Low Molecular Weight Protein-tyrosine Phosphatases Are Highly Conserved between Fission Yeast and Man*

Cdc25 protein phosphatase dephosphorylates tyrosine 15 of Cdc2, thereby activating CdcWcyclin B kinase, which then brings about mitosis. A fission yeast (Schiz-osaccharomyces pombe) cDNA expression library was screened for clones that rescue cdc25-22. In addition to the cdc25+ and pyp3' protein-tyrosine phosphatase genes, a third gene was discovered. This gene, named stpl' (small mosine phosphatase), encodes a -17.6-kDa protein that is -42% identical to members of an unusual class of small (-18 kDa) cytosolic phosphatases previ- ously known to exist only in mammalian species. "he biological functions of these proteins are unknown, but they have vigorous protein-tyrosine phosphatase activity in vitro and have a sequence motif, Cys-X,-Arg, that is present at the active sites of all known types of protein-tyrosine phosphatases. Sequence homology between s. pombe Stpl and high in the active site region the Rescue of cdc25-22 by overproduction of Stpl protein to an ability of Stpl to dephosphorylate tyrosine 15 of Cdc2. of stpl' no obvious type. The that

* This work was supported by grants from the National Institutes of Health (to P. R.) and Imperial Cancer Research Fund (to S. M.). 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankrM/EMBL Data Bank with accession numbeds) L33929. I/ To whom correspondence and reprint requests should be addressed: Tel.: 619-554-8273; Fax: 619-554-6165. cue of temperature sensitive mutations of cdc25  by gene overexpression has proven to be a productive approach.
The niml+ gene, encoding a protein kinase that functions as a mitotic inducer by phosphorylating and inactivating Weel, was discovered as a multicopy suppressor of cdc25-22 (14-18). The same screen also led to the discovery of pyp3', a gene encoding a protein-tyrosine phosphatase that is able to dephosphorylate tyrosine 15 of Cdc2 (19). Inactivation of Cdc25 is rescued by inactivation of Weel, but this is dependent on the ability of Pyp3 to counteract Mikl. In this report, we describe the discovery of a third gene that rescues cdc25-22 when overexpressed. This gene, stpl+, encodes a protein that is highly similar to members of a family of mammalian low M , proteintyrosine phosphatases. The biological functions of mammalian low M, protein-tyrosine phosphatases are unknown, but the discovery of a fission yeast homolog suggests that they are likely to be important.

EXPERIMENTAL PROCEDURES
Isolation of S. pombe cDNA Clones That Rescue cdc25-22-A S. pombe cDNA library constructed in pREP3Xho was kindly provided by Drs. Bruce Edgar and Chris Norbury. The vector pREP3Xho is derived from pREP3, in which expression of inserts is under the control of the thiamine-repressible nmtl' promoter (20). This library was transformed into S. pombe strain PRlOll (cdc25-22 leu132 ura4-Dl8 h' pWEE1-10 integrated) in which a copy of pWEE1-10 had been integrated at the weel' locus. Plasmid pWEE1-10 contains weel ' and ura4' (3). Addition of an extra copy of weel' eliminates the background of spontaneous loss of function mutations of weel' that suppress cdc25-22. Plasmids were recovered from 21 transformants that exhibited co-segregation of the ability to grow at 36 "C and on media lacking leucine (the Saccharomyces cereuisiae LEU2 gene in pREP3Xho rescues leu1-32). Eight of these transformants continued cell division at 36 "C in media with or without thiamine, the remaining 13 transformants continued cell division at 36 "C only in media lacking thiamine. Plasmids were analyzed by restriction endonuclease mapping and identified either as cDNA clones of cdc25+, pyp3+, or stpl'. None of the plasmids contained a cDNAclone ofniml'. AO.9-kb' BamHI-SalI DNAfragment containing stpl' cDNA and flanking vector sequences was subcloned from pREP3Xho into pBluescript SK and entirely sequenced on both DNA strands. The 5'-end of the cDNA was proceeded by a -40-nucleotide poly-CT tract. We have observed similar poly-CT tracts in other clones from this cDNA library, we presume that it was manufactured during the cDNA cloning process.
Disruption of stp1'-A 9-kb DNA fragment containing stpl' was isolated from a genomic library of S. pombe DNAmade in pFL2O. The stpl' cDNA in a 0.9-kb BamHI-Sal1 DNA fragment was as a probe for colony hybridization. A 2.1-kb EcoRI-XhoI fragment containing stpl' was subcloned into pTZ19. A -1.6-kb Sal1 fragment containing ural', derived from pCRII:ural+:SalI, was inserted into the XhoI site located in the EcoRI and XhoI and the DNA used to transform a diploid strain made stpl' open reading frame. The resulting plasmid was digested with by mating FWP172 (leu132 ural-Dl8 ade6-M210 h') to FWP165 (leu132 ura4-Dl8 ade6-MZ16 h-). Two stable uracil prototrophic trans-The abbreviations used are: kb, kilobaseb); PCR, polymerase chain reaction.

27996
Fission Yeast Low M , Protein-tyrosine Phosphatase formants were analyzed by Southern DNAhybridization to confirm that stpl::uru4+ had replaced one stpl+ locus. Random spore analysis produced uracil prototrophic and auxotrophic colonies a t a ratio of -1:l. PCR was used to confirm the stpl::uru4+ genotype of the uracil prototrophs. The primers were: STP1, 5'"lTYGTNTGYYTNGGNAAYAT-(A/C/T)TG-3' and STP3, 5'-CGRAA(A/GPT)ATRTARTCRAA-3', where R = A + G, Y = T + C, and N = A + G + C + T. STPl primer encodes the protein sequence FVCLGNIC (amino acids 9-16 of Stpl protein), whereas the complement of STP3 primer encodes the protein sequence FDYIFA (amino acids 85-90 of Stpl). These primers flank theXhoI site used to insert ura4'. PCR amplification generated the expected 0.25-kb fragment from stpl' haploids and -1.9-kb DNA fragment from stp1::urul' haploids (data not shown). Degenerate primers corresponding to sequences that are highly conserved amongst low M , proteintyrosine phosphatases were employed in an attempt to amplify additional homologs from fission yeast. However, PCR performed at an annealing temperature of 45 "C failed to amplify any sequences in the 0.25 kb range from stp1::urud' haploids (data not shown).

RESULTS
Isolation of S. pombe cDNA Clones That Rescue cdc25-22-An expression library containing S. pombe cDNA inserts was screened for plasmids that rescue cdc25-22. The library was constructed in the vector pREPSXho, in which the expression of cDNA inserts is under the control of the thiamine-repressible nmtl' promoter (20, 21). There is a low level of expression from the nmtl' promoter when cells are grown in media containing thiamine, therefore we screened for rescue of cdc25-22 at 36 "C both in the presence and absence of thiamine.
All transformants isolated in the screen performed in the presence of thiamine (i.e. nmtl promoter repressed) grew well at 36 "C ( Fig. 1). Cells typically had a wee phenotype, dividing at half the size of wild type. Plasmids were recovered from eight colonies of this class. Restriction enzyme mapping revealed that they all contained cDNA clones of cdc25+ (12). The size of the insert ranged from -2.4 to 3.1 kb, having -0.4 to 1.1 kb of 5"untranslated sequence (Fig. 2). These clones would be expected to express full-length Cdc25 protein. Interestingly, when grown in the absence of thiamine (ie. nmtl promoter derepressed), these transformants formed very small colonies (Fig.  1). These colonies contained a mixture of elongated cells that had a cell division cycle (cdc) arrest phenotype, and very wee cells that appeared to be undergoing mitotic catastrophe. These observations suggest that very strong overexpression of cdc25' is toxic.
All other transformants grew at 36 "C only in the absence of thiamine (Fig. 1). Plasmids were recovered from 13 of these transformants. Restriction enzyme mapping revealed that the cDNA inserts were derived from three genes. Four transformants had plasmids containing a truncated cDNA of cdc25' (Fig. 2). These contracts would be expected to express the Cterminal -25-kDa catalytic domain of Cdc25. Previous studies have shown that expression of this region of Cdc25 protein is sufficient to rescue cdc25-22 (11). Seven of the remaining nine plasmids contained pyp3+ (Fig. 2). All of the pyp3' clones were long enough to encode full length Pyp3 protein (19). These plasmids rescued cdc25-22 quite effectively, in agreement with previous studies.
The two remaining plasmids contained the same -0.8-kb insert (Fig. 2). DNA sequence analysis indicated that these clones contained one open reading frame encoding a 156-amino acid protein with a predicted M, of 17,371 Da (Fig. 3 A ) . This gene was named stpl'.
The stpl' Gene Encodes a Low M, Protein Drosine Phosphatase Homolog That Is Highly Conserved between Fission Yeast and Man-We expected that this screen might possibly identify additional protein-tyrosine phosphatases that rescued cdc25-22 by virtue of their ability to dephosphorylate tyrosine 15 of Cdc2. Initial inspection of S t p l sequence did not suggest containing full-length cDNA clones of cdc25', pyp3+, or stpl'. Cell were grown on selective media (E") with or without thiamine. Cell patches were grown for 1 day a t permissive temperature (25 "C), and then cells were streaked out and grown either at 25 "C or a t the restrictive temperature of 36 "C. All four transformant strains grew equally well in media containing thiamine a t 25 "C. Derepression of the nmtl promoter caused pREP3XhoI-cdc25' transformants to grow more slowly a t 25 and 36 "C. Only transformants expressing cdc25' were able to consistently form colonies a t 36 "C in media containing thiamine.
that it was likely to be a homolog of Cdc25 or Pyp3. However, we did notice that Stpl had the sequence CX,R (residues 11-17) that is absolutely conserved amongst all known tyrosine and dual-specificity protein phosphatases (22,23). CX,R forms part of the active site; several mutagenesis studies have shown that both the cysteine and arginine residues are essential for catalytic function. A protein sequence alignment program detected a low but apparently significant degree of homology between Stpl and the catalytic domain of Cdc25. Comparison of the entire Stpl protein to a portion of the catalytic domain of Cdc25 produced an alignment with -15% identity (Fig. 3B). Importantly, the two proteins were most alike in the region of the putative active site, being -27% identical in a 30-amino acid stretch encompassing the CX,R motif.
A sequence homology search of a protein sequence data base identified a human protein that was highly similar Stpl, being one amino acid longer and 42% identical (24). The greatest similarity is in the active site region. The two proteins are -67% identical in the 30-amino acid stretch mentioned above (Fig. 3C). The human protein is a low molecular weight (-18 kDa) cytosolic phosphatase. This class of proteins have vigorous in vitro hydrolytic activity on phosphotyrosine-containing protein or peptide substrates but poor activity on Ser/Thr-phosphorylated proteins (25-27). Several names have been given to this enzyme, it is now generally referred to as low M, protein- tyrosine phosphatase. The gene name stpl' was derived from small mosine Ehosphatase.

Fission Yeast Low M, Protein-tyrosine Phosphatase
Disruption of stp1'-We disrupted the genomic stpl+ locus in order to determine the phenotype caused by loss of S t p l activity. The ura4+ marker was inserted into the XhoI site of a cloned copy of stpl' (Fig. 4.4). The XhoI site is located at the 34th codon of the stpl' open reading frame, immediately downstream of the active site (Fig. 4A). The stpl::ura4+ construct was used to replace genomic stpl+ in one chromosome of a diploid strain having uru4-D18/ura4-D18 (Fig. 4B). Random spore analysis produced equal numbers of uracil prototrophs and auxotrophs, indicating that stpl' is not normally essential for viability. Southern analysis confirmed that the uracil prototrophs indeed had a copy of stpl::ura4+ substituted for stpl+, this was also confirmed by PCR. The Asptl cells appeared wild type as judged by colony size and cell length when grown on YES (yeast extracffglucose) and EMM (defined minimal) plates. Growth at temperature extremes (20 and 37 "C) and high osmotic media (1 M sorbitol) was unaffected by Asptl. The abilities of cells lacking S t p l to conjugate, sporulate, and germinate also appeared to be unimpaired.
We also examined the effect of Asptl in strains having defects in Pyp3 or Cdc25 activity. As shown in Table I, Asptl had no effect in cdc25-22 or Apyp3 or backgrounds. Disruption of Asptl also had no effect in cdc25-22 Apyp3 double mutants. These data indicate that at normal levels of expression, Stpl has very little impact on the activation of Cdc2.
Mapping of stp1'"Hybridization of an stpl probe to filters containing P1 and cosmid libraries of S. pombe genomic DNA revealed that stpl' is located near the centromere chromosome 1, very close to the genetic marker pmal Maier et al., 1992;Hoheisel et al., 1993). DISCUSSION The aim of this investigation was to identify additional elements of the mitotic control in fission yeast. Earlier studies had identified a protein kinase gene, niml+, and a protein-tyrosine phosphatase gene, pyp3+, as high copy suppressors of cdc25-22. Disruption of niml+ in a wild type background leads to a significant delay in the onset of mitosis (17). This result, coupled with the demonstration that Niml specifically inactivates Wee1 kinase, demonstrate that Niml has an important role in the mitotic control (14,16,18 4. Disruption of sfpl+. A, a -1.9-kb DNA fragment containing S. pombe uru4' was inserted into theXhoI site located in the stpl' open reading frame. The stpl::uru4+ DNA was used to transform a uru4-D18/ uru4-Dl8 ude6-M210/ude6-M216 diploid strain, selecting for stable uracil prototrophs. B, Southern analysis confirmed that the diploid strain had both the wild type 2.5-kb EcoRI-BumHI stpl' fragment and the 4.4-kb EcoRI-BumHI stpl::uru4+ fragment. PCR analysis of viable stpl' and stpl::uru4+ progeny also confirmed the gene disruption (data not shown).
Disruption ofstpl' does not affect cell size a t division TARIX I Cells were in YES media a t 30 "C to mid log phase. slight impact on the mitotic control in a wild type background, but loss of Pyp3 activity results in a dramatic mitotic delay in cells that are partially compromised for Cdc25 function. It is likely that Pyp3 normally carries out a small amount of Cdc2 tyrosyl dephosphorylation in vivo, although Cdc25 is perhaps not its primary target in the cell. As described in this report, we have carried this screen one step further and have identified a third gene, s t p l + , that rescues cdc25-22 when overexpressed. S t p l protein is very similar to an unusual class of mammalian tyrosine phosphatase, thus it is likely that Stpl has some ability to dephosphorylate tyrosine 15 of Cdc2. Disruption of stpl' has no detectable impact on the mitotic control, even in a cdc25-22 Apyp3 background. Thus, there is no compelling reason to believe that Stpl normally has any role to play in the mitotic control.
Low M , protein-tyrosine phosphatases were hitherto known only to exist in mammalian species. The human and fission yeast enzymes are remarkably similar, they differ in length by only one amino acid and are -42% identical. This structural similarity suggests that the two proteins are functional analogs, although this remains to be proven. Site-directed mutagenesis and kinetic studies have proven that the Cys and Arg residues in the CXsR motif are essential for the catalytic activity of bovine low M , protein-tyrosine phosphatase (26). The in vivo function mammalian low M , protein-tyrosine phosphatase is unknown; indeed it remains to be proven whether these enzymes function as protein-tyrosine phosphatases in vivo. The fact that overexpression of fission yeast Stpl rescues the cdc25-22 protein-tyrosine phosphatase mutation suggests that they do.
Disruption of s t p l + caused no obvious phenotype. The site of the insertion separated the active site domain from the rest of the protein, thus it is very likely that this disruption is a null mutation. It is typical to find that protein phosphatases in fission yeast are encoded by multigene families having redundant functions (28, 29); it is likely that low M , protein-tyrosine phosphatases are not an exception. The finding that low M , protein-tyrosine phosphatases are highly similar in structure between fission yeast and mammalian species suggests these enzymes have important functions. The fortuitous discovery of the stpl' gene in fission yeast presents the opportunity to undertake a genetic approach toward this problem. The major role that protein tyrosine phosphorylation plays in regulating cell division and proliferation would justify such an investigation.