(lh3) Glycosyltransferase and Type XVIII Collagen Are Critical for

Neuron 50, 683–695, June 1, 2006 ª2006 Elsevier Inc.
DOI 10.1016/j.neuron.2006.04.024
The Myotomal diwanka (lh3) Glycosyltransferase
and Type XVIII Collagen Are Critical
for Motor Growth Cone Migration
Valerie A. Schneider1 and Michael Granato1,*
1
University of Pennsylvania School of Medicine
Department of Cell and Developmental Biology
Philadelphia, Pennsylvania 19104
Summary
The initial migration of motor growth cones from the
spinal cord into the periphery requires extrinsic
cues, yet their identities are largely unknown. In zebrafish diwanka mutants, motor growth cones are motile
but fail to pioneer into the periphery. Here, we report
on the positional cloning of diwanka and show that it
encodes LH3, a myotomally expressed multifunctional
enzyme with lysyl hydroxylase and glycosyltransferase domains. Cloning, expression analysis, and ubiquitous overexpression of other LH family members reveals that only diwanka (lh3) possesses a critical role
in growth cone migration. We show that this unique
role depends critically on the LH3 glycosyltransferase
domain, and provide compelling evidence that diwanka (lh3) acts through myotomal type XVIII collagen, a ligand for neural-receptor protein tyrosine
phosphatases that guide motor axons. Together, our
results provide the first genetic evidence that glycosyltransferase modifications of the ECM play a critical
role during vertebrate motor axon migration.
Introduction
During development, growth cones migrate over great
distances to reach synaptic targets. In order to make appropriate pathfinding decisions during this migration,
they integrate guidance information provided by a variety of signals, including attractants and repellents, axonally expressed adhesion molecules, and components
of the extracellular matrix (ECM; reviewed in Dickson
[2002]; Krull and Koblar, 2000). It is now also clear that
the signaling activities of many of these guidance cues
are dependent, at least in part, upon their interactions
with one another, as well as with ECM proteoglycans (reviewed in Huber et al. [2003]; Yu and Bargmann, 2001). In
particular, heparan sulfate and chondroitin sulfate proteoglycans (HSPGs and CSPGs, respectively) bind several families of guidance cues and regulate their activities (Ethell et al., 2001; Geisbrecht et al., 2003). For
example, HSPGs are required for the effective binding
and repulsive signaling of Slit ligands through Robo receptors (Hu, 2001; Johnson et al., 2004; Liang et al.,
1999; Steigemann et al., 2004). Similarly, the interactions
of various HSPG and CSPG proteoglycans with Semaphorin 5A render this guidance cue attractive or repulsive to migrating growth cones (Kantor et al., 2004).
An emerging theme is that modifications of HSPG and
CSPG core proteins by ER-localized enzymes, including
glycosyltransferases, are critical for their specific role in
*Correspondence: [email protected]
axonal guidance (reviewed in Holt and Dickson [2005];
Lee and Chien, 2004). For example, a neural-specific
knockout of mouse EXT1, a key glycosyltransferase enzyme required for HSPG GAG chain synthesis, causes
retinal axons to misproject within the optic tract (Inatani
et al., 2003). Similarly, the zebrafish glycosyltransferases Ext2 and Extl3 both play central roles for the projection and appropriate sorting of retinal axons within
the optic tract (Lee et al., 2004). The specific nature of
the axonal defects observed in these mutants indicates
that the interactions mediated by complex carbohydrate
chains are required to direct very specific steps of axonal guidance and that the precise spatial and temporal
regulation of proteoglycan GAG chain biosynthesis
may play a critical role in this process. This view is supported by genetic studies on three C. elegans genes encoding enzymes that introduce additional HSPG modifications (Bulow and Hobert, 2004).
In contrast to HSPG-modifying enzymes, less is
known about the roles that glycosyltransferase enzymes
regulating simpler carbohydrate modifications (i.e., Nand O-linked) might play in axon guidance. In many instances, the expression of such carbohydrate moieties
is developmentally regulated and correlates with specific changes in axon behavior, consistent with the
idea that these moieties interact with guidance cues to
influence specific aspects of growth cone migration
(Henion et al., 2005; Iglesias et al., 1996; Tai and Zipser,
2002). For example, as motor axons enter the plexus region in the developing vertebrate limb, polysialic acid
(PSA) expression on axonally localized NCAM is upregulated (reviewed in Bruses and Rutishauser [2001]). Moreover, this carbohydrate modification is necessary for
these axons to defasciculate and rearrange into dorsally
and ventrally directed nerves (Allan and Greer, 1998;
Bruses and Rutishauser, 2001). Likewise, the downregulated expression of Gal-b1,3GalNAc containing glycoconjugates in the mesoderm of the developing limb occurs just prior to, and is required for, the migration of the
motor axons into this tissue (Oakley and Tosney, 1991).
Such observations demonstrating that carbohydrate
modifications exhibit temporally and spatially restricted
expression patterns and direct specific steps of growth
cone migration suggest that regulation of the glycosyltransferase enzymes catalyzing their synthesis may be
critical to growth cone migration. However, the specific
role of glycosyltransferase enzymes in motor growth
cone migration has not yet been examined.
We previously showed that the zebrafish diwanka
gene is required for the pioneering of motor axon growth
cones from the spinal cord into the periphery (Zeller and
Granato, 1999; Zeller et al., 2002). In diwanka mutant
embryos, all primary motor axons are affected, either
failing to exit the spinal cord or arresting their migration
soon thereafter. diwanka activity is provided by a subset
of myotomal adaxial cells that initially delineate the presumptive motor path but migrate away as motor growth
cones enter the myotome. Together, these results led us
to hypothesize that diwanka activity in adaxial cells induces modifications in the extracellular environment of
Neuron
684
Figure 1. Positional Cloning of the diwanka
Locus
(A) BAC (italics) and PAC (plain) clones in the
vicinity of the locus. Genetic markers used to
initiate a chromosomal walk are blue. The zebrafish lh3 genomic locus contains 19 exons
spanning 28 kb.
(B) SS, signal sequence. ER, ER retention
signal.
(C) LH proteins add hydroxyl groups to Lys
residues in proteins with -X-Lys-Gly- triplets.
Only LH3 adds galactose and glucose moieties to the modified Lys residue.
(D) Dendrogram showing the relationship of
zebrafish LH3 to other LH family members.
The percentage of amino acid identity is
shown. z, zebrafish; c, chick; r, rat; h, human;
m, mouse; ce, C. elegans; dm, Drosophila.
(E–G) SV-2 labeling of motor axon trajectories
in wild-type, diwanka mutant, and diwanka
mutant embryos rescued with lh3-eGFP
mRNA. Dashed lines, somite boundaries.
Bracket, SV-2 labeling of other axons within
the spinal cord.
Scale bar: 20 mm.
the presumptive motor path, enabling motor growth
cones to pioneer migration into the periphery (Zeller
and Granato, 1999).
Here, we report on the positional cloning of the
diwanka gene and show that it encodes LH3, a peripherally expressed, multifunctional enzyme with lysyl
hydroxylase and glycosyltransferase domains. We
demonstrate that growth cone migration depends critically on the Diwanka (LH3) glycosyltransferase domain,
demonstrating for the first time that this process requires temporally and spatially controlled carbohydrate
modifications. Furthermore, we provide strong evidence that diwanka (lh3) acts in adaxial cells through
a presumptive substrate, type XVIII collagen, a multiplexin collagen known to bind receptor protein tyrosine
phosphatase s (Aricescu et al., 2002). Given the wellestablished role of receptor protein tyrosine phosphatases in motor axon guidance, we propose that diwanka
(lh3)-dependent carbohydrate moieties are required for
the signaling between proteins on the surface of the
myotome, such as type XVIII collagen, with neural proteins, possibly receptor protein tyrosine phosphatases,
thereby guiding growth cones from the spinal cord into
the periphery.
Results
Positional Cloning Reveals that diwanka Encodes
lh3 (plod3)
We previously mapped the diwanka locus to a 0.6 cM
(w380 kb) interval on chromosome 23 (Zeller et al.,
2002). Molecular-genetic mapping narrowed the diwanka genetic interval to 0.07 cM (w50 kb; Figure 1A)
and sequencing of BAC and PAC clones spanning this
interval revealed the presence of two predicted genes.
One of these genes, procollagen lysine 2-oxoglutarate
5-dioxygenase 3 (Plod3, known as lh3), was expressed
in the myotome during the period of axonogenesis and
thus represented an excellent candidate. Sequence
analysis revealed that each of the three diwanka alleles
(diwtv205a, diwts286, and diwtz290) contains a mutation in
the lh3 gene. Two alleles (diwtz290 and diwts286) are nonsense mutations resulting in the production of truncated
proteins (W447* and Q608*; Figure 1B), while the
diwtv205a allele contains an A-to-G mutation in the conserved splice acceptor sequence of exon 5. Analysis of
diwanka (lh3) cDNA from diwtv205a mutant embryos indicates that this mutation causes the skipping of exon 5,
resulting in a subsequent frameshift at amino acid 156
diwanka Guides Motor Axon Migration
685
Figure 2. Diwanka Mutants Exhibit Defects in
Type II Collagen Secretion
(A–C) Cross-sections of 24 hpf embryos
labeled with a type II collagen antibody
(II-II6B3).
(A) In wild-type embryos, type II collagen
is detected primarily in the notochord
sheath (white arrow) and floor plate (yellow
arrowheads). Note expression in small
punctae within notochord cells (white
arrowhead).
(B) In diwanka mutant embryos, the thickness of the notochord sheath is reduced and type II collagen is retained intracellularly (white arrowhead).
(C) Rescue of type II collagen secretion in diwanka mutant embryo injected with lh3-eGFP mRNA.
Scale bar: 10 mm.
and a translational termination codon after amino acid
185. We previously demonstrated that the three mutant
diwanka (lh3) alleles exhibit equal phenotypic strength
and that they behave like amorphic alleles when placed
in combination with a deficiency allele (Zeller and Granato, 1999; Zeller et al., 2002). Consistent with this, we
find that proteins encoded by all three mutant diwanka
(lh3) alleles lack the carboxyterminal ER retention signal,
suggesting that Diwanka (LH3) functions predominantly
in the ER.
To confirm that mutations in the lh3 gene cause the diwanka phenotype, we tested the ability of lh3 to restore
growth cone migration in mutant embryos (Figures 1E–
1G). An mRNA encoding an LH3-eGFP fusion protein
was injected at the one-cell stage into embryos produced from the mating of two diwanka heterozygote
fish. Injected embryos were scored at 24 hr postfertilization (hpf) for motor axonal phenotypes using the SV-2
antibody. In uninjected clutches, an average of 27.2%
of the embryos displayed the characteristic diwanka
motor axonal phenotype (n = 238 embryos; Figure 1F),
while in clutches injected with synthetic lh3-eGFP
mRNA, the percentage of mutant embryos dropped to
0.30% (n = 338; Figure 1G), showing that lh3 mRNA fully
restores growth cone migration. These results, in conjunction with the mapping and sequence analysis, demonstrate that lh3 is the gene affected in diwanka mutants.
LH3 belongs to a family of evolutionarily conserved
proteins required for the co- and posttranslational modification of several secreted proteins, many of which are
ECM components. All LH family members contain
a characteristic endoplasmic reticulum (ER) retention
signal and a lysyl hydroxylase domain that catalyzes
the addition of hydroxyl groups to lysine residues at
the Y position in -X-Y-Gly- repeat units, a motif that generates the helical secondary structure found in collagens
and other proteins (Figures 1B and 1C; Kivirikko and Pihlajaniemi, 1998; Suokas et al., 2003). LH3 is the only family member to also exhibit extracellular enzymatic activities (Salo et al., 2006). Conceptual translation of the 2.2
kb diwanka (lh3) transcript predicts it to encode a 730
amino acid protein with 65% identity to human and
mouse LH3, compared to 59% identity to either zebrafish LH1 or LH2 (Figure 1D). The N-terminal third of the
Diwanka (LH3) protein encodes a glycosyltransferase
domain that is found in all LH3 proteins and distinguishes them from LH1 and LH2 family members
(Figure 1B; Heikkinen et al., 2000; Rautavuoma et al.,
2002; Wang et al., 2002a). This domain possesses two
glycosyltransferase activities that catalyze the subsequent modification of hydroxylysine residues to galactosylhydroxylysine and glucosylgalactosylhydroxylysine
(Figure 1C).
Similar to Other lh3 Orthologs, diwanka Is Required
for Collagen Secretion
Mutations in let-268, the gene encoding the single C. elegans LH ortholog, and knockouts of mouse lh3 show
both genes play a critical role in the biosynthesis of
type IV collagen, an integral component of basement
membranes (Norman and Moerman, 2000; Rautavuoma
et al., 2004; Ruotsalainen et al., 2006). In both C. elegans
let-268 and mouse lh3 mutant embryos, type IV collagen
is not secreted and instead accumulates intracellularly.
We therefore examined whether diwanka (lh3) also plays
a role in collagen modifications. We previously demonstrated that collagen 2a1 mRNA expression is normal
in diwanka mutant embryos (Zeller and Granato, 1999),
and we now examined type II collagen protein expression. In wild-type zebrafish embryos at 24 hpf, type II
collagen protein is predominantly localized to the sheath
surrounding the notochord (Figure 2A). Within notochordal cells, type II collagen expression is restricted
to small punctae that likely represent the presence of
this protein in compartments of the secretory pathway.
In contrast, in 100% of diwanka mutants, type II collagen
labeling in the notochord sheath is reduced, and considerable amounts of protein are detected within the notochordal cells (n = 51; Figure 2B). Injection of lh3-eGFP
mRNA restored type II collagen secretion in 100% of genotypically mutant embryos (n = 11; Figure 2C). Thus, as
in C. elegans and the mouse, diwanka (lh3) plays a critical role in collagen secretion.
The ECM Encountered by Motor Growth Cones
Is Unaffected in diwanka (lh3) Embryos
Identification of the collagen-secretion defect raised the
concern that axonal defects observed in diwanka (lh3)
mutants might be secondary to severe structural defects in the ECM through which growth cones migrate.
Once they exit from the spinal cord and enter the periphery, motor growth cones migrate along the medial surface of the myotome, adjacent to the notochord (Bernhardt et al., 1998). To determine whether mutations in
diwanka (lh3) affect the ECM encountered by motor
growth cones, we examined the expression of several
ECM components present along the motor path.
We found that CSPGs, laminin, HSPGs, and Galb1,3GalNAc-O-Ser/Thr containing glycoproteins were
Neuron
686
Figure 3. ECM Components on the Medial
Myotomal Surface Are Unaffected in diwanka
Embryos
(A and B) CSPG expression, 24 hpf. Dashed
circle, outline of notochord sheath. Arrows
point to the motor path.
(E and F) CSPG (red) and znp-1 (green) labeling of embryos in (A) and (B). Motor axons
extend only along the lateral edge of CSPG
expression.
(C and D) Laminin expression, 24 hpf. Arrows,
medial somite surface. Arrowheads, notochordal sheath.
(G and H) Laminin (red) and znp-1 (green) labeling of embryos in (C) and (D). Motor axons
associate with only the medial somite surface.
(I–N) Ultrastructure of the motor path in 18 hpf
embryos.
(I and L) Low-magnification view of the ECM
between the notochord (n) and medial somite
(s) in wild-type (I) and diwanka (L) embryos.
(J) Notochord sheath in wild-type embryo,
consisting of a trilayered matrix. Collagen fibrils are detected only in the middle layer
(red arrowheads).
(K) Medial myotomal surface of wild-type embryo. The extracellular surface is covered
with loosely organized matrix (arrow).
(M) Notochordal sheath in diwanka embryo.
The middle layer lacks organized collagen fibrils (red arrowheads, compare to [J]), while
the basement membrane (black arrowheads)
and outer layer (blue arrow) appear unaffected.
(N) Medial myotomal surface of diwanka embryo. The associated matrix (arrow) appears
normal (compare to [K]).
Scale bars: (A)–(H), 10 mm; (I) and (L), 500 nm;
(J), (K), (M), and (N), 100 nm.
all expressed in cell types flanking the motor path and
that their expression patterns in diwanka (lh3) mutant
embryos were indistinguishable from those in wildtype siblings (Figures 3A–3H and see Figure S1 in the
Supplemental Data available with this article online).
We also found that migration of sclerotomal cells and
neural crest cells, which travel along the same path as
motor growth cones (Eisen and Weston, 1993; MorinKensicki and Eisen, 1997), is unaffected in diwanka mutant embryos (Figure S1; Zeller and Granato, 1999).
Lastly, we used electron microscopy to examine the
motor path between the notochord and the medial myotomal surface at the ultrastructural level. At 18 hpf, the
zebrafish notochordal sheath consists of a trilayered
matrix, while the medial surface of the somite, which
serves as the substrate for motor growth cone migration, is covered with a loosely organized matrix (Figures
3I–3K; Bernhardt et al., 1998; Cerda et al., 2002). As predicted by the type II collagen staining (Figure 2), the central notochordal sheath layer in diwanka embryos lacks
assembled collagen fibrils, while the two other notochordal layers, including the outermost layer that faces
the motor path, appear unaffected (Figures 3J and
3M). Moreover, the appearance of the myotomal matrix
on which motor growth cones migrate is indistinguishable from that in wild-type embryos (Figures 3K and
3N), demonstrating that mutations in diwanka (lh3) do
not overtly disrupt the ECM encountered by motor
axons. Combined, our results demonstrate that diwanka
(lh3) is not required for ECM development or integrity in
the environment of the peripheral motor path. Rather, diwanka (lh3) acts specifically to enable motor growth
cone pioneering of the peripheral motor path.
The diwanka (lh3) Gene Is Transiently Expressed
in Myotomal Adaxial Cells
To obtain insights into how diwanka (lh3) influences
growth cone migration, we examined the expression
pattern of the diwanka (lh3) gene at various stages of development. Maternal diwanka (lh3) mRNA is ubiquitously
expressed until the high stage (3.3 hpf), at which point it
becomes significantly downregulated. A low level of
ubiquitous expression is then maintained until gastrulation (6 hpf), when diwanka (lh3) mRNA becomes concentrated in axial mesoderm (Figure S2). Throughout the period of axon outgrowth (14–20 somites), diwanka (lh3)
expression in the notochord is strong, while a low level
of expression is detected throughout the myotome.
Fluorescent in situ hybridization coupled with F59 antibody labeling not only confirms myotomal lh3 expression during these stages but also demonstrates its enhanced expression in adaxial cells (Figures 4D–4F). At
24 hpf, lh3 expression in the notochord persists, but myotomal expression (including adaxial cells) is undetectable (Figures 4G–4I and 4L). Thus, consistent with our
previous results demonstrating that diwanka (lh3)
diwanka Guides Motor Axon Migration
687
Figure 4. lh mRNA Expression in Zebrafish
Embryos
(A and B) Cross-sectional views of lh3 expression just prior to (14 somites) and during
(20 somites) motor axon migration. In addition to expression in the notochord and floor
plate, lower levels of lh3 are detected in the
myotome (*) throughout these stages.
(C) Cross-sectional view of lh3 expression at
24 hpf. lh3 transcripts are no longer detectable in the myotome but are found in sclerotome (yellow arrow) and in lateral cells (red
arrow).
(D–F) Just prior to axon extension, myotomal
lh3 expression is enhanced in adaxial cells
(white arrows) and sclerotome (yellow arrow).
(G–I and L) Once motor axons have extended
into the periphery, lh3 expression is seen in
the sclerotome (yellow arrow) but is no longer
detectable in adaxial cells. High-magnification view of box in (I) demonstrates lh3 expression in cells (white arrow) lateral to adaxial cells (red arrow), possibly epithelium and/
or dermamyotome (Devoto et al., 2006).
(J and K) Cross-sectional views of lh1 (J) and
lh2 (K) expression during somitogenesis (14
somite stage). Note the high level of lh2 expression in the myotome.
Scale bars: 20 mm.
activity in adaxial cells is critical for growth cone migration (Zeller and Granato, 1999), we now find low and
transient diwanka (lh3) expression in these cells.
A Unique Role for diwanka (lh3) in Growth Cone
Migration
The observation that diwanka (lh3) expression is spatially and temporally regulated suggests that this gene
has a specific developmental role critical for growth
cone migration. Alternatively, the specific axonal defects observed in diwanka (lh3) mutants could occur if
this protein has a general ‘‘housekeeping’’ role and other
LH family members compensate for the loss of diwanka
(lh3) in tissues other than the myotome. This latter hypothesis predicts that diwanka (lh3) is the only LH family
member expressed in the myotome and that forced expression of other LH family members in this tissue
should rescue the axonal diwanka (lh3) phenotype. To
distinguish between these two scenarios, we first
cloned the genes encoding the two other zebrafish LH
family members, LH1 and LH2, and examined their spatial expression patterns (cloning of these genes and
a detailed description of their expression patterns will
be published elsewhere). At stages of motor axon outgrowth, lh1 mRNA is detected in the notochord, but
not in the myotome (Figure 4J). In contrast, lh2 is expressed at high levels in the notochord, floor plate,
and throughout the myotome (Figure 4K). Thus, before
and during the stages of motor axon migration, the
lh3 and lh2 genes are both expressed in the myotome.
This strongly suggests that the axonal defects observed
in diwanka mutants are not caused by a lack of LH activity in myotomal tissues, including adaxial cells.
To further test the idea that diwanka (lh3) plays
a unique role, we examined the ability of lh1 or lh2 to rescue motor axon migration when overexpressed
throughout diwanka mutant embryos. We made constructs in which eGFP or the myc epitope was fused in
frame to the C termini of LH1 and both splice variants
of LH2. Transfection of these constructs into HeLa cells
confirmed that these fusion proteins are expressed at
levels comparable to the LH3-eGFP fusion protein, and
their colocalization with calnexin, an ER-resident protein, demonstrates their proper subcellular localization
(Figures 5A–5I). We injected mRNA encoding these fusion proteins at the one-cell stage into embryos from diwanka heterozygote crosses and analyzed motor axon
migration at 24 hpf. In contrast to lh3-eGFP (Figure 2G),
we found that the lh1-eGFP and lh2-5xmyc constructs,
although ubiquitously expressed at high levels, were unable to rescue axon migration (Figure 5J). Together,
these data strongly suggest that diwanka (lh3) plays
a distinct and unique role in growth cone migration.
The Diwanka (LH3) Glycosyltransferase Domain Is
Required and Sufficient for Growth Cone Migration
We next asked which features of the Diwanka (LH3) protein are responsible for its unique role in growth cone migration. While all LH proteins have lysyl hydroxylase activity, only LH3 proteins possess an additional domain
having glycosyltransferase activities. To determine if
the LH3 glycosyltransferase activities underlie its unique
Neuron
688
Figure 6. The LH3 Glycosyltransferase Domain Is Sufficient to
Restore Motor Axon Migration
Figure 5. lh3 Plays a Unique Role in Motor Growth Cone Migration
(A–I) HeLa cells transfected with lh3-eGFP, lh1-eGFP, and lh25xmyc constructs.
(A, D, and G) Fluorescent detection of tagged constructs in HeLa
cells.
(B, E, and H) Immunolocalization of calnexin, a resident ER protein.
(C, F, and I) Colocalization of the fusion proteins and calnexin.
(J) Only lh3-eGFP mRNA injection significantly rescues abnormal axonal trajectories in embryos obtained from a diwanka heterozygote
cross. Axonal projections in ten hemisegments were analyzed in
each embryo. n = hemisegments. Embryos in this experiment were
not genotyped. Results are expressed as the mean of multiple injection experiments 6 1 SD (t test; **p < 0.005). Results from injections
of the two lh2 splice variants were indistinguishable and therefore
combined.
ability to rescue motor axon migration, we used site-directed mutagenesis to change the cysteine residue at
position 133 to isoleucine and the aspartate residues
at positions 176, 178, 179, and 180 to alanines, mutations shown previously to reduce LH3 glycosyltransferase activities without disrupting lysyl hydroxylase activity (Wang et al., 2002b). This mutant construct, DXDD,
failed to restore motor axon migration in mutant embryos (Figure 6), suggesting that the glycosyltransferase
domain is critical for diwanka (lh3) function.
To further test if the specific ability of Diwanka (LH3) to
rescue motor axon migration errors stems from its
unique glycosyltransferase domain, we constructed a
chimeric protein in which we replaced the first 256 amino
acids of the LH1-eGFP fusion protein with the first 254
amino acids of LH3 (LH3/1-eGFP). Previous studies
have demonstrated that an equivalent N-terminal frag-
(A) Quantitative analysis of motor axon projections in genotypically
mutant diwanka embryos injected with various lh3 fusion constructs. Ten hemisegments were analyzed in each embryo; n = hemisegments. Results are expressed as the mean of multiple injection
experiments 6 1 SD (t test; *p < 0.02; **p < 0.005; ***p < 0.0001).
Two hundred fifty picograms of each mRNA was injected, with the
exception of lh3A-eGFP (500 pg) and lh3AER-eGFP (600 pg).
(B and C) Uninjected wild-type and diwanka mutant embryos.
(D–G) diwanka mutant embryos injected with the lh3 fusion constructs indicated at panel bottoms.
(H–J) Cross-sectional views of 24 hpf embryos stained for type II collagen (red) and motor axons (green).
(H) Wild-type embryo.
(I) diwanka mutant embryo.
(J) diwanka mutant embryo injected with lh3AER-eGFP mRNA in
which motor axon migration, but not notochordal type II collagen secretion, has been rescued.
White arrowheads indicate the position of the horizontal myoseptum. Scale bars: 20 mm (B) and 10 mm (H).
ment of human LH3 retains glycosyltransferase activities
but completely lacks lysyl hydroxylase activity (Rautavuoma et al., 2002). We injected lh3/1-eGFP mRNA encoding this chimeric fusion protein at the one-cell stage
into embryos from a diwanka heterozygote cross, analyzed motor axons in all embryos at 24 hpf, and then genotyped the embryos. Injection of lh3/1-eGFP mRNA restored growth cone migration in genotypically mutant
embryos, reducing the percentage of hemisegments
exhibiting mutant axonal phenotypes from 95.8% to
1.0% (Figure 6; t test; p < 0.0001), a rescue statistically
equivalent to that achieved with injection of the lh3eGFP mRNA construct (0.45%, t test; p < 0.0001). These
diwanka Guides Motor Axon Migration
689
results demonstrate that the N-terminal glycosyltransferase domain of Diwanka (LH3) plays a critical role in
motor growth cone migration.
We next asked if the glycosyltransferase domain is
sufficient for diwanka (lh3)-mediated growth cone migration. We first generated a mutant form of diwanka
(lh3) predicted to lack lysyl hydroxylase activity. Site-directed mutagenesis was used to change the aspartate
residue at position 661 of the LH3-eGFP fusion protein
to alanine (LH3D661A), a mutation shown to reduce human and mouse LH3 lysyl hydroxylase activities to 5%–
10% of normal levels without reducing glycosyltransferase activities (Heikkinen et al., 2000; Ruotsalainen et al.,
2006). Transfection into HeLa cells confirmed that this
mutant construct is expressed at high levels and is correctly localized to the ER (data not shown). Injection of
this construct rescued motor axon migration errors in diwanka embryos, reducing the percentage of affected
hemisegments from 95.8% to 13.9% (Figure 6; t test;
p < 0.005). These results strongly suggest that the activity of the lysyl hydroxylase domain is dispensable for the
role of Diwanka (LH3) in growth cone migration and that
the glycosyltransferase domain is sufficient to mediate
this process.
Since we could not exclude the possibility that residual lysyl hydroxylase activity present in the LH3D661A
protein functionally contributes to its ability to restore
growth cone migration, we created a version of Diwanka
(LH3) lacking the entire lysyl hydroxylase domain. Injection of an mRNA encoding the diwanka (lh3) glycosyltransferase domain without the ER retention signal did
not rescue axon guidance errors, consistent with our
finding that the proteins encoded by all three mutant diwanka (lh3) alleles lack the ER retention signal (Figure 6,
LH3A-eGFP). In contrast, injection of mRNA encoding
the glycosyltransferase domain fused to the ER retention
signal (LH3AER-eGFP) partially rescued growth cone
migration, reducing the percentage of hemisegments
displaying motor axon errors in genotypically mutant diwanka embryos from 95.8% to 50.1% (Figure 6; t test; p <
0.02). Moreover, we found that rescue of the axonal phenotype occurred independently of the type II collagen
secretion phenotype, as a notable fraction of lh3AEReGFP injected diwanka embryos (9.5%) exhibited strong
motor axonal rescue (over 70% hemisegments rescued),
while still displaying notochordal collagen secretion defects (Figures 6H–6J). Together, our studies demonstrate that motor growth cone migration critically depends on the Diwanka (LH3) glycosyltransferase
domain and reveals a previously unrecognized requirement for hydroxylysine-linked carbohydrate modifications of components along the peripheral motor path.
Type XVIII Collagen Plays a Critical Role in Motor
Growth Cone Migration
To identify through which signaling pathway myotomal
diwanka (lh3) controls growth cone migration, we
sought to identify adaxially expressed LH3 substrates.
The protein targets of LH enzymes are predominantly
collagens, which, in addition to acting as structural proteins, also function as ligands for transmembrane signaling proteins such as integrins, DDR tyrosine kinases,
and receptor protein tyrosine phosphatases (RPTPs;
Aricescu et al., 2002; Halfter et al., 1998; Heino, 2000;
Shrivastava et al., 1997; Vogel et al., 1997). Using expression analysis, dominant-negative, and morpholino
knockdown approaches, we asked if these signaling
pathways play a role in diwanka-mediated motor axonal
migration. While we failed to find functional evidence for
DDR signaling or signaling downstream of integrins to
be essential for motor axon migration (data not shown),
we found that type XVIII collagen plays a critical role in
this process.
Type XVIII collagen (c18) belongs to the multiplexin
group of nonfibrillar collagens. c18 forms homotrimers
characterized by interrupted triple helical motifs and
a unique C-terminal endostatin domain (ES) that has
been shown to inhibit endothelial cell migration (Myllyharju and Kivirikko, 2004). In addition, c18 has a cysteine-rich domain (CRD) homologous to that of WNT binding frizzled receptors and a unique thrombospondin
(TSPN) domain, commonly found in the ectodomain of
semaphorins and other guidance receptors (Figure 7G;
Elamaa et al., 2002). We focused on c18 because it contains multiple -X-Lys-Gly- motifs diagnostic for LH substrates, because it is expressed in adaxial cells (Haftek
et al., 2003), and because it has recently been identified
as a ligand for receptor protein tyrosine phosphatases,
which have key roles in motor axon guidance (reviewed
in Ensslen-Craig and Brady-Kalnay [2004]). We cloned
the full length zebrafish collagen 18a1 (col18a1) cDNA
(details of cloning will be published elsewhere) and examined mRNA expression during the stages of motor
axon migration. Just prior to axonogenesis, col18a1 is
very weakly expressed throughout the myotome and is
strongly expressed in adaxial cells (Figures 7A–7C).
Once motor growth cones exit the spinal cord and adaxial cells begin to migrate radially, however, adaxial expression is no longer detectable (Figures 7D–7F). Thus,
just prior to the migration of motor axons into the periphery, col18a1 and diwanka (lh3) are coexpressed transiently in adaxial cells, consistent with the idea that in
these cells c18 is an LH3 substrate (Figures 4 and 7).
To determine if c18 plays a critical role in motor
growth cone migration, we used an antisense morpholino oligonucleotide (MO) approach. We designed
a splice-blocking MO that induces a premature stop codon, eliminating 463 amino acids from the C terminus,
including the endostatin domain (Figures 7G and 7H). Injection with the maximal dose of the splice-blocking MO
(13 ng; higher doses result in early lethality) causes a partial reduction in levels of col18a1 transcripts, resulting in
morphologically normal embryos (Figure 7H). Antibody
labeling at 24 hpf revealed that blocking c18 only had
a mild effect on adaxial cells but a pronounced effect
on motor axon migration in about 60% of the injected
embryos (n = 146). In these embryos, motor axons
stalled shortly after exiting from the spinal cord, similar
to what is observed in diwanka embryos (Figures 7I–
7M; up to 16% of all hemisegments are affected; average stall = 11.8%; t test; p < 0.005). Thus, the fact that
collagens are well documented substrates for LH enzymes, paired with the spatial and temporal coexpression of lh3 and col18a1 in adaxial cells just prior to the
onset of motor axonal migration, and combined with
identical axonal loss of function phenotypes strongly
suggest that Diwanka (LH3) acts through c18 to control
motor axon migration.
Neuron
690
Figure 7. col18a1 Is Required for Motor Axon
Migration
(A–F) Cross-sectional views of col18a1 expression and adaxial cells at 16 (A–C) and
20 (D–F) somite stages.
(A–C) Similar to diwanka (lh3), myotomal
col18a1 expression is enhanced in premigratory adaxial cells (white arrowheads). Dashed
oval, outline of spinal cord.
(D–F) col18a1 expression in myotomal cells is
progressively eliminated corresponding to
the position of the adaxial cells (white arrow).
(G) Domain structure of c18. Red arrow, position of the MO-induced truncation.
(H) Normal splicing is shown by black lines
and MO-induced aberrant splicing is shown
in red. Red arrowhead, position of MO. Black
arrows, position of PCR primers to detect
col18a1 transcripts. In red are the 33 cryptic
residues preceding a termination codon.
Gel, RT-PCR on uninjected and col18a1
MO-injected embryos. Note the presence of
both normal (upper band) and aberrant (lower
band) transcripts in MO-injected embryos.
(I) Quantitative analysis of motor axon projections in embryos injected with col18a1 MO.
Ten hemisegments were analyzed in each
embryo; n = hemisegments. Results are expressed as the mean of multiple injection experiments 6 1 SD (t test; *p < 0.005).
(J and L) Lateral views of motor axons in 24
hpf wild-type (J) and col18a1 MO-injected
(L) embryos. Yellow asterisks, motor axons
stalled along the early motor path.
(K and M) Lateral views of slow muscle fibers
in same embryos as (J) and (L). Slow muscle
fibers in MO-injected embryos, though
slightly ruffled and occasionally broken at
somite boundaries (white arrows), have
grossly normal morphologies.
Scale bars: 20 mm.
Discussion
Pathfinding of vertebrate motor axons from the spinal
cord into the periphery is controlled by a variety of extrinsic cues (reviewed in Schneider and Granato
[2003]). One less-studied group of such cues includes
glycoconjugates whose expression patterns are correlated with routes of growth cone migration (Davies
et al., 1990; Oakley and Tosney, 1991). A number of recent studies have demonstrated that glycosyltransferase-mediated modifications of ECM components play
a critical role for many guidance decisions (reviewed in
Lee and Chien [2004]), but a role for glycosyltransferases in vertebrate motor axon pathfinding has not
been described. We now show that myotomal expression of the diwanka (lh3) glycosyltransferase is essential
for the pioneering of motor growth cones into the periphery. Furthermore, we identify a bona fide substrate
for Diwanka (LH3), adaxially expressed type XVIII collagen (c18), which functions as a ligand for neuronally expressed receptor protein tyrosine phosphatases. Our
findings provide the first genetic in vivo evidence that
precise spatiotemporal carbohydrate modifications of
substrates like c18 play a critical role for peripheral migration of motor growth cones. Thus, it is tempting to
speculate that diwanka (lh3)-dependent modifications
of c18 in adaxial cells enable this protein to interact
with motor axonally expressed receptor protein tyrosine
diwanka Guides Motor Axon Migration
691
Figure 8. Model
(A) Prior to adaxial cell (yellow) and motor growth cone (red) migration, Diwanka (LH3) catalyzes the addition of sugar residues (blue circles) to
substrate proteins (purple squares). These diwanka-modified proteins are subsequently secreted from the myotome into the future motor path.
(B) As adaxial cells begin their lateral migration, the diwanka-modified proteins along the motor path signal motor growth cones to exit the spinal
cord and pioneer migration into the periphery.
(C) In diwanka mutants, protein substrates are not appropriately glycosylated, and hence motor growth cones stall as they encounter the
periphery.
(D) In adaxial cells, c18 is glycosylated by diwanka (lh3) and subsequently secreted into the ECM along the motor path. Motor growth cone
receptors (possibly RPTPs) bind LH3-modified c18 and signal axon migration.
phosphatases, whose role in motor axon guidance is
well established (Figure 8).
diwanka (lh3) Plays Independent Roles in
Notochordal Development and Axonal Migration
In addition to its expression in the myotome (discussed
below), diwanka (lh3) is highly transcribed in the notochord, where it is essential for type II collagen modifications that lead to secretion into the notochordal sheath.
This finding is consistent with the phenotypes observed
for mutations in other LH family members. For example,
mutations in the C. elegans LH gene let-268 disrupt
secretion of type IV collagen, a major component of
basement membranes (Norman and Moerman, 2000).
Likewise, mutations in human LH1 and LH2 cause
Ehlers-Danlos (VI) and Bruck syndromes, respectively,
which are both characterized by defects in collagen biosynthesis (Hyland et al., 1992; van der Slot et al., 2003).
Finally, mouse embryos lacking LH3 die around embryonic day 9.5 with defects in type IV collagen secretion
that result in the development of an unstable basement
membrane (Rautavuoma et al., 2004; Ruotsalainen et al.,
2006). Thus, LH proteins play a critical role in collagen
modifications and basement membrane stability. While
notochordal collagen secretion is clearly affected in diwanka (lh3) mutants, other cell types expressing lh3,
such as the myotome, develop an apparently intact
basement membrane (Figure 3). This is different from
what is seen in LH3 null mice and might be due to compensation by maternally deposited diwanka (lh3) mRNA
or myotomal expression of other LH family members
(e.g., LH2; Figure 4).
Motor axons pioneer into the periphery along the medial surface of the myotome, adjacent to the notochord.
In diwanka mutants, the lack of mature type II collagen in
the notochordal sheath might therefore be hypothesized
to be the primary cause of the motor axonal defects.
However, our data and that of others show that the
role of diwanka (lh3) in the notochord is independent
of its role for motor axonal migration. Genetic studies
have demonstrated that the zebrafish notochord and
floor plate are dispensable for the migration of primary
motor growth cones into the periphery (Greenspoon
et al., 1995; Hatta, 1992). Furthermore, chimeric analyses have shown that only wild-type-derived adaxial
cells, but not notochord cells, efficiently restore axon
migration (Zeller and Granato, 1999). Consistent with
these findings we show here that diwanka (lh3) axonal
migration errors can be rescued independently of the
notochordal collagen defect, and vice versa (Figure 6
and data not shown). Thus, we conclude that diwanka
(lh3) plays a role in the notochord separate from its
role in growth cone migration.
The Role of the diwanka (lh3) Glycosyltransferase
in Motor Growth Cone Migration
Among members of the vertebrate LH protein family,
only LH3 proteins have an N-terminal glycosyltransferase domain. This domain was initially isolated as a fragment of purified LH3 that retains the glycosyltransferase
activities, capable of catalyzing subsequent modifications of hydroxylysines in -X-Lys-Gly- triplets to galactosylhydroxylysine and glucosylgalactosylhydroxylysine
in vitro (Figure 1C; Heikkinen et al., 2000; Rautavuoma
et al., 2002; Wang et al., 2002a; Wang et al., 2002b).
Our results here show that ubiquitous overexpression
of the two other zebrafish lh genes, which encode proteins lacking glycosyltransferase domains, fails to rescue axonal defects. Moreover, lh3 constructs carrying
point mutations predicted to abolish glycosyltransferase activity fail to restore axonal migration, suggesting
that the glycosyltransferase activity is essential for axonal migration. This does not, however, exclude a critical
role for the lysyl hydroxylase domain, as LH3-mediated
carbohydrate modifications can only be added to hydroxylysine residues. It is therefore likely that the axonal
rescue observed by overexpression of the LH3 glycosyltransferase domain depends upon myotomal lysyl hydroxylase activity provided by LH2. This is consistent
with previous findings that there is no strict protein substrate specificity for the lysyl hydroxylase activities of
the different LH family members (Risteli et al., 2004;
Wang et al., 2000). Nonetheless, our data show that
diwanka (lh3) plays a specific role in motor axon guidance and that it exerts its unique function through its
glycosyltransferase domain.
Neuron
692
Type XVIII Collagen (c18) Might Function
as a Link between diwanka (lh3) and Axonal
Guidance Receptors
We had previously shown that diwanka (lh3) is not required for motor axon extension in general but for control of a distinct step of migration, namely the exit of motor growth cones from the spinal cord into the periphery
(i.e., the myotome; Zeller and Granato, 1999). Our analyses show that diwanka (lh3) is transiently expressed in
the myotome and in adaxial cells, consistent with
a model in which dorsal adaxial cells, prior to their own
migration, effect changes in the environment of the presumptive peripheral motor path (Zeller and Granato,
1999). Our data now suggest that diwanka (lh3) exerts
its function through carbohydrate modifications of c18
(and other substrate proteins) produced by dorsal adaxial cells. Just prior to the exit of motor axons from the
spinal cord, diwanka (lh3) substrates, including c18,
are deposited along the future peripheral motor path,
into the ECM of the medial myotomal surface. There
they might signal to motor axonally expressed receptors, thus enabling growth cones to pioneer onto the peripheral path (Figure 8).
The predominant substrates of LH enzymes are collagens and multiple studies have shown that several different types of collagen are capable of modulating
axon growth and migration (Ackley et al., 2001; Lein
et al., 1991). Our analyses reveal that col18a1 expression
in adaxial cells coincides temporally with that of diwanka (lh3), just prior to when motor axons enter into
the periphery. Furthermore, partial MO knockdown of
col18a1 mRNA mimics the very unique diwanka axonal
phenotype (diwanka is the only mutant out of 1200 isolated in the large scale Tübingen screen to exhibit motor
axonal stalling at the exit point). The observation that
MO knockdown of c18 has a less severe effect on motor
axon migration than is found in diwanka (lh3) mutants
may be attributed to the fact that MO injection produced
only a partial knockdown of correctly spliced col18a1
mRNA, as well as the possible existence of additional
LH3 substrates involved in motor axon migration. Regardless, our studies provide the first evidence of
a role for this multifunctional collagen in early steps of
motor growth cone migration.
Collagens are known ligands for transmembrane signaling proteins such as integrin receptors and DDR tyrosine kinases (Heino, 2000; Shrivastava et al., 1997; Vogel
et al., 1997). Notably, studies have implicated collagen
signaling through both receptor types in axon extension,
and the carbohydrate moieties are essential for its signaling through DDR tyrosine kinases (Bhatt et al.,
2000; Lein et al., 1991; Vogel et al., 1997). c18 is also considered an HSPG, which have recently been identified as
ligands for receptor protein tyrosine phosphatases
(RPTPs; Aricescu et al., 2002; Halfter et al., 1998). These
transmembrane receptors have long been known to mediate axon guidance and, more recently, RPTPs expressed in vertebrate motor axons have been shown
to bind myotomally expressed ligands (Ensslen-Craig
and Brady-Kalnay, 2004; Sajnani-Perez et al., 2003).
Thus, c18 may signal through an axonally expressed receptor such as RPTP, thereby enabling growth cones to
pioneer onto the peripheral motor path (Figure 8). In
sum, we demonstrate for the first time that motor growth
cones require temporally and spatially controlled carbohydrate modifications to pioneer into the periphery and
that the diwanka (lh3) glycosyl transferase mediates this
process likely through collagens substrates that function as ligands for neuronal signaling receptors. Further
studies are required to reveal the precise mechanisms
by which diwanka (lh3) links ECM modification with signaling at motor growth cones.
Experimental Procedures
diwanka (lh3) Cloning
Genomic clones spanning the diwanka interval were isolated from
P1-artificial (PAC) and bacterial artificial chromosome (BAC) zebrafish libraries and sequenced by the Sanger Centre (Hinxton, UK).
Open reading frames within the diwanka interval were predicted
with Genscan (http://genes.mit.edu/GENSCAN.html) and Fgenes
(http://www.softberry.com/berry.phtml). These sequences were
used to BLAST the NCBI zebrafish EST database to identify cDNA
sequences. One predicted gene had very high homology to EST
BI877565, which contained the entire coding sequence and limited
50 and 30 UTR of lh3. Additional 50 UTR sequence was determined
by 50 -RACE (GeneRacer, Invitrogen). SIM4 was used to predict the
intron-exon boundaries of lh3 (http://pbil.univ-lyon1.fr/sim4.php).
RT-PCR (Superscript II, Invitrogen) was performed on mRNA prepared from pools of 24 hpf diwtv205a, diwts286, and diwtz290 mutant
and Tü wild-type embryos (Trizol, Invitrogen) using multiple primer
sets encompassing the entire lh3 sequence. PCR products were
subcloned and sequenced. All sequences were confirmed through
two or more independent PCRs. Genomic DNA from diwtv205a and
Tü embryos was used as a template to amplify a partial sequence
from intron 4. For details of genotyping, see Supplemental Experimental Procedures.
Fish Maintenance and Breeding
Zebrafish embryos were generated by natural mating and maintained as described by Mullins et al. (1994). Rescue experiments
were performed with all three diwanka alleles in the Tü genetic background.
In Situ Hybridization
Colorimetric in situ hybridizations were performed according to
Odenthal and Nusslein-Volhard (1998), using probes complementary to the lh1, 2, 3, and pax9a coding sequences. Fluorescent
in situ hybridizations were performed essentially as described in
Downes et al. (2002), using a hybridization and wash temperature
of 67ºC. Additionally, a 30 min wash of 0.3% H2O2 in 13 PBS at
room temperature followed by five five-minute washes in 13 PBS
were performed prior to the blocking step. Embryos were washed
extensively in TNT buffer prior to whole-mount immunohistochemistry. Probes complementary to the col18a1 endostatin domain and
lh3 full-length coding sequences were used.
Whole-Mount Immunocytochemistry and Lectin Staining
Embryos in which type II collagen and motor axons were labeled
were fixed in Carnoy’s Fixative and stained as described in Barresi
et al., (2000). For all other antibodies and lectins, staining was performed as described as in Zeller and Granato (1999), with the exception that embryos stained for laminin were permeabilized with 10 mg/
ml Proteinase K for 2 min. Antibodies used were as follows: II-II6B3
(1:10 dilution, Developmental Studies Hybridoma Bank), SV-2 (1:50,
DSHB), F59 (1:10, DSHB), laminin (1:100, Sigma), CSPG (CS-56,
1:200, Sigma), HSPG (10E4, 1:200, Seikagaku Corp.), and znp-1
(1:200, DSHB). Primary antibodies were detected with AlexaFluor
conjugated secondary antibodies (1:500, Molecular Probes).
AlexaFluor 594 conjugated lectin PNA (1:500, Molecular Probes)
and fluorescein conjugated jacalin (1:500, Vector Labs) were used
to visualize Gal-b1,3GalNAc-containing glycoproteins, respectively.
Embryos were imaged with LSM510 (Zeiss) and LCS (Leica) confocal
microscopes.
diwanka Guides Motor Axon Migration
693
Generation of lh Expression Constructs
To detect expression of injected LH constructs, proteins were
tagged with eGFP or five copies of the myc epitope. Using PCR,
the stop codons of LH1, 2, and 3 were mutated, and in-frame fusions
of these tags were added to their C termini. All tagged constructs
were cloned into the pCS2 + expression vector. For details of cloning, see Supplemental Experimental Procedures.
mRNA Injection
mRNA was transcribed from linearized expression constructs using
the SP6 mMessage mMachine kit (Ambion). mRNA was diluted in
DEPC-H2O and 23 injection buffer (0.5% phenol red/240 mM KCl/
40 mM HEPES [pH 7.5]). All embryos were injected at the one-cell
stage with approximately 250 pg mRNA unless otherwise noted.
HeLa Cell Transfection and Immunofluorescence
HeLa cells were maintained in DMEM with 10% fetal bovine serum,
penicillin, streptomycin, and fungizone. Transfections were performed with Fugene 6 (Roche), according to manufacturer’s instructions. The primary antibodies used were calnexin (1:200, Stressgen)
and anti-myc (9E10, 1:500). The secondary antibodies used were Cy3-conjugated goat anti-mouse IgG (1:1000, Jackson Immunoresearch), AlexaFluor 633-conjugated goat anti-rabbit IgG (1:250, Molecular Probes), and FITC-conjugated donkey anti-rabbit IgG (1:500,
Jackson Immunoresearch).
Electron Microscopy
Embryos were fixed in 2.5% glutaraldehyde + 2% paraformaldehyde
in 0.1 M sodium cacodylate buffer (NaCaC), postfixed in 2% osmium
tetraoxide in 0.1 M NaCaC, and then enbloc-stained in aqueous 2%
uranyl acetate. Using graded alcohol and propylene oxide, embryos
were dehydrated and then embedded in EPON 812. Blocks were
cured at 70ºC for 72 hr. Seventy to seventy-five nanometers thick
sections were cut using a Leica ultratome FCS microtome. Thin sections were mounted on formvar-coated single-slot grids, poststained in 1% uranyl acetate in 50% ethanol, followed by a counter
stain in bismuth subnitrite. In another series, embryos were fixed as
described in Morrison-Graham et al. (1990), using cetyl pyridinium
chloride and ruthenium red to stabilize and enhance contrast of
GAGs. Sections were observed on JEOL JEM 1010 Electron Microscope.
Antisense MO Analysis
The splice-blocking MO (50 -TAACAACCTACCTGGATAGAGCCTT30 ) was designed against the col18a1 splice donor site of exon 32
and is predicted to generate a transcript in which exon 31 and 33
are joined, causing a frameshift-induced truncation 32 amino acids
downstream in exon 35. This eliminates 463 amino acids from the
C terminus of the full-length protein, including the last collagenous
and NC1/Endostatin domains. Thirteen nanograms of MO was
injected into embryos at the one-cell stage. For RT-PCR analysis
following MO injection, cDNA template was prepared from RNA
extracted from five embryos at 24 hpf. Primers were as follows: 50 ACCCGCTGTGCCCTTGGA-30 (forward) and 50 -CCCGCTCACCCTT
CATCTCAT-30 (reverse).
Supplemental Data
Supplemental Data for this article can be found with this article online at http://www.neuron.org/cgi/content/full/50/5/683/DC1/.
Acknowledgments
We are grateful to Neelima Shah at the University of Pennsylvania
Biomedical Imaging Core for assistance with electron microscopy
and Sarah Robertson and Margaret Chou for assistance with HeLa
cell transfections. We thank members of the Granato laboratory
for helpful discussions and advice on the manuscript. This work
was supported by an NRSA postdoctoral fellowship to V.A.S. and
grants from the National Science Foundation and the National Institute of Health to M.G.
Received: May 9, 2005
Revised: March 19, 2006
Accepted: April 13, 2006
Published: May 31, 2006
References
Ackley, B.D., Crew, J.R., Elamaa, H., Pihlajaniemi, T., Kuo, C.J., and
Kramer, J.M. (2001). The NC1/endostatin domain of Caenorhabditis
elegans type XVIII collagen affects cell migration and axon guidance. J. Cell Biol. 152, 1219–1232.
Allan, D.W., and Greer, J.J. (1998). Polysialylated NCAM expression
during motor axon outgrowth and myogenesis in the fetal rat.
J. Comp. Neurol. 391, 275–292.
Aricescu, A.R., McKinnell, I.W., Halfter, W., and Stoker, A.W. (2002).
Heparan sulfate proteoglycans are ligands for receptor protein tyrosine phosphatase sigma. Mol. Cell. Biol. 22, 1881–1892.
Barresi, M.J., Stickney, H.L., and Devoto, S.H. (2000). The zebrafish
slow-muscle-omitted gene product is required for Hedgehog signal
transduction and the development of slow muscle identity. Development 127, 2189–2199.
Bernhardt, R.R., Goerlinger, S., Roos, M., and Schachner, M. (1998).
Anterior-posterior subdivision of the somite in embryonic zebrafish:
implications for motor axon guidance. Dev. Dyn. 213, 334–347.
Bhatt, R.S., Tomoda, T., Fang, Y., and Hatten, M.E. (2000). Discoidin
domain receptor 1 functions in axon extension of cerebellar granule
neurons. Genes Dev. 14, 2216–2228.
Bruses, J.L., and Rutishauser, U. (2001). Roles, regulation, and
mechanism of polysialic acid function during neural development.
Biochimie 83, 635–643.
Bulow, H.E., and Hobert, O. (2004). Differential sulfations and epimerization define heparan sulfate specificity in nervous system development. Neuron 41, 723–736.
Cerda, J., Grund, C., Franke, W.W., and Brand, M. (2002). Molecular
characterization of Calymmin, a novel notochord sheath-associated
extracellular matrix protein in the zebrafish embryo. Dev. Dyn. 224,
200–209.
Davies, J.A., Cook, G.M., Stern, C.D., and Keynes, R.J. (1990). Isolation from chick somites of a glycoprotein fraction that causes collapse of dorsal root ganglion growth cones. Neuron 4, 11–20.
Devoto, S.H., Stoiber, W., Hammond, C.L., Steinbacher, P., Haslett,
J.R., Barresi, M.J., Patterson, S.E., Adiarte, E.G., and Hughes, S.M.
(2006). Generality of vertebrate developmental patterns: evidence
for a dermomyotome in fish. Evol. Dev. 8, 101–110.
Dickson, B.J. (2002). Molecular mechanisms of axon guidance. Science 298, 1959–1964.
Downes, G.B., Waterbury, J.A., and Granato, M. (2002). Rapid in vivo
labeling of identified zebrafish neurons. Genesis 34, 196–202.
Eisen, J.S., and Weston, J.A. (1993). Development of the neural crest
in the zebrafish. Dev. Biol. 159, 50–59.
Elamaa, H., Peterson, J., Pihlajaniemi, T., and Destree, O. (2002).
Cloning of three variants of type XVIII collagen and their expression
patterns during Xenopus laevis development. Mech. Dev. 114, 109–
113.
Ensslen-Craig, S.E., and Brady-Kalnay, S.M. (2004). Receptor protein tyrosine phosphatases regulate neural development and axon
guidance. Dev. Biol. 275, 12–22.
Ethell, I.M., Irie, F., Kalo, M.S., Couchman, J.R., Pasquale, E.B., and
Yamaguchi, Y. (2001). EphB/syndecan-2 signaling in dendritic spine
morphogenesis. Neuron 31, 1001–1013.
Geisbrecht, B.V., Dowd, K.A., Barfield, R.W., Longo, P.A., and
Leahy, D.J. (2003). Netrin binds discrete subdomains of DCC and
UNC5 and mediates interactions between DCC and heparin.
J. Biol. Chem. 278, 32561–32568.
Greenspoon, S., Patel, C.K., Hashmi, S., Bernhardt, R.R., and Kuwada, J.Y. (1995). The notochord and floor plate guide growth cones
in the zebrafish spinal cord. J. Neurosci. 15, 5956–5965.
Haftek, Z., Morvan-Dubois, G., Thisse, B., Thisse, C., Garrone, R.,
and Le Guellec, D. (2003). Sequence and embryonic expression of
Neuron
694
collagen XVIII NC1 domain (endostatin) in the zebrafish. Gene Expr.
Patterns 3, 351–354.
Halfter, W., Dong, S., Schurer, B., and Cole, G.J. (1998). Collagen
XVIII is a basement membrane heparan sulfate proteoglycan.
J. Biol. Chem. 273, 25404–25412.
Hatta, K. (1992). Role of the floor plate in axonal patterning in the zebrafish CNS. Neuron 9, 629–642.
Heikkinen, J., Risteli, M., Wang, C., Latvala, J., Rossi, M., Valtavaara,
M., and Myllyla, R. (2000). Lysyl hydroxylase 3 is a multifunctional
protein possessing collagen glucosyltransferase activity. J. Biol.
Chem. 275, 36158–36163.
Heino, J. (2000). The collagen receptor integrins have distinct ligand
recognition and signaling functions. Matrix Biol. 19, 319–323.
Henion, T.R., Raitcheva, D., Grosholz, R., Biellmann, F., Skarnes,
W.C., Hennet, T., and Schwarting, G.A. (2005). Beta1,3-N-acetylglucosaminyltransferase 1 glycosylation is required for axon pathfinding by olfactory sensory neurons. J. Neurosci. 25, 1894–1903.
Holt, C.E., and Dickson, B.J. (2005). Sugar codes for axons? Neuron
46, 169–172.
Morrison-Graham, K., Bork, T., and Weston, J.A. (1990). Association
between collagen and glycosaminoglycans is altered in dermal extracellular matrix of fetal Steel (Sld/Sld) mice. Dev. Biol. 139, 308–
313.
Mullins, M.C., Hammerschmidt, M., Haffter, P., and Nusslein-Volhard, C. (1994). Large-scale mutagenesis in the zebrafish: in search
of genes controlling development in a vertebrate. Curr. Biol. 4, 189–
202.
Myllyharju, J., and Kivirikko, K.I. (2004). Collagens, modifying enzymes and their mutations in humans, flies and worms. Trends
Genet. 20, 33–43.
Norman, K.R., and Moerman, D.G. (2000). The let-268 locus of Caenorhabditis elegans encodes a procollagen lysyl hydroxylase that is
essential for type IV collagen secretion. Dev. Biol. 227, 690–705.
Oakley, R.A., and Tosney, K.W. (1991). Peanut agglutinin and chondroitin-6-sulfate are molecular markers for tissues that act as barriers to axon advance in the avian embryo. Dev. Biol. 147, 187–206.
Odenthal, J., and Nusslein-Volhard, C. (1998). fork head domain
genes in zebrafish. Dev. Genes Evol. 208, 245–258.
Huber, A.B., Kolodkin, A.L., Ginty, D.D., and Cloutier, J.F. (2003). Signaling at the growth cone: ligand-receptor complexes and the control of axon growth and guidance. Annu. Rev. Neurosci. 26, 509–563.
Rautavuoma, K., Takaluoma, K., Passoja, K., Pirskanen, A., Kvist,
A.P., Kivirikko, K.I., and Myllyharju, J. (2002). Characterization of
three fragments that constitute the monomers of the human lysyl hydroxylase isoenzymes 1–3. The 30-kDa N-terminal fragment is not
required for lysyl hydroxylase activity. J. Biol. Chem. 277, 23084–
23091.
Hyland, J., Ala-Kokko, L., Royce, P., Steinmann, B., Kivirikko, K.I.,
and Myllyla, R. (1992). A homozygous stop codon in the lysyl hydroxylase gene in two siblings with Ehlers-Danlos syndrome type VI. Nat.
Genet. 2, 228–231.
Rautavuoma, K., Takaluoma, K., Sormunen, R., Myllyharju, J., Kivirikko, K.I., and Soininen, R. (2004). Premature aggregation of type IV
collagen and early lethality in lysyl hydroxylase 3 null mice. Proc.
Natl. Acad. Sci. USA 101, 14120–14125.
Iglesias, M., Soler, R.M., Ribera, J., Esquerda, J.E., and Comella, J.X.
(1996). The carbohydrate N-acetylglucosamine is involved in the
guidance of neurites from chick ciliary ganglion neurons through
the extracellular matrix of rat skeletal muscle fiber. Neurosci. Lett.
207, 81–84.
Risteli, M., Niemitalo, O., Lankinen, H., Juffer, A.H., and Myllyla, R.
(2004). Characterization of collagenous peptides bound to lysyl hydroxylase isoforms. J. Biol. Chem. 279, 37535–37543.
Hu, H. (2001). Cell-surface heparan sulfate is involved in the repulsive guidance activities of Slit2 protein. Nat. Neurosci. 4, 695–701.
Inatani, M., Irie, F., Plump, A.S., Tessier-Lavigne, M., and Yamaguchi, Y. (2003). Mammalian brain morphogenesis and midline axon
guidance require heparan sulfate. Science 302, 1044–1046.
Johnson, K.G., Ghose, A., Epstein, E., Lincecum, J., O’Connor, M.B.,
and Van Vactor, D. (2004). Axonal heparan sulfate proteoglycans
regulate the distribution and efficiency of the repellent slit during
midline axon guidance. Curr. Biol. 14, 499–504.
Kantor, D.B., Chivatakarn, O., Peer, K.L., Oster, S.F., Inatani, M.,
Hansen, M.J., Flanagan, J.G., Yamaguchi, Y., Sretavan, D.W., Giger,
R.J., and Kolodkin, A.L. (2004). Semaphorin 5A is a bifunctional axon
guidance cue regulated by heparan and chondroitin sulfate proteoglycans. Neuron 44, 961–975.
Kivirikko, K.I., and Pihlajaniemi, T. (1998). Collagen hydroxylases
and the protein disulfide isomerase subunit of prolyl 4-hydroxylases.
Adv. Enzymol. Relat. Areas Mol. Biol. 72, 325–398.
Krull, C.E., and Koblar, S.A. (2000). Motor axon pathfinding in the peripheral nervous system. Brain Res. Bull. 53, 479–487.
Lee, J.S., and Chien, C.B. (2004). When sugars guide axons: insights
from heparan sulphate proteoglycan mutants. Nat. Rev. Genet. 5,
923–935.
Lee, J.S., von der Hardt, S., Rusch, M.A., Stringer, S.E., Stickney,
H.L., Talbot, W.S., Geisler, R., Nusslein-Volhard, C., Selleck, S.B.,
Chien, C.B., and Roehl, H. (2004). Axon sorting in the optic tract requires HSPG synthesis by ext2 (dackel) and extl3 (boxer). Neuron 44,
947–960.
Lein, P.J., Higgins, D., Turner, D.C., Flier, L.A., and Terranova, V.P.
(1991). The NC1 domain of type IV collagen promotes axonal growth
in sympathetic neurons through interaction with the alpha 1 beta 1
integrin. J. Cell Biol. 113, 417–428.
Ruotsalainen, H., Sipila, L., Vapola, M., Sormunen, R., Salo, A.M.,
Uitto, L., Mercer, D.K., Robins, S.P., Risteli, M., Aszodi, A., et al.
(2006). Glycosylation catalyzed by lysyl hydroxylase 3 is essential
for basement membranes. J. Cell Sci. 119, 625–635.
Sajnani-Perez, G., Chilton, J.K., Aricescu, A.R., Haj, F., and Stoker,
A.W. (2003). Isoform-specific binding of the tyrosine phosphatase
PTPsigma to a ligand in developing muscle. Mol. Cell. Neurosci.
22, 37–48.
Salo, A.M., Wang, C., Sipila, L., Sormunen, R., Vapola, M., Kervinen,
P., Ruotsalainen, H., Heikkinen, J., and Myllyla, R. (2006). Lysyl hydroxylase 3 (LH3) modifies proteins in the extracellular space, a novel
mechanism for matrix remodeling. J. Cell Physiol. 207, 644–653.
Schneider, V.A., and Granato, M. (2003). Motor axon migration:
a long way to go. Dev. Biol. 263, 1–11.
Shrivastava, A., Radziejewski, C., Campbell, E., Kovac, L., McGlynn,
M., Ryan, T.E., Davis, S., Goldfarb, M.P., Glass, D.J., Lemke, G., and
Yancopoulos, G.D. (1997). An orphan receptor tyrosine kinase family
whose members serve as nonintegrin collagen receptors. Mol. Cell
1, 25–34.
Steigemann, P., Molitor, A., Fellert, S., Jackle, H., and Vorbruggen,
G. (2004). Heparan sulfate proteoglycan syndecan promotes axonal
and myotube guidance by slit/robo signaling. Curr. Biol. 14, 225–
230.
Suokas, M., Lampela, O., Juffer, A.H., Myllyla, R., and Kellokumpu,
S. (2003). Retrieval-independent localization of lysyl hydroxylase in
the endoplasmic reticulum via a peptide fold in its iron-binding domain. Biochem. J. 370, 913–920.
Tai, M.H., and Zipser, B. (2002). Sequential steps of carbohydrate
signaling mediate sensory afferent differentiation. J. Neurocytol.
31, 743–754.
Liang, Y., Annan, R.S., Carr, S.A., Popp, S., Mevissen, M., Margolis,
R.K., and Margolis, R.U. (1999). Mammalian homologues of the Drosophila slit protein are ligands of the heparan sulfate proteoglycan
glypican-1 in brain. J. Biol. Chem. 274, 17885–17892.
van der Slot, A.J., Zuurmond, A.M., Bardoel, A.F., Wijmenga, C.,
Pruijs, H.E., Sillence, D.O., Brinckmann, J., Abraham, D.J., Black,
C.M., Verzijl, N., et al. (2003). Identification of PLOD2 as telopeptide
lysyl hydroxylase, an important enzyme in fibrosis. J. Biol. Chem.
278, 40967–40972.
Morin-Kensicki, E.M., and Eisen, J.S. (1997). Sclerotome development and peripheral nervous system segmentation in embryonic zebrafish. Development 124, 159–167.
Vogel, W., Gish, G.D., Alves, F., and Pawson, T. (1997). The discoidin
domain receptor tyrosine kinases are activated by collagen. Mol.
Cell 1, 13–23.
diwanka Guides Motor Axon Migration
695
Wang, C., Valtavaara, M., and Myllyla, R. (2000). Lack of collagen
type specificity for lysyl hydroxylase isoforms. DNA Cell Biol. 19,
71–77.
Wang, C., Luosujarvi, H., Heikkinen, J., Risteli, M., Uitto, L., and Myllyla, R. (2002a). The third activity for lysyl hydroxylase 3: galactosylation of hydroxylysyl residues in collagens in vitro. Matrix Biol. 21,
559–566.
Wang, C., Risteli, M., Heikkinen, J., Hussa, A.K., Uitto, L., and Myllyla, R. (2002b). Identification of amino acids important for the catalytic activity of the collagen glucosyltransferase associated with the
multifunctional lysyl hydroxylase 3 (LH3). J. Biol. Chem. 277, 18568–
18573.
Yu, T.W., and Bargmann, C.I. (2001). Dynamic regulation of axon
guidance. Nat. Neurosci. Suppl. 4, 1169–1176.
Zeller, J., and Granato, M. (1999). The zebrafish diwanka gene controls an early step of motor growth cone migration. Development
126, 3461–3472.
Zeller, J., Schneider, V., Malayaman, S., Higashijima, S., Okamoto,
H., Gui, J., Lin, S., and Granato, M. (2002). Migration of zebrafish spinal motor nerves into the periphery requires multiple myotomederived cues. Dev. Biol. 252, 241–256.
Accession Numbers
The accession number for zebrafish lh3 has been deposited in
GenBank (DQ019227).