Two Glycogen Synthase Isoforms in Saccharomyces cereuisiae Are Coded by Distinct Genes That Are Differentially Controlled*

In previous work, we identified a Saccharomyces cerevisiae glycogen synthase gene, GSYl, which codes for an 85-kDa polypeptide present in purified yeast glycogen synthase (Farkas, I., Hardy, T. A., DePaoli-Roach, A. A., and Roach, P. J. Biol. Chem. We have now cloned another gene, GSY2, which encodes a second S. cerevisiae glycogen synthase. The GSY2 sequence predicts a protein of 704 residues, molecular weight 79,963, with 80% identity to the protein encoded by GSYl. Amino acid sequences obtained from a second polypeptide of 77 kDa present in yeast glycogen synthase preparations matched those predicted by GSY2. GSYl resides on chromosome VI, and GSY2 is located on chromosome XII. Disruption of the GSYl gene produced a strain retaining about 85% of wild type glycogen synthase activity at stationary phase, while disruption of the GSY2 gene yielded a strain with only about 10% of wild type enzyme activity. The level of glycogen synthase activity in yeast cells disrupted for GSYl increased in stationary phase, whereas the activity re- mained at a constant low level in cells disrupted for GSY2. Disruption of both genes resulted in a viable haploid that totally lacked glycogen synthase activity and was defective in glycogen deposition. In conclu-sion, utilized ethanol precipitation from cell extracts followed by digestion with a-amylase and amyloglucosidase (Sigma). Glycogen was esti-mated from the amount of glucose released (26). In the assay of water- soluble and insoluble glycogen, yeast cells were extracted with hot alkali as described by Gunja-Smith et al. (19). Qualitative glycogen determinations were made by staining cells with iodine vapor on YPD plates. Other Methods-Protein was quantitated by the method of Brad- ford (27), using bovine serum albumin as standard. Tetrad analysis was performed by standard methods of yeast genetics (18).

In previous work, we identified a Saccharomyces cerevisiae glycogen synthase gene, G S Y l , which codes for an 85-kDa polypeptide present in purified yeast glycogen synthase (Farkas,  265, 20879-20886). We have now cloned another gene, GSY2, which encodes a second S. cerevisiae glycogen synthase. The GSY2 sequence predicts a protein of 704 residues, molecular weight 79,963, with 80% identity to the protein encoded by G S Y l . Amino acid sequences obtained from a second polypeptide of 77 kDa present in yeast glycogen synthase preparations matched those predicted by GSY2. GSYl resides on chromosome VI, and GSY2 is located on chromosome XII. Disruption of the G S Y l gene produced a strain retaining about 85% of wild type glycogen synthase activity at stationary phase, while disruption of the GSY2 gene yielded a strain with only about 10% of wild type enzyme activity. The level of glycogen synthase activity in yeast cells disrupted for G S Y l increased in stationary phase, whereas the activity remained at a constant low level in cells disrupted for GSY2. Disruption of both genes resulted in a viable haploid that totally lacked glycogen synthase activity and was defective in glycogen deposition. In conclusion, yeast expresses two forms of glycogen synthase with activity levels that behave differently in the growth cycle. The GSY2 gene product appears to be the predominant glycogen synthase with activity linked to nutrient depletion.
Glycogen synthase (EC 2.4.1.11) catalyzes the formation of a-1,4-glycosidic bonds in glycogen in a wide variety of species. Mammalian glycogen synthase is regulated by multisite phosphorylation and exists as at least two different isoforms in *This research was supported by National Institutes of Health Grants DK42576 and DK27221. The Biochemistry Biotechnology Facility is supported in part by the Indiana University School of Medicine Diabetes Research and Training Center, (Grant DK20542). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. muscle and liver (1, 2). cDNAs encoding human muscle (3), rabbit muscle (4), and rat liver (5) glycogen synthases have been cloned and the corresponding amino acid sequences deduced. Glycogen synthase in Saccharomyces cerevisiae has not been as extensively characterized. The yeast enzyme is known to undergo reversible phosphorylation (6,7), and some glycogen-deficient mutants have been shown to have increased activity of CAMP-dependent protein kinase due either to a defective regulatory subunit (8,9) or to increased cAMP levels (10). Thus, the cAMP pathway is implicated in the regulation of glycogen metabolism. However, glycogen accumulation may also be regulated by mechanisms independent of cAMP (11,12). Many details of the regulation remain to be established, and no glycogen-deficient mutant with defective glycogen synthase has been characterized so far.
Our yeast glycogen synthase preparations contained two polypeptides of M, 77,000 and 85,000 (13), in agreement with the results of Peng et al. (14) but differing from other reports of a single subunit type (15,16). We recently cloned a glycogen synthase gene, GSYI, from S. cereuisiue that encodes the 85-kDa polypeptide. Disruption of GSYl resulted in a viable haploid with significant residual glycogen synthase activity. Also, Southern hybridization of genomic DNA with a GSYl probe revealed a second fragment hybridizing at low stringency (13). We therefore postulated the existence of a second glycogen synthase gene. Understanding the control of glycogen synthesis in yeast obviously requires characterization of both glycogen synthase genes. Here we report cloning of a second gene, GSY2, and analysis of mutants with one or both genes selectively disrupted.
Growth of Yeas-Yeast were grown in YPD medium containing 1% Bacto-yeast extract (w/v), 2% peptone (Sigma) (w/v), and 2% dextrose (w/v) (18). To monitor the growth curve, 250-ml cultures (YPD) were inoculated with precultures grown to stationary phase and grown at 30 "C with shaking. The initial cell concentration was about 2 X 10" cells/ml. Growth was followed by measuring the turbidity of culture at 640 nm. Aliquots producing -150 pl of pelleted cells were removed at various times for analysis. For determination of water-soluble and water-insoluble glycogen, with or without glutaraldehyde fixation (19), 50-ml cultures (YPD) were inoculated with cells from 5-ml saturated cultures and grown for 16 h. Two 20-ml aliquots of each culture were pelleted.
Cloning of the GSY2 Gene-Screening of a yeast genomic library constructed in EMBL3a and kindly supplied by Dr. Mike Snyder, Yale University (20), was performed with a GSYl probe looking for differential hybridization at low versus high stringency. A 2.3-kb' SpeI-NdeI fragment of GSYl containing the coding region of the gene was labeled by random priming (U. S. Biochemical Corp. kit) using [ C Y -~~P ]~C T P .
(1 x SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0.) Approximately 30,000 recombinants were screened. Eight clones were selected, which hybridized strongly at 58 "C but more weakly at 68 "C. The eight clones were related to each other as judged by restriction mapping. One was selected for further analysis. A 3.8-kb SalI-SalI fragment, which hybridized with the probe, was subcloned into pBluescript KS' (Stratagene), and a 2,914-base pair segment covering the coding region of the gene was sequenced from both strands.
DNA and Protein Sequence Analysis-pBluescript vector containing the subcloned 3.8-kb SalI-Sal1 fragment of GSY2 inserted in both orientations was digested with NotI. After filling in the 5' restriction overhangs with a-thio-dNTPs, the plasmids were cut with SmaI, and nested unidirectional deletions were created by exonuclease 111 and Mungbean nuclease (Ex0 III/Mungbean deletion kit, Stratagene). Nucleotide sequence was determined by the dideoxy chain termination method (21) using a T7 (17-mer) primer. The sequencingprojects and sequence alignments were managed using the PCGene program (Intelligenetics). Protein sequence data base searching was achieved using the FASTA algorithm and the data base maintained by Dr. Mark Goebl at Indiana University.
DNA Hybridization Analysis-DNA isolated (18) from the wild type strain YPH52 and from mutants containing the gsylAl and/or the gsy2A1 mutation was digested with EcoRI and/or EcoRV and electrophoresed on a 0.8% agarose gel. After transfer to nitrocellulose, the filter was hybridized with the 1.1-kb Sad-EcoRV fragment of GSY2. The fragment was labeled by random priming (U. S. Biochemical Corp. kit) using [cx-~'P]~CTP and hybridized at 58 "C in 6 X SSPE, 10 X Denhardt's solution, 0.1% SDS, 0.05% sodium pyrophosphate. Washing was carried out in 6 X SSC, 0.1% SDS, 0.05% sodium pyrophosphate at 60 "C prior to autoradiography.
For chromosomal analysis of GSYl and GSY2, chromosomes from yeast strain AB 1050 were separated by contour-clamped homogeneous electric field electrophoresis (22) (CHEF-DRII System, Bio-Rad) on a 1% agarose gel. The switch times were 60 s for 15 h followed by 90 s for 4 h. Electrophoresis was performed at 200 V. The DNA was transferred to a nitrocellulose filter that was hybridized with a 2.3kb EcoRI-SalI fragment of GSY2. After removal of the probe with hot 0.1 X SSPE and 0.1% SDS, the filter was rehybridized to a 2.3kb SpeI-NdeI fragment of GSYI. To confirm chromosome identification, a 0.8-kb EcoRV-EcoRV fragment of CDC4 from plasmid pCDC4-35 (23) and rDNA in the plasmid pYlrA12 (24) were used as probes for chromosomes VI and XII, respectively. Hybridization was performed at 58 "C as before.
Determination of Glycogen and Glycogen Synthase Activity-For correlation of growth with glycogen content and glycogen synthase activity, aliquots of yeast cultures were harvested by centrifugation at 1,700 X g. Cell pellets were suspended in 1 ml of cold homogenization buffer containing 50 mM Tris-C1 (pH 7.5), 1 mM EDTA, 100 mM sodium fluoride, 3 mM dithiothreitol, 0.1 mM 1-chloro-3-tosylamido-7-amino-2-heptanone, 5 mM benzamidine, 0.25 pg/ml leupeptin, 0.5 pg/ml aprotinin. Cell pellets were harvested by centrifugation and resuspended in 300 pl of homogenization buffer. The cells were broken by vortexing with an equal volume of glass beads in four 30-s pulses. The homogenized mixture was centrifuged at 8,000 X g for 2 min, and the supernatant was used for the assay of glycogen synthase and glycogen. Total glycogen synthase activity was measured in the presence of glucose-6-P by the method of Thomas et al. (25). One unit of activity represents the incorporation of 1 pmol of glucose from UDPglucose into glycogen per min at 30 "C. Glycogen measurements ' The abbreviations used are: kb, kilobase; SDS, sodium dodecyl sulfate.
utilized ethanol precipitation from cell extracts followed by digestion with a-amylase and amyloglucosidase (Sigma). Glycogen was estimated from the amount of glucose released (26). In the assay of watersoluble and insoluble glycogen, yeast cells were extracted with hot alkali as described by . Qualitative glycogen determinations were made by staining cells with iodine vapor on YPD plates.
Other Methods-Protein was quantitated by the method of Bradford (27), using bovine serum albumin as standard. Tetrad analysis was performed by standard methods of yeast genetics (18).

RESULTS
Cloning of GSY2-A yeast genomic library was screened using GSYl as probe as described under "Experimental Procedures." Eight clones were identified that hybridized strongly at low stringency but much more weakly at high stringency. Because of previous Southern hybridization of genomic DNA, we considered these clones to be candidates for a second gene encoding glycogen synthase. One clone was chosen for sequencing. A 3.8-kb SalI-Sal1 fragment was subcloned, and a segment of 2,914 base pairs was sequenced from both strands (Fig. 1). The 2,115-base pair open reading frame (Fig. 2) is 74% identical to GSYl and encodes a protein of 704 amino acids with 80% identity to the GSYl protein. We propose to call the gene GSY2 and the corresponding protein glycogen synthase-2. The GSYl gene product would then be glycogen synthase-1. The overall identity of yeast glycogen synthase-2 to rabbit muscle and rat liver glycogen synthases is 48 and 45%, respectively (Fig 3). In the untranslated sequence preceding the initiator ATG, T residues occur instead of A residues at position -1 to -5 unlike what is found in highly expressed yeast genes (28). Also, the codon bias index of 0.24 calculated according to Bennetzen and Hall (29) shows little codon bias characteristic of highly expressed genes.
Chromosomal Localization of GSYl and GSY2"To identify the chromosomes on which GSYl and GSY2 reside, fragments of the genes were hybridized with S. cereukiue chromosomes separated by pulsed field electrophoresis and transferred to nitrocellulose. Fourteen of the 16 chromosomes were clearly separated, with chromosomes V and VI11 running as a poorly resolved doublet (not shown). A 2.3-kb SpeI-NdeI fragment of GSYl hybridized to chromosome VI. The identity of chromosome VI was confirmed by its hybridization to a fragment of CDC4, which is known to reside in chromosome VI (30). A 2.3-kb EcoRI-SalI fragment of GSY2 hybridized to the same chromosome as a rDNA probe. Therefore, GSY2 is located on chromosome XII, which is known to contain rDNA sequences (30).
Disruption of GSY2"For disruption of the GSY2 gene, pRS303 vector (31) containing the HIS3 marker gene was used. An EcoRV-ApaI fragment bearing the 5' region of GSY2 was introduced into pRS303 followed by insertion of the 3' Sad-EcoRV fragment of the gene into the corresponding sites of the modified plasmid. The resulting pRS303-GSY2Al construct (Fig. 4)  the plasmid with EcoRV resulted in a linear DNA in which most of the coding sequence was replaced by pRS303 sequences, including the HIS3 marker. This linearized DNA was used for one-step gene replacement (32). Haploid wild type (YPH52) and gsylA1 mutant yeast cells were transformed with the linearized pRS303-GSY2Al DNA, using the LiAc method (33). Selection was made for His+ transformants. Southern hybridization was used to confirm that GSY2 was replaced by the disrupted gene (Fig. 5). DNA from wild type and from haploid mutant strains were digested with appropriate restriction enzymes and probed with a 1.1-kb SacI-EcoRV fragment containing a segment of the coding region of GSY2. DNA isolated from wild type and gsyl A1 mutant cells contained a 3.5-kb EcoRV-EcoRV fragment (Fig. 5,  tracks 1 and 3 ) , as predicted from the restriction map of GSY2. Similarly, the 2.6-kb hybridizing fragments in the EcoRV-EcoRI digests of wild type (Fig. 5, track 2) and gsyl A1 mutant (Fig. 5, track 4 ) strains were indicative of the presence of GSY2. In the gsy2Al mutant and in the double mutant, these fragments were absent, and a 6.2-kb fragment appeared (Fig. 5, tracks 5-8), as expected from the strategy for the disruption of GSY2. Thus, haploid yeast strains were generated in which either or both of the glycogen synthase genes were defective. Phenotype of Yeast Cells Defective in Glycogen Synthase Genes-Disruption of GSY2 in both wild type (YPH52) and in gsylA1 haploid strains resulted in viable mutants. Therefore, GSY2 itself is not an essential gene nor is the presence of at least one of the two glycogen synthase genes an obligate to stationary phase (Fig. 6). No significant difference in the doubling time during exponential growth was observed, and growth profiles of wild type and mutant strains were similar. Glycogen synthase activity increased and glycogen accumulated in haploid wild type and gsyl A1 mutant yeast approaching stationary phase, although the levels of glycogen synthase and glycogen were somewhat lower in the mutant without functional GSYl. In the gsy2Al cells, however, the level of glycogen synthase remained at a constant low level throughout, about 10% that of wild type cells in stationary phase. A small increase in glycogen content was observed, but the increase was significantly lower than in wild type cells entering the stationary phase. In the strain defective for both GSYl and GSY2, no glycogen synthase activity was detected at any stage in the growth cycle, and very little or no glycogen accumulated. The experiment was also performed using a homozygous gsyl A1 gsy2Al diploid strain (described below) with similar results as the gsyl A1 gsy2Al haploid (data not shown). The phenotype for glycogen accumulation was also demonstrable using iodine staining to detect glycogen' (Fig.  7).  had suggested that the water-soluble and -insoluble glycogen fractions obtained after treating cells with hot alkali represented two different pools of glycogen. They proposed that insoluble glycogen was associated with a component of the cell wall (19). An obvious possibility was that the two glycogen synthases defined above might be correlated with the synthesis of ''soluble'' or ''insoluble'' glycogen pools. To test this hypothesis, we determined the conrequirement for viability. No obvious morphological differ-* The iodine staining procedure made it simple to investigate ence between wild type and mutant cells was apparent. To glycogen accumulation after longer periods of growth, and at about determine the specific effects of the loss of GSYl and GSY2 2-4 days we consistently saw a slightly higher intensity of stain in type and mutant strains were grown under the of glycogen deposits over a longer period seemed to be enhanced by the gsylA1 mutant as compared with wild type. Thus, maintenance Same conditions (YPD) and monitored for &ol glycogen the disruption of GSYI, a seemingly paradoxical result unless the synthase activity, and glycogen content during the approach expression of GSYI was somehow linked to glycogen degradation.  Table I). T o avoid overestimation of soluble glycogen due to enzymatic hydrolysis during cell homogenization, cells were pretreated with glutaraldehyde in one set of determinations. In gsylA1 mutant cells, the concentration of total and insoluble glycogen was only slightly lower than in wild type cells. In gsy2A1 mutant cells, however, the concentration of  gsylA1 (tracks 3 and 4 ) , gsy2A1 (tracks 5 and 6 ) , or gsylAl gsy2A1 (tracks 7 and 8 ) mutant yeast strains were digested with EcoRV (tracks I , 3, 5, and 7) or with EcoRV and EcoRI (tracks 2, 4, 6, and 8 ) and analyzed as described under "Experimental Procedures." The filter was hybridized with a 1.1-kb SacI-EcoRV fragment of GSY2.
both the soluble and insoluble glycogen was significantly lower than in YPH52 cells. Glycogen determination without glutaraldehyde fixation gave a higher estimate of the proportion of soluble glycogen. However, there was no abolition of either the soluble or insoluble pools of glycogen associated with the disruption of either of the glycogen synthase genes.
To test whether GSYI and GSY2 are required for sporulation, we generated a homozygousgsylA1 gsy2Al diploid strain as follows. A GSYl/gsylAl GSY2/gsy2Al diploid produced by sequential disruption a t either locus in the YP diploid was sporulated in McClary's medium (1% potassium acetate, 0.25% yeast extract, 0.1% glucose). Tetrads were dissected and scored for uracil and histidine prototrophy and analyzed by iodine staining. All glycogen synthase phenotypes were recovered and correlated with the different levels of glycogen staining shown in Fig. 7. Two gsyl A1 gsy2A1 haploids of opposite mating type were selected and crossed to generate homozygous gsyl A1 gsy2A1 diploids. The diploid mutant cells were identical to the gsylAl gsy2A1 haploid in terms of glycogen accumulation (Fig. 7). The mutant diploids and the parental YP strain were then sporulated in McClary's medium and the efficiency of sporulation scored by microscopy after 4 days a t 24 "C. Analysis of -600 cells of each strain indicated an efficiency of 27.4% for the YP strain and 14.7% for the mutant. Thus, cells defective in both glycogen synthase genes could sporulate with an efficiency not vastly different from wild type cells. We also considered the possibility that glycogen deposits were important for the germination of spores. When tetrads were micromanipulated onto YPD plates, the asci from the mutant and wild type diploids showed a similar capacity to form colonies (data not shown).
Glycogen can be converted to trehalose (34,35), which has been reported to accumulate in yeast cells upon exposure to FIG. 6. Changes in the levels of glycogen synthase and glycogen during growth of haploid wild type (YPH52) and mutant yeast strains disrupted for one or two glycogen synthase genes. YPD medium was inoculated with precultures as described under "Experimental Procedures." Panel A, absorbance a t 640; panel R, glycogen levels; panel C, total glycogen synthase activity.

FIG. 7.
Glycogen staining of wild t y p e and mutant yeasts disrupted in glycogen synthase genes. Yeast strains were patched onto a YPD plate, incubated a t 30 "C for 24 h, and exposed to iodine vapor. I, YPH52; 2, gsylAl (IFI); 3, gsy2AZ (IF2); 4, gsyZAl gsy2AZ haploid (IF3); 5, YP; 6, homozygousgsyZAZ gsy2AZ diploid (IF4). heat (36). One hypothesis is that glycogen reserves might influence trehalose availability and hence the ability of the cell to withstand heat shock. Therefore, we compared the thermotolerance of wild type and mutant strains. No significant difference was found in the heat shock resistance of wild type and glycogen-deficient cells exposed for 30 min to 55 and 50 "C on YPD plates or in YPD liquid cultures, respectively (results not shown).

DISCUSSION
The results show unequivocally that there are two genes in S. cereuisiae, GSYl and GSY2, that code for glycogen syn- thase. Preparations of the enzyme from wild type yeast contained two polypeptides, M, 85,000 and 77,000. We had previously concluded that the 85-kDa polypeptide was encoded by GSYI, since enzyme purified from yeast defective in GSYl contained only the 77-kDa species (13). Nonetheless, amino acid sequences of several try-ptic peptides derived from the 77-kDa polypeptide matched the deduced GSYl protein sequence. It is now clear that these peptide sequences either came from regions of complete identity between the two protein sequences or else contained ambiguities in the assignments that embraced the small sequence differences. One peptide sequence (ATYQNEVDILD) (Fig. 2) did not match the glycogen synthase-1 sequence and had only 7 of 11 identities. This sequence matches perfectly the protein sequence predicted by GSY2, and we therefore conclude that the GSY2 gene encodes the 77-kDa polypeptide. From our failure to obtain the NH2-terminal sequence from the intact 77-kDa polypeptide, we had suggested that it may be modified at the NH2 terminus (13 of the corresponding polypeptides as judged by SDS-polyacrylamide gel electrophoresis. As noted, glycogen synthase-2 might be post-translationally modified at the NH2 terminus leading to aberrant electrophoretic mobility. Glycoge.1 synthase-1 has apparent M , of 85,000, significantly greater than its predicted mass of 80,378 daltons. Phosphorylation is known to reduce the electrophoretic mobility of many proteins, including mammalian glycogen synthase (17), and could contribute to the unpredicted mobility differences between the two glycogen synthase subunits. However, phosphorylation and dephosphorylation in vitro have not been shown to alter electrophoretic mobility of either of the polypeptides (10). Limited proteolysis of glycogen synthase-2 could also explain why its electrophoretic mobility is higher than predicted but, if the NH2 terminus is blocked, would have to be at the COOH terminus. Glycogen synthase-2, with predicted PI of 5.9, is acidic, like yeast glycogen synthase-1 and the mammalian glycogen synthases (3-5). Furthermore, the COOH-terminal region has a notable net negative charge, again in common with all the other glycogen synthases for which sequences are known (3-5). The COOH terminus of glycogen synthase-2 is very similar to that of glycogen synthase-1 and is the part of the molecule that differs most from the mammalian enzymes. This region is implicated in the regulation of the enzyme by phosphorylation and may contain sites for phosphorylation by CAMP-dependent protein kinase (13). In the glycogen synthase-1 sequence, three potential COOH-terminal sites for this protein kinase were predicted (13) based on a minimal recognition sequence of R/K-X-X-S. Interestingly, the presence of an extra residue in the glycogen synthase-2 sequence would destroy one of these potential sites (Ser-660 of glycogen synthase-2) leaving as candidates Ser-650 and Ser-662. Biochemical and genetic experiments to define the exact phosphorylation sites are under way.
Though many yeast mutants with aberrant glycogen accumulation have been identified, none of these has, to our knowledge, involved defective glycogen synthase. The presence of two quite similar genes, of course, can account for this failure. Having cloned both the GSYl and GSY2 genes, we were in a position to examine the viability and phenotypes of mutant yeast defective in one or both genes. Yeast lacking functional versions of both genes were viable and yet lacked detectable glycogen synthase activity, even at stationary phase. Glycogen content was very low (Table I). Since glycogen was determined from the assay of glucose released by hydrolysis, the small values measured may be due to the lack of specificity for glycogen of amyloglycosidase. The absence of measurable glycogen synthase activity in the double mutant argues that, at least under the growth conditions used, no other gene for glycogen synthase was expressed. We consider it probable that yeast contains only two genes encoding glycogen synthase, although one cannot rigorously exclude the possibility that another gene might be functional under other untested circumstances. In any event, our results with the double mutants demonstrated that loss of glycogen synthase activity and defective glycogen accumulation were not associated with any major impairment in viability, ability to sporulate or recover from sporulation, or survival and rate of growth after heat shock. We conclude that glycogen is not essential for S. cereuisiae under the limited number of conditions tested. However, the ability of the organism to accumulate glycogen and to do so in a carefully regulated manner makes it difficult to view glycogen deposition as a redundant process. Future work will seek conditions in which defective glycogen accumulation correlates with impaired performance of the cells.
Our studies of strains selectively disrupted in one of the GSY genes permit us to speculate on differences in the regulation of the two genes. Yeast containing only the GSYl gene express a low but constant level of glycogen synthase activity throughout the growth cycle and are severely restricted in their ability to accumulate glycogen upon entry into stationary phase. Though our analyses to date have been only of enzyme activity, it appears as if the GSYl gene is expressed constitutively. In yeast that express only glycogen synthase-2, glycogen deposition is only slightly less than in the wild type, and large increases in glycogen synthase activity are observed prior to the onset of stationary phase, as with the wild type. Thus, GSY2 encodes the predominant glycogen synthase, in keeping with the relative proportions of 77-and 85-kDa polypeptides in the purified enzyme samples.
Furthermore, we hypothesize that expression of GSY2 is controlled by environmental signals such as nutrient deprivation. Why yeast contain glycogen synthase genes under different controls remains unexplained.