Dual influence of the yeast Cat1p (Snf1p) protein kinase on carbon source-dependent transcriptional activation of gluconeogenic genes by the
regulatory gene
CAT8
Dual influence of the yeast Cat1p (Snf1p) protein kinase on carbon source-dependent transcriptional activation of gluconeogenic genes by the regulatory gene CAT8
Antje
Rahner
,
Anja
Schöler
,
Erika
Martens
,
Boris
Gollwitzer
and
Hans-Joachim
Schüller*
Institut für Mikrobiologie, Biochemie und Genetik, Lehrstuhl Biochemie, Universität Erlangen/Nürnberg, Staudtstraße 5, D-91058
Erlangen
,
Germany
Received March 1, 1996;
Revised and Accepted April 26, 1996
ABSTRACT
The CSRE (carbon source-responsive element) is a sequence motif responsible for the
transcriptional activation of gluconeogenic structural genes in
Saccharomyces cerevisiae
. We have isolated a regulatory gene,
DIL1
(derepression of isocitrate lyase, =
CAT8
), which is specifically required for derepression of CSRE-dependent genes. Expression of
CAT8
is carbon source regulated and requires a functional Cat1p (Snf1p) protein
kinase. The derepression defect of
CAT8
in a
cat1
mutant could be suppressed by a mutant Mig1p repressor protein. Derepression of
CAT8
also requires a functional
HAP2
gene, suggesting a regulatory connection between respiratory and gluconeogenic
genes. Carbon source-dependent protein-CSRE complexes detected in a gel retardation analysis with wild-type extracts were absent in
cat8
mutant extracts. However, similar experiments with an epitope-tagged
CAT8
gene product in the presence of tag-specific antibodies gave evidence against a direct binding of Cat8p to the
CSRE. A constitutively expressed
GAL4
-
CAT8
fusion gene revealed a carbon source-dependent transcriptional activation of a UAS
GAL
-containing reporter gene. Activation mediated by Cat8p was no longer
detectable in a
cat1
mutant. Thus, biosynthetic control of
CAT8
as well as transcriptional activation by Cat8p requires a functional Cat1p
protein kinase. A model proposing
CAT8
as a specific activator of a transcription factor(s) binding to the CSRE is
discussed.
INTRODUCTION
The utilization of non-fermentable carbon sources such as lactate, ethanol or acetate by the
yeast
Saccharomyces cerevisiae
requires the coordinate biosynthesis of gluconeogenic enzymes. However, sugar
phosphates required for protein glycosylation and generation of the cell wall
can be more easily produced in the presence of fermentable carbon sources.
Thus, genes encoding gluconeogenic enzymes are strictly regulated at the
transcriptional level by the complex glucose repression network (reviewed in
1
-
4
). This regulatory system also affects genes of alternative sugar metabolism (
SUC
,
MAL
or
GAL
genes), respiration and peroxisomal [beta]-oxidation. The
CAT1
gene (=
SNF1
,
CCR1
) encodes a serine/threonine-specific protein kinase which is essential for the derepression of most
glucose-repressible genes (
5
-
7
). The
CAT3
(=
SNF4
) gene product is physically associated with Cat1p and may function as a
stimulatory subunit of the protein kinase (
8
). Since
cat1
and
cat3
mutants fail to derepress glucose-repressible enzymes, pleiotropic growth defects on various carbon sources
(raffinose, maltose, galactose or ethanol) can be observed. Some of these
defects could be suppressed by the
mig1
(=
cat4
,
ssn1
) mutation, leading to a glucose-insensitive expression of
SUC
,
MAL
and
GAL
genes (
9
-
12
).
MIG1
encodes a zinc finger protein that binds to a GC-rich motif upstream of several glucose-repressible genes (
13
). Under conditions of glucose derepression, Cat1p may lead to a deactivation of
the Mig1p repressor by a currently unknown mechanism.
Isocitrate lyase (encoded by the
ICL1
gene), as a key enzyme of the glyoxylate cycle (considered as a subpathway of
gluconeogenesis from C
2
substrates), is regulated by transcriptional repression/derepression (
14
,
15
), phosphorylation (
16
) and proteolytic degradation (
17
). We have previously identified the CSRE (carbon source-responsive element), an upstream activation site (UAS) responsible for
transcriptional derepression of the
ICL1
gene (
18
). Similar sequence motifs were found upstream of the fructose-1,6-bisphosphatase gene
FBP1
(
18
,
19
), the phosphoenolpyruvate carboxykinase gene
PCK1
(
20
) and the acetyl-CoA synthetase gene
ACS1
(
21
). Thus, the CSRE can be considered as a pathway-specific
cis
-element of genes involved in non-fermentative metabolism. Gene activation by a CSRE requires
functional
CAT1
and
CAT3
genes (
18
). The carbon source-dependent binding of a protein factor, designated Ang1 (activator of non-fermentative growth) to the CSRE could no longer be observed in
cat1
and
cat3
mutants. We thus wished to identify
trans
-acting factors responsible for the signal transduction from pleiotropic regulators to the CSRE. In this paper we describe the
characterization of the
DIL1
(=
CAT8
) gene and identify its gene product as a carbon source-regulated,
CAT1
-dependent transcriptional activator.
MATERIALS AND METHODS
Yeast strains and media
Strains of
S.cerevisiae
used in this work are listed in Table
1
. Synthetic complete medium have been described elsewhere (
7
). Repressed cells were grown with 2% glucose until the mid log phase. For
derepressing conditions, cells were directly grown with 0.2% glucose.
Alternatively, repressed cells were harvested in the early log phase and
subsequently transferred to derepression medium for 8 h.
MAT
a
ura3 leu2 trp1 his3 ade1 gal4
[Delta]
gal80
[Delta]
GAL1-lacZ
YJOZ.cat1
MAT
a
ura3 leu2 trp1 his3 ade1 gal4
[Delta]
gal80
[Delta]
GAL1-lacZ cat1
::
HIS3
YJOZ.cat1mig1
MAT
a
ura3 leu2 trp1 his3 ade1 gal4
[Delta]
gal80
[Delta]
GAL1-lacZ cat1
::
HIS3 mig1-
[Delta]
2
::
LEU2
Isolation of
dil
mutants
Cells of the wild-type strain JS92.32-11 containing an integrated
ICL1
-
lacZ
reporter gene were mutagenized with 1% ethylmethanesulfonate for 60 min (rate
of survival ~90%). Mutants unable to grow on synthetic complete medium with 3% ethanol
(SCE) were further assayed on X-gal-containing plates, thereby screening for
ICL1
activation defects. In order to distinguish pleiotropic mutations affecting
utilization of various carbon sources (such as
cat1
/
snf1
or
snf2
;
7
,
22
) from those specific for CSRE-mediated gene expression, activation-deficient mutants were finally plated on medium containing
raffinose, maltose or galactose. Mutants exhibiting the expected phenotype (no
growth on SCE, reduced expression of the
ICL1
-
lacZ
reporter gene, normal growth with raffinose or galactose) were further
investigated.
Gene isolation and plasmid constructions
For the isolation of the
DIL1
gene by mutant complementation, strain JS94.9-9 was transformed with a plasmid library based on YEp24 (
23
). Plasmid YEp24-DIL1 was isolated from transformants showing restored growth with ethanol as the sole carbon source. Plasmid
YEp13-DIL1 was isolated similarly.
Procedures for recombinant DNA followed established protocols (
24
). Episomal and integrative reporter plasmids containing
ICL1
-
lacZ
fusions (pJS310 and pJS330, respectively) have been described (
14
). For the synthetic promoter construct pAS74 (CSRE-
ICL1
-
lacZ
), the
ICL1
upstream region (-527 to -164) was deleted and replaced by the oligonucleotide ICLSM1
containing the CSRE of
ICL1
(5'-tcgaggatcCCATTCATCCGctagca-3'; authentic promoter sequence shown in capital
letters). Reporter plasmids pJS151 (
FBP1
-
lacZ
;
25
), pJS154 (
ADH2
-
lacZ
) and pJS334 (
POT1
-
lacZ
) contained at least 1 kb of the respective upstream regions. The
CAT8
-
lacZ
fusion construct pJS427 was obtained by ligation of a 1.4 kb
Eco
RI-
Hin
dIII fragment containing 1.2 kb of the
CAT8
upstream region into the
lacZ
fusion plasmid YEp356 (
26
). For the construction of a [Delta]
cat8
null mutation, a 2.5 kb
Eco
RI-
Kpn
I fragment from YEp24-DIL1 was subcloned into pGEM7 (Promega). Subsequently, a 1.45 kb
Pst
I-
Bgl
II fragment representing the
CAT8
promoter and the N-terminus of its reading frame was replaced by the
HIS3
gene, giving the
cat8-
[Delta]
1
::
HIS3
construct pEM4 (cf. Fig.
1
). For the heterologous expression of the zinc cluster domain of Cat8p, a 0.85
kb fragment (encoding amino acids 1-281) was amplified by PCR and subsequently ligated into the GST fusion
plasmid pGEX-2TK (Pharmacia) to give pAR1. The
GAL4
-
CAT8
fusion constructs pAR2-pAR5 were obtained by cloning the fragments shown in Figure
1
into plasmid pY1 (
27
). The
CAT8
variant
CAT8
-
FLAG
was constructed by PCR-mediated addition of a sequence encoding the FLAG epitope (DYKDDDDK). The
resulting plasmid pEM6 (2[mu]m
URA3 CAT8
-
FLAG
) contains an extended
CAT8
gene under native promoter control, epitope-tagged at the C-terminus.
Gel retardation analysis
Labeling of promoter fragments and gel retardation experiments were carried out
as described previously (
18
). Protein extracts were prepared from repressed or derepressed transformants of
proteinase-deficient strains C13-ABY.S86 and ABYS.[Delta]dil1, respectively. The synthetic DNA fragment OAS12
containing the CSRE of the
ICL1
promoter has been described (
18
). The anti-FLAG antibody M2 was used for supershift studies.
Miscellaneous procedures
[beta]-Galactosidase assays were performed with crude extracts of yeast
transformants as described (
18
). Transformation of yeast strains followed established procedures (
28
). Isolation of plasmids and chromosomal DNA from yeast has been described (
29
). Verification of the
cat8
null mutation was done by Southern blot hybridization (
24
). Synthetic oligonucleotides were purchased from MWG Biotech (Ebersberg,
Germany). For the immunoblot analysis (
30
) of protein extracts from
CAT8
-
FLAG
transformants, the monoclonal anti-FLAG antibody M2 (Kodak/IBI) was used. The ECL chemiluminescence system
(Amersham) was used for immunodetection.
RESULTS
Isolation of
dil1
mutants
We have previously described the CSRE as an essential activating promoter element upstream of structural genes of gluconeogenesis (preliminary consensus sequence CCRTYCRTCCG;
18
; modified). At least expression of the isocitrate lyase gene
ICL1
depends almost completely on a functional CSRE. Thus, we considered an
ICL1
-
lacZ
fusion as a suitable reporter construct for the isolation of mutants
specifically affected in CSRE-dependent gene expression. Following the procedure described above (see
Materials and Methods), we isolated nine representatives of the
dil1
complementation group (derepression of isocitrate lyase). In these mutants,
ICL1
-
lacZ
derepression was reduced to 1-14% of the wild-type level (not shown). Expression of reporter genes dependent on
the
FBP1
promoter or a CSRE-containing synthetic minimal promoter was similarly affected. On the other
hand, enzymes of alternative sugar utilization, such as invertase, maltase or
galactokinase, were normally derepressed in
dil1
mutants (data not shown).
Isolation of the
DIL1
gene and its identity with
CAT8
The
DIL1
wild-type gene was isolated by functional complementation of the growth defect
of a
dil1
mutant on ethanol medium. Two distinct plasmids recovered from the
transformants contained inserts overlapping for 4.8 kb. Restriction mapping of
this DNA segment (Fig.
1
) revealed a fragment pattern similar to that of the recently isolated
CAT8
gene necessary for efficient expression of the fructose bisphosphatase gene
FBP1
(
19
). Indeed, partial DNA sequencing of the
DIL1
gene proved its identity with
CAT8
. Thus, for reasons of uniformity, we shall use the designation
CAT8
instead of
DIL1
in the further text. The DNA sequence of
CAT8
predicts a binuclear zinc cluster motif at the N-terminus of the respective 1433 amino acid protein (
19
; residues 69-99), reminiscent of the Gal4p protein family. To date, more than 50
members of this family have been described in fungal systems (not shown; a
recent compilation of C
6
zinc cluster proteins is available upon request). Thus, Cat8p may be considered
as a DNA binding transcription factor directly or indirectly involved in
activation of CSRE-dependent structural genes (see below).
A
cat8
deletion mutation was constructed (cf. Fig.
1
) and characterized for its influence on structural genes involved in C
2
metabolism. As expected,
cat8
null mutants failed to grow on non-fermentable carbon sources, such as lactate, ethanol or acetate, while
maltose, raffinose or galactose were utilized normally. In the
cat8
null mutant, expression of
ICL1
-
lacZ
or
FBP1
-
lacZ
fusion constructs under derepressing conditions was reduced to 1-2% of the wild-type level (Table
2
). Expression of a reporter gene driven by a CSRE-containing minimal promotor was completely abolished in the
cat8
mutant. A partial influence of the
cat8
mutation was also found for the glucose-repressible alcohol dehydrogenase gene
ADH2
. This effect is independent of the
ADH2
positive regulator Adr1p, since
ADH2
expression is further decreased in a
cat8 adr1
double mutant (data not shown). Obviously, the positive regulators
CAT8
and
ADR1
function in distinct activating pathways. A complete loss of derepression has
been also reported for the phosphoenolpyruvate carboxykinase gene
PCK1
(
19
). On the other hand, a functional
CAT8
gene is not required for induction of the peroxisomal thiolase gene
POT1
by oleic acid (Table
2
). In summary,
CAT8
is necessary for the efficient transcriptional derepression of all currently
known CSRE-controlled structural genes, at least in part.
.
Expression of carbon source-regulated structural genes in a
cat8
null mutant
Reporter gene
a
Specific [beta]-galactosidase activity (nmol/mg/min)
Wild-type
[Delta]
cat8
::
HIS3
Repressed
Derepressed
Repressed
Derepressed
ICL1
-
lacZ
5
1600
2
20
CSRE-
ICL1
-
lacZ
2
1100
1
2
FBP1
-
lacZ
10
4600
9
160
ADH2
-
lacZ
200
12900
160
2300
POT1
-
lacZ
6
40 (210)
b
8
120 (390)
b
a
Episomal plasmids were transformed into strains JS91.15-23 (wild-type) and EMY1 ([Delta]
cat8
::
HIS3
), respectively. Transformants were grown in synthetic complete medium
containing 2% glucose (repressing conditions) or 0.2% glucose (derepressing
conditions).
b
Specific enzyme activities in parentheses refer to conditions of oleate
induction (0.2% glucose + 0.2% oleic acid).
.
Influence of increased
CAT8
gene dosage on
ICL1
-
lacZ
expression in wild-type and pleiotropic regulatory mutants
Genotype
a
Specific [beta]-galactosidase activity in transformants
YEp24
YEp24-DIL1
Repressed
Derepressed
Repressed
Derepressed
CAT1 MIG1
1.5
320
1.8
305
cat1 MIG1
1.1
3.7
1.2
2.8
CAT1 mig1
1.4
350
2.4
360
cat1 mig1
1.3
11
1.4
85
a
Recipient strains JS92.24-1 (wild-type), JS92.37-1 (
cat1
), JS92.25-1 (
mig1
) and JS92.36-1 (
cat1 mig1
) contained an integrated
ICL1
-
lacZ
reporter gene. Variation of
CAT8
gene dosage was achieved by transformation of strains with episomal plasmids
YEp24 (
URA3
) or YEp24-DIL1 (
URA3 CAT8
). Repressing and derepressing conditions were as defined in the legend to Table
2.
Influence of
CAT8
gene dosage on
ICL1
expression
In order to investigate the position of
CAT8
in the regulatory hierarchy of CSRE-dependent gene activation, a yeast strain containing an integrated
ICL1
-
lacZ
reporter gene was transformed with the episomal plasmid YEp24-DIL1 containing the entire
CAT8
gene. In addition, isogenic
cat1
,
mig1
and
cat1 mig1
derivatives of this strain were treated similarly. An increased
CAT8
gene dosage did not influence
ICL1
-
lacZ
expression in the wild-type, the
cat1
mutant or the
mig1
mutant, neither under repressing nor under derepressing conditions (Table
3
). However, a substantial increase in reporter gene expression was found in
CAT8
multi-copy transformants of the
cat1 mig1
strain under derepressing conditions (28% of the wild-type level), but not under repressing conditions. This
CAT8
dosage-dependent suppression of the derepression defect caused by the
cat1
mutation allowed a slow growth of the transformants even on synthetic complete
medium containing ethanol as the sole carbon source. For this suppression to
occur, the
MIG1
gene must be defective. Thus, the Mig1p repressor must fulfil some negative
function even under conditions of non-fermentative growth. In contrast to previous
CAT8
gene dosage studies (
19
), no suppression of growth defects of the
cat1
single mutant by episomal
CAT8
could be observed. This discrepancy may be due to the rich medium used by these
authors for growth assays. In conclusion, our data provide evidence for a
function of
CAT8
downstream of the pleiotropic Cat1p protein kinase system. Since we could not
detect
CAT8
-dependent gene dosage effects in a wild-type strain, positive regulators in addition to Cat8p presumably
affect CSRE-dependent structural genes.
Expression of
CAT8
in wild-type and regulatory mutants
Several transcription factors involved in various regulatory networks have been
shown to be regulated at the biosynthetic level (e.g. Gal4p and Gcn4p;
11
,
31
). We thus investigated the expression of a
CAT8
-
lacZ
reporter gene in wild-type and regulatory mutant strains under repressing and derepressing
conditions. In agreement with recent data (
19
), we found
CAT8
expression to increase ~45-fold in derepression medium (Table
4
). Repression of
CAT8
was significantly alleviated in the
mig1
mutant, while
CAT8
derepression required a functional
CAT1
gene. However,
mig1
turned out to be epistatic to
cat1
, since the
CAT8
derepression defect observed in the
cat1
mutant could be suppressed by a
mig1
mutation. This result can be easily explained by the Mig1p repressor binding
site identified upstream of
CAT8
(
19
). Thus, the Cat1p protein kinase system appears to be required for deactivation
of the Mig1p repressor.
.
Carbon source-regulated expression of
CAT8
in wild-type and regulatory mutants
Strain
a
Genotype
Specific [beta]-galactosidase activity
Repressed
Derepressed
JS89.27-3
Wild-type
20
920
JS87.11-22D
cat1
6
20
JS92.23-1
mig1
320
950
JS92.39-1
cat1 mig1
280
490
WAY.6-2B
hxk2
60
750
EMY1
cat8
60
3200
LGW1
hap2
15
180
a
Strains were transformed with the episomal plasmid pJS427 containing the
CAT8
-
lacZ
reporter gene. Repressing and derepressing growth conditions have been defined
in the legend to Table 2.
The
HXK2
structural gene encoding hexokinase PII has been shown as essential for glucose
repression of
SUC
and
MAL
genes (
32
). However, a
hxk2
null mutation did not cause a significantly increased
CAT8
expression under repressing conditions (Table
4
). Interestingly, derepression of
CAT8
was severely affected in a
hap2
mutant, indicating a regulatory connection between respiration and
gluconeogenesis. The sequence motif TGATTGGT upstream of
CAT8
(-199 to -192) agrees completely with the consensus binding site of the
heteromeric Hap2/3/4/5p complex (
33
), which may mediate a respiration-coupled activation of
CAT8
. The
CAT8
gene appears to be negatively autoregulated, since expression of the
CAT8
-
lacZ
reporter gene in a
cat8
mutant increases >3-fold under either growth conditions when compared with the isogenic wild-type. Presently, the
cis
-acting element responsible for this control and its possible regulatory
significance are unknown.
Importance of
CAT8
for protein-CSRE interactions
Cat8p functions as a
CAT1
-dependent transcriptional activator
A large number of transcription factors can be functionally separated into DNA
binding and transcriptional activation domains. We thus fused various parts of
CAT8
to the DNA binding domain of
GAL4
(controlled by the
ADH1
promoter) and assayed the expression of a
GAL4
-dependent reporter gene. As shown in Figure
1
, none of the tested
CAT8
subfragments could mediate transcriptional activation. On the other hand, a
strong activation was observed with the entire
CAT8
gene (plasmid pAR4). Interestingly, this activation was completely carbon
source dependent. In contrast, a
GAL4
DBD
-
INO2
fusion (
36
), used as a control, showed a constant activation under repressing and
derepressing conditions (Table
5
). Gene activation mediated by Cat8p was no longer observed in a
cat1
null mutant. This defect was not suppressed in a
cat1 mig1
double mutant. Interestingly,
INO2
-dependent activation was also significantly reduced, although not
completely abolished, in the
cat1
mutant. This result may explain the inositol auxotrophy caused by a
cat1
mutation in some strain backgrounds (
37
; Table
5
).
.
Transcriptional activation of a
GAL1
-
lacZ
reporter gene by a
GAL4
DBD
-
CAT8
fusion construct in the wild-type and pleiotropic regulatory mutants
Genotype
a
Specific [beta]-galactosidase activity (nmol/min/mg)
GAL4
DBD
GAL4
DBD
-
INO2
GAL4
DBD
-
CAT8
Inositol
Repressed
Derepressed
Repressed
Derepressed
Repressed
Derepressed
prototrophy
Wild-type
1
1
98
75
1
70
+
cat1
1
1
22
16
1
2
-
cat1 mig1
1
1
82
78
1
2
+
a
Recipient strains YJOZ (wild-type), YJOZ.cat1 (
cat1
) and YJOZ.cat1mig1 (
cat1 mig1
) contained an integrated
GAL1
-
lacZ
reporter gene. Centromere plasmids pY1 (
GAL4
DBD
), pSS23 (
GAL4
DBD
-
INO2
) and pAR4 (
GAL4
DBD
-
CAT8
), respectively, were used for transformation. Transformants were grown in
selective medium with 2% glucose and harvested in the mid log phase (repressed)
or transferred to a medium with 0.2% glucose for 8 h (derepressed). Inositol
prototrophy was tested on a synthetic complete medium lacking inositol (SCD-Ino).
These data prove (direct or indirect) transcriptional activation mediated by the
positive regulatory protein Cat8p. The loss of this function in the
cat1
mutant suggests a possible signal transduction pathway from the pleiotropically
acting Cat1p protein kinase system to the pathway-specific regulator Cat8p affecting CSRE-dependent structural genes.
DISCUSSION
The complex pattern of carbon source-dependent transcription in the yeast
S.cerevisiae
involves several pleiotropic as well as pathway-specific regulators. Previously, we identified the CSRE as a UAS element
upstream of the isocitrate lyase gene
ICL1
and other co-regulated genes of non-fermentative metabolism (
18
,
21
). In this work, we describe the
DIL1
gene as a transcriptional regulator specifically affecting CSRE-containing structural genes.
DIL1
turned out to be identical to the recently isolated
CAT8
gene, which is required for derepression of the gluconeogenic genes
FBP1
and
PCK1
(
19
). This identity was not surprising, since the promoters of
ICL1
,
FBP1
and
PCK1
all contain at least a single copy of a functional CSRE (
18
-
20
). Cat8p also functions as a positive regulator of the glucose-repressible
ADH2
gene. Although we did not find an
ADH2
promoter sequence element convincingly similar to the CSRE, the previously
described UAS2 region immediately upstream of the Adr1p binding site (
38
) may be considered as a putative target of Cat8p. On the other hand,
CAT8
is clearly not involved in the oleate induction pathway (mediated by the oleate-responsive element ORE;
39
,
40
), responsible for the efficient proliferation of peroxisomal structures.
Previous gel retardation experiments led to the identification of two CSRE
binding protein factors (Ang1 and Ang2) exclusively detectable in extracts from
derepressed yeast cells (
18
). The presence of a binuclear C
6
zinc cluster motif in the N-terminus of Cat8p strongly argues for its binding to DNA. Thus, Cat8p may
be considered as a CSRE binding factor possibly related to Ang1 and/or Ang2.
This idea would be consistent with the binding of several members of the Gal4p
family to sites containing a CCG sequence (
34
,
35
), which is also present within the CSRE. The carbon source-dependent binding pattern of Ang factors to the CSRE would be easily
explained by the regulation of
CAT8
at the biosynthetic level shown here and elsewhere (
19
). Although Ang-CSRE interactions could no longer be observed in a
cat8
null mutant, we were unable to demonstrate CSRE binding of a GST-Cat8p fusion protein produced in
E.coli
. Thus, Cat8p is necessary but not sufficient for protein binding to the CSRE.
Even extracts from derepressed yeast cells containing a full-length, epitope-tagged protein variant gave no evidence for Cat8p being at least a
part of an Ang factor. Although a failure of the antibody used in the
supershift experiment to detect the tag under native conditions cannot
completely be ruled out, we favor a model with
CAT8
being an indirect activator of CSRE-containing genes. Two additional findings support this view: first,
overexpression of
CAT8
by introduction of a multi-copy plasmid into a wild-type strain does not cause any gene dosage effects, neither under
repressing nor under derepressing conditions (Table
3
). Dosage effects on structural genes have been frequently observed when
directly activating transcription factors are overexpressed (e.g. in the case
of
GAL4
;
41
). Second, both UAS elements of the carbon source-regulated
FBP1
promoter require a functional
CAT8
gene (
19
; this work), although distinct protein factors bind to UAS1 and UAS2 (
25
,
42
). Such a result would not be expected when assuming the sole and direct binding
of Cat8p to both UAS elements. In conclusion, these data can be easily
explained by considering Cat8p as a transcriptional regulator of activating
proteins (Ang factors) finally binding to the CSRE and other UAS elements of
gluconeogenic genes. Thus, it will be necessary in the future to identify the
genes encoding Ang factors. According to our hypothesis (cf. Fig.
4
), their control regions may contain sequence elements recognized by the
putative DNA binding domain of Cat8p.
ACKNOWLEDGEMENTS
This paper is dedicated to Prof. Dr E.Schweizer on the occasion of his 60th
birthday. The work described was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. We wish to thank C.Denis, L.Guarente, S.Hohmann, S.Johnston,
K.Melcher, H.Ronne, I.Sadowski and D.Wolf for providing strains or plasmids
used in this work. We also thank E.Schweizer for kind support and B.Hoffmann
for excellent technical assistance.
REFERENCES
1 Johnston,M. and Carlson,M. (1992) In Jones,E.W., Pringle,J.R. and Broach,J.R. (eds), The Molecular and Cellular Biology of the Yeast Saccharomyces. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, Vol. 2, pp. 193-281.
