Use of Gene Replacement Transformation to Elucidate

Copyright 0 1992 by the Genetics Society of America
Use of Gene Replacement Transformationto Elucidate Gene Function in the
qa Gene Clusterof Neurospora crassa
Mary E. Case, Robert F. Geever' and David K. Asch
Department of Genetics, University of Georgia, Athens, GA 30602
Manuscript received September4, 1991
Accepted for publication December 17, 1991
ABSTRACT
Gene replacement by transformation, employing selective genetic recombination techniques,has
been used to delete or disrupt the qa-x, qa-y and qa-IS genes of the qa gene cluster of Neurospora
crassa. The growth characteristics of the strain carrying the deletion of the qa-y gene support earlier
of the qaevidence that this gene encodes
a quinic acid permease.The strain containing the deletion
IS gene (Aqa-IS)was examined with respect to quinic acid induction and carbon catabolite repression.
The Aqa-IS strain exhibits constitutive expression of the qa genes supporting earlier evidence that
the qa-IS gene codesfor a repressor. Severalof the qa genes continuedto be expressed at high levels
even in the presence of glucose in the Aqa-IS strain, which indicates that transcriptionof these genes
is not being affected directlyby a repressor moleculein the presence of glucose.
HE quinic acid (qa) gene cluster of Neurospora
crassa is a well characterized system to study
genetic regulation in a multicellular yet relatively simple organism. T h e qa gene cluster of N. crassa is
located on approximately 17.2 kb of DNA on Neurospora linkage group VI1 (GILESet al. 1985). T h e
DNAsequence of theentire qa cluster is known
(GEEVERet al. 1989). T h e cluster includes five structural genes and two regulatory genes. Three of the
structural genes, qa-2, qa-3 and qa-4 are involved in
the first three steps of quinic acidcatabolism (GILESet
a1 1985). A fourthgene, qa-y, is hypothesized to
encode a quinic acid permease (GEEVERet a1 1989)
based on its similarity to otherknown permeases and
by the phenotypeof a strain carrying a deletion the
of
qa-y gene. The other structural gene,qa-x, is defined
by a quinicacid inducible RNA. However,its function
is unknown. The other two genes of the qa cluster
(qa-IF and qa-IS) code for regulatory proteinson the
basis of genetic and molecular analysis (GILESet al.
1985; GEEVER
et al. 1989).
The genes ofthe qa gene cluster appear to be under
two levels ofgeneticregulation.
T h e first level is
mediated by the inducer quinic acid and by the productsof the qa regulatory genes. T h e qa-IF gene
encodes a protein of 816 amino acids (HUIET1984).
Molecular analysis indicates that the qa-IF protein is
an activator protein which plays a positive role in the
transcriptionof all the qa genes by bindingto a
conserved 16
bp
sequence
(GGRTAARYRYTTATCC) (BAUM, GEEVER
and GILES1987). A total of
14 binding sites for theqa-IF protein have been shown
T
' Current address: Department of Biological Sciences, University of Nevada-Las Vegas, 4504 Maryland Parkway, Las Vegas, Nevada 89154-4004.
Genetics 130: 729-736 (April, 1992)
to exist in the qa genecluster. T h e DNA binding
domain of the activator protein is located in the Nterminal183amino
acids. Several mutations have
been obtained in the qa-IF gene. T h e resulting mutants are recessive and noninducible which has lead to
the hypothesis that they produce defective activators.
T h e qa-IS gene encodes a protein of 18
9 aminoacids
(GEEVERet al. 1987). Thisproteinfunctions
as a
repressor molecule, preventing transcriptionof the qa
genes in the absence of quinic acid. Indirect evidence
suggests that the qa-IS protein is not a DNA-binding
protein but interacts with the qa-IF activator protein
to inhibit activator function (GILESet al. 1985, 1991;
J. AVALOS,R. GEEVER,M. E. CASEand N. H. GILES,
unpublished). Two types of mutations have been detected in the qa-IS gene: one produces mutants in
which the qa genes are noninducibleeven in the
presence of quinicacid and thesecond type produces
mutants in which the qa genes are constitutively expressed in the absence of inducer. Sequence evidence
indicates that the noninducible mutants are theresult
of missense mutations which cause the qa-IS gene
product to become a superrepressor that is unable to
respond to the presence of the inducer, quinic acid,
while the constitutivestrainscontain
nonsense or
frame shift mutations which render the qa-IS gene
product nonfunctional.
T h e second control circuit functions to repress transcription of the qa genes in response to the presence
of a preferred carbon source such as glucose or sucrose. Wild-type N . crassa strains grownon quinic acid
together with a preferred carbon source such as sucrose show a level of induction of the qa enzymes
about 1% of the levels of inductionobserved with
730
M. E. Case, R. F. Geever and D. K. Asch
quinic acid as a sole carbon source. T h e GAL system
in Saccharomyces cerevisiae shares many similarities in
terms of gene control with the qa gene cluster and is
also subjectto carboncatabolite repression (JOHNSTON
1987). Carbon repressionof the GAL genes is thought
to operate throughseveral mechanisms such as direct
action on the various GAL promoters, inhibition of
activator binding and inhibition
of galactose transport
(JOHNSTON
1987). However, the mechanism(s) of carboncataboliterepressionon
qa geneexpression is
unknown.
Gene replacement has been widely used in S. cerevisiae tostudygenefunction
(SCHERERand DAVIS
1979). Sincemost transforming DNA in N. crassa
integrates at ectopic sites by nonhomologous recombination (CASEet al. 1979; KINSEYand RAMBOSEK
1984) gene replacement in N . crassa has been more
difficult. There are two selectable markers in the qa
gene cluster, qa-2 and qa-lF, which can be used for
transformation. Double mutants qa-2, aro-9 and qaI F aro-9 are incapable of growth on minimal media
unless supplemented with aromatic amino acids thus
permitting qa-2+ and qa-1Fe transformants to be selected on minimal media. Preliminary reports describe
the inactivation of the qa-x gene (CASEand GEEVER
1987) and the deletion
of the qa-y gene (GEEVERet al.
1989) by transforming with a n inactivated or truncated selectable marker thereby selecting for the rare
integration by homologous recombination. A similar
approach has been used to select for am+ (glutamic
dehydrogenase) transformants(FREDERICK,
ASCHand
KINSEY 1989). FREDERICKand KINSEY (1990) also
employed this technique t o identify those sequences
upstream of the am gene required for full am gene
expression. This paperwill describe in more detail the
characteristics of the strains carrying an altered qa-x
gene or a deleted qa-y gene. To further our understanding of the role of the repressor protein in the
regulation of the qa gene cluster, a strain has been
constructed by gene replacement transformation in
which the qa-1S gene hasbeendeletedfromthe
genome. T h e characteristics of thisstrain,
which
makes no repressor protein, have been examined
with
respect to the rolesof both quinic acid induction and
carboncataboliterepressioninregulating
qa gene
expression. T h e results have important implications
for an understanding of genetic regulation in the qa
gene cluster of N. crassa.
MATERIALSANDMETHODS
Strains and media: N . crassa strains used in this study
are
listed in Table 1. N. crassa was grownonFriesminimal
medium supplemented with aromatic amino acids, p-aminobenzoic acid, inositoland methionine where required. N .
crassa crosses were made
on WESTERGAARD'S
crossing media
(WESTERGAARD
and MITCHELL 1947). All plasmidsused
were grownon Escherichia coli strain JM 101.
TABLE 1
Neurospora strainsused in this study
Strains
Genotype
N. crassa 74A
Wild type
N . crassa
qa-2, aro-9,inl,A
N. crassa
qa-lF, aro-9
N. crassa
met-7-a
N. crassa (128-5-3-2)
N. crassa (241-9-2-1)
N. crassa (227-3-5-10)
Aqa-1S
Aqa-y
qa-x-
Source
Authors stock
collection"
Authors stock
collectionb
Authors stock
collection'
Authors stock
collectiond
This work'
This work'
This workg
FGSC No. 2489.
FGSC No. 3953.
' FGSC No. 6904.
FGSC No. 4088.
Contains deletion of the qa-1S gene.
'Contains deletion of the qa-y gene.
g Contains disruption of the qa-x gene.
a
Construction of plasmids: The plasmid pRDlO (Figure
1A), which contains the disrupted qa-x gene, was constructed
by cleaving at a unique Hind111 site inthe qa-x coding region
and filling in the staggered ends using the Klenow fragment.
This creates a unique NheI site withinthe qa-x coding region.
The plasmid pAqa-IS wasconstructed in the following manner. The 5.7-kb BamHI-NheI fragment containing the qa-3
and qa-y genes was ligated to a 323-bp AvrZI- BamHI fragment from the plasmid pQa-IS. This hybrid fragment was
then ligated into the BamHI site of pQa-IF to forma plasmid
containing the intact qa-3, qa-y and qa-IF genes but deleting
the entire qa-IS gene (Figure 1A). The plasmid pAqa-y was
constructed by ligating a 2.6 kb XbaI-NcoI fragment containing the qa-3 gene to a BglII-XbaI fragment containing
the qa-IS and part of the qa-IF gene. This fragment was
then ligated into a XbaI cleaved vector. The resulting plasmid contained all of qa-3 and qa-IS and enough of qa-IF to
complement the qa-IF mutation; however the qa-y gene had
been deleted (Figure 1B).
Enzymology: Growth of N . crassa strains, induction procedures and assay conditions for catabolic dehydroquinase
weredescribedpreviously
(CASE,HAUTALA
and GILES
1977). Specific conditionsare noted in Table 2.
Neurospora transformation: N. crassa transformation
was performed as described earlier by CASEet al. (1979)
modified to use Novozyme 234 (VOLLMERand YANOFSKY
1986). Transformants wereisolatedtominimalmedium.
Stable isolates were then crossed to a met-7- aro-9+ strain
and qa-2+aro-9+met-7+or q a - I F aro-9+met-7+progeny were
selected for further analysis.
DNA and RNA isolation: N. crassa DNA was prepared
and TIMBERLAKE
(1984).
by the method of YELTON, HAMER
Plasmid DNA was prepared by the method of ISH-HOROWICZ and BURKE(1981). N. crassa total RNA was prepared
et al. (1990).
by the method of LINDGREN
DNA and RNA hybridization: Digested genomic DNA
(2 k g ) was fractionated on 0.8% agarose gelsfor 12-20 hr,
then transferred to nylon membranes (Nytran, Schleicher
and Schuell), hybridizedand rinsed as recommended by the
manufacturer. Total RNA was fractionated by electrophoresis through formaldehyde gels and transfered to Nytran,
hybridized and rinsed as recommended by the manufacturer. Probes were labeledwith "P by the method of FEINBERG and VOGELSTEIN
(1984).
73 1
Gene Replacement in N . crassa
...............
. . . . . . .
1.0 kb
A
H
0
0
-"
Hind111 Bgl I1 S m a I BamHI
BamHI
I
I
qa-x qa
I
I
qa-2
qa-3
1-
3
NheI
BamH
A Iv r I I B a m H I
I
I
I
I
I
Xba I
I
""
qa-4
qa-IF
qa-IS
-Y
B
XbaI
I
"+
qa-x
qa-2
qa-3
40-4
N c o I Bgl I1
I
4 -
90-Y
XbaI
I
I
"
qa-IF qa-IS
1
FIGUREI.-Plasmid construction. A, Construction of pRDlO (qa-x-) and pAqa-1s. T h e crosshatched area represents the fragment used
for the construction o f pRDIO. T h e Hind111 site present in the qa-x gene was filled in to create a unique Nhel site. the squared-in area
represents the fragments used in the construction of the plasmid pAqa-Is. The dotted areas above the line define the probes used in this
study. B, Construction o f pAqa-y. the heavy cross-hatched areas show the fragments used to form the plasmid pdqa-y.
S1 analysis: S1 analysis was performed by the procedure
(1977). To map the transcriptional start
o f BERKand SHARP
site of the ga-2 gene a "P-end-labeled SmaI-BgIII probe
from the ga-2 region (Figure 1A) was hybridized at 52" to
total RNA from strains containing the Aga-IS mutation
grown under various growthconditions.
Hybridizations
were then digested with SI nuclease and fractionated on an
7%denaturing polyacrylamide gel. Sequencing ladders were
prepared by the method of MAXAMand GILBERT
( 1 980).
RESULTS
Effect of disrupting of the qa-x gene: T h e qa-x
gene was originally identified as a quinic acid-inducible transcriptofunknownfunction
(PATEL et al.
198 1). T o examine the function of the
qa-x gene, the
plasmid pRDlO was constructed in which the qa-x
gene was disrupted by filling in a unique Hind111 site
in the qa-x coding region creating a new NheI site in
the amino-terminal end of the protein
as shown in
Figure 1A. This plasmid also contained thega-2 gene,
which could be truncated by cleaving at an SphI site
(Figure 1A). Strains containing qa-2 and aro-9 mutations were transformed using a fragment containing
the truncated ga-2 gene linked to the disrupted qa-x
gene. Since the ga-2 mutation in the recipient strain
had an A/T deletion in codon 8, the only way ga-2+
transformants could arise would be by homologous
recombination in theregionbetweenthemutation
and the SphI site (Figure 2A).
Transformants were selectedfortheir
ability to
grow on minimalmediumandcrossed
to a strain
M. E. Case, R. F. Geever and D. K. Asch
732
TABLE 2
Catabolic dehydroquinase (CDHQ) and quinic dehydrogenase
(QDH) activities in wild-type 74A, aro-9 and Aqa-y strains
C
B
A
13.1 Ib
-
I
--I
1
7.7 bb
Ukb--
Percent wild-type
activity
Strain
Wild-type 74A
Aqa-Y
aro-Yaro-9Aqa-y, aro-9Aqa-y, aro-5"
Conditions
QDH
Quinic acid
Quinic acid
Fries
Quinic acid
Fries 13
Quinic acid
CDHQ
100
10
23
118
25
29
3.2 kb
100
5
14
151
9
Cultures were grown for 24 hr at 25" C with vigorous shaking
in a Fries, sucrose medium. Cultures were harvested and throughly
rinsed to remove all sucrose. Harvested samples were then incubated for 6 hr with no carbon source (Fries alone) or with 0.3%
quinic acid as a sole carbon source.
A
B
FIGURE
2.-Homologous recombination events necessary to produce qa-2' or qa-lF+transformants. A, Homolgous recombination
event necessary to produce a transformant of pRDlO which is qa2+ and contains a disruption of the qa-x gene. The stripped box
represents the qa-x gene while the empty box represents the qa-2
gene. The arrowhead representsthe location of the mutation in the
qa-2 recipient strain. B,Homolgous recombination events necessary
to produce qa-IF transformants. This strategy was used to delete
both the qa-y and qa-IS genes. The stippled box represents the qaI F gene. The arrowhead indicates the location of the mutation in
the qa-IF recipient strain. The filled-box represents sequences
upstream of the qa-IF gene.
containing the closely linked met-7 mutation to obtain
homokaryotic progeny. T h e homokaryotic progeny
isolates could utilize both quinic acid and chlorogenic
acid as carbon sources. T h e only detected phenotype
of strains containing the disrupted qa-x gene is the
accumulation of an unidentified brown pigment following growth on quinic acid as acarbonsource.
Genomic DNA was isolated fromtheseprogeny
strains, digested with NheI, then probed by Southern
blots with a "P-labeled BglII-BamHI fragment containing the qa-2 codingregion(Figure
1A). In the
wild-type strain, this probe would hybridize to a large
NheI fragment greater than 12.0kb. In transformants
in which the qa-x gene had been disrupted the qa-2
probe would hybridize to a NheI fragment of 3.2 kb
in size because of the new NheI site created in the qax coding region. An example of this pattern is shown
in Figure 3A Several transformantswereobtained
which contained a disruptionof the qa-x gene indicat-
-
-
U bb
2
.
---
I
7.7 kb
- 5.2 bb
-
5 3 kb-
U
2
-
&Ikb
bb--
FIGURE3.-Genomic Southern blot of selected homokaryotic
transformants. DNAs in A were digested withNheI then probed
with the BglII-BamHI fragment containing the qa-2 gene (Figure
1A). DNAs in B and C were digested with BamHI and probed with
the three BamHI fragments indicated in Figure 1. A, Lane 1, 74A;
lane 2, a transformant of pRDlO containing the disruption of the
qa-x gene. B, Lane 1, 74A; lane 2, a transformant of the plasmid
pAqa-y containing the deletion of the qa-y gene. C, Lane 1, 74A;
lane 2, a transformantof the plasmid pAqa-IS containing a deletion
of the qa-IS gene. The same Southern patterns were observed in
nonhomokaryotic transformants.
ing that specific alterations could be introduced into
the qa gene cluster by transforming with a truncated
selectable marker. J. HOPPERand E. MATTES(personal
communication) have recently shown thatthe qa-x
gene has 31% homology with the GAL12 gene. This
gene appears to be involved in the dephosphorylation
of the GAL4 activator in the presence of a preferred
carbon source. However, we were unable to demonstrate that the disruption
of qa-x has any effect on the
level of qa-2 gene expression in the presence of glucose (data not shown).
Effect of deleting the
qa-y gene: T h e qa-y gene was
originally identified as a quinic acid-inducible transcript of unknown function (PATELet al. 1981). Subsequent sequence analysis revealed that the qa-y coding region showed a high degree of homology with
the qutD gene of Aspergdlus nidulans (GEEVERet al.
1989). T h e qutD gene is known to encode a quinic
acid permease on thebasis of genetic and physiological
et al. 1987). In order to define
data (WHITTINGTON
further the function of the qa-y gene in N. crassa a
plasmid (pbqa-y) was constructed which contained
intact qa-3 and qa-IS genes, a truncated qa-IF gene
and adeletion of the qa-y gene(Figure 1B). This
construct was transformed into a double mutant strain
of N. crassa containing the qa-IF aro-9 mutations.
Transformants were selected for their ability to grow
on minimal medium. The qa-IFmutation (128) in the
recipientstrain isin amino acid 1 18 of the gene
et al. 1989) and theplasmid pbqa-y contains
(GEEVER
qa-IF sequences through codon 613. Therefore, any
homologous recombination event between these two
points would produce the q a - l p phenotype (Figure
2B). Any other recombination event would leave the
recipient q a - I F a n dtherefore unable to activate the
qa genes to allow growth on minimal medium.
Transformants were crossed to a strain containing
Gene Replacement in N . crassa
tation, the strains containing the
Aqa-y deletion expressed the qa-2 and qa-3 genes at the same level as
strains containingthe aro-9 mutation alone. However,
the strains containing the Aqa-y, aro-9 mutations did
not show induction of the qa-2 and qa-3 genes in the
presence of exogenous quinic acid which does occur
in a strain containing only the aro-9 mutation (Table
A
Wild- type
" D +
qa-x qa-2 qa-4
-P+"
pa-3
qa-y
qa-IS
qa-IF
A qa-Y
"D+
q a - x q a - 2 911-4
B
+"
1
qa-1S
qa-3
qa-IF
2).
Wild- type
+-D+
90-x pa-2 pa-4
+"q aq- 3aq- ya - I S
Aqa-IS
"
w
4
-
733
1
q d F
-PC"---)
90-xqa-2qa-4qa-3qa-yqa-IF
FIGURE4.-The qa gene cluster after transformation with the
pAqa-y and pAqa-IS plasmids. A, Arrangement of the qa gene
cluster before and after transformation with the pdqa-y plasmid. B,
Arrangement of the 9a gene cluster after transformation with the
plasmid pAqu-IS.
the met-7 mutation to obtain homokaryotic progeny.
Genomic DNA from these progeny strains
wasisolated, digested with BamHI, and probed by Southern
blot with two "P-labeled BamHI fragmentsand a
BamHI-XbaI fragment from the qa gene cluster (Figure 1A). In the wild-type strain, 74A, these probes
would hybridize to bands at 7.7, 5.5 and 2.3 kb. If
the qa-y gene had been deleted in the transformant
strains, these probes would be expected to hybridize
to bands at 5.5 and2.3 kb; however, the 7.7-kb band
would have beenshifted to 5.2 kb. Several of the
transformants tested produced this pattern. An example is shown in Figure 3B. The arrangement of the
qa gene cluster before and after transformation with
pAqa-y is shown in Figure 4A.
If the qa-y gene encodes a permease, a strainwhich
has been deleted for qa-y should show very low levels
of induction of the various qa transcripts since quinic
acid would not be actively transported into the cell.
In the absence of the inducer, quinic acid, the qa-IF
protein would not be induced and therefore would
not be present at a sufficient level to activate normal
transcription of the qa genes. Consequently these
genes would be expressedonly at very low basal levels.
T h e amount of qa-2 and qa-3 gene expression was
determined in a strain containing the Aqa-y deletion
by assaying the levels of their gene products,catabolic
dehydroquinase and quinicdehydrogenase, respectively. Strains containing the Aqa-y deletion showed
only about 5-10% of the normal induction levels of
the qa-2 and qa-3 genes (Table
2). This level of expression is not sufficient for the strains to grow on quinic
acid as a sole carbon source. However,in the presence
of the internal inducer dehydroquinic acid (DHQ),
which accumulates in strains containing the aro-9 mu-
Effect ofdeletingthe
qa-IS gene: The plasmid
pAqa-IS was constructed as shown in Figure 1A. This
construct, which retainsfunctional qa-y andqa-IF
genes, was used to transform a strain containing the
qa-IF, aro-9 mutations to obtain strains in which the
qa-IS gene was deleted. Transformants were crossed
to a met-7 strain of N . crassa in order to obtain homokaryotic progeny.T o determine if the transformants
did, in fact, contain deletions of the qa-IS gene, genomic DNA was isolated from several of these progeny strains, digested with BamHI, andprobed by
Southern blot with the same "P-labeled fragments
from the qa gene cluster as used earlier (Figure 1A).
These probes would be expected to hybridize with
three genomic BamHI fragments in wild-type DNA at
7.7, 5.5 and 2.3 kb. In strains where deletion of the
qa-IS gene had occurred these fragments
would be
replaced by two fragmento 6.1 and 5.5 kb insize.
Several of the transformants hadthis pattern diagnostic of a deletion of the qa-IS gene. An example is
shown in Figure 3C. Representative strains containing
the Aqa-IS deletion were selected for further studies.
T h e arrangement of the qa gene cluster after deletion
of the qa-IS geneis shown in Figure 4B.
In theGAL system inyeast, strains which are deleted
for the GAL80 repressor gene show constitutive levels
of GAL gene expression. Expression of the various
GAL genes in these strains, however, is still strongly
repressed in the presence of glucose (FLICKand JOHNSTON 1990). In order to determine
if the Aqa-IS strain
behaved like the AgaZ80 strain in yeast, total RNA
was isolated from the Aqa-IS strain as well as wildtype 74A after growth for 6 hr with 2% glucose as a
sole carbon source, on 0.1 % quinic acid as a carbon
source, and in the presence of 0.1 % quinic acid and
2% glucose. T h e RNA was then fractionated by electorphoresis, transferred to a
nylon membrane, and
probed with the 32P-labeled BgZII-BamHI fragment
(Figure 1A) which contains the qa-2 coding region or
the "P-labeled H3 (histone) gene.As shown in Figure
5A, the qa-2 message is expressed in the Aqa-IS strain
under all three conditionstested, glucose (lane l),
glucose and quinic acid (lane 2),and quinic acid alone
(lane 3). In wild-type 74A, the qa-2 message is present
only when the strain is grown with quinic acid as the
sole carbon source (lane 6) not on glucose (lane 4)or
on quinic acid and glucose (lane 5 ) . This indicates that
the qa-2 gene is being expressed constitutively in the
M. E. Case, R. F. Geever and D. K. Asch
734
A
1
2
3
4
5
A
6
B
C
- 2746
B
1
2
3
4
5
6
FIGURE5.-A, Northern blot of total RNA from wild-type 74a
and a strain containing Aqa-IS deletion probed with the BglIIHamHl fragment containing the qa-2 coding region. Lane 1, total
RNA from the Aqa-IS strain grown for 6 hr with 2% glucose as a
carbon source; lane 2, total RNA extracted from the Aqa-IS strain
grown for 6 hr with 0.1 % quinic acid and 2% glucose as carbon
sources; lane 3, total RNA from the Aqa-IS strain grown for 6 hr
with 0.1% quinic acid as a sole carbon source; lane 4, total RNA
from wild-type 74A grown for 6 hr with 2% glucose as a carbon
source; lane 5 , total RNA from wild-type 74A grown for 6 hr with
0.1% quinic acid and 2% glucose as carbon sources; lane 6, total
RNA from wild-type 74A grown for 6 hr with 0.1 % quinic acid as
a sole carbon source. B.The Same lanes probed with the H3 histone
gene of N . crassa.
Aqa-IS strain and is not subject to the strong
catabolite
repression as is wild-type 74A. This isin contrast to
the situation with the gal genes in yeast where the gal
messages are repressed in the presence of a preferred
carbonsource
such as glucose in Aga180 strains.
Northern results similar to those in Figure 5A were
observed when the levels of the qa-x and qa-IF messages were determined in the Aqa-IS strain (data not
shown). Probing with the H3 geneindicated some
variation in the mRNA concentrations (Figure 5B).
T h e variation in the levels of the histone message
however can not account forthe repression of the qa2 message in the presence of glucose in wild-type 74A.
The qa-2 gene of N . crassa utilizes multiple transcriptional start sites which fall into two main groups
(TYLER
et al. 1984). One group is located about 120140 bp fromthe ATG start codon of qa-2 and is
activator independent. These start sites account for
the basal level transcripts which are produced under
of start sites
noninducing conditions. The other group
is activator dependent andaccounts forthe transcripts
produced under inducing conditions. These sites are
located 70- 100 bp upstream of the qa-2 start codon.
et al. 1984) proposes thatthe
One model (TYLER
binding of the p - I F activator protein upstreamof the
qa-2 gene allows transcription to begin at the activator-dependent start sites. T o determine if the activator-dependent transcriptional start sites for the qa-2
gene were being used in the presence of glucose, the
major transcriptional startsites for theqa-2 gene were
determined in the Aqa-IS strain grown with glucose
as a carbon source (Figure 6, lane B) and both quinic
acid and glucose present (Figure 6, lane C) and compared with the qa-2 start site in wild-type 74A grown
with quinic acid as a sole carbon source (Figure 6,
- 2800
FIGURE6.”Sl analysis. Total RNA from wild-type 74A and a
strain containing the Aqa-IS deletion was probed with a ’*P-labeled
BglIISmaI fragment (Figure 1A) to map the transcriptional start
site of the qa-2 gene. Lane A, total RNA from wild-type 74Agrown
for 6 hr with 0.1 % quinic acid as a sole carbon source; lane B, total
RNA from the Aqa-IS strain grown for 6 hr with 2%glucose as the
carbon source; lane C, total R N A from the Aqa-IS strain grown
with 2% glucose and 0.1 % quinic acid as carbon sources. Numbers
tothe right indicate positions in the qa gene cluster sequence
published earlier (GEEVERet al. 1989).
lane A). In all cases tested, the major inducible start
site for the qa-2 gene was being utilized.
DISCUSSION
While gene replacement and gene disruption have
long been useful tools in the study of gene function
in S. cereviseae, the fact that themajority of transformants in N . crassa arise via nonhomolgous, ectopic
integration has made such studies in N . crassa and
other filamentous fungi much more difficult. In this
study, three genes of the qa gene cluster(qa-x, qa-IS
and qa-y) have been deleted or disrupted by using the
“targeted” transformationtechnique.Southern
hybridization of the nonhomokaryotic transformants indicated thatonly gene replacement had occurred
since
there was no evidenceforectopicintegration.
We
have shown that this technique can be used to replace
or alter specific sequences on the N . crassa genome at
a distance of up to 8.0 kb from theselectable marker.
Our initial studies resultedin the disruption of the qax gene using the qa-2 gene as a selectable marker.
Recent data have shown that the qa-x gene has 31%
homology with GALI2, agenecontrollingcarbonregulated dephosphorylation of the GAL4 activator
protein
HOPPERand E. MATTES, personal communication). However, carbon mediatedrepression of
the qa-2 gene seemed unaffected by the disruption in
qa-x. Therefore, the role of the qa-x gene in quinic
acid catabolism remains unclear.
Subsequently, the qa-y and qa-IS genes were deleted
using the qa-IF geneasa selectable marker.This
involved the introductionof deletions endingas much
as 8.0 kb from the qa-lF gene. The qa-y gene had
been shown to have significant homology with the
u.
735
Gene Replacement in N . crassa
qutD gene of A. nidulans which is known to be a quinic
acid permease (WHITTINGTON
et al. 1987). GEEVERet
al. (1989) searchedthe GenBank data base and found
that the qa-y gene had a probable homology with a
glucose transporter of human hepatoma cells. Deletion of the qa-y gene in N . crassa produced a phenotype consistent with the interpretation that the qa-y
geneproduct is involved in quinic acid transport.
Strains containing the Aqa-y cannot grow on quinic
acid as a sole carbon source and have very low inducible levels of qa-2 and qa-3 enzyme activities. In strains
containing the aro-9 mutation, the qa genes are partially induced by the action of an internal inducer
(DHQ). In strains containing both the Aqa-y and the
aro-9 mutations equivalent partial induction occurs,
indicating that the qa genes can be inducedif internal
inducer is present in the cell. The evidence that Aqay strains have only low levels of inducible qa enzyme
activities is thus consistent with the view that the qa-y
gene encodes a quinic acid permease whose absence
in the Aqa-y strain precludes the normal uptake of
quinic acid. The low level of induction of the qa-2 and
qa-3 genes observed in the Aqa-y strain in the absence
of internal inducer is presumably due to the entry of
a low level ofquinic acid into thecell by other avenues.
The qa-IS gene was deleted completely by transforming a qa-IF mutant strain with a plasmid construct containing the qa-IF gene as a selectable marker
as well as intact qa-3 and qa-y genes but totally lacking
the qa-1S gene andits flanking regions. The resulting
transformants had lost the 3.8 kb of sequence from
linkage group VI1 which normally contains the qa-IS
gene. However, there was no rearrangement of the
surrounding restriction
fragments.
Transformants
containing the Aqa-IS deletion have constitutive levels
of expression of the qa-2, qa-IF and qa-x genes as
determined by mRNA analysis, which is consistent
with earlier data indicating that theqa-IS gene product is a repressor molecule (GEEVERet al. 1989).
The GAL system of S. cerevisiae and the qa system
of N . crassa share many similarities in gene regulation.
Both systems are under the control of an activator
protein which activates the transcription of the various
genes in the presence of substrate and a repressor
protein which inhibits activator function in the absence of inducer. Gene expression of both systems is
undercarboncataboliterepression
in thatgene
expression is severely repressed in the presence of a
preferred carbon sourceeven if the induceris present.
It has been shown that carbon repressionof the GALl
gene is mediated by specific sequences upstream of
GAL1 as well and by action on the GAL4 binding site
(FLICKand JOHNSTON 1990). In yeast strains containing a deletion of the GAL80 repressor gene, GALl
remains strongly repressedin the presence of glucose.
However, in N . crassa strains containing the Aqa-IS
deletion the mRNA levels of the qa-2, qa-x and qa-IF
transcripts are high in the presence of glucose. This
suggests that carbon repression of at least some of the
genes of the qa gene cluster may be operating by a
different mechanism(s), not involving either specific
sequences upstream of the qa genes or inhibition of
activator binding as in seen with GALl.
These results may indicate that the qa-IS repressor
protein plays a role in catabolite repression of the qa
gene cluster. However our present data does not rule
out other possible explanations such as control at the
level of quinic acid uptake. T o determine if the qa-IS
protein is involved in carbon repression outside the
qa gene cluster we have assayed the mRNA levels of
the glucose repressible gene (grg) (MCNALLYand
FREE 1988) in the Aqa-IS strain grown under conditions in which this gene would be carbon repressed.
Under these conditions grg transcription is repressed
normally in the Aqa-IS strain (M. E. CASEand D. K.
ASCH,unpublished). Therefore we feel that if the qaI S gene product plays any role in carbon repression
that it may be a target for a general carbon repressor
molecule and acts only to repress the qa gene cluster
in the presence of a preferred carbon source.
We gratefully acknowledge theexpert technical assistance of
JULIEBLACK,PAULETTEGEEVER,NANCY
NORTONandPHYLLIS
SMITH.Special thanks to NORMAN
GILFSfor critical reading of the
manuscript. This work was supported by U.S. Public Health Service
grant GM 28777 from the National Institutes of Health to M.E.C.
LITERATURE CITED
BAUM,J. A., R. GEEVER
and N. H. GILFS,I987 Expression of the
qa-IF activator protein: identification o f upstream binding sites
in the qa gene cluster and localization of the DNA binding
domain. Mol. Cell. Biol. 7: 1256-1266.
BERKA. J., and P. A. SHARP, 1977 Sizing and mapping of early
adenovirus mRNAs by gel electrophoresis of SI digested hybrids. Cell 12: 721-732.
CASE,M. E., and R. F. GEEVER, 1987 Gene replacement experiments within the qa gene cluster of Neurospora crassa. Fungal
Genet. Newslet. 3 4 17.
CASE, M. E., J. HAUTALA
A. H.N.
and
GILES,
1977 Characterization of qa-2 mutants of Neurospora c r a m
utilizing genetic, enzymatic and immunological techniques. J.
Bacteriol. 129: 166-172.
CASE,M. E., M. SCHWEIZER,
S. R KUSHNER and N.H. GILES,
1979 Efficient transformation of Neurospora crassa by utilizing hybrid plasmid DNA. Proc. Natl. Acad. Sci. USA 76: 52595263.
FEINBERC
A., and B. VOGELSTEIN, 1984 A technique for radiolabling DNA restriction fragments to high specific activity. Anal.
Biochem. 137: 266-267.
FLICK,J. S., and M . JOHNSTON,1990Two systems of glucose
repression of the GALl promotor in Saccharomyces cerevisiae.
Mol. Cell. Biol. 10: 4757-4769.
FREDERICK,
G . D., D. K. ASCHand J. A. KINSEY, 1989Useof
transformation to make targeted sequence alterations at the
am (GDH) locus of Neurospora. Mol. Gen. Genet. 217: 294300.
FREDERICK,
G. D., and J. A. KINSEY, 1990 Distant upstream
regulatory sequences control the level of expression of the am
736
M. E. Case, R. F. Geever and D. K. Asch
(GDH) locus ofNeurospora crassa. Curr. Genet. 1 8 53-58.
GEEVER,R. F., J. A. BAUM,M. E. CASE and N. H. GILES,
1987 Regulation of the qa gene cluster of Neurospora crassa.
Antonie van Leeuwenhoek J. Microbiol. 53: 343-348.
GEEVER,
R. F., L. HUIET,J. A. BAUM,B. M. TYLER,V. B. PATEL,
B. J. RUTLEGE,M. E. CASEand N. H. GILES, 1989 DNA
sequence organization and regulation of the qa (quinic acid)
gene cluster of Neurospora crassa. J. Mol. Biol. 207: 15-37.
GILES,N. H., M. E. CASE,J. BAUM,R. GEEVER,
L. HUIET,V. PATEL
and B. M. TYLER, 1985 Gene organization and regulation of
the qa (quinic acid) cluster of Neurospora crassa. Microbiol.
Rev. 4 9 338-358.
GILES,N. H., R. F. GEEVER,
D. K. ASCH,J. AVALOS
and M. E. CASE,
1991 Organization and regulation of the qa (quinic acid)
genes in Neurospora crassa and other fungi. J. Hered. 82: 1-7.
HUIET, R. L., 1984 Molecular analysisof the Neurospora pa-1
regulatory region indicates that two interacting genes control
qa gene expression. Proc. Natl. Acad.Sci.USA
81: 11741178.
ISH-HOROWICZ,
D., and J. F. BURKE,1981 Rapid and efficient
cosmid cloning. Nucleic Acids Res. 9 2989-2998.
JOHNSTON,
M., 1987 A model fungal regulatory mechanism: the
GAL genes of Saccharomyces cerevisiae. Microbiol. Rev.51: 458476.
KINSEY,
J. A., and J. RAMBOSEK,
1984 Transformation of Neurospora crassa with the cloned am (glutamate dehydrogenase)
gene. Mol. Cell. Biol. 4 117-122.
LINDGREN,
K. M., A. LICHENS-PARK,
J. L. LOROSand J. C. DUNLAP,
1990 A quick RNA mini-prep for Neurospora mycelial cultures. Neusorpora Newslet. 37: 21.
MAXAM,
A.M., and W. GILBERT,1980 Sequencing end labeled
DNA with base-specificchemical cleavages. Methods Enzymol.
6 5 499-560.
MCNALLY,
M. T., andS. J. FREE,1988 Isolation and characterization of a Neurospora glucose-repressible gene. Curr Genet. 14:
545-55 1.
PATEL,V. B., M. SCHWEIZER,
C. C. DYKSTRA,
S. R. KUSHNERand
N. H. GILES,1981 Genetic organization and transcriptional
regulation of the qa gene cluster of Neurospora crassa. Proc.
Natl. Acad. Sci. USA 78: 5783-5787.
SCHERER,
J., and R. W. DAVIS,1979 Replacement of chromosomal
segments with altered DNA sequences constructed in vitro.
Proc. Natl. Acad. Sci. USA 7 6 4951-4955.
TYLER,B. M., R. F. GEEVER,M. E.CASE and N. H. GILES,
1984 Cis-acting and trans-acting regulatory mutations define
two types of promoters controlled by the qa-IF gene of Neurospora. Cell 3 6 493-502.
WESTERGAARD, M.,
and H.K. MITCHELL, 1947 Neurospora. V. A
synthetic medium favoring sexual reproduction. Am. J. Bot.
3 4 573-577.
WHITTINCTON
H. A,, A. J. FRANCISO
DE SILVA,S. GRANT,H. LAMB
and A.R. HAWKINS,1987 Identification and isolation of a
putative permease gene in the quinic acid utilization (QUT)
gene cluster ofAspergillus nidulans. Curr. Genet. 12: 135-139.
VOLLMER,
S. J., and C. YANOFSKY,
1986 Efficient cloning of genes
of NeurosPora crassa Proc. Natl. Acad.Sci. USA 83: 48694873.
YELTON, M.
M.,
J. E. HAMER and W. E. TIMBERLAKE,
1984Transformation
of Aspergillus nidulans using a trpC
plasmid. Proc. Natl. Acad. Sci. USA 81: 1470-1474.
Communicating editor: R. H. DAVIS