The Yeast GLC7 Gene Required for Glycogen Accumulation Encodes a Type 1 Protein Phosphatase*

The glc7 mutant of the yeast Saccharomyces cereuisiae does not accumulate glycogen due to a defect in glycogen synthase activation (Peng, Z., Trumbly, R. J., and Reimann, E. M. (1990) J. Biol. Chern. 266, 13871-13877) whereas wild-type strains accumulate glycogen as the cell cultures approach stationary phase. We isolated the GLC7 gene by complementation of the defect in glycogen accumulation and found that the GLC7 gene is the same as the DIS2SZ gene (Oh-kura, H., Kinoshita, N., Miyatani, S., Toda, T., and Yanagida, M. (1989) Cell 67, 997-1007). The protein product predicted by the GLC7 DNA sequence has a sequence that is 81% identical with rabbit protein phosphatase 1 catalytic subunit. Protein phosphatase 1 activity was greatly diminished in extracts fromglc7 mutant cells. Two forms of protein phosphatase 1 were identified after chromatography of extracts on DEAE-cellulose. Both forms were diminished in the glc7 mutant and were partly restored by transformation with a plasmid carrying the GLC7 gene. Southern blots indicate the presence of a single copy of GLC7 in S. cereviaiae, and gene disruption

J., and Reimann, E. M. (1990) J. Biol. Chern. 266, 13871-13877) whereas wild-type strains accumulate glycogen as the cell cultures approach stationary phase. We isolated the GLC7 gene by complementation of the defect in glycogen accumulation and found that the GLC7 gene is the same as the DIS2SZ gene (Ohkura, H., Kinoshita, N., Miyatani, S., Toda, T., and Yanagida, M. (1989) Cell 67, 997-1007). The protein product predicted by the GLC7 DNA sequence has a sequence that is 81% identical with rabbit protein phosphatase 1 catalytic subunit. Protein phosphatase 1 activity was greatly diminished in extracts fromglc7 mutant cells. Two forms of protein phosphatase 1 were identified after chromatography of extracts on DEAEcellulose. Both forms were diminished in the glc7 mutant and were partly restored by transformation with a plasmid carrying the GLC7 gene. Southern blots indicate the presence of a single copy of GLC7 in S. cereviaiae, and gene disruption experiments showed that the GLC7 gene is essential for cell viability. The GLC7 mRNA was identified as a 1.4-kilobase RNA that increases 4-fold at the end of exponential growth in wild-type cells, suggesting that activation of glycogen synthase is mediated by increased expression of protein phosphatase 1 as cells reach stationary phase.
The control of glycogen metabolism is in general similar in yeast and mammalian cells. The net rate of glycogen accumulation is determined by the opposing actions of glycogen synthase and glycogen phosphorylase. The activities of these two enzymes are regulated by phosphorylation, with glycogen synthase inhibited (Rothman-Denes and Cabib, 1971) and * This work was supported in part by United States Public Health Service Grant HL36573 and by American Cancer Society Grant MV-450. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. phosphorylase activated (Fosset et al., 1971) by phosphorylation.
Glycogen synthase in mammalian cells is phosphorylated by a number of protein kinases, acting in a synergistic manner to regulate its activity. The highly phosphorylated forms of glycogen synthase require glucose 6-phosphate for maximal activity, whereas the dephosphorylated forms have activity independent of glucose 6-phosphate . The major protein phosphatases implicated in regulation of mammalian glycogen synthase are protein phosphatases 1 and 2A . Interconversion between phosphorylated and dephosphorylated forms of yeast glycogen synthase has been demonstrated, both in vivo and in vitro (Rothman-Denes and Cabib, 1970;Huang and Cabib, 1974;Peng et ai., 1990), but the protein kinases and phosphatases that regulate glycogen synthase in the yeast cell have not been identified. The cAMP pathway plays an important role in the control of glycogen accumulation in yeast. Mutations that increase CAMP-dependent protein kinase activity, by increasing cAMP levels for example, reduce glycogen accumulation; mutations that reduce CAMPdependent protein kinase activity, such as ras2 mutations that reduce cAMP leveb, induce hyperaccumulation of glycogen (Cannon et al., 1986). Glycogen accumulation in yeast can also be regulated by mechanisms independent of the cAMP pathway (Cameron et al., 1988).
We examined glycogen synthase activity in a collection of glycogen-deficient mutants, glcl -gk8 and found the greatest effect in the glc7 strain, where all of the enzyme was highly phosphorylated and dependent on Glc-6-P (Peng et al., 1990). In this report we show, by cloning and sequencing of the GLC7 gene, that GLC7 encodes protein phosphatase 1 and demonstrate that protein phosphatase 1 activity is decreased in the g k 7 strain.
Strains-All yeast strains used in this study are Saccharomyces cereuisiae of the S28& genetic background. A yeast mutant with reduced glycogen accumulation, gk7, was obtained from Dr. J. Pringle (University of Michigan). RTY370, a diploid wild-type strain, is the same as C276 (Lillie and Pringle, 1980). The glc7 mutant strains were derived from the diploid strain 22R1 (glc7/glc7) (Peng et al., 1990). The haploid strain RTY378-1A ( M A T a glc7) derived from sporulation of 22R1 was crossed with RTY214 ( M A T a cyc8-20 his4-519 leu2-3 leu2-112 u r d -5 2 ) . In the tetrads produced by sporulation of this diploid, the glycogen deficiency associated with glc7 segregated 2:2, and a glc7 u r d segregant, RTY398 ( M A T a g k 7 ku2-3 leu2-112 u r d -52), was chosen to facilitate the isolation of the GLC7 gene. RTY518 is RTY398 transformed with pRS316. RTY520 is RTY398 trans-
Phmids-Standard methods were used for the construction of plasmids (Maniatis et al., 1982). The yeast genomic DNA library in the centromere vector YCp50, which contains URA3 as a selectable marker, was a kind gift of Dr. M. D. Rose (Rose et al., 1987). The plasmid pZFl containing the GLC7 gene isolated from this library had approximately 8.5 kbl of insert DNA. The following plasmids ( Fig. 1) were derived directly from pZF1. A 7-kb HindIII fragment was subcloned into the Hind11 site of YCp50 to produce pZF2; the 2.6-kb XhoI fragment was subcloned into the XhoI site of pRS316, which contains URA3 as a selectable marker (Sikorski and Hieter, 1989), to yield pZF111; and a 3-kb XhoI fragment was ligated into the XhoI site of pRS316, to yield pZF135. pZF139 was created by digestion of pZFl with EcoRI followed by religation. The 2.1-kb BglII-XhoI fragment of pZF135 was subcloned into the XhoI and BamHI sites of pRS316 to yield pZF141. pZF135 was digested by EcoRI and religated to form pZF144 and digested by SalI and religated to yield pZF145. The 2.7-kb HindIII-XhoI fragment from pZF135 was transferred into the HindIII-XhoI sites of pRS316 to give pZF226 and into the HindIII-SalI sites of YIp5 (Botstein et al., 1979) to make pZF239. The plasmid pZF238 used for gene disruption was constructed by deleting the EcoRI site in the polylinker of pZF226 by digestion with HindIII and NotI, Klenow treatment, and religation and then subcloning the 1.5-kb EcoRI-BamHI fragment from YEp24 (Botstein et al., 1979) containing the URA3 gene into the EcoRI-BglII sites.
To generate subclones for DNA sequencing, the 3-kb XhoI fragment of pZFl was subcloned into the SalI site of the vector pBS(+) (Stratagene) in both orientations to produce pZF69 (+ orientation) and pZF 149 (-orientation). Nested series of deletions from the ends of insertion were prepared by sequential treatments with exonuclease I11 and S1 nuclease (Henikoff, 1984).
master plate onto a YEPD plate. The YEPD plate was put at 30 "C Iodine Staining-Yeast colonies were patched or replicated from a overnight, and 10 ml of 0.1% iodine and 1% potassium iodide solution were added to stain colonies for about 10 min at room temperature (Chester, 1967). Glycogen Synthose Assay-Activity of the glycogen synthase was measured with or without 7.2 mM Glc-6-P at 30 "C for 10 min in an assay mixture that also contained 4.4 mM UDP-['4C]glucose, 50 mM Tris-HC1, pH 7.5, 10 mM EDTA, 10 mM EGTA, 10 mg of purified (Peng et al., 1990) rabbit liver glycogen/ml, and enzyme in a total volume of 90 pl. Samples were diluted with dilution buffer, 50 mM Tris-HC1, pH 7.5.1 mM EDTA, 60 mg of oyster glycogen/ml, and 42 mM 0-mercaptoethanol. Reactions were terminated by spotting 75 p1 of the reaction mixture onto Whatman filter papers that were processed as described previously (Thomas et al., 1968).
Protein Phosphatase Assay-Phosphorylase phosphatase was assayed by measuring the release of 32P from [32P]phosphorylase a (Killilea et al., 1978;Peng et al., 1991). A 100-p1 assay contained 45 mM imidazole-HC1, pH 7.4 (at 25 "C), 6 mM theophylline, 1.8 mM dithiothreitol, 5 mM Tris-HC1, pH 7.4 (at 25 "C), 0.02 mM EDTA, 2 p~ [32P]phosphorylase a (monomer), protein phosphatase, and unless otherwise indicated, 3 nM okadaic acid to suppress protein phosphatase 2A activity. Co2+/trypsin-stimulated phosphatase activity was measured by preincubating the enzyme with 0.2 mM free CoC12 for 5 min at 30 "C, followed by addition of trypsin to a final concentration of 0.02 mg/ml for an additional 10 min (Schlender et al., 1986). Proteolysis was terminated by addition of soybean trypsin inhibitor to a concentration 4 times greater than trypsin. The phosphatase reaction was initiated by addition of 10 pl of untreated or Co2+/ trypsin-treated enzyme to 90 pl of reaction mix. After incubation at 30 "C for 20 min, the reaction was terminated by addition of trichloroacetic acid to 10% and acid-soluble 32P was measured. One unit of phosphatase activity releases 1 pmol of Pi from phosphorylase a/min. Chromatography on DEAE-cellulose 52-Extracts of yeast cells were prepared as follows. Yeast cells of wild-type strain and transformants with GLC7-carrying plasmids were grown in 200 ml of SD medium without uracil at 30 "C, and yeast cells of mutant strain were grown in 200 ml of SD + uracil. Yeast cells were harvested at early The abbreviations used are: kb, kilobase(s); EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; SDS, sodium dodecyl sulfate; MOPS, 4-morpholinepropanesulfonic acid. stationary phase ( A m = 1.6) by centrifuging at 4,000 X g in a GSA rotor for 10 min and washed with 200 ml of cold water. Frozen cells (0.5 g) were resuspended in 1 ml of buffer A (50 mM Tris-HC1 at pH 7.4,0.2 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride). Cells were lysed by vortexing for 4 x 1 min with 1 ml of glass beads, with cooling on ice between vortexing, and centrifuged at 15,000 X g for 15 min in a Beckman Microfuge to yield 1 ml of extract. Extracts were diluted to 4 ml and applied to a 3 ml DEAE-cellulose (DE52) column. The column was washed with 4 ml of buffer A and eluted with 20 ml of a linear salt gradient (0-0.5 M NaC1) in buffer A. Oneml fractions were collected and assayed for phosphorylase phosphatase activity.
Southern Blot of GLC7"An overnight culture of wild-type strain RTY370 (10 ml) was harvested, and a rapid method was used for preparing chromosomal DNA from yeast (Hoffman and Winston, 1987). Chromosomal DNA (10 pg) was digested with various endonucleases and applied to a 0.8% agarose gel. After DNA was transferred to the nitrocellulose filter, the filter was baked in an 80 'C vacuum oven for 2 h, prehybridized in 10 ml of DNA hybridization buffer (6 X SSC (1 X ssc: 0.15 M NaCl, 15 mM sodium citrate, pH 7.0), 0.1% SDS, 5 X Denhardt's solution, 100 pg of salmon sperm DNA/ml) at 55 "C for 3 h, and then hybridized overnight with a labeled 1.3-kb ChI-EcoRI fragment from the 3' end of GLC7 in 10 ml of DNA hybridization buffer.
DNA Sequencing-DNA samples for sequencing were prepared by alkaline lysis and acid phenol extraction (Weickert and Chambliss, 1989). After precipitation of cell debris and SDS with potassium acetate, an equal volume of acid phenol (equilibrated with 50 mM sodium acetate at pH 4.0) was added to the supernatant for extraction. The supernatant was then extracted with 1 volume of acid phenolxh1oroform:isoamyl alcohol (25:24:1) and then with 1 volume of ch1oroform:isoamyl alcohol (24:l). The DNA was precipitated by the addition of 2 volumes of 100% ethanol, washed with 70% ethanol, and then resuspended in 50 pl of TE solution (10 mM Tris, pH 8.0, 1 mM EDTA); 10 p1 (-1 pg) of DNA solution was used for sequencing. Both DNA strands were sequenced by the dideoxynucleotide chain termination method using Sequenase 2.0 (Tabor and Richardson, 1987).

RESULTS
Cloning the GLC7 Gene-We have found previously that yeast glc7 mutant strains show decreased dephosphorylation of glycogen synthase, causing diminished activation of this enzyme (Peng et al., 1990). A screening method for isolating the GLC7 gene exploited the fact that glc7 mutants do not accumulate glycogen, enabling mutant and wild-type colonies to be distinguished by staining with iodine (Chester, 1967). A yeast glc7 mutant strain RTY398 (MATa glc7 leu2 ura3) was transformed with a genomic DNA library in the centromere vector YCp50. Out of 16,000 Ura+ transformant colonies screened, eight (0.05%) gave a positive staining for glycogen. The plasmids extracted from these eight transformants gave identical restriction maps with an insert of 8.5 kb. One of these plasmids, pZF1, was retransformed into the gZc7 mutant strain RTY398. All of the transformant colonies accumulated glycogen to the levels similar to wild-type when assayed by iodine staining. This result confirmed that the complementation of the glc7 mutation was due to plasmid pZF1. One of these transformants was grown to stationary phase in YEPDrich medium to allow for loss of the plasmid and then plated onto YEPD. When these colonies were replicated onto SD plates without uracil and onto YEPD plates for iodine staining, colonies that failed to grow without uracil also were negative for glycogen after iodine staining. Therefore, complementation of the glc7 defect in glycogen accumulation was due to the presence of the gene on the plasmid.
The GLC7 gene was localized within the 8.5-kb insert by subcloning and testing glycogen accumulation with iodine staining. A restriction map of this region was derived from single and double restriction digests (Fig. 1). Different selected fragments were subcloned into the centromere vectors YCp50 or pRS316 and tested for complementation by trans- formation of the glc7 mutant RTY398. These experiments localized the GLC7 gene to the 2.7-kb HindIII-XhoI fragment on pZF226.

Yeast Protein Phosphatase
To demonstrate that we had cloned the GLC7 gene and not an extragenic suppressor, the cloned yeast DNA was used to direct integration of a plasmid into yeast chromosomal DNA. First, the 2.7-kb HindIII-XhoI fragment was subcloned into the integrating vector YIp5. Next, the resulting plasmid pZF239 was cut with SalI, which cuts once within the insert DNA. The diploid yeast strain RTY553, which is heterozygous for the gk7 mutation, was transformed with this DNA. One of the transformants, which had a single copy of pZF239 integrated into the chromosome as determined by Southern blotting (data not shown), was sporulated. Of the 23 tetrads that had regular segration of the URA3 gene, 22 were parental ditype for segregation of glc71urd and one was tetratype. This result shows that the cloned gene is identical with or very tightly linked to GLC7. The spores that were URA3' were glycogen-deficient, indicating that the plasmid had integrated at the chromosome containing the glc7 allele. The wild-type information on the plasmid could have been converted to the mutant sequence, which would be expected if the SalI site is close to the site of the mutation in the glc7 allele (Rothstein, 1991).
Deletion of the GLC7 Gene-A deletion was created in the GLC7 gene by replacing the 1.5-kb BglII-EcoRI fragment with the yeast URA3 gene. This deletion removes the entire intron and two-thirds of the coding sequence (see below). The deletion was inserted at the glc7 chromosomal locus by transforming the yeast strain RTY553 with the plasmid pZF238 that had been digested with Sac1 and XhoI. A transformant with a single copy of pZF238 integrated at the GLC7 locus was identified by Southern blotting and sporulated. Fourteen of the dissected tetrads produced two viable spores, one produced three viable spores, and two had one viable spore. Since all of the viable spores were urd-, with the exception of one of the spores of the three-spored ascus, the insertion of URA3 into the GLC7 gene prevents spore germination or cell proliferation.
Sequencing the GLC7 Gene-The nucleotide sequence of the 3-kb XhoI fragment, comprising 3,064 nucleotides, was determined by DNA sequencing (data not shown). The gene contains two long coding regions separated by a region with stop codons in all three reading frames. The open reading frames together code for a protein of 312 amino acids with strong homology to the type 1 protein phosphatases from several species. The putative GLC7 gene product was deduced by aligning the predicted amino acid sequence with those of other type 1 protein phosphatases. The alignment gives consensus sequences for exon-intron splice junctions at the appropriate positions. A defect in protein phosphatase 1 is consistent with the previous finding that glycogen synthase in the glc7 strain is phosphorylated more extensively than normal (Peng et al., 1990). The GLC7 sequence is virtually identical with the S. cerevisiae DIS2S1 gene, which was isolated by hybridization screening with the Schizosaccharomyces pombe 0152 gene involved in chromosomal disjunction (Ohkura et al., 1989). The first 204 nucleotides of the sequence are 97% identical with the 6 repeat sequence found at the ends of the TY1-transposable elements (Boeke et al., 1988).

Protein Phosphatase Activity of Wild-type and Mutant
Strains-Protein phosphatase activity was measured in yeast extracts to determine if the glc7 mutation of the gene for protein phosphatase 1 would be reflected by changes in the activity of protein phosphatase 1. In order to distinguish protein phosphatases 1 and 2A, we made use of two inhibitors: heat-stable protein inhibitor 2 and okadaic acid . Protein phosphatase 2A is not inhibited by inhibitor 2 but is almost completely inhibited by low concentrations of okadaic acid (2-5 nM) (Bialojan and Takai, 1988). Treatment of protein phosphatase 1 with Co2+ and trypsin can activate "cryptic" forms of protein phosphatase (Tung et al., 1984). These properties were used to optimize the detection of differences in the levels of protein phosphatases 1 and 2A in extracts of different strains of yeast. Phosphorylase a was used as the substrate since other phosphatases do not significantly dephosphorylate this substrate . Initially, we detected no significant difference in protein phosphatase 1 activity between unfractionated cell homogenates from wild-type and glc7 mutant strains (Peng et al., 1990). Subsequent experiments using activation with Co2+ and trypsin and sufficient dilution before assay in the presence of 3 nM okadaic acid, which suppresses phosphatase 2A, indicated that the glc7 mutant has about one-third as much protein phosphatase 1 as the wild-type and that activity is restored to normal after transformation with a plasmid carrying the GLC7 gene (Table I). The effects on protein phosphatase 1 were similar, irrespective of whether protein phosphatase 1 was defined as the activity sensitive to inhibitor 2 or as the activity remaining in the presence of 3 nM okadaic acid. No major changes in phosphatase 2A were detected with either assay. Since multiple forms of protein phosphatase 1 exist in mammalian cells (Cohen, 19891, these extracts were fractionated on DEAE-cellulose (DE52) to determine if multiple forms of protein phosphatase 1 are present and if these forms are differentially affected by the glc7 mutation. Fractions were treated with Co2+ and trypsin and assayed for the ability to dephosphorylate [32P]phosphorylase a in the presence of 3 nM okadaic acid. The data of Fig. 2 show the presence of two forms of protein phosphatase 1, both of which are markedly decreased in the glc7 mutant and restored after transformation. Treatment with Co2+ and trypsin increased activity

(TR).
about 5-fold in cell extracts and 3-fold or less in column fractions.
The GLC7 Gene Restores Glycogen Synthase Actiuity-To confirm that restoration of glycogen synthase activation is associated with increased phosphatase activity in the transformed glc7 strain, extracts of the wild-type, glc7 mutant, and glc7 mutant strains transformed with a plasmid carrying the GLC7 gene were assayed for glycogen synthase activity. As shown in Table I, two separate experiments demonstrated restoration of glycogen synthase activity after transformation. Southern Blot of Yeast Genomic DNA-Multiple genes encoding type 1 protein phosphatase have been found in fission yeast (Ohkura et al., 1989) and other organisms (Axton et al., 1990;Wadzinski et a/., 1990;Sasaki et al., 1990). A Southern blot of S. cereuisiae chromosomal DNA was performed to determine the number of type 1 protein phosphatase genes in this species. A restriction fragment containing most of the GLC7 coding region was hybridized to the blot under conditions of reduced stringency. As shown in Fig. 3, lanes 2 and 3 had a single band, confirming that there is a single gene in the yeast genome encoding a type 1 protein phosphatase.
Regulation of GLC7 mRNA-The level of glycogen synthase activity in yeast cells greatly increases during the transition from exponential growth to stationary phase (Lillie and Pringle, 1980; Cameron et al., 1988;Peng et al., 1990). T o see if Chromosomal DNA from wild-t.ype strnin HTYB'7O WRS prepared hv a rapid method (Hoffman and IVinston,198'7) and 10-pg portions were digested with different endonucleases and separated t)v elertrophoresis on a 0.8% agarose gel. The D N A WAS transferred to a nitrocellulose filter and hybridized by standard methods (Maniatin d  ab, 1982), except the hybridization conditions were 6 X SSC', 55 "C. The probe was a 1.3-kb Clal-EcoRI fragment comprising most of the second exon of the GLC7 gene. I m w I . a '7-kh Hind111 fragment and a 5.4-kb HindIII-Sal1 fragment of CLC7 were applied as control DNA (a small amount of the 1.6-kh HindllI-SnlI fragment of (;/,(' 7 was also present on the gel there is a relationship between the amount of GLC7 mRNA and glycogen synthesis, the amounts of GLC7 mRNA were determined a t different times of batch culture growth. As shown in Fig. 4, a Northern blot probed with a fragment from the 5' end of GLC7 revealed a 1.4 kb band present in polv(A)+ and total RNA. The level of GLC7 mRNA is constant during exponential growth and augments considerably near the end of exponential growth (panel I?). The relative amount of GLC7 mRNA increases about 4-fold at this time, as determined by scanning densitometry of the autoradiograph and rRNA stained with ethidium bromide. This increase in mRNA coincides with the increase in glycogen synthase activity ratio noted in previous studies (Rothman-Denes and Cabib, 1970;Peng et al., 1990).

DISCUSSION
Until recently, relatively little information has been available regarding the protein phosphatases that participate in the regulation of glycogen metabolism in S. cerpuisiap. Peng et al. (1991) showed that phosphatase 2A from yeast could activate yeast glycogen synthase in uitro. The experiments described here demonstrate that the glc7 mutation, which results in glycogen deficiency and excessive phosphorylation of glycogen synthase, is caused by a defect in the GLC7 gene that encodes a type 1 protein phosphatase, showing that protein phosphatase 1 can regulate yeast glycogen synthase in uiuo. Under the conditions of the experiments described here, it appears that protein phosphatase 1 is more important than phosphatase 2A for regulating glycogen synthase, since phosphatase 2A activity is normal in the mutant but the glycogen synthase activity ratio is much lower than normd. However, it is possible that phosphatase 2A is also involved in the regulation of the enzyme under other conditions and that glycogen synthase activity would be even lower in the absence of phosphatase 2A activity.
Our finding that a mutation in the yeast gene encoding protein phosphatase 1 decreases protein phosphatase 1 activity and results in glycogen deficiency is consistent with experiments showing that protein phosphatases 1 and 2A are R, repletion of GL('7 mRNA. Wild-t.ype strain RTYB70 WAS grown on YEPD medium, and portions were harvested at various intervals. Total RNA WAS purified from the cells by a rapid method (Schmitt et af., 1990), and poly(A)' RNA was purified as described (Edmonds et a!., 1971). Northern blotting using a 1% agarose gel in MOPS buffer and formaldehyde was performed as described previously (Trumbly, 1986). the major protein phosphatases regulating glycogen synthase in mammalian tissues. In some cells, the catalytic subunit of protein phosphatase 1 is associated with a larger regulatory subunit (G-subunit) which targets the phosphatase activity toward the enzymes of glycogen metabolism. I t was shown recently that stimulation of glycogen synthesis by insulin is mediated by the phosphorylation of the G-subunit, which activates the phosphatase activity toward glycogen synthase and phosphorylase kinase (Dent et al., 1990). The protein phosphatase 1 catalytic subunit also associates with inhibitor-2, whose activity may be regulated by phosphorylation by CAMP-dependent protein kinase. It is likely that the yeast protein phosphatase 1 is associated with analogous regulatory subunits, but these have not yet been identified. The presence of two peaks of activity of protein phosphatase 1 when yeast extracts are chromatographed on DEAE-cellulose suggests that at least one of these forms is a complex of the GLC7 gene product and a regulatory subunit. I t is possible that the first of the two peaks, which appears to be increased to a greater extent than the second peak after transformation (Fig. 2), represents the GLC7 gene product that is not complexed to a regulatory subunit. The increased activity of both peaks after transformation indicates that both contain the GLC7 gene product. Additional experiments are needed to identify the relationships between the two peaks of phosphatase activity and to determine the exact nature of the defect in the GLC7 gene. The phosphatase activity is altered in the glc7 mutant, but the precise nature of the glc7 mutation has not been determined. The GLC7 gene is required for cell viability, as shown in this study and recently by Clotet et ai. (1991), yet the glc7 mutant strains grow almost as well as wild-type strains. The g/c7 mutation may reduce the affinity of the catalytic subunit for a G-like subunit, specifically diminishing its role in regulating glycogen metabolism. Alternatively, the catalytic activity or expression level of the catalytic subunit may be reduced so as to affect glycogen metaholism but still be above a threshold for a deleterious effect on vital functions, including cell cycle regulation.
The nucleotide sequence of the GLC7 gene is virtually identical with that of the DIS2SI gene that was reported previously (Ohkura et al., 1989). The sole discrepancy in the coding region is at nucleotide 2209 in the GIX7 sequence, which is a G, where the DIS2S1 sequence has a T (position 1343). These would be translated to glycine in the GLC7 sequence and to valine in the IlIS2SI sequence. A glycine residue is present at this position in all other protein phosphatase 1 sequences reported (Sasaki et a/., 1990;Cohen et al., 1990;Shenolikar and Nairn, 1991). The nucleotide sequence reported here also significantly extends the previously reported 5' and 3' sequences. The first 204 bases of the sequence contain part of a A repeated element, which are found at the ends of TY1-transposable elements (Roeke et a/., 1988). This region is not required for GLC7 function, since the HindIII-XhoI fragment on pZF226 can complement the glc7 mutation (Fig. 3).
The complete sequences of many protein phosphatases have recently been derived from their cDNA sequences (Shenolikar and Nairn, 1991). Protein phosphatase 1 sequences from mammals, insects, and yeasts show a remarkable degree of conservation. Type 2A protein phosphatase sequences are also very highly conserved and show about 50% identity with t-ype 1 sequences (Orgad et al., 1990;. Few t.ype 2B sequences have been determined, and these show a significant but lower degree of similarity to the t-ype 1 and 2A proteins. Several type 2C sequences have been determined, and these show no sequence homology with the other classes of protein phosphatases. Two genes encoding type 1 protein phosphatases have been isolated from the fission yeast S. pomhe: DIS2 (= R W S I ) and SDS21 (Ohkura et al., 1989;Rooher and Beach, 1989). There are also duplicate genes encoding type 2A phosphatases in fission yeast (Kinoshita et al., 1990). The type 1 and type 2A protein phosphatases have essential but distinct roles in mitosis in fission yeast (Kinoshita et al., 1990). In the budding yeast S . cereuisiae, the protein phosphatase 2A gene SIT4 was first identified by its effect on transcription (Arndt et ~l . , 1989). Two different protein phosphatase 2A genes, PPH2I and PPH22, were isolated from S. cereuisiae by hybridization with the homologous rabbit cDNA (Sneddon et al., 1990). A single gene encoding a t-ype 1 protein phosphatase, DIS2.71, was isolated from S. cerwisiav using the fission yeast gene as a hybridization probe (Ohkura et a / . , 1989).
Glycogen accumulation occurs in yeast cultures as they approach stationary phase (Rothman-Denes and Cabib, 1970;Lillie and Pringle, 1980). At the same time, the percentage of glycogen synthase in the dephosphorylated form and the total amount of glycogen synthase is increased (Rothman-Denes and Cabib, 1970; Peng et a/., 1990). We have found that the GLC7 mRNA increases at the end of logarithmic growth, at the same time that glycogen synthase activity increases. An increase in protein phosphatase activity at this time may be responsible for the increase in the glycogen synthase activation. I t is also possible that the expression of the glycogen synthase and protein phosphatase genes are coregulated by the same metabolic signals to coordinate glycogen metabolism.