He et al. 2002) and the LDLR member, LRP3 (Takatsu et al. 2001). The GAT domain binds to the Arf family of GTP-binding protein, the hinge region binds to clathrin and the GAE domain binds to several accessory proteins (reviewed by Bonifacino 2004). Through these multiple interactions, GGAs are involved in the sorting of DXXLLcontaining receptors into coated vesicles. In this study, we describe for the first time the intracellular localization and trafficking of LRP9. We observed by confocal microscopy that LRP9 was predominantly localized at the TGN, where it co-localized with GGA proteins, and to a lesser extent, in endosomes. We also demonstrated that the VHS domain of GGA proteins bound to two functional DXXLL motifs in the cytoplasmic tail of LRP9. LRP9 is the first reported 233 example of a transmembrane protein with an internal GGA-binding sequence in addition to the usual C-terminal motif. Interestingly, mutation (LL → AA) of both DXXLL motifs prevented the interaction of LRP9 with GGA proteins and caused a redistribution of LRP9 in early endosomes and at the plasma membrane. These data demonstrated that LRP9 cycles between the TGN, endosomes and the plasma membrane and suggested that GGA proteins participate in the regulation of LRP9 trafficking. Materials and methods Antibodies and reagents Rabbit antibodies against mannosidase II were generously provided by Dr. Marilyn Gist Farquhar (University of California, San Diego, CA, USA). Anti-HA monoclonal antibodies (mAbs) were purchased from Covance (Berkley, CA, USA), anti-Myc mAbs from Cell Signaling Technology (Danvers, MA, USA), anti-FLAG mAbs from Sigma (St Louis, MO, USA) and anti-GM130 mAbs from BD Transduction Laboratories (Franklin Lakes, NJ, USA). Anti-GFP polyclonal antibodies (pAbs) were purchased from Molecular Probes (Eugene, OR, USA), anti-HA pAbs from Covance, anti-EEA1 pAbs from Affinity Bioreagents (Golden, CO, USA), and anti-TGN46 pAbs from Novus Biologicals (Littleton, CO, USA). Brefeldin A and monensin were obtained from Sigma. DNA constructs Expression vectors for GGA1-Myc, GGA2-Myc and GGA3-Myc were kindly provided by Dr. Juan Bonifacino (National Institutes of Health, Bethesda, MD, USA). pcDNA-FLAG-tagged furin was kindly provided by Dr. Richard Leduc (Université de Sherbrooke, QC, Canada). Mammalian expression vector pMKITNeo encoding the murine LRP9-HA was kindly provided by Dr. T. Kitamura (University of Tokyo, Japan). PCRbased mutagenesis was used to generate the LRP9 mutant constructs. To generate the LRP9-proximal mutant, Leu692, Leu693 and Leu694 in the C-terminal sequence DDVLLL were replaced by alanines. To generate the LRP9-distal mutant, Leu711 and Leu712 in the C-terminal sequence DEPLLA were replaced by alanines. To generate the double LRP9 mutant, Leu692, Leu693, Leu694, Leu711 and Leu712 in the two DXXLL motifs in the Cterminal sequences (DDVLLL and DEPLLA) were replaced by alanines. The sequences of the DNA constructs generated by PCR were systematically verified. Cell culture and transfection COS7 cells were obtained from Dr. Klaus Hahn (University of North Carolina, Chapel Hill, NC, USA) and HEK293T cells were obtained from Dr. Alexandra Newton (University of California, San Diego, CA, USA). The cells were grown in Dulbecco’s modified Eagle’s high glucose medium (Invitrogen, Carlsbad, CA, USA) containing 10% foetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT, USA), penicillin, and streptomycin. COS7 cells were transfected using Fugene6 transfection reagent (Roche Diagnostics, Indianapolis, IN, USA), while HEK cells were transfected using Lipofectamine 2000 transfection reagent (Invitrogen), both according to the manufacturers’ instructions. 234 Immunofluorescence COS7 cells were plated on coverslips. Twelve hours after transfection, the cells were fixed in 3% paraformaldehyde (PFA) in 100 mM phosphate buffer, pH 7.4 for 30 min, permeabilized with 0.1% Triton X-100 for 10 min, blocked with 10% goat serum for 30 min, and incubated with primary antibodies for 1 h at RT, followed by Alexa Fluor-594 or 488-conjugated antibodies (Molecular Probes) for 1 h at RT. Specimens were visualized using an inverted confocal laser-scanning microscope (FV1000, Olympus, Tokyo, Japan) with a PlanApo 60x/1.42 oil immersion objective (Olympus). Olympus Fluoview software version 1.6a was used for image acquisition and analysis. The images were further processed using Adobe Photoshop (Adobe Systems, San Jose, CA, USA). Glutathione S-transferase (GST) pull-down assays GST constructs encoding GGA1-VHSGAT (residues 5–314), GGA1-VHS (residues 5–147), GGA1-GAT (residues 148–314), GGA1-Hinge (residues 315–514), GGA1-GAE (residues 515–639) and GGA3-VHS (residues 1–146) were generated by standard PCR methods and subcloned into pGEX-KG (Amersham-Pharmacia Biotech, Piscataway, NJ, USA). GST fusion proteins were expressed in E. Coli BL21 and purified on glutathioneSepharose 4B beads (GE Healthcare, Piscataway, NJ, USA) according to the manufacturer’s instructions. Cells were lysed in buffer (50 mM Tris pH 7.4, 200 mM NaCl, 1% NP40 and complete protease inhibitors (Roche)) for 1 h at 4°C and then centrifuged at 13,000×g for 20 min. One mg of cleared lysate was incubated overnight at 4°C with 10 µg GST fusion proteins immobilized on beads. Beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. Bound proteins were analyzed by SDS-PAGE and detected by immunoblotting. Coimmunoprecipitation HEK cells were plated in 60 mm culture dishes and transfected with the various constructs. After 48 h, the cells were lysed in 50 mM Tris buffer (pH 7.4) containing 100 mM NaCl, 1% NP40 and protease inhibitors for 1 h at 4°C and then centrifuged at 13,000 rpm for 20 min. Cleared supernatants were incubated with primary antibodies overnight at 4° and then with protein A-sepharose (GE Healthcare, Piscataway, NJ, USA) or protein GSepharose (Zymed, San Francisco, CA, USA) for 1 h. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. Bound immune complexes were analyzed by SDS-PAGE and immunoblotting. Immunoblotting Protein samples were separated on 10% SDS-PAGE gels and transferred to nitrocellulose membranes (Perkin Elmer, Woodbridge, ON, Canada). Membranes were blocked in Tris-buffered saline (20 mM Tris–HCl, pH 7.5, 150 mM NaCl) containing 0.1% Tween 20 and 5% nonfat dry milk and incubated with primary antibodies for 2 h at RT and then with horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (Bio-Rad, Richmond, ON, Canada) and enhanced chemiluminescence detection (Pierce Chemical, Rockford, IL, USA). 235 Results LRP9 localizes in the trans-Golgi network and early endosomes in COS7 cells The cellular distribution of LRP9 has never been investigated. To characterize the intracellular localization of this receptor, we compared the distribution of LRP9 with several specific intracellular markers by confocal microscopy. HA-tagged LRP9 expressed in COS7 cells was not detected at the cell surface but was primarily localized in the juxtanuclear compartment as well as in some scattered vesicles (Fig. 1a, d, g). The juxtanuclear distribution of LRP9 did not significantly co-localize with α-mannosidase II (Fig. 1a–c) or galactosyl transferase (Fig. 1d–f), which are markers of the medial- and trans-Golgi cisternae, respectively. However, extensive co-localization with furin (Fig. 1g– i) and TGN46 (Fig. 2e–h), which are markers of the TGN, was observed. To determine whether LRP9 is a Golgi resident protein or a cargo protein in transit, COS7 cells transfected with LRP9-HA were treated for 1 h with cycloheximide, a protein synthesis inhibitor. This treatment did not alter the distribution of LRP9-HA, which suggested that LRP9, which has a half-life of 4 h, remains in the Golgi for a long period of time and is not just in transit to the plasma membrane (data not shown). Figure 1 - Comparison of LRP9 distribution with different Golgi markers in COS7 cells. COS7 cells transfected with LRP9-HA alone (a–c) or together with galactosyl transferaseYFP (d–f) or furin-FLAG (G-I) were fixed, permeabilized and immunostained using antiHA (a, d, g), anti-mannosidase II (b), anti-GFP (e), anti-FLAG (h) antibodies. Stained cells were examined by confocal fluorescence microscopy. Yellow colour in the merged images (c, f, i) indicates co-localization. Scale bar 10 µm. 236 Figure 2 - Effect of brefeldin A and monensin treatments on LRP9 intracellular localization. 237 COS7 cells transfected with LRP9-HA were incubated in either the absence (a–h) or presence of 10 µg/µl brefeldin A (i–p) or 10 µM monensin (q–x) for 60 min. Cells were fixed, permeabilized and immunostained using anti-HA (b, f, j, n, r, v), anti-GM130 (a, i, q) and anti-TGN46 (e, m, u) antibodies. Stained cells were examined by confocal fluorescence microscopy. Yellow colour in the merged images indicates co-localization. Last column show ×2.5 magnified views of the boxes. Scale bar 10 µm. To confirm that LRP9-HA specifically localized to the TGN compartment, experiments were performed with brefeldin A (BFA) and monensin, two Golgi perturbing agents. BFA is a fungal metabolite known to inhibit coat formation on the Golgi membranes, thereby disrupting the cisternal stacks leading to a redistribution of the cis-, medial- and trans-cisternae in the endoplasmic reticulum, leaving the TGN to coalesce into a spherical mass near the microtubule organizing centre (MTOC; (Banting and Ponnambalam 1997; Chardin and McCormick 1999; Watson and Pessin 2000). After BFA treatment, LRP9-HA showed a concentrated vesicular distribution that colocalized with the TGN marker TGN46 (Fig. 2m–p). This differs from the cis-Golgi marker GM130 (Fig. 2i– l) which remained associated with dispersed punctate structures as previously described (Nakamura et al. 1995; Xu et al. 2002). Monensin is an ionophore that disrupts the acidic pH of luminal compartments and causes the accumulation of some Golgi proteins such as TGN38/46 and GPP130 in swollen vesicles that has been defined as derivative of the TGN (Schaub et al. 2006) or endosomes (Linstedt et al. 1997; Puri et al. 2002). Monensin treatment of COS7 cells led to the relocalization of LRP9-HA and TGN46 to swollen vesicles (Fig. 2u–x) while GM130 staining remained in a characteristic perinuclear Golgi pattern (Fig. 2q–t). These results support the TGN localization of LRP9. Confocal microscopy analysis of the intracellular localization of LRP9-HA also showed a partial colocalization with EEA1, an early endosome marker (Fig. 3a–c). These LRP9-EEA1labelled vesicles mainly localized in the Golgi area and not at the periphery of the cells (Fig. 3a–c). The presence of LRP9 in EEA1-labelled endosomes was confirmed in COS7 cells transfected with an established Rab5 GTPase-deficient mutant (Rab5Q79L) that greatly enlarges early endosomes due to increased homo- and heterotypic fusion (Stenmark et al. 1994) (Fig. 3d–f). These observations suggest that LRP9 predominantly localizes to the TGN and, to a lesser extent, to early endosomes at steady state in COS7 cells. LRP9 co-localizes with GGA proteins in COS7 cells GGA proteins play a role of molecular adaptors in the transport and retrieval of cargo proteins between the TGN and endosomal compartments (Puertollano et al. 2001; Zhu et al. 2001; Ghosh et al. 2003; Bonifacino 2004). GGA proteins directly bind to acidic cluster/dileucine (DXXLL) motifs located at the extreme end of the C-termini of cargo proteins (Bonifacino 2004; He et al. 2005). Since LRP9 localized to the TGN and endosomes (Figs. 1–3) and since the C-terminal domain of LRP9 possesses a DXXLL motif similar to the one found in transmembrane proteins reported to bind GGAs (LRP9distal; Fig. 5a, b), we postulated that GGA proteins might interact with LRP9 and consequently regulate its intracellular trafficking. To test this hypothesis, we first examined 238 the co-localization of GGA isoforms with LRP9 by confocal microscopic imaging. This was performed using COS7 cells expressing HA-tagged LRP9 together with Myc-tagged GGA1, GGA2 or GGA3. GGA1-Myc displayed both juxtanuclear and vesicular distributions that substantially overlapped those of LRP9-HA (Fig. 4a–c). A similar distribution pattern was observed for GGA3-Myc (Fig. 4g–i) whereas partial colocalization was observed between GGA2-Myc and HA-LRP9 (Fig. 4d–f). These results demonstrated that GGA proteins broadly co-localize with LRP9 and suggested that GGA may be involved in the TGN-endosome transport of LRP9 through a potential interaction with its C-terminal domain. Figure 3 - LRP9 partially co-localizes with EEA1 on endosomes. COS7 cells were co-transfected with HA-tagged LRP9 alone (a–c) or together with Rab5Q79L, a Rab5 GTPase-deficient mutant that creates enlarged early endosomes (d–f). Cells were fixed, permeabilized and immunostained using anti-HA (a, d) and anti-EEA1 (b, e) antibodies. Stained cells were examined by confocal fluorescence microscopy. Yellow colour in the merged images (c, f) indicates co-localization. Scale bar 10 µm. 239 Figure 4 - LRP9 co-localizes with GGA proteins at the TGN in COS7 cells. COS7 cells co-transfected with HA-tagged LRP9 (a, d, g) together with Myc-tagged GGA1 (b), GGA2 (e) or GGA3 (h) were fixed, permeabilized and immunostained using anti-HA (a, d, g) or anti-Myc (b, e, h) antibodies. Stained cells were examined by confocal fluorescence microscopy. Yellow colour in the merged images (c, f, i) indicates colocalization. Scale bar 10 µm. 240 Figure 5 - LRP9 interacts with GGA proteins via a proximal and distal DXXLL motif in its cytoplasmic tail. a Schematic representation of the cytoplasmic tail of LRP9 containing two DXXLL motifs: proximal and distal (bold face letters). Underlined leucines were substituted by alanines. b Alignment of the C-terminal amino acids of transmembrane proteins that contain acidic cluster/dileucine signals in their cytosolic tails. The two crucial leucine residues are shown in red and the crucial aspartate is shown in blue. The position of the transmembrane domain (Tm) and the number of residues before and/or after the signals is indicated. CD-MPR, cation-dependent mannose-6-phosphate receptor; CI-MPR, cation-independent mannose-6phosphate receptor; LRP, low-density lipoprotein receptor related protein. c–d Coimmunoprecipitations of wild-type and mutant LRP9 with GGA proteins. Lysates from HEK cells transfected with wild-type (WT) or DXXLL mutant LRP9-HA [proximal (Pr), 241 distal (Di) or double (Do)] and GGA1-Myc, GGA2-Myc or GGA3-Myc were immunoprecipitated with anti-HA (c, lanes 2–13), anti-Myc (d, lanes 2–13) or control (Ctl) (c, d, lanes 1 and 14) antibodies and then immunoblotted with anti-Myc or anti-HA to detect GGA proteins and LRP9, respectively. Asterisk denotes longer exposure to show GGA1-Myc bands with same intensity as GGA3-Myc and GGA2-Myc. It highlights the fact that the interaction between LRP9-WT and GGA1-Myc is weaker than with GGA2Myc and GGA3-Myc. Note that the level of expression of Myc-GGA3 is approximatively six times the level of endogenous GGA3 proteins. GGA proteins interact with two DXXL motifs in the cytoplasmic tail of LRP9 To determine whether GGA proteins bind to LRP9, we first tested whether GGA proteins co-immunoprecipitate with LRP9 in HEK293 cells transfected with both HAtagged wild-type LRP9 (WT) and Myc-tagged GGA1, GGA2 or GGA3. When the immunoprecipitations were carried out using anti-HA IgGs, GGA1, GGA2 and GGA3 coimmunoprecipitated with LRP9-HA (Fig. 5c, lanes 2, 6, 10). However, GGA2-Myc and GGA3-Myc bind better to LRP9-HA than GGA1-Myc. The same results were obtained when the immunoprecipitation was carried out using anti-Myc IgGs. HA-LRP9 coimmunoprecipitated with GGA1-Myc, GGA2-Myc and GGA3-Myc (Fig. 5d, lanes 2, 6, 10), thus confirming the interaction of GGA proteins and LRP9 in cultured cells. Since an inactivating mutation of the dileucine residues of the DXXLL sequence has previously been reported to abrogate the interaction with GGA proteins (Nielsen et al. 2001; Puertollano et al. 2001; Takatsu et al. 2001; Zhu et al. 2001; He et al. 2002), leucines 711 and 712 were replaced by alanines in the DXXLL motif at the extreme C-terminus of the cytoplasmic tail of HA-LRP9 (named distal DXXLL, Fig. 5a). Interactions between the various GGA-Myc proteins and this LRP9 mutant were then investigated using the coimmunoprecipitation assays described above with either anti-HA (Fig. 5c) or anti-Myc IgGs (Fig. 5d). The interaction between LRP9 and GGA1 was almost completely abolished (Fig. 5c, d, lane 4) by this mutation. However, the interactions between LRP9 and GGA2 and GGA3 were not affected (Fig. 5c, d, lanes 8, 12). These results suggested that the distal DXXLL motif is not the only key element that contributes to the interaction with GGA proteins. After further examination of the C-terminal sequence of LRP9, another DXXLL motif was identified thirteen amino acids upstream from the distal DXXLL (Fig. 5a). This new DXXLL motif was composed of two overlapping DXXLL motifs (D689 D690 V691 L692 L693 L694) and was called the proximal DXXLL motif (Fig. 5a, b). Alanines were substituted for Leu692, Leu693 and Leu694 in the proximal DXXLL motif and interactions between GGA proteins and the mutated LRP9 were analyzed by immunoprecipitation using either anti-HA IgGs (Fig. 5c) or anti-Myc IgGs (Fig. 5d). As shown in Fig. 5c and d (lanes 3, 7, 11), the mutations introduced into the proximal DXXLL did not abrogate the interaction of LRP9 with GGA1, GGA2 or GGA3. This indicated that both the proximal and distal DXXLL motifs might be responsible for the binding of GGA proteins to LRP9. To confirm this, leucines (Leu692, Leu693, Leu694, Leu711 and Leu712) were substituted by alanines in both the proximal and distal DXXLL motifs of LRP9 cytoplasmic tail (Fig. 5a). This double DXXLL mutation completely inhibited the interaction of LRP9 with the 242 three GGA proteins (Fig. 5c, d; lanes 5, 9, 13), indicating that the C-terminal domain of LRP9 contained two DXXLL motifs that were able to bind to GGA proteins. This discovery is of particular interest because LRP9 is the first example of a transmembrane protein with an internal GGA-binding sequence in addition to the usual C-terminal motif. Figure 6 - LRP9 interacts with the VHS domain of GGA proteins. a Diagram of GGA1 showing the VHS (Vps27, Hrs, Stam) domain, the GAT (GGA and TOM) domain, the hinge segment and the GAE (γ-adaptin ear) domain (Bonifacino 2004). b–c The indicated GST-fusion proteins immobilized on glutathione beads were incubated with lysates from HEK cells expressing wild-type (WT) LRP9-HA or cation dependant mannose-6-phosphate receptor–GFP (CD-MPR-GFP, b) or LRP9-HA in which the two DXXLL motifs were mutated (Double, c). Bound proteins were separated by SDS-PAGE and immunoblotted with anti-HA or anti-GFP antibodies to detect LRP9 and CD-M6PR, respectively. We next determined which domain of GGA proteins mediates the interaction with LRP9 using pull-down assays. Glutathione-S-transferase (GST) fusion proteins comprising the various GGA1 domains (Fig. 6a) were incubated with lysates from HEK293 cells expressing HA-tagged wild-type LRP9 (Fig. 6b). As a positive control, the pull-down-assay were also performed with lysates from cells expressing GFP-tagged cation-dependant mannose-6-phosphate receptor (CD-M6PR-GFP) that contains a DXXLL motif reported to interact with the VHS domain of GGA1 (Puertollano et al. 2001). Similar to CD-M6PR243 Figure 7 - Mutations in both the proximal and distal DXXLL motifs of LRP9 affect its intracellular distribution. a–i In COS7 cells transfected with GGA1-Myc together with wild-type LRP9-HA (WT, a) or LRP9-HA in which the distal DXXLL motif (Distal, d) or proximal DXXLL motif (Prox., g) were mutated, LRP9-HA co-localized extensively with GGA1-Myc (b, e, h) in the juxtanuclear Golgi region (arrowheads, c, f, i) and on vesicular structures (inset, c, f, i). j–l In cells transfected with GGA1-Myc and LRP9-HA in which the two DXXLL motifs (double, j) were mutated, LRP9 was mainly present on the cell surface (arrows, j) and in peripheral punctate structures (inset, J) and did not significantly co-localize with GGA1244 Myc in the Golgi region (arrowhead, l) or on vesicular structures (inset, l). Cells were fixed, permeabilized and immunostained using anti-HA (a, d, g, j) or anti-Myc (b, e, h, k) antibodies. Stained cells were examined by confocal fluorescence microscopy. Yellow colour in the merged images (c, f, i, l) indicates co-localization. Insets show ×2 magnified views of cytoplasmic regions. Scale bar 10 µm. GFP, LRP9-HA bound strongly to GST-GGA1-VHS or GST-GGA1-VHSGAT and weakly to GST-GGA1-GAT (Fig. 6b). In contrast, no interaction was observed with GST, GSTGGA1-Hinge and GST-GGA1-GAE (Fig. 6b). Similar results were obtained with GST fusion proteins comprising the various GGA3 domains (data not shown). These results demonstrated that LRP9 interacts predominantly with the VHS domain of GGA proteins. The importance of the DXXLL motifs in LRP9 interaction with the VHS domain of GGA was next examined. To this end, interactions of GST-GGA1-VHS with the double DXXLL mutants of LRP9-HA were compared to wild-type LRP9-HA. As shown in Fig. 6c, mutation of both DXXLL motifs completely inhibited the interaction of LRP9 with the VHS domain of GGA1. Similar results were obtained with GST-GGA3-VHS (data not shown). These results indicate that the DXXLL motifs of LRP9 interact specifically with the VHS domain of GGA proteins. This is in agreement with previous studies showing that VHS is the specific GGA domain that binds to the DXXLL motif of various of transmembrane proteins such as sortilin, LRP3, β-secretase and M6PR (Puertollano et al. 2001; Takatsu et al. 2001; He et al. 2002; Bonifacino 2004). Mutations in both the proximal and distal DXXLL motifs induce a redistribution of LRP9 to the plasma membrane and peripheral endosomes To investigate the role of the DXXLL motifs in LRP9 trafficking, the subcellular distributions of the proximal, distal and double DXXLL mutants were compared to wildtype LRP9. COS7 cells were transfected with Myc-tagged GGA1 together with HA-tagged wild-type LRP9 or the mutant DXXLLs and were analyzed by confocal microscopy. The distributions of LRP9-HA mutated in the distal- (Fig. 7d) or proximal-DXXLL (Fig. 7g) motifs were identical to the wild-type LRP9-HA (Fig. 7a). They all extensively colocalized with GGA1 in the juxtanuclear Golgi region (Fig. 7a–i, arrowheads) and in cytoplasmic vesicles (insets, Fig. 7a–i). In contrast, the double DXXLL mutant mainly localized at the cell surface and in peripheral punctate structures (Fig. 7j) and did not colocalize with GGA1 (Fig. 7l, inset). Similar data were obtained with cells expressing Myctagged GGA2 and Myc-tagged GGA3 (data not shown). We noticed that the localization of Myc-GGA1 was also modified in cells expressing the double DXXLL mutant. GGA1 was less concentrated in the Golgi region in these cells (Fig. 7k) compared to cells expressing LRP9-WT (Fig. 7b), distal (Fig. 7e) or proximal mutants (Fig. 7h). This could be explained by the fact that cargo proteins help recruit GGAs onto membranes (Hirst et al. 2007). Indeed, it has been reported that the overexpression of CD8 chimeras with cytoplasmic tails containing DXXLL-sorting signals, which bind to GGAs, increases the localization of all three GGAs to juxtanuclear membranes (Hirst et al. 2007). In the same vein, we suggest that the overexpression of LRP9 double DXXLL mutants that cannot bind GGA proteins would decrease GGA labelling in the Golgi region compared to the cells that overexpress LRP9 that can bind GGA proteins. 245 To better define the subcellular localization of the double DXXLL mutant, the cells were transfected with HA-tagged wild-type LRP9 or the DXXLL mutants and were doublelabelled with endogenous EEA1, an early endosome marker. Figure 8a–i shows that the wild-type LRP9 (Fig. 8a–c) and the distal DXXLL (Fig. 8d–f) and proximal DXXLL (Fig. 8g–i) mutants did not co-localize with peripheral EEA1-labelled structures and were not present at the plasma membrane. In contrast, the double DXXLL mutant was present at the plasma membrane and co-localized with EEA1 in peripheral punctate structures (Fig. 8j–l). In summary, mutations of both DXXLL motifs, which inhibited the interaction with GGA proteins, shifted the distribution of LRP9 from the TGN to the plasma membrane and peripheral EEA1-labelled endosomes. These observations suggested that GGA proteins participate in the intracellular trafficking of LRP9. Figure 8 - Redistribution of LRP9 double DXXLL mutant to the plasma membrane and peripheral EEA1-positive compartments. 246 a–i In COS7 cells transfected with wild-type LRP9-HA (WT, a–c) or LRP9-HA in which the distal DXXLL motif (Distal, d–f) or proximal DXXLL motif (Prox., g–i) were mutated, LRP9-HA was not present in peripheral EEA1-labelled endosomes (inset, c, f, i). j–l In cells transfected with LRP9-HA in which the two DXXLL motifs were mutated (Double, j– l), LRP9-HA was mainly present on the cell surface (arrows, j, l) and in peripheral punctate structures (arrowheads, j) that significantly co-localized with EEA1 (inset, l). Cells were fixed, permeabilized and immunostained using anti-HA (a, d, g, j) and anti-EEA1 (b, e, h, k) antibodies. Stained cells were examined by confocal fluorescence microscopy. Yellow colour in the merged images indicates co-localization. Merged images (c, f, i, l) show ×2 magnified views of the boxes in a–l. Scale bar 10 µm. Discussion LRP9 is a distant member of the LDLR superfamily that has been reported to be expressed in a number of tissues, notably kidney, liver, lung and heart tissue and to be involved in the internalization of ApoE-enriched βVLDL (Sugiyama et al. 2000). However, the intracellular distribution, subcellular targeting signals and physiological role of LRP9 remain to be clearly demonstrated. In this study, we characterized for the first time the intracellular localization and trafficking of LRP9. Confocal microscopy studies showed that LRP9 was not present at the PM [as confirmed by biotinylation assays (data not shown)] but was predominantly distributed to the juxtanuclear region where it co-localized with TGN markers such as TGN46 and GGA proteins. Localization to the TGN was supported by the action of brefeldin A and monensin, two agents that promote Golgi disorganization. These drugs induced a co-redistribution of LRP9 and TGN46 into the same intracellular structures, consistent with a TGN localization of LRP9. In addition, a part of LRP9 staining was detected in vesicles that overlapped with early endosome markers. Thus, our data suggest that LRP9 predominantly localizes to the TGN and, to a lesser extent to early endosomes, at steady state. GGA proteins are known to interact with acidic cluster/dileucine (DXXLL) motifs in the cytoplamic tail of transmembrane proteins (Nielsen et al. 2001; Puertollano et al. 2001; Takatsu et al. 2001; Zhu et al. 2001; He et al. 2002; Misra et al. 2002). Interestingly, we noted that the cytoplasmic tail of LRP9 contained two DXXLL sequences, one at the extreme end of the cytoplasmic tail (distal motif relative to the transmembrane domain, residues 708–712) and another, thirteen amino acids upstream (proximal motif relative to the transmembrane domain, residues 689–694). Using immunoprecipitation assays, we demonstrated that LRP9 interacts strongly with GGA2 and GGA3 and more weakly with GGA1. Simultaneous mutations in both DXXLL motifs completely inhibited the interaction of LRP9 with the three GGA proteins. However, the substitution of LL by AA in the proximal DXXLL motif did not abrogate the interaction of LRP9 with the three GGA proteins. In addition, the LL → AA mutation in the distal DXXLL motif did not alter the interaction of LRP9 with GGA2 and GGA3, although it almost completely abrogated the interaction with GGA1. Therefore, both DXXLL motifs bound to GGA proteins but had different affinities for GGA1. Indeed, the distal motif bound the three GGA proteins (albeit 247 more weakly to GGA1) whereas the proximal DXXLL motif bound mainly GGA2 and GGA3. Significant differences in the affinity of GGA proteins for acidic dileucine motifs have been previously reported. Although both CD-M6PR and CI-M6PR contain similar DXXLL motifs near the C-terminus of their cytoplasmic tails (Fig. 5b), CI-M6PR binds to GGA proteins with higher affinity than the CD-M6PR (Doray et al. 2002). There is also a difference in the relative affinity of the DXXLL motif in the CD-M6PR cytoplasmic tail for the different GGAs. Puertollano et al. (2001) showed that CD-MPR bound well to GGA1, but only poorly to GGA3, and not at all to GGA2. In contrast, the cytoplasmic tail of CIMPR bound to the VHS domains of all three GGAs with about equal affinity. Comparison of the C-terminal residues of these two M6PR (Fig. 5b) revealed a number of nonconservative differences in the acidic cluster/dileucine motifs and the flanking residues. Mutation analysis revealed that several residues immediately upstream and downstream from the DXXLL motif, as well as the XX residues of this motif, significantly impact the specificity of the interactions of GGA proteins with these motifs (Puertollano et al. 2001; Zhu et al. 2001; Doray et al. 2002; He et al. 2002; He et al. 2003). Alignment of the proximal and distal acidic cluster dileucine motifs of LRP9 with the amino acid sequences of the cytoplasmic tail of other transmembrane proteins containing a DXXLL motif (Fig. 5b) showed differences in residues in position +1 and +2 as well as in several residues upstream and downstream from the proximal and distal DXXLL motifs. The importance of these residues is now under investigation, and should provide new insights into selectivity towards specific interactions with GGA proteins and the trafficking mechanisms of DXXLL-mediated sorting signals. LRP9 is the first example of a transmembrane protein with an internal GGA-binding sequence in addition to the usual C-terminal motif. Using various protein sequence data banks such as TrEMBL and Swiss-Prot, we were unable to identify other proteins with two DXXLL motifs, suggesting that LRP9 might be the only transmembrane protein with two functional DXXLL motifs. The VHS domain of GGA proteins has been identified as the binding site of the DXXLL motifs of LRP9. This is in agreement with previous reports defining VHS as the specific GGA domain that binds to the DXXLL motif (Bonifacino 2004; He et al. 2002; Nielsen et al. 2001; Puertollano et al. 2001; Takatsu et al. 2001). LRP9 can now be added to the long list of proteins with DXXLL-type sorting signals that bind to the VHS domain of the GGAs including sortilin, LRP3, β-secretase and M6PR (Bonifacino 2004). Previous studies have reported that interactions of GGA proteins with a DXXLL motif in the cytoplasmic tail of various receptors mediate their anterograde or retrograde transport between the TGN and endosomes (Huse et al. 2000; Tikkanen et al. 2000; Tortorella et al. 2007). These studies, together with our observation that LRP9 interacted and co-localized with GGA proteins, prompted us to investigate whether GGA proteins were involved in the intracellular transport of LRP9. Mutations (LL → AA) in the proximal or distal DXXLL motif did not affect the intracellular distribution of LRP9. Therefore, proximal and distal DXXLL motifs alone seemed to be sufficient for the proper trafficking of LRP9 since they both had the capacity to bind GGA proteins. In addition, since we did not detect any substantial differences between the single DXXLL mutants, we concluded that these two motifs might play redundant roles in the trafficking of LRP9. Mutations in 248 both DXXLL motifs that completely inhibited the interaction of LRP9 with the three GGA proteins resulted in a redistribution of the receptor from the TGN to EEA1-labelled endosomes and the plasma membrane. Similar mutations in the DXXLL motifs in the cytoplasmic tails of other receptors such as MPR and mepapsin (BACE1) have also been reported to cause a redistribution of these receptors from the TGN to endosomes and the plasma membrane (Huse et al. 2000; Tikkanen et al. 2000; He et al. 2005; Tortorella et al. 2007), thus reinforcing our observations. Furthermore, downregulation of GGA proteins or overexpression of dominant negative GGA mutants cause an accumulation of MPR in endosomal compartments (Ghosh et al. 2003; Puertollano and Bonifacino 2004; He et al. 2005; Wahle et al. 2005). These receptor redistributions have been explained by a sorting defect at different levels: (1) TGN sorting: GGA proteins normally retain DXXLL receptors in the TGN and prevent premature exit from the Golgi apparatus to early endosomes or from the Golgi apparatus via the secretory pathway to the cell surface where they would then be rapidly internalized into early endosomes or (2) endosomal sorting: GGA proteins are involved in early endosome-to-TGN retrieval of receptors containing a DXXLL motif. These explanations and our results provide support for a potential role for GGA adaptor proteins in sorting LRP9 between the TGN and endosomal compartments in both anterograde and retrograde directions. Our data thus suggested that LRP9 cycles between the TGN and early endosomes at steady state. LRP9 can also reach the cell surface since the double DXXLL mutant relocalized at the plasma membrane. However, since there was no significant labelling of wild-type LRP9 at the plasma membrane, we propose that either only a minor fraction of wild-type LRP9 is transported to the cell surface or its presence at the plasma membrane is very transient because the receptor is quickly internalized. Interestingly, Doray et al. (2007) have recently reported that a chimera protein containing the cytoplasmic tail of LRP9 undergoes rapid endocytosis. Furthermore, like our observations, these authors showed that a double DXXLL mutant (LL → AA) accumulates at the cell surface, impairing LRP9 internalization. They reported that the proximal and distal dileucine motifs of LRP9 bind γ/σ1 AP1 and α/σ2 AP2 hemicomplexes and that the intracellular distribution of an LRP9 mutant that is unable to bind AP1 is not affected whereas the LRP9 mutant that is unable to bind to both AP2 and AP1 is redistributed to the plasma membrane (Doray et al. 2007). They concluded that there is a correlation between α/σ2 AP2 hemicomplex binding and LRP9 chimera receptor internalization mediated by dileucine-based sorting signals. Since current evidence does not support the involvement of GGA proteins in the internalization steps, we suggest that the retention of the double DXXLL mutant at the cell surface we observed (Figs. 7, 8) might be due to the loss of the interaction between LRP9 and AP2. The DXXLL motifs of LRP9 thus appear to be required for sorting at two different steps in LRP9 trafficking: (1) internalization from the plasma membrane via an interaction with AP2 and (2) transport between the endosomes and the TGN via an interaction with the GGA proteins. Further studies are now underway to characterize the exact trafficking itinerary and kinetics of LRP9 and to identify the precise molecular machinery involved in the intracellular sorting of LRP9. Only a few other members of the LDLR family have been reported to bind GGA proteins. LRP3 and LRP12, the two other members of the new LRP9 subgroup of the 249 LDLR superfamily, possess a single DXXLL motif in their cytoplasmic tail, suggesting that they have the capacity to interact with GGA proteins. Indeed, LRP3 has already been shown to interact with GGA proteins (Takatsu et al. 2001). However, the intracellular trafficking and function of LRP3 and LRP12 remain uncharacterized. SorLA (sorting protein related receptor; LR11) is another distant member of the LDLR family shown to mediate the uptake of ApoE-rich lipoproteins in vitro (Taira et al. 2001), to interact with GGA1 and GGA2 (Jacobsen et al. 2002) and to shuttle between the Golgi and endosomes (Jacobsen et al. 2001; Andersen et al. 2005; Offe et al. 2006). Interestingly, SorLA/LR11 has been recently described as a sorting receptor that regulates the intracellular transport and processing of the amyloid precursor protein (APP) (Andersen et al. 2005; Shah and Yu 2006; Schmidt et al. 2007) pointing to an involvement in Alzheimer’s disease. To date, the only ligand identified for LRP9 is apoE-βVLDL (Sugiyama et al. 2000). However, since LRP9 is not mainly localized at the cell surface at steady state, it is likely to serve other functions inside the cell. Different lines of evidence give hints towards a possible function of LRP9. First, the intracellular distribution of LRP9 at steady state. We observed that only a minor fraction of the receptor is expressed on the cell surface, whereas the majority is found in TGN and endosomes. Secondly, LRP9 cytoplasmic domain contains signal sequence involved in adaptor protein binding for endocytosis and Golgiendosome sorting. We and others have recently demonstrated that LRP9 interacts with AP1, AP2 (Doray et al. 2007) and GGA proteins (this paper). Thirdly, many receptors containing a DXXLL motif or a GGA-binding site, such as MPR (Tikkanen et al. 2000; Puertollano et al. 2001; Tortorella et al. 2007), Sortilin (Nielsen et al. 2001) and SorLA (Andersen et al. 2005; Nielsen et al. 2001), are sorting receptors that traffic between the TGN, endosomes and the cell surface. It would appear from the above that LRP9 might be a candidate sorting receptor, targeted for transport by ligands in the synthetic pathway as well as on the surface membrane. Investigations are presently underway to test this hypothesis and identify intracellular cargos for LRP9. Furthermore, there is increasing evidence of a prominent role for certain members of the LDLR superfamily in signal transduction pathways and cell physiology regulation (Strickland et al. 2002; May et al. 2007). This signalling occurs through cooperation with other cell surface molecules that associate directly or indirectly with LDL receptor family members and cytoplasmic adaptor molecules that interact to their cytoplasmic tails. The fact that signalling domains are present in the cytoplasmic tail of LRP9 suggests that it might also play a role in signal transduction. This hypothesis is supported by the work of Battle et al. (2003), who reported that various proteins involved in signal transduction interact with a juxtamembrane region of LRP12 that is conserved in LRP3 and LRP9. Further investigation is needed to clarify the biological functions of LRP9 and the other members of this LDLR subfamily. In conclusion, our study is the first to characterize the intracellular localization and trafficking of LRP9, an unusual member of the LDLR family. Its predominantly subcellular localization in the TGN and endosomes, together with its capacity to bind GGA proteins via two DXXLL motifs, suggest that LRP9 cycles between the TGN and endosomes through a GGA dependent-trafficking mechanism. These observations are important first steps in understanding the role of this uncharacterized receptor and constitute evidences pointing to a function of LRP9 in protein sorting and transport. 250 Acknowledgments We would like to thank Juan Bonifacino, Richard Leduc and T. Kitamura for DNA constructs and Marilyn Gist Farquhar for the generous gifts of reagents. We would also like to thank Eric Chevet, Catherine Denicourt and Richard Leduc for their constructive comments on the manuscript. This work was supported by grants from the Canadian Institutes for Health Research and a Canada Research Chair to C.L. References Andersen OM, Reiche J, Schmidt V, Gotthardt M, Spoelgen R, Behlke J, von Arnim CA, Breiderhoff T, Jansen P, Wu X (2005) Neuronal sorting protein-related receptor sorLA/LR11 regulates processing of the amyloid precursor protein. Proc Natl Acad Sci USA 102:13461–13466 Banting G, Ponnambalam S (1997) TGN38 and its orthologues: roles in post-TGN vesicle formation and maintenance of TGN morphology. Biochim Biophys Acta 1355:209–217 Battle MA, Maher VM, McCormick JJ (2003) ST7 is a novel low-density lipoprotein receptor-related protein (LRP) with a cytoplasmic tail that interacts with proteins related to signal transduction pathways. Biochemistry 42:7270–7282 Bonifacino JS (2004) The GGA proteins: adaptors on the move. Nat Rev Mol Cell Biol 5:23–32 Bork P, Beckmann G (1993) The CUB domain. A widespread module in developmentally regulated proteins. J Mol Biol 231:539–545 Chardin P, McCormick F (1999) Brefeldin A: the advantage of being uncompetitive. Cell 97:153–155 Christensen EI, Birn H (2002) Megalin and cubilin: multifunctional endocytic receptors. Nat Rev Mol Cell Biol 3:256–266PubMed Doray B, Bruns K, Ghosh P, Kornfeld S (2002) Interaction of the cation-dependent mannose 6-phosphate receptor with GGA proteins. J Biol Chem 277:18477–18482 Doray B, Lee I, Knisely J, Bu G, Kornfeld S (2007) The gamma/sigma1 and alpha/sigma2 hemicomplexes of clathrin adaptors AP-1 and AP-2 harbor the dileucine recognition site. Mol Biol Cell 18:1887–1896 Ghosh P, Kornfeld J (2004) The GGA proteins: key players in protein sorting at the trans-Golgi network. Eur J Cell Biol 83:257–262 Ghosh P, Griffith J, Geuze H, Kornfeld S (2003) Mammalian GGAs act together to sort mannose-6-phosphate receptors. J Cell Biol 163:755–766 He X, Chang WP, Koelsch G, Tang J (2002) Memapsin 2 (beta-secretase) cytosolic domain binds to the VHS domains of GGA1 and GGA2: implications on the endocytosis mechanism of memapsin 2. FEBS Lett 524:183–187PubMed He X, Li F, Chang WP, Tang J (2005) GGA proteins mediate the recycling pathway of memapsin 2 (BACE). J Biol Chem 280:11696–11703 He X, Zhu G, Koelsch G, Rodgers KK, Zhang XC, Tang J (2003) Biochemical and structural characterization of the interaction of memapsin 2 (beta-secretase) cytosolic domain with the VHS domain of GGA proteins. Biochemistry 42:12174– 12180 251 Herz J (2001) The LDL receptor gene family: (un) expected signal transducers in the brain. Neuron 29:571–581 Hirst J, Seaman MN, Buschow SI, Robinson MS (2007) The role of cargo proteins in GGA recruitment. Traffic 8:594–604 Huse JT, Pijak DS, Leslie GJ, Lee VM, Doms RW (2000) Maturation and endosomal targeting of beta-site amyloid precursor protein-cleaving enzyme. The Alzheimer’s disease beta-secretase. J Biol Chem 275:33729–33737 Hussain MM, Strickland DK, Bakillah A (1999) The mammalian low-density lipoprotein receptor family. Annu Rev Nutr 19:141–172 Jacobsen L, Madsen P, Jacobsen C, Nielsen MS, Gliemann J, Petersen CM (2001) Activation and functional characterization of the mosaic receptor SorLA/LR11. J Biol Chem 276:22788–22796 Jacobsen L, Madsen P, Nielsen MS, Geraerts WP, Gliemann J, Smit AB, Petersen CM (2002) The sorLA cytoplasmic domain interacts with GGA1 and -2 and defines minimum requirements for GGA binding. FEBS Lett 511:155–158 Kikuchi A, Yamamoto H, Kishida S (2007) Multiplicity of the interactions of Wnt proteins and their receptors. Cell Signal 19:659–671 Linstedt AD, Mehta A, Suhan J, Reggio H, Hauri HP (1997) Sequence and overexpression of GPP130/GIMPc: evidence for saturable pH-sensitive targeting of a type II early Golgi membrane protein. Mol Biol Cell 8:1073–1087PubMed May P, Woldt E, Matz RL, Boucher P (2007) The LDL receptor-related protein (LRP) family: an old family of proteins with new physiological functions. Ann Med 39:219–228 Misra S, Puertollano R, Kato Y, Bonifacino JS, Hurley JH (2002) Structural basis for acidic-cluster-dileucine sorting-signal recognition by VHS domains. Nature 415:933–937 Nakamura N, Rabouille C, Watson R, Nilsson T, Hui N, Slusarewicz P, Kreis TE, Warren G (1995) Characterization of a cis-Golgi matrix protein, GM130. J Cell Biol 131:1715–1726 Nielsen MS, Madsen P, Christensen EI, Anders Nykjær Gliemann J, Kasper D, Pohlmann R, Petersen CM (2001) The sortilin cytoplasmic tail conveys Golgi– endosome transport and binds the VHS domain of the GGA2 sorting protein. EMBO J 20:2180–2190 Offe K, Dodson SE, Shoemaker JT, Fritz JJ, Gearing M, Levey AI, Lah JJ (2006) The lipoprotein receptor LR11 regulates amyloid beta production and amyloid precursor protein traffic in endosomal compartments. J Neurosci 26:1596–1603 Puertollano R, Bonifacino JS (2004) Interactions of GGA3 with the ubiquitin sorting machinery. Nat Cell Biol 6:244–251 Puertollano R, Aguilar RC, Gorshkova I, Crouch RJ, Bonifacino JS. (2001) Sorting of mannose 6-phosphate receptors mediated by the GGAs. Science 292:1712–1716 Puri S, Bachert C, Fimmel CJ, Linstedt AD (2002) Cycling of early Golgi proteins via the cell surface and endosomes upon lumenal pH disruption. Traffic 3:641–653 Schaub BE, Berger B, Berger EG, Rohrer J (2006) Transition of galactosyltransferase 1 from trans-Golgi cisterna to the trans-Golgi network is signal mediated. Mol Biol Cell 17:5153–5162 252 Schmidt V, Sporbert A, Rohe M, Reimer T, Rehm A, Andersen OM, Willnow TE. (2007) SorLA/LR11 regulates processing of amyloid precursor protein via interaction with adaptors GGA and PACS-1. J Biol Chem 282:32956–32964 Shah S, Yu G (2006). sorLA: sorting out APP. Mol Interv 6(74–76):58 Stenmark H, Parton RG, Steele-Mortimer O, Lutcke A, Gruenberg J, Zerial M (1994) Inhibition of rab5 GTPase activity stimulates membrane fusion in endocytosis. Embo J 13:1287–1296PubMed Strickland DK, Gonias SL, Argraves WS. (2002) Diverse roles for the LDL receptor family. Trends Endocrinol Metab 13:66–74 Sugiyama T, Kumagai H, Morikawa Y, Wada Y, Sugiyama A, Yasuda K, Yokoi N, Tamura S, Kojima T, Nosaka T et al (2000) A novel low-density lipoprotein receptor-related protein mediating cellular uptake of apolipoprotein E-enriched beta-VLDL in vitro. Biochemistry 39:15817–15825 Taira K, Bujo H, Hirayama S, Yamazaki H, Kanaki T, Takahashi K, Ishii I, Miida T, Schneider WJ, Saito Y (2001) LR11, a mosaic LDL receptor family member, mediates the uptake of ApoE-rich lipoproteins in vitro. Arterioscler Thromb Vasc Biol 21:1501–1506 Takatsu H, Katoh Y, Shiba Y, Nakayama K. (2001) Golgi-localizing, gamma-adaptin ear homology domain, ADP-ribosylation factor-binding (GGA) proteins interact with acidic dileucine sequences within the cytoplasmic domains of sorting receptors through their Vps27p/Hrs/STAM (VHS) domains. J Biol Chem 276:28541–28545 Tikkanen R, Obermuller S, Denzer K, Pungitore R, Geuze HJ, von Figura K, Honing S (2000) The dileucine motif within the tail of MPR46 is required for sorting of the receptor in endosomes. Traffic 1:631–640 Tortorella LL, Schapiro FB, Maxfield FR. (2007) Role of an acidic cluster/dileucine motif in cation-independent mannose 6-phosphate receptor traffic. Traffic 8:402– 413 Wahle T, Prager K, Raffler N, Haass C, Famulok M, Walter J (2005) GGA proteins regulate retrograde transport of BACE1 from endosomes to the trans-Golgi network. Mol Cell Neurosci 29:453–461 Watson RT, Pessin JE. (2000) Functional cooperation of two independent targeting domains in syntaxin 6 is required for its efficient localization in the trans-golgi network of 3T3L1 adipocytes. J Biol Chem 275:1261–1268 Willnow TE (1999) The low-density lipoprotein receptor gene family: multiple roles in lipid metabolism. J Mol Med 77:306–315 Xu Y, Martin S, James DE, Hong W (2002) GS15 forms a SNARE complex with syntaxin 5, GS28, and Ykt6 and is implicated in traffic in the early cisternae of the Golgi apparatus. Mol Biol Cell 13:3493–3507 Yamamoto TDC, Brown MS, Schneider WJ, Casey ML, Goldstein JL, Russell DW (1984) The human LDL receptor: a cysteine-rich protein with multiple Alu sequences in its mRNA. Cell 39:27–38 Zhu Y, Doray B, Poussu A, Lehto VP, Kornfeld S. (2001) Binding of GGA2 to the lysosomal enzyme sorting motif of the mannose 6-phosphate receptor. Science 292:1716–1718 253 ANNEXE 2 Calnuc Binds to LRP9 and Affects its Endosomal Sorting. Brodeur, J., Larkin, H., Boucher, R., Gagnon, H., Chayer St-Louis, S., and Lavoie, C. Traffic 10(8):1098-114 (2009) Travaux de maîtrise IF = 6,255 254 Calnuc Binds to LRP9 and Affects its Endosomal Sorting Brodeur, J., Larkin, H., Boucher, R., Gagnon, H., Chayer St-Louis, S., and Lavoie, C. Abstract Calnuc is an ubiquitous Ca2+-binding protein found in the cytoplasm where it binds different Gα subunits, in the Golgi lumen where it constitutes a major Ca2+ storage pool, and outside the cell. We identified LDLR-related protein 9 (LRP9) as the first transmembrane protein shown to interact directly with Calnuc. LRP9 is a member of a new subfamily of the LDLR superfamily that cycles between the trans-Golgi network (TGN) and endosomes through a mechanism dependent on clathrin adaptor GGA proteins. The aim of the present study was to characterize the interaction between Calnuc and LRP9. Various biochemical assays showed that the N-terminus of Calnuc interacts with an arginine-rich region in the cytosolic tail of LRP9. Confocal microscopy showed that Calnuc colocalizes with LRP9 at the surface of the TGN and early endosomes. Depletion of Calnuc by small interfering RNA (siRNA) missorted LRP9 in the late endosome/lysosome compartments and enhanced its lysosomal degradation. This phenotype was rescued by the expression of siRNA-resistant wild-type Calnuc as well as cytoplasmic Calnuc, indicating that the cytoplasmic pool of Calnuc is involved in LRP9 endosomal sorting to prevent the delivery of LRP9 to lysosomes. This is the first report showing that Calnuc plays a role in receptor trafficking. Keywords Calnuc; endosomal sorting; GGA; LDLR-related protein 9; lysosome; nucleobindin; trans-Golgi network Introduction Calnuc, also known as nucleobindin, is an ubiquitous (1–3), EF-hand calciumbinding protein that is conserved from flies to humans, suggesting that it has important biological functions (3,4). Indeed, the modulation of Calnuc expression is associated with various pathologies, including cancer (5–10), lupus (11,12) and Alzheimer's (13). Calnuc is a multifunctional protein that contains a signal peptide (SP), a putative DNA-binding domain, a leucine zipper and two EF-hand motifs that possess the ability to bind Ca2+ (14). It is also a multicompartmental protein, with Golgi luminal, extracellular and cytosolic pools. This is not unique to Calnuc because other proteins with a SP, such as calreticulin, also localize in different compartments (15). The Golgi pool is localized in the lumen of cis-Golgi cisternae, where it constitutes the major Ca2+-binding protein (3) and a major Ca2+ storage pool (16). After a long period of retention in the Golgi, Calnuc is secreted via constitutive and constitutive-like pathways (17). The role of the extracellular pool is poorly characterized but has been involved in bone matrix maturation (18). Information on the functions of cytosolic Calnuc is just beginning to emerge. To date, a few partners that interact with this pool of Calnuc have been identified (Gαi subunits, necdin) (3,13,19), but the functional roles of these interactions are unknown. 255 In the present study, we carried out a yeast-two-hybrid screening to identify new partners that interact with Calnuc to better understand the intracellular roles of Calnuc. One of these new interacting proteins was LDLR-related protein 9 (LRP9). LRP9 is a type I membrane protein and a member of a new subfamily of the LDLR superfamily that includes two other members, LRP3 and LRP12 (previously named ST7) (20). Like all LDLR superfamily proteins, this new subgroup contains LDLRA domain repeats in their extracellular domains (21,22) that function as ligand-binding sites (23). However, several unique structural features distinguish that the new LRP subfamily from prototypical members of the LDLR superfamily (20), including a relatively large cytoplasmic tail that contains a proline-rich (P-rich) domain and acidic dileucine (DXXLL) motif, does not possess EGF-like repeats and YWTD-containing domains and has CUB domains in the extracellular region. These structures are found in a diverse array of functionally unrelated extracellular proteins (24) and are involved in ligand binding (25). Little is known about LRP9 apart from the fact that it is expressed in various tissues, may be involved in apolipoprotein E internalization (26), and is mainly localized in the trans-Golgi network (TGN) and endosomes rather than the plasma membrane (27,28). Two acidic cluster/dileucine (DXXLL) motifs in the cytoplasmic tail of LRP9 interact with clathrin adaptor GGA proteins and are involved in transporting it between the TGN and endosomes (27,28). This suggests that LRP9 may play a role in the trafficking of ligands between these intracellular compartments much like the mannose 6-phosphate receptor, which also contains a GGA-binding site and traffics between the TGN and endosomes (29– 31). LRP9 is the first transmembrane protein shown to interact directly with Calnuc. In the present study, we investigated the nature of the interaction between Calnuc and LRP9 and characterized the role of Calnuc in LRP9 intracellular trafficking. Results LDLR-related protein 9 (LRP9) is a novel Calnuc-interacting protein A yeast-two-hybrid system was employed to screen for new protein(s) that interact with Calnuc. We used the conserved N-terminus of Calnuc (residues 25–172, Figure 1A) as bait and an adult human kidney cDNA library as the source of interacting proteins. Several clones were isolated, sequenced and analyzed. Two had an identical 1400-bp insert and shared 100% homology with a portion of the cytoplasmic tail of human LDL-related protein 9 (LRP9) (Figure 1B). To confirm the interaction between Calnuc and LRP9, we performed yeast one-on-one interaction experiments. Yeast strain AH109 was transformed with vectors expressing various portions of Calnuc fused to the Gal4-DNA-binding domain together with a vector expressing the Gal4-DNA activation domain fused to the C-terminal portion of human LRP9 fished out by the yeast-two-hybrid screen (residues 556–713) (Table 1). The clones containing the N-terminal region of Calnuc (Figure 1A) and LRP9556−−713 grew on histidine-deficient plates and had β-galactosidase activity (Table 1). However, little or no interaction was detected when C-terminal portions of Calnuc (with or without the EF-hands ; Figure 1A) or the PXA domain (control) were used as bait, 256 indicating that the interaction was specific to the N-terminal region of Calnuc (Table 1). Full-length Calnuc could not be used in this assay because it self-activated the yeast reporter genes. Figure 1. Domains and motifs of Calnuc and the cytoplasmic tail of LRP9. A) Schematic representation of Calnuc showing the N-terminal SP, the putative DNAbinding domain (basic region), the two EF-hands (EF−1 and EF−2) and the leucine zipper motif. B) Comparison of the amino acid sequences of the cytoplasmic tail of LRP9 and the clone fished out by yeast-two-hybrid (TH52p). YXXϕ and the acidic cluster dileucine (DXXLL) motifs, two putative sorting signals, an R-rich domain and a P-rich, which are potential protein–protein interactions sites, are indicated in the sequences. m, mouse; h, human; TM, transmembrane. Table 1. Calnuc and LRP9 interaction in yeast-two-hybrid. 257 LRP9 interacts with the N-terminus of Calnuc We next verified the in vitro and in vivo interactions between Calnuc and LRP9 using pull-down and immunoprecipitation assays. 35S-labeled in vitro-translated LRP9556−−713 bound to glutathione S-transferase (GST)-Calnuc-full-length and GSTCalnuc-N-terminus but not to GST-Calnuc-C-terminus or GST alone (Figure 2A). Similarly, GST-Calnuc-full-length and GST-Calnuc-N-terminus were able to pull-down full-length LRP9 from the lysate of HEK293 cells that overexpressed hemagglutinin (HA)tagged LRP9 (Figure 2B). However, no interactions were detected with GST-Calnuc-Cterminus or GST alone (Figure 2B). To confirm that the interaction between Calnuc and LRP9 also occurs in vivo, we performed immunoprecipitation experiments on HEK293 cells transfected with green fluorescent protein (GFP)-tagged Calnuc and HA-tagged LRP9. When we performed the immunoprecipitation with anti-HA immunoglobulin (IgG), Calnuc-GFP coprecipitated with HA-LRP9 (Figure 2C). Similarly, when we used anti-GFP IgG, LRP9-HA co-precipitated with Calnuc-GFP in cells cotransfected with both proteins (Figure 2C). Calnuc-GFP in the cell lysate is visible as two bands, the lower of which likely represents an N-terminal cleavage product (Figure 2C). It is interesting to note that the lower band did not bind LRP9 (Figure 2C), which provides further support for the importance of the N-terminal portion of Calnuc in the interaction with LRP9. The conformation of Calnuc has been reported to change when it associates with calcium. However, the amount of Calnuc that coimmunoprecipitated with LRP9 was the same whether calcium was present or not (data not shown). These findings suggest that the Nterminal region of Calnuc (residues 25–172) interacts with the cytoplasmic tail of LRP9, both in vitro and in vivo and that the interaction is calcium independent. Figure 2. Calnuc interacts with LRP9. A) In vitro interaction of Calnuc N-terminus with the cytoplasmic tail of LRP9. The in vitro-translated 35S-labeled C-tail of LRP9 bound to GST-Calnuc-N-terminus (N-term) and GST-Calnuc-full-length (FL) but not to GST-Calnuc-C-terminus (C-term) or GST alone. GST proteins (10 μg each) immobilized on glutathione beads were incubated with in vitrotranslated human LRP9 (residues 556–713). Bound proteins were separated by SDS-PAGE 258 and detected by autoradiography. Input equals 7% of the total in vitro-translated product. B) In vivo interaction of Calnuc N-terminus with HA-tagged LRP9-full-length. LRP9-HA bound to GST-Calnuc-N-terminus (N-term) and GST-Calnuc-full-length (FL) but not to GST-Calnuc-C-terminus (C-term) or GST alone. GST proteins (30 μg each) immobilized on glutathione beads were incubated with lysate from HEK cells transfected with mouse LRP9-HA. Bound proteins were separated by SDS-PAGE and detected with anti-HA antibodies. C) Coimmunoprecipitations of LRP9-HA with Calnuc-GFP protein. Lysates from HEK cells transfected with LRP9-HA and Calnuc-GFP were immunoprecipitated with anti-HA (lane 2), anti-GFP (lane 4) or control (lanes 1 and 3) antibodies and then immunoblotted with anti-HA or anti-GFP to detect LRP9 and Calnuc, respectively. CalnucGFP is seen as two bands, the lower of which probably represents an N-terminal cleavage product. Note that only the 95 kDa band is coimmunoprecipitated. Quantification of the band intensity revealed that approximately 2% of LRP9-HA immunoprecipitated with Calnuc-GFP and <1% of the total Calnuc-GFP were recovered in the LRP9-HA immunoprecipitate. Nevertheless, coimmunoprecipitations in both directions were highly specific. Figure 3. Calnuc interacts with LRP3 but not with LRP1. Lysates from HEK cells transfected with GFP or Calnuc-GFP together with (A) LRP3FLAG or (B) LRP1-Myc were immunoprecipitated with anti-GFP antibodies and then immunoblotted with anti-FLAG, anti-Myc, or anti-GFP to detect LRP3, LRP1 and Calnuc, respectively. 259 To verify the specificity of the Calnuc–LRP9 interaction, we next looked at whether Calnuc bound to LRP1 and LRP3, two other members of the LDLR family. LRP3 is a member of the same LDLR subfamily as LRP9, whereas LRP1 is a member of a different subfamily (20). Immunoprecipitations with anti-GFP IgG were performed on HEK293 cells expressing GFP or GFP-Calnuc together with FLAG-tagged LRP3 or Myc-tagged LRP1. As shown in Figure 3, LRP3 bound to Calnuc, whereas LRP1 did not display any appreciable binding to Calnuc. These results suggest that Calnuc interacts with LRP9 and LRP3, a related family member, perhaps via a common binding domain. LRP9 interacts with endogenous Calnuc We next looked at whether endogenous Calnuc interacts with LRP9 in HEK293 cells transfected with HA-tagged LRP9. When we performed the immunoprecipitation with anti-Calnuc, a fraction of LRP9-HA coimmunoprecipitated with Calnuc, confirming the occurrence and specificity of the LRP9 interaction with endogenous Calnuc in intact cells (Figure 4A). Because endogenous Calnuc is found in both cytosolic and membrane fractions (3), we next looked at whether the interaction with LRP9 occurs specifically with the cytosolic pool of endogenous Calnuc. Using subcellular fractionation, we prepared a cytosolic fraction from HeLa cells. As observed previously (3), endogenous Calnuc was distributed in both membrane (P100) and cytosolic (S100) fractions (Figure 4B, lanes 1–2). However, the transmembrane receptor mannose−6-phosphate receptor (MPR) was detected mainly in the P100 fraction, whereas the cytosolic protein EEA1 was only detected in the S100 fraction (Figure 4B, lanes 1–2). The ability of Calnuc in the S100 fraction to interact with the C-tail of LRP9 was next assessed using pull-down binding assays. GST-LRP9-Ctail (GST-LRP9-CT1 in Figure 5A) was able to pull-down endogenous cytosolic Calnuc (Figure 4B, lane 4). However, only little interaction was detected with GST alone (Figure 4B, lane 3). These results indicated that endogenous cytosolic Calnuc interacts with the LRP9 C-tail. Calnuc interacts with an arginine-rich (R-rich) region in the cytoplasmic tail of LRP9 The cytoplasmic tail of LRP9 contains many putative protein–protein interaction sites (Figure 5A). To identify the specific region within LRP9 required for the interaction with Calnuc, several deletion mutants of the C-terminal tail of LRP9 were generated (Figure 5A). The ability of these mutants to interact with Calnuc was assessed using pulldown binding assays. GST-LRP9-CT1, CT3 and CT4 pulled down Calnuc from the lysate of HEK293 cells that expressed Calnuc-GFP (Figure 5B). In contrast, no interaction was detected with GST-LRP9-CT2, CT5, CT6, or GST alone (Figure 5B). The complete Cterminus of LRP9 (GST-CT) could not be used because the fusion protein is degraded when expressed in bacteria. The only domain present in all the deletion mutants that bound to Calnuc was the R-rich domain of the cytoplasmic tail of LRP9, indicating that it is major determinant for the interaction between LRP9 and Calnuc. Interestingly, this R-rich region is also present in LRP3 but not in LRP1. To confirm the importance of this region for the interaction with Calnuc in vivo, a LRP9 mutant lacking the R-rich domain (ΔR-LRP9) was generated and tested for its ability to coimmunoprecipitate with Calnuc. Surprisingly, we observed that the ΔR-LRP9 mutant had not lost its capacity to interact with Calnuc (data 260 not shown). This suggests that the R-rich domain of LRP9 might be important but not crucial for the interaction with Calnuc in vivo. Figure 4. Endogenous Calnuc binds to the cytoplasmic tail of LRP9. A) Coimmunoprecipitation of LRP9-HA with endogenous Calnuc. Lysates from HEK cells transfected with LRP9-HA or a control vector (pCMV-HA) were immunoprecipitated with anti-Calnuc antibodies and then immunoblotted with anti-HA (LRP9) or anti-Calnuc antibody. Quantification of the band intensity revealed that <1% of the total LRP9-HA coprecipitated with endogenous Calnuc. B) Interaction of cytoplasmic endogenous Calnuc with LRP9-C-tail. Endogenous Calnuc is found in approximately equal amounts in both membrane (P100, lane 1) and cytosolic (S100, lane 2) fractions obtained by subcellular fractionation of HeLa cells. Calnuc from the S100 fraction bound to GST-LRP9-C-tail (GST-CT1 described in Figure 5) (lane 4) but bound only poorly to GST alone (lane 3). GST proteins (10 μg each) immobilized on glutathione beads were incubated with the S100 fraction from HeLa cells. Input (lane 2) equals 4% of the total S100 used for pull-down. Bound proteins were separated by SDS-PAGE and detected with anti-Calnuc antibody. Calnuc and LRP9 colocalize on Golgi cisternae To determine whether LRP9 colocalizes with Calnuc, we compared the intracellular distribution of HA-tagged LRP9 with endogenous Calnuc and GFP-tagged Calnuc in COS7 cells using confocal microscopy. LRP9-HA was mainly detected in the juxtanuclear region and surrounding vesicles (Figure 6A,D,G). This is in agreement with our previous findings showing that LRP9 localizes mainly in the TGN and, to a lesser extent, in endosomes (27). 261 Endogenous Calnuc was concentrated in the Golgi region (Figure 6B). Indeed, previous immunofluorescence (IF) and immunoelectronmicroscopy studies precisely localized Calnuc in the cis-Golgi cisternae (3). This suggests that endogenous Calnuc is mainly concentrated and detected in the lumen of the cis-Golgi, whereas the cytoplasmic pool of endogenous Calnuc is undetectable by IF. There was only partial colocalization between LRP9-HA and endogenous Calnuc because both proteins were located in different stacks of the Golgi apparatus (Figure 6C). In cells overexpressing Calnuc-GFP (Figure 6D–F), two intracellular pools of Calnuc were detected by IF, as previously reported (3,17,32). A major pool was localized in the lumen of the cis-Golgi cisternae and a minor pool in the cytoplasm (Figure 6E). After double labeling, LRP9-HA partially overlapped with CalnucGFP in the Golgi cisternae (Figure 6F) as observed with endogenous Calnuc (Figure 6C). Given that Calnuc interacted with the cytoplasmic tail of LRP9, we next asked whether the specific cytosolic pool of Calnuc and LRP9 colocalized. To address this question, we conducted confocal analyses of COS7 cells transfected with LRP9-HA and Calnuc-GFP without its SP (ΔSP-Calnuc-GFP). This deletion resulted in the expected defects in targeting Calnuc in the lumen of the Golgi and its accumulation in the cytoplasm (32). ΔSP-Calnuc-GFP was detected in the cytoplasm and on the plasma membrane (Figure 6H) and clearly colocalized with LRP9-HA on the surface of the Golgi cisternae and surrounding vesicles (Figure 6I). Together, these IF and biochemical interaction assays suggested that the cytoplasmic pool of Calnuc interacts with the C-terminal tail of LRP9 on the cytoplasmic side of the TGN cisternae. Figure 5. Calnuc interacts with the R-rich region of LRP9. A) Schematic representation of the various GST-LRP9 proteins used for the pull-down assays. R-rich, arginine-rich domain; P-rich, proline-rich domain; PDZb, PDZ-binding domain motif; DXXLL, acidic dileucine motif. B) The deletion mutants of LRP9 shown in A were bound to glutathione beads and incubated with lysate from HEK cells transfected with Calnuc-GFP. Bound proteins were separated by SDS-PAGE and detected with antiGFP antibody. Calnuc-GFP bound to the deletion mutants spanning residues 92–115 containing the R-rich region (CT1, CT3, CT4) but not to the mutants lacking this region (CT2, CT5, CT6) or to GST alone. The additional Calnuc-GFP bands detected in the cell lysate by the polyclonal anti-GFP antibody may probably correspond to an N-terminal cleavage product. Note that only the 95 kDa band was pulled down. 262 Figure 6. Calnuc and LRP9 colocalize on the Golgi apparatus. In COS7 cells transfected with (A–C) LRP9-HA alone or with (D–I) wild-type CalnucGFP, LRP9-HA was found in the juxtanuclear region and surrounding vesicles (inset, A,D,G). A–C) Endogenous Calnuc was found mainly in the Golgi region (inset, B) where it partially overlapped with LRP9-HA (inset, C). D–F) Calnuc-GFP was found mainly inside the Golgi cisternae (inset, E) and, to a lesser extent, in the cytoplasm (E). The merged image (yellow) shows a partial overlap between LRP9-HA and Calnuc-GFP in the Golgi cisternae (inset, F). G–I) COS7 cells transfected with LRP9-HA and Calnuc devoid of its SP (ΔSP-Calnuc-GFP). Deleting the signal sequence of Calnuc (residues 2–25) resulted in the expected defect in targeting in the Golgi lumen and an accumulation in the cytoplasm. ΔSP-Calnuc-GFP was also detected on the plasma membrane (arrows, H) and the cytosolic side of the Golgi cisternae (arrowhead and inset, H) where it colocalized with LRP9-HA (yellow, inset I). Twelve hours after the transfection, the cells were fixed, permeabilized and immunostained using (A, D, G) anti-HA, (B) anti-Calnuc and (E, H) anti-GFP antibodies. The labeled cells were examined by confocal fluorescence microscopy. Scale bar, 10 μm. 263 Calnuc colocalizes with LRP9 on early endosomes A minor pool of LRP9 has been previously reported to be present in early endosomes (27,28). To determine whether Calnuc is present on the surface of endosomes and whether it colocalizes with LRP9 on this organelle, we performed confocal IF on COS7 cells expressing ΔSP-Calnuc-GFP alone or with HA-LRP9. We started by comparing the subcellular localization of ΔSP-Calnuc-GFP with the early endosome marker EEA1 (Figure 7A–C). ΔSP-Calnuc-GFP was detected on the plasma membrane and in the Golgi region as well as on some fine, punctate structures distributed throughout the cytoplasm (Figure 7A) that partially overlapped with EEA1 (Figure 7C). This colocalization was consistent with the presence of Calnuc on early endosomes. Figure 7. Calnuc and LRP9 colocalize on early endosomes. A–C) In COS7 cells transfected with ΔSP-Calnuc-GFP (cytoplasmic form of Calnuc) alone, ΔSP-Calnuc-GFP was distributed on the plasma membrane (arrow, A), the cytosolic side of the Golgi (arrowhead, A), and small vesicular structures in the cytoplasm (inset, A). EEA1 was found on early endosomes distributed throughout the cell (B). Merged image showing occasional overlap in the vesicular distribution of ΔSP-Calnuc-GFP and EEA1 (inset C). D–F) COS7 cells were cotransfected with ΔSP-Calnuc-GFP, HA-tagged LRP9 and Rab5Q79L, a Rab5 GTPase-deficient mutant that creates enlarged early endosomes. 264 LRP9-HA was present in enlarged endosomes (E). ΔSP-Calnuc-GFP was present in the cytoplasm (D), on the plasma membrane (arrow, D) and on the enlarged endosomes (inset, D) labeled with LRP9-HA (inset, F). G–I) COS7 cells were transfected with ΔSP-CalnucGFP and LRP9-HA in which the two DXXLL motifs were mutated (LRP9−2DXXAA). The mutation induced a redistribution of LRP9 to the plasma membrane (arrow, H) and early endosomes (inset, H). ΔSP-Calnuc-GFP was present in the cytoplasm, on the cytosolic side of the Golgi (arrowhead, G), and together with the HA-LRP9−2DXXAA mutant on the plasma membrane (arrow, I) and endosomes (inset, I). Twelve hours after the transfection, the cells were fixed, permeabilized and immunostained using (E, H) anti-HA, (B) anti-EEA1 and (A,D,G) anti-GFP antibodies. The stained cells were examined by confocal fluorescence microscopy. The yellow color in the merged images (C, F, I) indicates colocalization. Scale bar, 10 μm. To determine whether Calnuc colocalized with LRP9 on endosomes, we enriched LRP9 in early endosomes and compared its localization with that of Calnuc. We first trapped LRP9-HA in enlarged endosomes (Figure 7E) by transfecting COS7 cells with an established Rab5 GTPase-deficient mutant (Rab5Q79L) that greatly enlarges early endosomes because of increased homo- and heterotypic fusion (33). LRP9-labeled enlarged vesicles have been previously shown to colocalize with EEA1 (27). ΔSP-Calnuc-GFP colocalized with LRP9-HA on subregions of the enlarged endosomes (Figure 7F), suggesting that they are present together on early endosomes. To confirm this finding, we next transfected COS7 cells with HA-tagged LRP9−2DXXAA, an LRP9 mutant in which the leucines of both DXXLL motifs (Figure 1B) are substituted by alanines. This mutant, which is incapable of binding GGA proteins, mislocalized on the plasma membrane and early endosomes (Figure 7H) (27,28). Figure 7(G–I) shows that ΔSP-Calnuc-GFP colocalized with LRP9−2DXXAA on the plasma membrane and peripheral early endosomes. From these experiments, we concluded that a fraction of Calnuc is present on endosomes where it partially colocalizes with LRP9, suggesting that the cytoplasmic pool of Calnuc can also bind LRP9 on early endosomes. 265 Figure 8. Knocking down Calnuc affects the intracellular distribution of LRP9. LRP9 localized mainly in the TGN region in control cells (B), but knocking down Calnuc induced a redistribution of LRP9 to peripheral vesicles (D). HeLa cells were transfected with (A, B) control or (C, D) Calnuc siRNA for 3 days. LRP9-HA cDNA was transfected for 12 h before the IF experiment. The cells were fixed, permeabilized and immunostained using (A, C) anti-Calnuc and (B, D) anti-HA antibodies. Scale bar, 10 μm. Calnuc knockdowns redistribute LRP9 to late endosomes/lysosomes To investigate the functional role of Calnuc in LRP9 trafficking, we studied the subcellular distribution of LRP9 in Calnuc-depleted cells. HeLa cells were transfected with control or Calnuc small interfering RNA (siRNA) together with HA-tagged LRP9 and were analyzed by confocal microscopy. Calnuc was significantly (≥ 90%) depleted 72 h after transfection with specific siRNA, as observed in the IF (Figure 8) and immunoblotting experiments (Figure 10). In Calnuc knockdown cells, LRP9 was mainly localized in peripheral punctate structures (Figure 8D) rather than in the TGN (Figure 8B). This did not appear to be because of gross changes in TGN morphology, because the localization of TGN46, a TGN marker, was not altered (Figure S1). 266 Figure 9. LRP9 redistributes in late endosomes/lysosomes in Calnuc-depleted cells. 267 A) Comparison of LRP9 and EEA1 distribution in Calnuc-depleted cells. LRP9-HA was localized in the juxtanuclear region and surrounding vesicles in HeLa cells transfected with control siRNA (a). Some of the LRP9-containing vesicles colocalized with EEA1 in the Golgi area (inset, c) but not at the periphery of the cells. In cells transfected with Calnuc siRNA (d–f), LRP9-HA redistributed into peripheral vesicles (d) but did not colocalize mainly with EEA1 (inset, f). Arrowheads indicate the Golgi region. B) Comparison of LRP9 and LAMP−2 distribution in Calnuc-depleted cells. LRP9-HA was found in the Golgi region and surrounding vesicles in HeLa cells transfected with control siRNA (a) and was relatively rare in the LAMP−2-labeled compartment (b–c). In Calnuc siRNAtransfected cells (d–f), LRP9-HA was mainly found in peripheral punctate structures (d) and significantly colocalized with LAMP−2 (inset, f). The cells were fixed, permeabilized and immunostained using anti-HA, anti-EEA1 and anti-LAMP−2 antibodies. The stained cells were examined by confocal fluorescence microscopy. The yellow color in the merged images indicates colocalization. Scale bar, 10 μm. C) Quantification of the degree of overlap between LRP9-HA and EEA1 or LAMP−2 in cells treated with control or Calnuc siRNA as described in (A) and (B). The histogram depicts relative colocalization R values calculated as described in Materials and Methods. Values are significant at p < 0.001 (**) for control versus Calnuc knockdown cells. Data represent means ± SD (n≥10) for each condition. Figure 10. Knocking down Calnuc affects the levels of LRP3 and LRP9. The steady-state levels of LRP3 and LRP9 but not LRP1 were reduced in Calnuc-depleted cells. HeLa cells were transfected with control or Calnuc siRNA for 3 days. The cells were also transfected with LRP1-Myc, LRP3-FLAG or LRP9-HA cDNA for 12 h before the experiment. Cells were lysed and proteins were separated by SDS-PAGE and detected with specific anti-Myc, anti-FLAG, anti-HA, anti-Calnuc and anti-actin antibodies. Actin served as a loading control. Calnuc was below the detection level in samples treated with Calnucspecific siRNA. 268 To better define the specific vesicular compartment in which LRP9 redistributes in the absence of Calnuc, the cells were transfected with control or Calnuc siRNAs and the localization of LRP9-HA was compared to endosomal markers by confocal microscopy (Figure 9). Colocalization was evaluated quantitatively by measuring Pearson correlation coefficient (R) values, which are standard measures of colocalization (34). Because we previously observed that LRP9 can redistribute to early endosomes (27), we first compared the localization of LRP9 to that of the early endosome marker EEA1 (Figure 9A). In control cells, as previously reported (27), LRP9 partially colocalized with EEA1 on vesicles in the Golgi area (R = 0.29 ± 0.07) [Figure 9A(a–c),C]. In Calnuc-depleted cells, the colocalization of EEA1 and LRP9-HA did not increase (R = 0.26 ± 0.05) [Figure 9A(d– f),C]. We next compared the colocalization of LRP9-HA and LAMP−2, a late endosome/lysosome marker. Figure 9B(a–c) shows that LRP9 partially colocalized with LAMP−2-labeled structures in control cells (R = 0.33 ± 0.06). However, the colocalization of LRP9 with LAMP−2 on vesicles significantly increased in Calnuc knockdown cells (R = 0.53 ± 0.09) [Figure 9B(d–f),C]. In summary, the depletion of Calnuc redistributed LRP9 from the TGN to peripheral LAMP−2 labeled vesicles. Knocking down Calnuc thus caused missorting of LRP9 to late endosomes/lysosomes. These observations suggested that Calnuc participates in the intracellular trafficking or sorting of LRP9. Calnuc knockdown increases the lysosomal turnover of LRP9 Because Calnuc depletion shifted LRP9 towards late endosomes/lysosomes, the level of LRP9 present in cell lysates of control- and siRNA-treated cells was next examined by immunoblotting (Figure 10). Following a 72-h treatment with Calnuc siRNA, the levels of LRP9 were significantly reduced compared with control cells (Figure 10). Because Calnuc interacted with LRP3 but not with LRP1 (Figure 3), we also examined the effect of knocking down Calnuc on the level of these two LRPs. Interestingly, the depletion of Calnuc decreased the amount of LRP3 but had no effect on LRP1 (Figure 10). These results suggest that the interaction with Calnuc is important for the proper sorting of LRP3 and LRP9. We reasoned that the reduction in LRP9 levels might be because of the degradation of the receptor in lysosomes. To test this hypothesis, we performed cycloheximide chase experiments in which the levels of LRP9 in control- and siRNA-treated cells were examined at different times after inhibiting protein synthesis (Figure 11A). In mock-treated cells, LRP9 had a half-life of 4 h, whereas in Calnuc siRNA-treated cells, the half-life of LRP9 was reduced to <2 h (Figure 11A). To determine whether this degradation occurred in lysosomes, we incubated the siRNA-treated cells with the lysosomal inhibitors leupeptin and E64. In the absence of lysosomal inhibitors, 48% of the LRP9 was degraded in Calnuc knockdown cells, whereas, in the presence of lysosomal inhibitors, only 13% of the LRP9 was degraded (Figure 11B). Treatments with lysosomal inhibitors thus significantly prevented a decrease in LRP9 levels (Figure 11B). Taken together, these observations suggest that depleting Calnuc increases the delivery of LRP9 to lysosomes. 269 Figure 11. LRP9 is degraded in lysosomes when Calnuc is depleted. A) The half-life of LRP9 is shorter in Calnuc-depleted cells than in control cells. HeLa cells were transfected with control or Calnuc siRNA for 3 days, and with LRP9-HA cDNA for 12 h before the experiment. Cells were incubated with 40 μg/mL of cycloheximide, an inhibitor of protein synthesis, for different periods of time. Cells were lysed and analyzed by immunoblotting. B) LRP9 was degraded in lysosomes in Calnuc-depleted cells. Control and Calnuc-suppressed cells transfected with LRP9-HA [as described in (A)] were treated or not with the lysosomal inhibitors leupeptin (1 mg/mL) and E64 (10 μg/mL) for 3 h. The cells were lysed and the proteins were separated by SDS-PAGE and detected with specific anti-Calnuc, anti-HA (LRP9), anti actin or anti-tubulin antibodies. Tubulin and actin were used as loading controls. (Bottom) Bar graph showing the quantification of LRP9 degradation in cells treated as described in (B) and expressed as a percentage of LRP9 present in control cells. Results are shown as means ±sd (n = 3). *p < 0.05 (compared with control cells). Calnuc siRNA phenotype is rescued by Calnuc and ΔSP-Calnuc overexpression To confirm that the consequence of Calnuc depletion on LRP9 sorting is because of a lack of Calnuc binding to the LRP9 cytoplasmic domain, we performed rescue experiments with siRNA-resistant forms of wild-type Calnuc and cytosolic Calnuc (ΔSP270 Calnuc). Rat Calnuc (rCalnuc) was used as a rescue construct because it differs in many nucleotides from human siRNA target sequences, rendering it siRNA resistant. We first compared the distribution of LRP9-HA in Calnuc siRNA-treated cells overexpressing GFP, GFP-tagged wild-type rCalnuc, or GFP-tagged ΔSP-rCalnuc (Figure 12A,B). In control (GFP) cells, LRP9 was mainly localized in peripheral punctate structures that strongly colocalized with LAMP−2 (R = 0.53 ± 0.08) [Figure 12A(a–d),B]. However, in cells rescued with rCalnuc-GFP or ΔSP-rCalnuc-GFP, LRP9-HA was mainly concentrated in the TGN region and very little LRP9 was observed in LAMP−2-labeled vesicles [Figure 12A(e–h,i–l)]. The colocalization values (R) of LRP9 with LAMP−2 decreased significantly to 0.28 ± 0.02 and 0.30 ± 0.05 in rCalnuc-GFP- and ΔSP-rCalnuc-GFPrescued cells, respectively (Figure 12B). These values were similar to those observed in control siRNA-treated cells (R = 0.33 ± 0.06, Figure 9C). As expected, partial colocalization between LRP9-HA and Calnuc-GFP or ΔSP-Calnuc-GFP was observed in the Golgi region of these rescued cells (Figure S2) Figure 12. Expression of cytoplasmic Calnuc restores the normal distribution of LRP9. 271 A) Comparison of LRP9 and LAMP−2 distribution in Calnuc-depleted cells rescued with GFP (a–d), siRNA-resistant versions of full-length Calnuc-GFP (e–h) or cytoplasmic Calnuc-GFP (ΔSP-Calnuc-GFP) (i–l). LRP9-HA was mainly found in peripheral vesicles (b) that colocalized with LAMP−2 in cells transfected with GFP (inset, d). However, LRP9-HA was found in the Golgi region and surrounding vesicles (f and j) that were occasionnaly labeled with anti-LAMP−2 in cells rescued with Calnuc-GFP (inset, h) or ΔSP-Calnuc-GFP (inset, l). The cells were fixed, permeabilized and immunostained using anti-HA and anti-LAMP−2 antibodies. The expression of GFP proteins was monitored using GFP fluorescence and was pseudocolored blue. The stained cells were examined by confocal fluorescence microscopy. The yellow color in the merged images indicates colocalization between LRP9-HA and LAMP−2. Scale bar, 10 μm. B) Quantification of the degree of overlap between LRP9-HA and LAMP−2 in cells treated as described above. The histogram depicts relative colocalization R values calculated as described in Materials and Methods. Values are significant at p < 0.001 (**) for control (GFP) versus Calnuc rescue cells. Data represent means ± SD (n≥10) for each condition. We next looked at whether the degradation of LRP9 was reversed when Calnuc was restored in Calnuc siRNA-treated cells. In cells rescued with rCalnuc-GFP or ΔSP-rCalnucGFP, the levels of LRP9 increased by 44.8 and 48.9%, respectively, compared with cells rescued with control GFP (Figure 13A,B). In summary, the redistribution and degradation of LRP9 in lysosomes observed in Calnuc-depleted cells could be significantly reversed by expressing siRNA-resistant wild-type Calnuc or cytosolic Calnuc. These results suggested that the cytoplasmic pool of Calnuc participates in the intracellular trafficking/sorting of LRP9 and provided support for a direct role for the interaction of Calnuc with the cytoplasmic tail of LRP9. Figure 13. Expression of cytoplasmic Calnuc rescues the lysosomal degradation of LRP9. 272 Steady-state levels of LRP9 are increased in Calnuc-depleted cells rescued with siRNAresistant full-length Calnuc-GFP or cytoplasmic Calnuc-GFP (ΔSP-Calnuc-GFP) unlike cells rescued with control GFP. HeLa cells treated with Calnuc siRNA were transfected with GFP, full-length rCalnuc-GFP or ΔSP-rCalnuc-GFP for 3 days, and then with LRP9HA cDNA for 12 h before the experiment. Cells were lysed, and proteins were separated by SDS-PAGE and detected with specific anti-HA, anti-GFP, anti-Calnuc and anti-actin antibodies. Actin served as a loading control. Endogenous human Calnuc was below the detection level in samples treated with Calnuc-specific siRNA. B) Bar graph showing the quantification of LRP9 degradation in cells treated as described in A) and expressed as a percentage of LRP9 present in control (GFP) cells. Results are shown as means ± SD (n = 3). *p< 0.05 [compared with control (GFP) cells]. Discussion In the present study, we used a yeast-two-hybrid assay to search for new interacting partners for Calnuc and to better understand its intracellular functions. We identified LDLR-related protein 9 (LRP9) as the first transmembrane protein that directly interacts with Calnuc. We also demonstrated that Calnuc plays a role in the retrieving of LRP9 from endosomes to the TGN. Our results demonstrated that the interaction between Calnuc and LRP9 involved the N-terminus domain of Calnuc and an R-rich region in the cytoplasmic domain of LRP9. In vivo and in vitro pull-down and yeast-two-hybrid assays indicated that the conserved Nterminal region of Calnuc was essential for the interaction with the cytoplasmic tail of LRP9. Furthermore, an analysis of LRP9 C-terminal tail truncation mutants identified an Rrich domain as the Calnuc-binding determinant. Our yeast-two-hybrid results are in agreement with this because the R-rich region was present in both clones fished out by the yeast-two-hybrid approach. Interestingly, we noted that an R-rich region was present in the cytoplasmic tail of LRP3, a member of the same LRP subfamily as LRP9 (20) that interacted with Calnuc, but was absent from LRP1, a member of another subfamily of LRP that did not interact with Calnuc. This is consistent with the notion that the region bearing the R-rich domain is responsible for the interaction with Calnuc. However, an LRP9 mutant lacking the R-rich region (ΔR-LRP9) still coimmunoprecipitated Calnuc in vivo, suggesting that the R-rich domain of LRP9 is important but not crucial for the interaction with Calnuc in vivo. Calnuc might also interact indirectly, via other partners, with LRP9HA, or with endogenous LRP9 or LRP3, which might form oligomers with ΔR-LRP9. Indeed, we have data suggesting that LRP9 and LRP3 form homo- and hetero-oligomers (C. Lavoie, Université de Sherbrooke, Sherbrooke, unpublished results). R-rich domains, also called basic amphiphilic segments, have been previously identified in the cytoplasmic tail of other receptors such as EGFR. This region has been reported to be the binding site for calmodulin (CaM), another EF-hand calcium-binding protein (35). CaM regulates EGFR tyrosine kinase activity, trafficking and signaling (36–39). The CaM-binding site is an R-helical structure composed of a cluster of positively charged amino acids arranged on one side of the R-helix, whereas most of the hydrophobic amino acids are on the opposite 273 side (35). A helical-wheel projection predicted that the R-rich regions of LRP9 and LRP3 fold in the same kind of helical structure (C. Lavoie, unpublished results). These observations suggest that Calnuc might bind and modulate the trafficking and signaling of other receptors (e.g., EGFR or other members of the LDLR superfamily). This is presently under investigation. The interaction between Calnuc and LRP9 is likely physiological because our confocal microscopic studies showed that the cytosolic form of Calnuc colocalized with LRP9 predominantly on the TGN and, to a lesser extent, on early endosomes. Indeed, in cells expressing the Rab5QL or LRP9−2DXXAA mutants, which both induced the redistribution of LRP9 on early endosomes, Calnuc clearly colocalized with LRP9 on this compartment. Because Calnuc colocalized with LRP9 on the TGN and early endosomes, we propose that it is at this location that Calnuc participates in the trafficking of LRP9. The involvement of Calnuc in the retrieval of LRP9 from endosomes to the TGN is supported by the phenotype elicited by siRNA-mediated interference of Calnuc expression. This treatment resulted in the redistribution of LRP9 into late endosomes/lysosomes, enhanced lysosomal degradation of LRP9 and, consequently, decreased steady-state levels of this receptor. Importantly, treatment with lysosomal protease inhibitors restored LRP9 levels in Calnuc-depleted cells, unequivocally demonstrating that the reduction in LRP9 levels was because of mistargeting and degradation in lysosomes. These data point to a role for Calnuc in the recycling of LRP9 from endosomes. These effects were specific because the silencing of Calnuc had no effect on the distribution and levels of TGN46, another TGN protein, indicating no obvious perturbation in Golgi organization and increased the turnover of LRP3 but not of LRP1, two members of different LRP subfamilies that do and do not interact with Calnuc, respectively. These results strongly indicate that Calnuc is directly involved in retrograde transport from endosomes of specific receptors rather than a generalized effect on membrane protein dynamics. The specificity of the phenotype observed with Calnuc depletion was also corroborated by rescue experiments using an siRNA-resistant version of Calnuc. Indeed, the expression of siRNA-resistant wild-type and cytoplasmic Calnuc restored the distribution of LRP9 in the TGN and decreased its redistribution and degradation in lysosomes in Calnuc-depleted cells. Therefore, the cytoplasmic pool of Calnuc emerged as the functional pool involved in the endosomal sorting of LRP9. This provided further support for a direct role in the interaction of Calnuc with LRP9 cytoplasmic tail. The most logical interpretation of these observations is that Calnuc keeps LRP9 from going to lysosomes by mediating its removal from maturing endosomes. The cytoplasmic LRP9 determinant recognized by Calnuc might encompass a ‘lysosomal avoidance’ signal functionally similar to that in the cytosolic domain of the MPR (40). However, the Trp/Phe-Leu-Met/Val or Phe-Trp motifs defined as endosome-toTGN retrieval motifs for MPR and sortilin (40,41) are not present in LRP9. This indicates that other determinants within LRP9 direct its sorting from endosomes. Our results suggest that Calnuc recognizes determinants in the R-rich region of the cytoplasmic domain of LRP9 and facilitates LRP9 sorting into transport vesicles destined for the TGN. Calnuc, GGA and adaptor protein (AP1/AP2) bind to distinct domains in the cytoplasmic tail of LRP9 and are involved in different trafficking steps. Deletion of 274 dileucine signals that interact with both GGAs and AP1/AP2 cause an accumulation of LRP9 on early endosomes and the plasma membrane (27,28), whereas siRNA depletion of Calnuc increases the lysosomal delivery of LRP9. Interestingly, the effects of Calnuc depletion are similar to the retromer subunits depletion on MPR trafficking (42). Retromers are molecules that retrieve receptors such as MPR, sortilin and SorLA from maturing endosomal intermediates or late endosomes to the TGN and whose depletion results in a dramatic reduction of the half-life of these receptors (42–44). This suggests that Calnuc function is similar to retromers in the efficient endosomal sorting and retrograde transport of LRP9. These findings provide evidence that GGA and Calnuc may act in concert to promote the transport of LRP9 from the TGN and their return from endosomes. LRP9 shares trafficking similarities with the sortilin, SorLA and MPR receptors, which transport cargo from the TGN to the endosomal/lysosomal system (29,43–46). These receptors predominate in the TGN and endosomes, whereas only a minor fraction (5–10%) is present on the cell surface at any given time (28,44,47). They contain GGA and AP1/AP2-binding sites, which are crucial for endocytosis and Golgi-endosome sorting (27– 29,43,48–51) and are delivered from early to late endosomes prior to recycling to TGN for another round of transport. Their passage through late endosomal intermediates is important for the dissociation of cargo from the receptors that occurs at low pH (52,53). Proteins such as retromers (54,55), AP1 (56), TIP47 (57,58) and PACS−1 (59) participate in this retrieval step. We have identified Calnuc as a key player in the endosomal sorting of LRP9 for proper retrieval to the TGN. The physiological role(s) of LRP9 are far from clear but, given the finding described above, it likely involved in trafficking of ligands between the TGN and endosomes. Investigations are presently underway to identify potential intracellular cargos for LRP9. LRP9 thus most likely cycles between the TGN, endosomes and lysosomes through a GGA- and Calnuc-dependent trafficking mechanism. Further work is currently in progress to define the precise mechanism by which Calnuc regulates LRP9 trafficking and to determine whether Calnuc regulates the transport of other receptors. These studies will help clarify the biological functions of Calnuc and LRP9. Materials and Methods Antibodies and reagents Anti-HA monoclonal antibodies (mAbs) and anti-LAMP−2 mAbs were purchased from Covance and the Developmental Studies Hybridoma Bank, respectively. Anti-GFP polyclonal antibodies (pAbs) were purchased from Molecular Probes, anti-HA pAbs from Covance, anti-EEA1 pAbs from Affinity Bioreagents, anti-TGN46 pAbs from Novus Biologicals, and anti-Calnuc pAbs from Aviva System Biology. Rabbit antibodies against CI-MPR were generously provided by Dr Thomas Braulke (University of Hamburg, Hambourg, Germany). DNA constructs Mammalian expression vector pMKITNeo encoding the murine LRP9-HA was kindly provided by Dr T. Kitamura (University of Tokyo, Japan). The LRP9 double 275 DXXLL mutant construct has been described previously (27). Mammalian expression vector pcDNA3.1 containing rCalnuc-GFP fusion protein or rat ΔSP-Calnuc (Calnuc without SP) is described elsewhere (32). Yeast-two-hybrid screen and mating The MATCHMAKER GAL4 Two-Hybrid System 3 (Clontech) was used. The Nterminus of Calnuc cDNA (residues 25–168) was cloned into the pGBKT7 vector to generate the bait construct expressing the Calnuc-GAL4 DNA-binding domain fusion protein. Saccharomyces cerevisiae strain AH109, which contains the HIS3, ADE2 and MEL1 reporter genes, was transformed with the bait plasmid using the lithium acetate method described in the MATCHMAKER instruction manual (Clontech). Transformants containing the bait plasmid were selected by plating the cells on minimal SD agar lacking tryptophan (SD/-trp). A MATCHMAKER human kidney cDNA library (Clontech) pretransformed in yeast strain Y187 was used for mating. To detect Calnuc-interacting proteins, AH109-Calnuc-N-term cells were incubated overnight with a Y187-kidney library. The mating mixture was plated on medium stringency SD/-leu/-trp/-his/5 mm 3amino triazole agar plates. After incubating the plates at 30^C for several days, the colonies were streaked on SD/-leu/-trp and SD/-leu/-trp/-his agar plates. These yeast colonies were subjected to a higher stringency test (i.e., they were replica plated onto SD/-leu/-trp/-his/ade agar plates). The colonies that grew well were restreaked to assess β-galactosidase (βGal) activity using a colony lift assay (CLONTECH Laboratories). Plasmid DNA was isolated from the colonies using Zymoprep yeast plasmid miniprep kits (Zymo Research). Prey plasmids were purified by transforming E. coli KC8 by electroporation with the plasmid DNA isolated from yeast. The prey plasmids were isolated using QIAprep spin plasmid DNA miniprep kits (Qiagen) and were sequenced. cDNA fragments were identified by BLAST analysis against GenBank. One of the clones (TH52) contained a portion of the C-terminus of LRP9 (residues 556–713). To confirm the interaction of Calnuc with LRP9 C-tail (TH52 clone), bait plasmids encoding truncated forms of Calnuc (N-term, C-term or EF-hand+C-term) were engineered, and AH109 was transformed with the bait plasmids together with the LRP9 C-tail pACT2 prey plasmids. AH109 transformed with empty prey and bait plasmid (pACT2 and pGBKT7, respectively) were used as negative controls. To assess the specificity of the interaction between the prey proteins and Calnuc, the interactions of an unrelated bait protein (the PXA domain) with the preys was also tested. All the transformants were plated on quadruple drop-out agar (SD/-leu/-trp/-his/-ade). Growth on this medium indicated that the bait proteins interacted with the prey protein of interest. The colonies were also tested for β-galactosidase (β-Gal) activity using a colony lift assay to confirm a positive interaction. Cell culture and transfection COS7 cells were obtained from Dr Klaus Hahn (University of North Carolina, Chapel Hill, NC, USA) and HEK293T cells were obtained from Dr Alexandra Newton (University of California, San Diego, CA, USA). HeLa cells were purchased from ATCC. The cells were grown in Dulbecco's modified Eagle's high glucose medium (Invitrogen) containing 10% FBS (Hyclone Laboratories), penicillin and streptomycin. COS7 cells were 276 transfected using Fugene6 transfection reagent (Roche Diagnostics), and the HEK and HeLa cells were transfected with Lipofectamine 2000 transfection reagent (Invitrogen), both according to the manufacturers’ instructions. Immunofluorescence COS7 cells were plated on coverslips. Twelve hours after the transfection, the cells were fixed in 3% paraformaldehyde in 100 mm phosphate buffer, pH 7.4, for 30 min, permeabilized with 0.1% Triton X−100 for 10 min, blocked with 10% goat serum for 30 min, and incubated with primary antibodies for 1 h at room temperature, followed by Alexa Fluor−594 or 488-conjugated antibodies (Molecular Probes) for 1 h at room temperature. The specimens were visualized using an inverted confocal laser-scanning microscope (FV1000, Olympus) equipped with a PlanApo 60×/1.42 oil immersion objective (Olympus). Olympus Fluoview software version 1.6b was used for image acquisition and analysis. The images were further processed using Adobe Photoshop (Adobe Systems). To quantify the relative colocalization in multiple images (n = 10–15 cells), the Pearson correlation coefficients (R values) were calculated for each image set for pixels above the calculated thresholds using Olympus Fluoview v1.6b colocalization software. The Pearson coefficient is a statistical appraisal of how well a linear equation describes the relationship between two variables for a measured function and is commonly used in image colocalization analysis. Cell fractionation (S100/P100 centrifugation) Cytosolic and membrane fractions were prepared as described previously (3). HeLa cells were washed in cold PBS, scraped into cold homogenization buffer [50 mm Tris, pH 7.4, 150 mm NaCl, 5 mm ethylenediaminetetraacetic acid (EDTA) with protease inhibitors], and centrifuged for 5 min at 500 ×g. The pellet was homogenized by 40 passages through a 25 G 5/8 needle. Nuclei and unbroken cells were removed by centrifugation (1000 ×g for 10 min). The post-nuclear supernatant was collected and centrifuged at 100,000 ×g for 1 h at 4^C. The soluble or cytosolic fraction (S100) was collected, and the pellet or membrane fraction (P100) was resuspended using a pestle in an equal volume of buffer. The same volumes (60 μL) of membrane and cytosolic fractions were analyzed by SDS-PAGE and immunoblotting. GST pull-down assays Full-length rCalnuc cDNA or Calnuc fragments containing the N-terminus (residues 1–168) or C-terminus (residues 335–459) were amplified and subcloned in pGEX-KG (Amersham Biosciences). GST fusion proteins were expressed in E. coli BL21 and purified on glutathione-Sepharose 4B beads (Pharmacia) according to the manufacturer's instructions. 35S-labeled in vitro translation products of pCDNA3.1-human LRP9 Cterminus (residues 556–713) were prepared using the TNT T7 rabbit reticulocyte Quick Coupled Transcription/Translation system (Promega) in the presence of [35S]methionine (1000 Ci/mmol, in vivo cell labeling grade; Amersham Biosciences). For the pull-down assays, GST fusion proteins (10 μg) immobilized on beads were incubated with in vitrotranslated products in 20 mm Tris–HCl buffer (pH 7.4) containing 150 mm NaCl, 3 mm EDTA and 0.1% NP−40 in the presence of protease inhibitors for 3 h at 4^C and were washed four times with the same buffer. For the lysate GST pull-down assays, 1 mg of cell 277 lysate or cytosolic fraction (S100) in buffer [50 mm Tris, pH 7.4, 150 mm NaCl, 1% NP−40, 5 mm EDTA and complete protease inhibitors (Roche)] was incubated overnight at 4^C with 10 or 30 μg of GST fusion proteins immobilized on glutathione-Sepharose 4B beads. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. The bound proteins were separated by SDS-PAGE and detected by autoradiography or immunoblotting. Coimmunoprecipitation HEK cells were plated in 60-mm culture dishes and transfected with the various constructs. After 48 h, the cells were lysed in 50 mm Tris buffer (pH 7.4) containing 150 mm NaCl, 1% NP−40 and protease inhibitors for 1 h at 4^C and then centrifuged at 13,000 ×g for 20 min. The cleared supernatants were incubated with primary antibodies overnight at 4^C and then with protein A-sepharose (GE Healthcare) or protein G-Sepharose (Zymed) for 1 h. The beads were washed three times in lysis buffer and boiled in Laemmli sample buffer. Bound immune complexes were analyzed by SDS-PAGE and immunoblotting. Immunoblotting The protein samples were separated on 8 or 10% SDS-PAGE gels and transferred to nitrocellulose membranes (Perkin Elmer). The membranes were blocked in Tris-buffered saline (20 mm Tris–HCl, pH 7.4, 150 mm NaCl) containing 0.1% Tween 20 and 5% nonfat dry milk and incubated with primary antibodies for 1 h at room temperature and then with horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG (Bio-Rad) for 45 min at room temperature and enhanced chemiluminescence detection reagent (Pierce Chemical). RNA interference Scrambled RNA oligos (scramble II duplex) and Calnuc siRNA were purchased from Dharmacon Research. HeLa cells were transfected with a final concentration of 100 nM siRNA duplex using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. The cells were analyzed 72 h after the transfection. HA-LRP9 was transfected with Fugene 12 h before the IF experiments. Reversal of the phenotype was attempted by transfecting HeLa cells with cDNA encoding rCalnuc-GFP, rat ΔSP-CalnucGFP, or GFP alone using Fugene 5 h after they were transfected with human Calnuc siRNA. Cycloheximide chase Control- and siRNA-treated cells were incubated at 37^C in complete DMEM containing 25 mm Hepes buffer, pH 7.4 and 40 μg/mL of cycloheximide (Sigma-Aldrich). At the corresponding time points, the cells were lysed as described above and analyzed by SDS-PAGE and immunoblotting. Treatment with lysosomal inhibitors Control- and siRNA-treated cells were incubated for 3–5 h at 37^C in complete DMEM containing 1 mg/mL of leupeptin (Roche Diagnostics) and 10 μg/mL of E64 (Sigma-Aldrich). The cells were collected, lysed and analyzed by SDS-PAGE and immunoblotting. 278 Statistical analysis Experiments were performed in triplicate and results are expressed as means ± SD. Statistical significance between various samples was assessed using the Student t-test. A p < 0.05 was considered significant. Acknowledgments We are grateful to Toshio Kitamura for the LRP9 construct, Takahiro Fujino for the LRP3 construct, to Bradley Hyman for the LRP1 construct, to Thomas Braulke for the rabbit CI-MPR antibodies, and to Marilyn Gist Farquhar for the generous gifts of reagents. We would also like to thank Radu Stan for the advices with the yeast-two-hybrid screening. The LAMP−2 monoclonal antibody developed by J. Thomas August and James E.K. Hildreth was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa. This work was supported by grants from the Canadian Institutes for Health Research and a Canada Research Chair to C.L. References 1 2 3 4 5 6 7 8 9 Wendel M, Sommarin Y, Bergman T, Heinegard D. Isolation, characterization and primary structure of a calcium-binding 63-kDa bone protein. J Biol Chem 1995;270: 6125–6133. Miura K, Hirai M, Kanai Y, Kurosawa Y. Organization of the human gene for nucleobindin (NUC) and its chromosomal assignment to 19q13.2–q13.4. Genomics 1996;34: 181–186. Lin P, Le-Niculescu H, Hofmeister R, McCaffery JM, Jin M, Hennemann H, McQuistan T, De Vries L, Farquhar MG. The mammalian calcium-binding protein, nucleobindin (CALNUC), is a Golgi resident protein. J Cell Biol 1998;141: 1515– 1527. Kawano J, Kotani T, Ogata Y, Ohtaki S, Takechi S, Nakayama T, Sawaguchi A, Nagaike R, Oinuma T, Suganuma T. CALNUC (nucleobindin) is localized in the Golgi apparatus in insect cells. Eur J Cell Biol 2000;79: 208–217. Kubota T, Miyauchi M, Miura K, Hirokawa G, Awaya A, Miyasaka N, Kurosawa Y, Kanai Y, Maruyama K. Upregulation of nucleobindin expression in human-activated lymphocytes and non-Hodgkin's lymphoma. Pathol Int 1998;48: 22–28. De Vos S, Hofmann WK, Grogan TM, Krug U, Schrage M, Miller TP, Braun JG, Wachsman W, Koeffler HP, Said JW. Gene expression profile of serial samples of transformed B-cell lymphomas. Lab Invest 2003;83: 271–285. Howell BG, Solish N, Lu C, Watanabe H, Mamelak AJ, Freed I, Wang B, Sauder DN. Microarray profiles of human basal cell carcinoma: insights into tumor growth and behavior. J Dermatol Sci 2005;39: 39–51. Chen Y, Lin P, Qiu S, Peng XX, Looi K, Farquhar MG, Zhang JY. Autoantibodies to Ca2+ binding protein Calnuc is a potential marker in colon cancer detection. Int J Oncol 2007;30: 1137–1144. Gronborg M, Kristiansen TZ, Iwahori A, Chang R, Reddy R, Sato N, Molina H, Jensen ON, Hruban RH, Goggins MG, Maitra A, Pandey A. Biomarker discovery 279 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 from pancreatic cancer secretome using a differential proteomic approach. Mol Cell Proteomics 2006;5: 157–171. Tamura K, Furihata M, Tsunoda T, Ashida S, Takata R, Obara W, Yoshioka H, Daigo Y, Nasu Y, Kumon H, Konaka H, Namiki M, Tozawa K, Kohri K, Tanji N et al. Molecular features of hormone-refractory prostate cancer cells by genome-wide gene expression profiles. Cancer Res 2007;67: 5117–5125. Kanai Y, Miura K, Uehara T, Amagai M, Takeda O, Tanuma S, Kurosawa Y. Natural occurrence of Nuc in the sera of autoimmune-prone MRL/lpr mice. Biochem Biophys Res Commun 1993;196: 729–736. Iizuka H, Kubota T, Satake F, Hirokawa G, Miura K, Kurosawa Y, Miyasaka N, Kanai Y. Enhanced expression of nucleobindin in lymphatic organs of lupus-prone mice. Lupus 1997;6: 365–370. Lin P, Li F, Zhang YW, Huang H, Tong G, Farquhar MG, Xu H. Calnuc binds to Alzheimer's beta-amyloid precursor protein and affects its biogenesis. J Neurochem 2007;100: 1505–1514. De Alba E, Tjandra N. Structural studies on the Ca2+-binding domain of human nucleobindin (calnuc). Biochemistry 2004;43: 10039–10049. Johnson S, Michalak M, Opas M, Eggleton P. The ins and outs of calreticulin: from the ER lumen to the extracellular space. Trends Cell Biol 2001;11: 122–129. Lin P, Yao Y, Hofmeister R, Tsien RY, Farquhar MG. Overexpression of CALNUC (nucleobindin) increases agonist and thapsigargin releasable Ca2+ storage in the Golgi. J Cell Biol 1999;145: 279–289. Lavoie C, Meerloo T, Lin P, Farquhar MG. Calnuc, an EF-hand Ca(2+)-binding protein, is stored and processed in the Golgi and secreted by the constitutive-like pathway in AtT20 cells. Mol Endocrinol 2002;16: 2462–2474. Petersson U, Somogyi E, Reinholt FP, Karlsson T, Sugars RV, Wendel M. Nucleobindin is produced by bone cells and secreted into the osteoid, with a potential role as a modulator of matrix maturation. Bone 2004;34: 949–960. Taniguchi N, Taniura H, Niinobe M, Takayama C, Tominaga-Yoshino K, Ogura A, Yoshikawa K. The postmitotic growth suppressor necdin interacts with a calciumbinding protein (NEFA) in neuronal cytoplasm. J Biol Chem 2000;275: 31674– 31681. Battle MA, Maher VM, McCormick JJ. ST7 is a novel low-density lipoprotein receptor-related protein (LRP) with a cytoplasmic tail that interacts with proteins related to signal transduction pathways. Biochemistry 2003;42: 7270–7282. Hussain MM, Strickland DK, Bakillah A. The mammalian low-density lipoprotein receptor family. Annu ReV Nutr 1999;19: 141–172. Strickland DK, Gonias SL, Argraves WS. Diverse roles for the LDL receptor family. Trends Endocrinol Metab 2002;13: 66–74. Yamamoto T, Davis CG, Brown MS, Schneider WJ, Casey ML, Goldstein JL, Russell DW. The human LDL receptor: a cysteine-rich protein with multiple Alu sequences in its mRNA. Cell 1984;39: 27–38. Bork P, Beckmann G. The CUB domain. A widespread module in developmentally regulated proteins. J Mol Biol 1993;231: 539–545. Christensen EI, Birn H. Megalin and cubilin: multifunctional endocytic receptors. Nat Rev Mol Cell Biol 2002;3: 256–266. 280 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Sugiyama T, Kumagai H, Morikawa Y, Wada Y, Sugiyama A, Yasuda K, Yokoi N, Tamura S, Kojima T, Nosaka T, Senba E, Kimura S, Kadowaki T, Kodama T, Kitamura T. A novel low-density lipoprotein receptor-related protein mediating cellular uptake of apolipoprotein E-enriched beta-VLDL in vitro. Biochemistry 2000;39: 15817–15825. Boucher R, Larkin H, Brodeur J, Gagnon H, Thériault C, Lavoie C. Intracellular trafficking of LRP9 is dependent on two acidic cluster/dileucine motifs. Histochem Cell Biol 2008;130: 317–327. Doray B, Knisely JM, Wartman L, Bu G, Kornfeld S. Identification of acidic dileucine signals in LRP9 that interact with both GGAs and AP−1/AP−2. Traffic 2008;9: 1551–1562. Puertollano R, Aguilar RC, Gorshkova I, Crouch RJ, Bonifacino JS. Sorting of mannose 6-phosphate receptors mediated by the GGAs. Science 2001;292: 1712– 1716. Tikkanen R, Obermuller S, Denzer K, Pungitore R, Geuze H, Von Figura K, Honing S. The dileucine motif within the tail of MPR46 is required for sorting of the receptor in endosomes. Traffic 2000;1: 631–640. Tortorella LL, Schapiro FB, Maxfield FR. Role of an acidic cluster/dileucine motif in cation-independent mannose 6-phosphate receptor traffic. Traffic 2007;8: 402–413. Weiss TS, Chamberlain CE, Takeda T, Lin P, Hahn KM, Farquhar MG. Galpha i3 binding to calnuc on Golgi membranes in living cells monitored by fluorescence resonance energy transfer of green fluorescent protein fusion proteins. Proc Natl Acad Sci USA 2001;98: 14961–14966. Stenmark H, Parton RG, Steele-Mortimer O, Lutcke A, Gruenberg J, Zerial M. Inhibition of rab5 GTPase activity stimulates membrane fusion in endocytosis. EMBO J 1994;13: 1287–1296. Manders E, Verbeek F, Aten J. Measurement of colocalization of objects in dualcolour confocal images. J Microsc 1993;169: 375–382. Martin-Nieto J, Villalobo A. The human epidermal growth factor receptor contains a juxtamembrane calmodulin-binding site. Biochemistry 1998;37: 227–236. Li E, Hristova K. Role of receptor tyrosine kinase transmembrane domains in cell signaling and human pathologies. Biochemistry 2006;45: 6241–6251. Tebar F, Villalonga P, Sorkina T, Agell N, Sorkin A, Enrich C. Calmodulin regulates intracellular trafficking of epidermal growth factor receptor and the MAPK signaling pathway. Mol Biol Cell 2002;13: 2057–2068. Aifa S, Aydin J, Nordvall G, Lundstrom I, Svensson SP, Hermanson O. A basic peptide within the juxtamembrane region is required for EGF receptor dimerization. Exp Cell Res 2005;302: 108–114. Llado A, Timpson P, Vila de Muga S, Moreto J, Pol A, Grewal T, Daly RJ, Enrich C, Tebar F. Protein kinase Cdelta and calmodulin regulate epidermal growth factor receptor recycling from early endosomes through Arp2/3 complex and cortactin. Mol Biol Cell 2008;19: 17–29. Rohrer J, Schweizer A, Johnson KF, Kornfeld S. A determinant in the cytoplasmic tail of the cation-dependent mannose 6-phosphate receptor prevents trafficking to lysosomes. J Cell Biol 1995;130: 1297–1306. 281 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 Seaman MN. Identification of a novel conserved sorting motif required for retromermediated endosome-to-TGN retrieval. J Cell Sci 2007;120: 2378–2389. Arighi CN, Hartnell LM, Aguilar RC, Haft CR, Bonifacino JS. Role of the mammalian retromer in sorting of the cation-independent mannose 6-phosphate receptor. J Cell Biol 2004;165: 123–133. Nielsen MS, Gustafsen C, Madsen P, Nyengaard JR, Hermey G, Bakke O, Mari M, Schu P, Pohlmann R, Dennes A, Petersen CM. Sorting by the cytoplasmic domain of the amyloid precursor protein binding receptor SorLA. Mol Cell Biol 2007;27: 6842– 6851. Mari M, Bujny MV, Zeuschner D, Geerts WJ, Griffith J, Petersen CM, Cullen PJ, Klumperman J, Geuze HJ. SNX1 defines an early endosomal recycling exit for sortilin and mannose 6-phosphate receptors. Traffic 2008;9: 380–393. Nielsen MS, Madsen P, Christensen EI, Nykjaer A, Gliemann J, Kasper D, Pohlmann R, Petersen CM. The sortilin cytoplasmic tail conveys Golgi-endosome transport and binds the VHS domain of the GGA2 sorting protein. EMBO J 2001;20: 2180–2190. Ghosh P, Dahms NM, Kornfeld S. Mannose 6-phosphate receptors: new twists in the tale. Nat Rev Mol Cell Biol 2003;4: 202–212. Jacobsen L, Madsen P, Jacobsen C, Nielsen MS, Gliemann J, Petersen CM. Activation and functional characterization of the mosaic receptor SorLA/LR11. J Biol Chem 2001;276: 22788–22796. Honing S, Sosa M, Hille-Rehfeld A, Von Figura K. The 46-kDa mannose 6phosphate receptor contains multiple binding sites for clathrin adaptors. J Biol Chem 1997;272: 19884–19890. Nielsen MS, Madsen P, Christensen EI, Nykjær A, Gliemann J, Kasper D, Pohlmann R, Petersen1 CM. The sortilin cytoplasmic tail conveys Golgi-endosome transport and binds the VHS domain of the GGA2 sorting protein. EMBO J 2001;20: 2180– 2190. Jacobsen L, Madsen P, Nielsen MS, Geraerts WP, Gliemann J, Smit AB, Petersen CM. The sorLA cytoplasmic domain interacts with GGA1 and −2 and defines minimum requirements for GGA binding. FEBS Lett 2002;511: 155–158. Canuel M, Lefrancois S, Zeng J, Morales CR. AP−1 and retromer play opposite roles in the trafficking of sortilin between the Golgi apparatus and the lysosomes. Biochem Biophys Res Commun 2008;366: 724–730. Sahagian GG, Distler J, Jourdian GW. Characterization of a membrane-associated receptor from bovine liver that binds phosphomannosyl residues of bovine testicular beta-galactosidase. Proc Natl Acad Sci USA 1981;78: 4289–4293. Hoflack B, Kornfeld S. Purification and characterization of a cation-dependent mannose 6-phosphate receptor from murine P388D1 macrophages and bovine liver. J Biol Chem 1985;260: 12008–12014. Collins BM. The structure and function of the retromer protein complex. Traffic 2008;9: 1811–1822. Bonifacino JS, Hurley JH. Retromer. Curr Opin Cell Biol 2008;20: 427–436. Meyer C, Zizioli D, Lausmann S, Eskelinen EL, Hamann J, Saftig P, Von Figura K, Schu P. mu1A-adaptin-deficient mice: lethality, loss of AP−1 binding and rerouting of mannose 6-phosphate receptors. EMBO J 2000;19: 2193–2203. 282 57 58 59 Diaz E, Pfeffer SR. TIP47: a cargo selection device for mannose 6-phosphate receptor trafficking. Cell 1998;93: 433–443. Krise JP, Sincock PM, Orsel JG, Pfeffer SR. Quantitative analysis of TIP47-receptor cytoplasmic domain interactions: implications for endosome-to-trans Golgi network trafficking. J Biol Chem 2000;275: 25188–25193. Scott GK, Fei H, Thomas L, Medigeshi GR, Thomas G. A PACS−1, GGA3 and CK2 complex regulates CI-MPR trafficking. EMBO J 2006;25: 4423–4435. 283 ANNEXE 3 Macrocyclic cell penetrating peptides: a study of structure-penetration properties. Traboulsi, H., Larkin, H., Bonin, M.-A., Volkov, L., Lavoie, C., and Marsault, É. Bioconjug Chem. (2015) Travaux en collaboration IF= 4,58 284 Macrocyclic cell penetrating peptides: a study of structure-penetration properties. Traboulsi, H., Larkin, H., Bonin, M.-A., Volkov, L., Lavoie, C., and Marsault, É. Abstract Arginine-rich cell penetrating peptides are short cationic peptides able to cross biological membranes despite their peptidic character. In order to optimize their penetration properties and further elucidate their mechanisms of cellular entry, these peptides have been intensively studied for the last two decades. Although several parameters are simultaneously involved in the internalization mechanism, recent studies suggest that structural modifications influence cellular internalization. Particularly, backbone rigidification, including macrocyclization, was found to enhance proteolytic stability and cellular uptake. In the present work, we describe the synthesis of macrocyclic arginine-rich cell penetrating peptides and study their cellular uptake properties using a combination of flow cytometry and confocal microscopy. By varying ring size, site of cyclization, and stereochemistry of the arginine residues, we studied their structure–uptake relationship and showed that the mode and site of cyclization as well as the stereochemistry influence cellular uptake. This study led to the identification of a hepta-arginine macrocycle as efficient as its linear nona-arginine congener to enter cells. Introduction Cell-penetrating peptides (CPPs), originally known as protein transduction domains, are peptides consisting of fewer than 35 amino acids capable of penetrating cells despite their peptidic character. Initially identified as specific transduction domains embedded in protein sequences, they represent a natural mechanism of cellular targeting and entry.(1-4) Since their discovery, the cell penetrating properties of CPPs have been harnessed to internalize otherwise impermeable cargos into cells, including small molecules, macromolecules (e.g., proteins, nucleic acids), and nanoparticles.(5-11) The majority of CPPs contain a high relative abundance of positively charged amino acids such as lysine 285 (Lys) or arginine (Arg), or display alternating patterns of polar/charged amino acids and nonpolar, hydrophobic amino acids.(9, 12, 13) Arginine-rich CPPs have been used successfully to deliver a broad diversity of biologically active macromolecules intracellularly, although their mechanisms of cellular entry are still under investigation.(9, 12, 14-17) In the process, the guanidinium cationic groups of the Arg residues are crucial for efficient cell penetration, a role attributed to their capacity to interact with the various anion types localized on the plasma membrane such as the polar heads of phospholipids or the sulfate groups of glycosaminoglycans.(13, 18, 19) Polyarginines are readily synthesized and represent the structurally simplest CPPs. They have been studied extensively for their ability to penetrate through cell membranes. It was found that cellular uptake can be achieved with oligoarginines composed of 5–15 residues.(14, 20) In particular, nona-arginine (R9) was shown to possess improved cell penetration efficiency compared to TAT peptides.(9, 20) Thus, several studies have utilized R8 and R9 as reference tools to import into cells a variety of biological molecules including siRNA, anticancer drugs, small molecules, proteins, peptides, and oligonucleotides.(10, 11) The mechanisms by which CPPs translocate across cell membranes remain a subject of investigation. Certain CPPs, when attached to small cargos, translocate passively across the plasma membrane of cells.(21) However, an increasing number of studies have shown that cellular uptake of cationic CPPs, when conjugated to macromolecules or used at low concentrations, is temperature-sensitive, suggesting the involvement of energy-dependent processes.(17, 22-25) To better delineate the mechanisms at play, Matsushita et al. recently applied siRNA library screening to identify potential partners involved in the cellular entry of polyarginine peptides. This led to the identification of the involvement of cotransporter gene SLC4A4 and the trafficking regulator gene COPA in cellular entry. Knowing that COPA plays an important role in early endosome maturation, these results suggest that cellular entry of polyarginine involves at least two steps, such as initial binding to the cell surface, followed by endosomal entry.(26) Since polyarginines are the most commonly used CPPs, their optimization has been investigated. Replacement of (l)-Arg residues by (d) analogues and backbone modifications have produced protease-resistant analogues with improved translocation properties.(12, 13) Major structural modifications aimed at improving cellular penetration and proteolytic stability, while retaining acceptable cytotoxicity, have met with various degrees of success. In particular, it has been demonstrated that increased rigidity and static display of guanidine groups is beneficial for cellular entry properties.(12, 13, 27-32) In this context, macrocyclization is a well established approach to simultaneously rigidify a peptide backbone, modulate its structure–activity, improve its proteolytic stability, and reduce the entropic cost of macrocycle–target interactions.(33-36) These attributes are attractive for improving the properties of CPPs;(37-44) thus, macrocyclization and backbone rigidification of CPPs was expected to improve cellular uptake and stability, as well as to provide a means to discriminate among different cell types. Indeed, it has been proposed that guanidinium groups are forced into maximally distant positions by 286 cyclization, increasing membrane contacts and leading to enhanced cellular penetration.(28, 31, 45, 46) To date, there is no systematic study of the features favorable for cellular uptake in macrocyclic polyarginines (Figure 1). In order to better understand the potential structure– penetration relationship, we synthesized several analogues in this series, with the goal of better delineating how structural variations like cyclization, stereochemistry, or endocyclic vs exocyclic display of arginine residues influence cellular uptake. Compounds presented in this study were functionalized with a fluorescent label in the form of a fluorescein isothiocyanate (FITC) moiety. Their cellular penetration properties were assessed using flow cytometry and results were confirmed using fluorescence confocal microscopy, with the hypothesis that stereochemistry and mode of cyclization could influence cellular uptake. Figure 1. Representative structure of cyclic polyarginines (n = 4 and 6–9); (l) stereochemistry given as an example. Results Synthesis Analogues were synthesized as described in Figure 2 and Supporting Information (SI) Figure S1. Briefly, synthesis relied on Fmoc chemistry and macrocyclization was performed via the formation of a lactam bridge between a Lys and a Glu residue. These two residues were previously protected as Alloc and Allyl ester, respectively, then deprotected under Pd catalysis immediately prior to cyclization. Every macrocycle was functionalized with FITC and a spacer (β-Ala, except for R9 which was functionalized as an aminohexanoate as per previous works). 287 Figure 2. Synthetic scheme for the preparation of cyclic polyarginine analogues. A total of 23 analogues (4 linear and 19 cyclic) were synthesized for this study, as summarized in Table 1. Entries 1–4 (Table 1) are reference compounds aimed at comparing cellular uptake with published results; entries 5–15 are side-chain to side-chain macrocycles containing 4–9 homochiral (l)Arg or (d)Arg residues or heterochiral (l,d)Arg residues. It should be kept in mind that between homo- and heterochiral macrocycles, the projection of side chains with respect to the plane of the macrocycle differs. Entries 16–19 are bicyclic macrocycles built as described in SI Figure S1, with the goal of assessing the influence of additional rigidification on cellular uptake; finally, entries 20–23 are macrocyclic polyarginines (3–6 residues) bearing 1–4 exocyclic arginine residues built to assess the influence of endocyclic vs exocyclic residues on cellular uptake. Compounds were synthesized on solid phase synthesis and purified to >95% purity (UPLC-UV). The cellular uptake of new compounds was first assessed by flow cytometry, then confirmed using confocal microscopy as described below. Table 1. Sequences of Linear and Macrocyclic Peptides Used in This Study. Lower case letters indicate (d) and upper case letters (l) stereochemistry. Residues between brackets are included in the macrocycle, closed by macrolactamization between the Glu (E) and Lys (K) side chains. 288 Flow Cytometry Flow cytometry analysis was used to quantify the cellular uptake of the new constructs in HeLa cells. Uptake measurements were based on changes in the mean cellular fluorescence following incubation with FITC-labeled peptides, relatively to linear control R9. Knowing that conventional flow cytometry cannot distinguish between intracellular and cell surface fluorescence, trypsin treatment was applied (Trypsin-EDTA 0.05%, 5 min) to remove cell surface-bound peptides prior to analysis. The efficiency of trypsin treatment for macrocycles was confirmed since treatment of Cyc-R4, which contains only 4 Arg residues, showed a very weak emission signal comparable to the signal of autofluorescence, confirming that membrane-bound cyclic peptides were indeed detached from the cell surface (Figure 3B). Additionally, it was confirmed that the uptake of Cyc-R9 is concentration-dependent (Supporting Information, Figure S2). Figure 3. Flow cytometry analysis of HeLa cells treated with (A) linear reference compounds; (B) all (l) macrocycles; (C) all (d) macrocycles; and (D) alternating (l,d) macrocycles. Each value represents the average of three experiments, and values are normalized with respect to control compound R9. Linear peptides R6–R9 were used as reference compounds and demonstrated an increasing uptake with increasing arginine contents (Figure 3A), consistently with previous literature reports.(14) In order to better understand the influence of structural variations on cellular uptake of macrocyclic constructs, we next analyzed the uptake of the cyclic analogues of the polyarginine peptides (Figure 3B). Cyc-R4 and Cyc-R6 showed a very weak emission signal (commensurate with the autofluorescence signal), suggesting little or no uptake. A significant increase in uptake was noted for Cyc-R8 and Cyc-R9. Importantly, Cyc-R9 displayed 40% higher uptake than its linear analogue R9. In order to better understand the influence of stereochemistry on internalization by HeLa cells, we prepared both homochiral (l) macrocycles, homochiral (d) macrocycles, as well as heterochiral (l,d) macrocycles, then compared their uptake by flow cytometry (Figure 3B–D). It is interesting to note that cycloocta-arginine macrocycles were 289 particularly sensitive to the influence of stereochemistry, as opposed to their cycloheptaarginine counterparts. Indeed, homochiral (l) analogue Cyc-R8 displayed higher levels of uptake (125% the value obtained for linear R9) than analogues Cyc-r8 and Cyc-(l,d)-R8 (63% and 75% the value reached by R9, respectively). In fact, Cyc-R8 demonstrated improved penetration compared to the standard linear R9 peptide. Similar observations were made for cyclonona-arginines. For example, Cyc-R9 was internalized twice as much as Cyc-r9 and 50% more than its linear analogue. Finally, Cyc-R9 macrocycle displayed an improved uptake compared to all the other cyclic or linear analogues (Figure 3B). To confirm that the fluorescence observed in flow cytometry is intracellular and not associated with cell surface, uptake experiments of R9, Cyc-R9, and Cyc-r9 were repeated in the presence of Trypan Blue to quench extracellular fluorescence (Figure 4, white vs black bars).(47) No major loss of fluorescence was observed in the presence of Trypan Blue in the case of R9 and Cyc-R9, indicating that the observed fluorescence is indeed intracellular. In the case of Cyc-r9, 20% of the fluorescence was quenched in the presence of Trypan Blue, suggesting that a small portion of the compounds remain membranebound. However, when Triton-X100 was added to permeabilize the cells, extensive quenching was observed, as expected if the fluorescence is intracellular (Figure 4, gray bars). Figure 4. Flow cytometry analysis of HeLa cells treated with R9, Cyc-R9, and Cyc-r9 peptides in the presence and absence of Trypan blue and Triton-X100. Each value represents the average of three experiments. In order to better understand how further rigidification influences cellular uptake, we next explored the effect of double cyclization, spacing between the two macrocycles, and the respective contribution of macrocyclic and linear groups on cellular uptake. This was performed on analogues bearing 6 Arg derivatives, to assess whether increased structural rigidification could bring these analogues to levels of uptake similar to those observed with linear R8 or R9 (see Table 1 for structures). Indeed, it was demonstrated that rigid scaffolds bearing as little as 4 or 5 guanidine groups could be as efficient as linear R9 to mediate cellular uptake.(27-29, 31, 32) Macrobicyclic hexa-arginine derivatives were synthesized using a similar approach to that of macrocycles with a single ring (see SI Figure S1). In some cases, spacers composed of 1–3 βAla residues were incorporated between the two 290 rings, in order to introduce some flexibility to the peptide and potentially impart some αhelical contents.(48) For comparison, we present in Figure 5A the results obtained by flow cytometry of Cyc-R6 and the different bicyclic peptides bearing 6 Arg residues. One can observe that in the whole series of hexa-arginine derivatives, the mode of cyclization and the presence of a linker had little influence on cellular uptake, which remained low in all cases. Bicyclization of hexa-arginine had either no effect, or a small tendency toward a negative effect. We next investigated the respective impact of endo- vs exocyclic Arg residues on cellular uptake. Given the poor results obtained in the previous hexa-arginine bicyclic derivatives, this was performed on hepta-arginine derivatives (Figure 5B). In this series, we compared a macrocycle containing 7 endocyclic Arg groups (Cyc-R7) to analogues bearing 6 endocyclic and 1 exocyclic Arg (Cyc-R6-R1), 5 endocyclic and 2 exocyclic Arg (CycR5-R2), 4 endocyclic and 3 exocyclic Arg (Cyc-R4-R3), and finally 3 endocyclic and 4 exocyclic Arg (Cyc-R3-R4) groups. Thus, the number of Arg residues was kept constant in this series. Flow cytometry results (Figure 5B) confirmed the possibility to significantly increase cellular uptake via the structural modification of these peptides. The optimal structure, Cyc-R4-R3, contains 4 endocyclic and 3 exocyclic Arg residues. Ultimately, the latter reached the same level of internalization as the standard linear R9 derivative, despite the fact that it possesses 7 instead of 9 Arg residues. In the linear series (Figure 3A), the internalization of linear R7 was ∼40% of that of linear R9. These results were consistent with those observed in confocal microscopy (see below). Figure 5. Flow cytometry analysis of HeLa cells treated with (A) bicyclic and (B) 7 arginine monocyclic peptides. Each value represents an average of three experiments, and values are normalized with respect to control compound R9. Confocal Microscopy The cellular uptake of selected linear and macrocyclic peptides was subsequently assessed by confocal microscopy on HeLa cells. Toward this end, cells were incubated with compounds (5 μM) for 30 min at 37 °C, then imaged by confocal microscopy after fixation on coverslips, or using live cells. The images of fixed HeLa cells treated with linear 291 peptides R6–R9 are shown in Figure 6 (top panels) as controls. Cells treated with linear R6 polyarginine showed little, if any, intracellular fluorescence, while treatment with R7–R9 showed increasing levels of fluorescence with increasing number of Arg residues under the same imaging conditions, as previously reported.(14) Similarly to linear peptides, internalization of macrocyclic peptides showed strongly diffused signals for peptides containing 7, 8, and 9 arginine residues (Cyc-R7, Cyc-R8, Cyc-R9), consistent with a predominantly cytoplasmic distribution (Figure 6, middle panels). As observed with linear peptides, increasing the number of Arg residues was associated with increased uptake from Cyc-R6 to Cyc-R8. However, there appeared to be no significant increase between Cyc-R8 to Cyc-R9 (Figure 6, middle panels). In parallel, the macrocyclic peptides possessing an exocyclic arginine chain showed significant variations in cellular uptake, whereby Cyc-R3-R4 displayed a stronger diffused signal (Figure 6, bottom panels) compared with the other peptides of this family (Cyc-R4-R3, Cyc-R5-R2, Cyc-R6-R1). Figure 6. Confocal microscopy images of HeLa cells treated with FITC-labeled linear and macrocyclic peptides (in green). Nuclei are marked in blue. Scale bar, 10 μm. The results of internalization of Cyc-r7, Cyc-r8, and Cyc-r9, as well as those of Cyc(l,d)-R6, Cyc-(l,d)-R7, and Cyc-(l,d)-R9, were also confirmed by confocal microscopy (SI Figures S3, S4). Selected linear and macrocyclic peptides were also tested using live cell imaging (SI Figure S5), given that fixation may lead to artifactual cell permeabilization and uptake.(49) 292 Essentially, results obtained with live cell imaging were consistent with those obtained with fixed cells, indicating that fixation did not alter cellular penetration. Discussion The aim of this study was to better understand how the mode of macrocyclization, the stereochemistry, and the endo vs exocyclic display of Arg in macrocyclic polyarginines influence cellular uptake. Indeed, it was previously reported, on one hand, that a minimal number of 8–9 arginine residues is required for efficient cell entry of linear arginine-rich peptides, yet it is also known that structural rigidification of the molecule could improve cellular penetration and allow a reduction in the number of exposed guanidine motifs, down to 4 or 5 guanidine units.(27-29, 31, 32) Flow cytometry was used to quantify the extent of cellular uptake in the various series, and confocal microscopy was used to confirm uptake and assess preliminarily the intracellular distribution. It should be kept in mind that the only charged units in these peptides are the guanidinium groups since Rink amide resin was used to deliver C-terminal amide peptides. First, results from flow cytometry of the control linear peptides confirmed that increasing the number of arginine increased cell internalization by 20% to 30% per arginine residue, as previously reported (Figure 3A).(14) We further asked how structural modifications to the macrocycles could influence cellular uptake. Indeed, reported molecular modeling suggested that a subset of the side chain guanidinium groups of these transporters might be required for transport involving contact with a common surface such as a plasma membrane or cell surface receptor, suggesting the possibility to develop the structure–uptake relationship and possibly cellular specificity.(27-29, 31, 46, 50) Spacers such as glycine have been introduced between arginine amino acids, showing that the spacing between the arginine residues in a set of linear peptides could influence uptake.(50) Consequently, by cyclizing polyarginine peptides, we expected to increase the distance between the guanidinium groups by way of conformational restrictions, without the addition of spacer elements and increase in molecular weight. Uptake measurements using flow cytometry on monocyclic peptides (Figure 3B) showed a dramatic effect on the cellular uptake of octa- and nona-arginine derivatives. The cyclic form of octa-arginine was not only able to penetrate into cells more efficiently than the linear form, it also exhibited superior uptake than control linear nonaarginine. This result is in agreement with literature describing the cyclic arginine-rich cellpenetrating peptides showing enhanced cellular uptake relative to their linear and more flexible counterpart.(31, 38) It has been proposed that guanidinium groups are forced into maximally distant positions by cyclization and this orientation increases membrane contacts leading to enhanced cell penetration. Interestingly, macrocyclic analogues of (l)Arg gave much higher penetration than those of (d)-Arg congeners or those alternating (l) and (d) stereochemistry (Figure 3B–D). This result is in agreement with the works of Verdurmen et al., who reported that cationic (l)-cell-penetrating peptides are taken up more efficiently than their (d)-counterparts in MC57 fibrosarcoma and HeLa cells but not in Jurkat T leukemia cells.(46) 293 To further rigidify structure and improve cellular uptake, we examined bicyclic hexaarginine analogues (Figure 5). Despite increased structural rigidification, cellular uptake of these variously double-cyclized analogues remained very low (Figure 5A). Subsequently, the position of cyclization within the macrocycle was varied. Based on the disappointing results obtained in the hexa-arginine series, we focused this part of the study on hepta-arginine derivatives. Thus, hepta-arginine macrocycles containing 7, 6, 5, 4, or 3 endocyclic and 0, 1, 2, 3, or 4 exocyclic Arg residues, respectively, were synthesized (Table 1). Flow cytometry data showed an important effect of changing the position of macrocyclization on cellular uptake. Indeed, all the hepta-arginine cyclic peptides with exocyclic arginine residues possessed enhanced cellular internalization compared to CycR7 with 7 endocyclic Arg residues. Interestingly, the cyclic form Cyc-R4-R3 was able to penetrate cells more efficiently than the other hepta-arginine derivatives, and possessed an uptake comparable to that displayed by linear nona-arginine (Figure 5B). This important result confirms that structural modification of the peptide enhances cellular uptake per Arg residue in this macrocyclic series; however, a clear relationship remains to be identified. We performed circular dichroism analyses of the peptides presented in this manuscript (SI Figures S6–S8); however, they did not lead to any conclusive relationship between apparent secondary structure and cellular uptake. Results from confocal microscopy performed on fixed or live cells (Figure 6 and SI Figures S3–S5) were similar and demonstrated qualitatively that 7 Arg residues are necessary to reach a decent level of uptake. Conclusion We have synthesized diversely cyclized macrocyclic oligoarginine analogues in order to better understand the influence of conformational restrictions on cellular penetration in HeLa cells. Similarly to linear peptides, our results showed that increasing arginine content in the macrocycles enhanced cellular uptake. We also demonstrated that incorporation of 6 arginines in bicyclic peptides was not sufficient to enhance cellular uptake. Finally, we explored the effect of stereochemistry and mode of cyclization and provided the evidence of an important effect on cellular uptake. In this series, the most efficient macrocycle contained 4 endocyclic and 3 exocyclic Arg residues. We are currently investigating additional macrocyclic analogues of the hepta-arginine family, as well as further details on the mechanisms of cellular uptake and the potential of this new macrocycle to mediate the cellular uptake of functional cargos. Experimental Section Peptide Synthesis Linear and cyclic peptides were synthesized using standard solid-phase synthesis with Fmoc chemistry on a Tribute peptide synthesizer (Protein Technologies Inc.) with IR activation. The protected amino acids Fmoc-l-Arg(Pbf)-OH, Fmoc-d-Arg(Pbf)-OH, Fmocl-Glu(OAll)-OH, Fmoc-l-Lys(Alloc)-OH, and Fmoc-βAla-OH were purchased from ChemImpex International and used with no further purification. The protected peptide 294 resins used to synthesize the macrocycles were prepared using 0.25 g of Rink amide resin from Rapp Polymere (0.23 mmol of NH2/g of resin) by first coupling to the resin, previously deprotected with piperidine/DMF (1:1). Couplings were performed using HATU (1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate), using ∼20 min for all amino acids as monitored by UV. Upon complete formation of the protected linear peptide, resins were washed with dimethylformamide (DMF), isopropanol, and dichloromethane (DCM) and dried in vacuo. At this step, the Alloc and Allyl protecting groups of Lys and Glu residues were simultaneously removed. A solution of dimedone (15 equiv) in 10 mL dry tetrahydrofuran (THF) was added under argon to the resin. Subsequently, 0.45 equiv of Pd(PPh3)4 was added to the resin and the reaction was mechanically stirred for 3 h under Ar. Then, the resin was washed with DMF, isopropanol, and DCM. To perform macrocyclization, the resin was suspended in 20 mL DMF, followed by side-chain to side-chain cyclization of the peptide by addition of PYBOP (Benzotriazol-1yl-oxytripyrrolidinophosphonium hexafluorophosphate, 10 equiv), and DIPEA (4 equiv) overnight. The last step of synthesis was to attach the FITC fluorophore by deprotecting the peptide resin with piperidine/DMF (1:1), followed by reaction with FITC in DCM/Pyridine (7:3) during 30 min. Because FITC conjugates are susceptible to Edman degradation, we used β-alanine as a spacer.(51) The bicyclic peptides were prepared using the standard solid phase method described in SI Figure S1. Peptides were cleaved from the resin and the Pbf protecting groups were simultaneously removed using a mixture of TFA/TIPS/H2O (95:2.5:2.5) for 5 h. The resin was filtered, washed with 1 mL of the cleavage solution, and then the crude peptide recovered by precipitation with cold diethyl ether to give an orange powder that was purified by preparative HPLC (Waters Autosampler 2707, Quaternary gradient module 2535, UV detector 2489, fraction collector WFCIII) equipped with an ACE5 C18 column (250 × 21.2 mm, 5 μm spherical particle size) and water + 0.1% TFA and acetonitrile as eluents. The purification was monitored at 245 nm and the fractions corresponding to the major peak were collected, pooled, and lyophilized. Purities and molecular weights of the corresponding peptides were determined using an Acquity H-Class UPLC-MS system with PDA UV and SQD2Mass detectors equipped with a C18 column (2.1 × 50 mm, 1.7 μm spherical particle size column). All peptides possessed UV purity >92% (for characterization details, please refer to Supporting Information). Confocal Microscopy HeLa cells were purchased from the American Type Culture Collection (Manassas, VA, USA). Approximately 5 × 104 cells were plated on coverslips in 35 mm culture dishes. The cells were grown for 24 h in Dulbecco’s modified Eagle’s high glucose medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT, USA), and 1% Penicillin–Streptomycin–Glutamine solution (Invitrogen, Carlsbad, CA, USA). The day of the experiment, cells were treated by adding 5 μM FITC-labeled peptide to the culture medium and incubated for 30 min at 37 °C in the presence of 5% CO2. Cells were then washed twice with phosphate buffered saline (PBS) and fixed for 30 min in 3% paraformaldehyde (PFA) in 100 mM phosphate 295 buffer, pH 7.4. PFA was quenched for 10 min using a solution of 50 mM ammonium chloride and nuclei were marked using 1 μg/mL Hoechst 33342 stain for 10 min. The specimens mounted on slides using mounting medium (1% n-propyl gallate in glycerol/PBS (1:1)) and visualized using an inverted confocal laser-scanning microscope (FV1000, Olympus, Tokyo, Japan) equipped with a PlanApo 60×/1.42 oil immersion objective (Olympus, Tokyo, Japan). Olympus Fluoview software version 1.6a was used for image acquisition and analysis. The images were further processed using Adobe Photoshop (Adobe Systems, San Jose, CA, USA). FITC-labeled peptides and nuclei were excited at 488 and 405 nm and detected with 500–600 and 425–475 nm band-pass filters, respectively. Flow Cytometry Approximately 5 × 104 HeLa cells were plated on 35 mm culture dishes. After 48 h, cells were treated with the FITC-labeled peptides (5 mM) as described above. Cells were washed once with PBS and taken off using Trypsin-EDTA 0.05% (GIBCO, Invitrogen, Carlsbad, CA, USA). Finally, cells were resuspended in PBS buffer containing 10 mg/mL propidium iodide (PI) to establish the live gate to exclude debris and dead cells. A minimum of 10 000 gated events by sample were acquired and analyzed by a FACScan cytometer (Becton Dickinson, Mountain View, CA) equipped with a 15 mW argon ion laser tuned at 488 nm. The emitted fluorescences were split and collected as follows: FITC 530 ± 15 nm (green), PI 585 ± 21 nm (orange). Flow cytometry analyses were performed in the presence of Trypan Blue to ensure extracellular fluorescence quenching. Cells were treated with peptides in the conditions described above. Signal was then acquired a first time in cytometry. Cells were then incubated with Trypan Blue (0.05% w/v in PBS) for 3 min before the second signal acquisition. Finally, the same cells were permeabilized with 0.05% (v/v) Triton-X100 and fluorescence was measured one last time by cytometry. For live cell experiments, cells were cultured and treated with peptides as described above. The coverslips were mounted directly onto slides using mounting medium and visualized within the next 10 min. Acknowledgment This work was financially supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). H.L. acknowledges NSERC for a graduate fellowship. E. Marsault is a member of the FRQNT-funded Proteo network and the FRQS-funded Réseau Québécois de Recherche sur le Médicament (RQRM). References 1. Frankel, A. D. and Pabo, C. O. (1988) Cellular uptake of the tat protein from human immunodeficiency virus Cell 55, 1189– 1193 296 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Green, M. and Loewenstein, P. M. (1988) Autonomous functional domains of chemically synthesized human immunodeficiency virus tat trans-activator protein Cell 55, 1179– 1188 Rudolph, C., Plank, C., Lausier, J., Schillinger, U., Muller, R. H., and Rosenecker, J. (2003) Oligomers of the arginine-rich motif of the HIV-1 TAT protein are capable of transferring plasmid DNA into cells J. Biol. Chem. 278, 11411– 11418 Perez, F., Joliot, A., Bloch-Gallego, E., Zahraoui, A., Triller, A., and Prochiantz, A. (1992) Antennapedia homeobox as a signal for the cellular internalization and nuclear addressing of a small exogenous peptide J. Cell. Sci. 102 (Pt 4) 717– 722 Derossi, D., Chassaing, G., and Prochiantz, A. (1998) Trojan peptides: the penetratin system for intracellular delivery Trends Cell. Biol. 8, 84– 87 Kondo, E., Saito, K., Tashiro, Y., Kamide, K., Uno, S., Furuya, T., Mashita, M., Nakajima, K., Tsumuraya, T., Kobayashi, N., and M 2012, Tumour lineage-homing cell-penetrating peptides as anticancer molecular delivery systems Nat. Commun. 3, 951– 954 Kwon, E. J., Liong, S., and Pun, S. H. (2010) A truncated HGP peptide sequence that retains endosomolytic activity and improves gene delivery efficiencies Mol. Pharmacol. 7, 1260– 1265 Fonseca, S. B., Pereira, M. P., and Kelley, S. O. (2009) Recent advances in the use of cell-penetrating peptides for medical and biological applications Adv. Drug Delivery Rev. 61, 953– 964 Tung, C. H. and Weissleder, R. (2003) Arginine containing peptides as delivery vectors Adv. Drug Delivery Rev. 55, 281– 294 Cerrato, C. P., Lehto, T., and Langel, U. (2014) Peptide-based vectors: recent developments Biomol. Concepts 5, 479– 488 Copolovici, D. M., Langel, K., Eriste, E., and Langel, U. (2014) Cell-penetrating peptides: design, synthesis, and applications ACS Nano 8, 1972– 1994 Wender, P. A., Mitchell, D. J., Pattabiraman, K., Pelkey, E. T., Steinman, L., and Rothbard, J. B. (2000) The design, synthesis, and evaluation of molecules that enable or enhance cellular uptake: peptoid molecular transporters Proc. Natl. Acad. Sci. U.S.A. 97, 13003– 13008 Wender, P. A., Galliher, W. C., Goun, E. A., Jones, L. R., and Pillow, T. H. (2008) The design of guanidinium-rich transporters and their internalization mechanisms Adv. Drug Delivery Rev. 60, 452– 472 Mitchell, D. J., Kim, D. T., Steinman, L., Fathman, C. G., and Rothbard, J. B. (2000) Polyarginine enters cells more efficiently than other polycationic homopolymers J. Pept. Res. 56, 318– 325 Eiriksdottir, E., Mager, I., Lehto, T., El Andaloussi, S., and Langel, U. (2010) Cellular internalization kinetics of (luciferin-) cell-penetrating peptide conjugates Bioconjugate Chem. 21, 1662– 72 Madani, F., Lindberg, S., Langel, U., Futaki, S., and Graslund, A. (2011) Mechanisms of cellular uptake of cell-penetrating peptides J. Biophys. 2011, 1– 10 Duchardt, F., Fotin-Mleczek, M., Schwarz, H., Fischer, R., and Brock, R. A. (2007) comprehensive model for the cellular uptake of cationic cell-penetrating peptides Traffic 8, 848– 866 297 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. Naik, R. J., Chandra, P., Mann, A., and Ganguli, M. (2011) Exogenous and cell surface glycosaminoglycans alter DNA delivery efficiency of arginine and lysine homopeptides in distinctly different ways J. Biol. Chem. 286, 18982– 18993 Wallbrecher, R., Verdurmen, W. P., Schmidt, S., Bovee-Geurts, P. H., Broecker, F., Reinhardt, A., van Kuppevelt, T. H., Seeberger, P. H., and Brock, R. (2014) The stoichiometry of peptide-heparan sulfate binding as a determinant of uptake efficiency of cell-penetrating peptides Cell. Mol. Life Sci. 71, 2717– 2729 Futaki, S., Suzuki, T., Ohashi, W., Yagami, T., Tanaka, S., Ueda, K., and Sugiura, Y. (2001) Arginine-rich peptides. An abundant source of membrane-permeable peptides having potential as carriers for intracellular protein delivery J. Biol. Chem. 276, 5836– 5840 Rydstrom, A., Deshayes, S., Konate, K., Crombez, L., Padari, K., Boukhaddaoui, H., Aldrian, G., Pooga, M., and Divita, G. (2011) Direct translocation as major cellular uptake for CADY self-assembling peptide-based nanoparticles PLoS One 6, e25924 Fittipaldi, A., Ferrari, A., Zoppe, M., Arcangeli, C., Pellegrini, V., Beltram, F., and Giacca, M. (2003) Cell membrane lipid rafts mediate caveolar endocytosis of HIV-1 Tat fusion proteins J. Biol. Chem. 278, 34141– 34149 Mager, I., Eiriksdottir, E., Langel, K., El Andaloussi, S., and Langel, U. (2010) Assessing the uptake kinetics and internalization mechanisms of cell-penetrating peptides using a quenched fluorescence assay Biochim. Biophys. Acta 1798, 338– 343 Vives, E. (2003) Cellular uptake [correction of uptake] of the Tat peptide: an endocytosis mechanism following ionic interactions J. Mol. Recog. 16, 265– 271 Pae, J., Saalik, P., Liivamagi, L., Lubenets, D., Arukuusk, P., Langel, U., and Pooga, M. (2014) Translocation of cell-penetrating peptides across the plasma membrane is controlled by cholesterol and microenvironment created by membranous proteins J. Controlled Release 192, 103– 113 Tsumuraya, T. and Matsushita, M. (2014) COPA and SLC4A4 are required for cellular entry of arginine-rich peptides PLoS One 9, e86639 Elson-Schwab, L., Garner, O. B., Schuksz, M., Crawford, B. E., Esko, J. D., and Tor, Y. (2007) Guanidinylated neomycin delivers large, bioactive cargo into cells through a heparan sulfate-dependent pathway J. Biol. Chem. 282, 13585– 13591 Geisler, I. and Chmielewski, J. (2009) Cationic amphiphilic polyproline helices: sidechain variations and cell-specific internalization Chem. Biol. Drug Des. 73, 39– 45 Okuyama, M., Laman, H., Kingsbury, S. R., Visintin, C., Leo, E., Eward, K. L., Stoeber, K., Boshoff, C., Williams, G. H., and Selwood, D. L. (2007) Small-molecule mimics of an alpha-helix for efficient transport of proteins into cells Nat. Meth. 4, 153– 159 Wender, P. A., Cooley, C. B., and Geihe, E. I. (2012) Beyond Cell Penetrating Peptides: Designed Molecular Transporters Drug Discovery Today Technol. 9, e49– e55 Lattig-Tunnemann, G., Prinz, M., Hoffmann, D., Behlke, J., Palm-Apergi, C., Morano, I., Herce, H. D., and Cardoso, M. C. (2011) Backbone rigidity and static presentation of guanidinium groups increases cellular uptake of arginine-rich cellpenetrating peptides Nat. Commun. 2, 453– 458 298 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. Prochiantz, A. (2007) For protein transduction, chemistry can win over biology Nat. Meth. 4, 119– 120 White, C. J. and Yudin, A. K. (2011) Contemporary strategies for peptide macrocyclization Nat. Chem. 3, 509– 524 Hill, T. A., Shepherd, N. E., Diness, F., and Fairlie, D. P. (2014) Constraining cyclic peptides to mimic protein structure motifs Angew. Chem., Int. Ed. 53, 13020– 13041 Delorbe, J. E., Clements, J. H., Whiddon, B. B., and Martin, S. F. (2010) Thermodynamic and structural effects of macrocyclization as a constraining method in protein-ligand interactions ACS Med. Chem. Lett. 1, 448– 452 Marsault, E. and Peterson, M. L. (2011) Macrocycles are great cycles: applications, opportunities, and challenges of synthetic macrocycles in drug discovery J. Med. Chem. 54, 1961– 2004 Qian, Z., Liu, T., Liu, Y. Y., Briesewitz, R., Barrios, A. M., Jhiang, S. M., and Pei, D. (2013) Efficient delivery of cyclic peptides into mammalian cells with short sequence motifs ACS Chem. Biol. 8, 423– 431 Mandal, D., Nasrolahi Shirazi, A., and Parang, K. (2011) Cell-penetrating homochiral cyclic peptides as nuclear-targeting molecular transporters Angew. Chem., Int. Ed. 50, 9633– 9637 Hu, Y., Liu, X., Sinha, S. K., and Patel, S. (2014) Translocation thermodynamics of linear and cyclic nonaarginine into model DPPC bilayer via coarse-grained molecular dynamics simulation: implications of pore formation and nonadditivity J. Phys. Chem. B 118, 2670– 2682 Nischan, N., Herce, H. D., Natale, F., Bohlke, N., Budisa, N., Cardoso, M. C., and Hackenberger, C. P. (2014) Covalent attachment of cyclic TAT peptides to GFP results in protein delivery into live cells with immediate bioavailability Angew. Chem., Int. Ed. 53, 1– 5 Oh, D., Nasrolahi Shirazi, A., Northup, K., Sullivan, B., Tiwari, R. K., Bisoffi, M., and Parang, K. (2014) Enhanced cellular uptake of short polyarginine peptides through fatty acylation and cyclization Mol. Pharmacol. 11, 2845– 2854 Qian, Z., LaRochelle, J. R., Jiang, B., Lian, W., Hard, R. L., Selner, N. G., Luechapanichkul, R., Barrios, A. M., and Pei, D. (2014) Early endosomal escape of a cyclic cell-penetrating peptide allows effective cytosolic cargo delivery Biochemistry 53, 4034– 4046 Cascales, L., Henriques, S. T., Kerr, M. C., Huang, Y. H., Sweet, M. J., Daly, N. L., and Craik, D. J. (2011) Identification and characterization of a new family of cellpenetrating peptides: cyclic cell-penetrating peptides J. Biol. Chem. 286, 36932– 36943 Wallbrecher, R., Depre, L., Verdurmen, W. P., Bovee-Geurts, P. H., van Duinkerken, R. H., Zekveld, M. J., Timmerman, P., and Brock, R. (2014) Exploration of the design principles of a cell-penetrating bicylic peptide scaffold Bioconjugate Chem. 25, 955– 964 Duchardt, F., Ruttekolk, I. R., Verdurmen, W. P., Lortat-Jacob, H., Burck, J., Hufnagel, H., Fischer, R., van den Heuvel, M., Lowik, D. W., and Vuister, G. W. 2009, Cell-penetrating peptide derived from human lactoferrin with conformationdependent uptake efficiency J. Biol. Chem. 284, 36099– 36108 299 46. Verdurmen, W. P., Bovee-Geurts, P. H., Wadhwani, P., Ulrich, A. S., Hallbrink, M., van Kuppevelt, T. H., and Brock, R. (2011) Preferential uptake of L- versus D-amino acid cell-penetrating peptides in a cell type-dependent manner Chem. Biol. 18, 1000– 1010 47. Hed, J., Hallden, G., Johansson, S. G., and Larsson, P. (1987) The use of fluorescence quenching in flow cytofluorometry to measure the attachment and ingestion phases in phagocytosis in peripheral blood without prior cell separation J. Immunol. Methods 101, 119– 125 48. Oh, D., Darwish, S. A., Shirazi, A. N., Tiwari, R. K., and Parang, K. (2014) Amphiphilic bicyclic peptides as cellular delivery agents ChemMedChem 9, 2449– 2453 49. Richard, J. P., Melikov, K., Vives, E., Ramos, C., Verbeure, B., Gait, M. J., Chernomordik, L. V., and Lebleu, B. (2003) Cell-penetrating peptides. A reevaluation of the mechanism of cellular uptake J. Biol. Chem. 278, 585– 590 50. Rothbard, J. B., Kreider, E., VanDeusen, C. L., Wright, L., Wylie, B. L., and Wender, P. A. (2002) Arginine-rich molecular transporters for drug delivery: role of backbone spacing in cellular uptake J. Med. Chem. 45, 3612– 3618 51. Julien, M. H. A., Mauras, A., Puget, K., Amblard, M., Martinez, J., and Subra, G. (2009) N-terminus FITC labeling of peptides on solid support: the truth behind the spacer Tetrahedron Lett. 50, 260– 263 300
