A Novel Gene, STT4, Encodes a Phosphatidylinositol 4-Kinase in the PKCl Protein Kinase Pathway of Saccharomyces cerevisia.e*

A staurosporine-sensitive mutation (sttl) in yeast has been found in the PKCl gene that encodes a protein kinase C homologue (Yoshida, S., Ikeda, E., Uno, I., and Mitsuzawa, H. (1992) Mol. Gen. Genet. 231, 337-344). We report here another staurosporine-sensitive mutant, stt4, which shows very similar phenotypes to that of the sttl mutant. The stt4 temperature-sensitive mutant ar- rests mostly in G2m phase at 37 "C, and the stt4 deletion mutant shows an osmoremedial phenotype. Staurospo- rine sensitivity of the stt4 mutant was suppressed by overexpression of PKClIS!l"l, indicating genetic inter- action between stt4 and pkcllsttl. The nucleotide sequence of STT4 predicts a hydrophilic protein composed of 1,900 amino acid residues, with 26% sequence identity to the yeast VPS34 gene product and 27% to the catalytic subunit of mammalian phosphatidylinositol (PI) 3-ki-nase, respectively. Cell homogenates of the sft4 deletion mutant show normal PIS-kinase activity but lack most of the PI4-kinase activity that is detected in the wild-type. We conclude that Sm4 encodes a yeast PI4-kinase that functions in the PKCl protein kinase pathway. Eukaryotic

A staurosporine-sensitive mutation (sttl) in yeast has been found in the PKCl gene that encodes a protein kinase C homologue (Yoshida, S., Ikeda, E., Uno, I., and Mitsuzawa, H. (1992) Mol. Gen. Genet. 231, 337-344). We report here another staurosporine-sensitive mutant, stt4, which shows very similar phenotypes to that of the sttl mutant. The stt4 temperature-sensitive mutant arrests mostly in G2m phase at 37 "C, and the stt4 deletion mutant shows an osmoremedial phenotype. Staurosporine sensitivity of the stt4 mutant was suppressed by overexpression of PKClIS!l"l, indicating genetic interaction between stt4 and p k c l l s t t l . The nucleotide sequence of STT4 predicts a hydrophilic protein composed of 1,900 amino acid residues, with 26% sequence identity to the yeast VPS34 gene product and 27% to the catalytic subunit of mammalian phosphatidylinositol (PI) 3-kinase, respectively. Cell homogenates of the sft4 deletion mutant show normal PIS-kinase activity but lack most of the PI4-kinase activity that is detected in the wild-type. We conclude that Sm4 encodes a yeast PI4-kinase that functions in the PKCl protein kinase pathway.
Eukaryotic cell proliferation is directed through many signal transduction systems responding to hormones, growth factors, and neurotransmitters. Receptor-mediated activation of inositol phospholipid metabolism is one of the typical signal transducing pathways (1,2). Phosphatidylinositol (PI)' kinase plays an important role as the first committed enzyme in this pathway. Historically, three forms of PI kinase (type I, 11, and 111) have been distinguished on the basis of inhibition by detergents, such as Triton X-100, and adenosine (3,4). The type I PI kinase phosphorylates the D-3 position of the inositol ring, whereas the type I1 and I11 enzymes phosphorylate the D-4 position (5). All three products of the type I PI3-kinase (PI-3P, PI-3,4P2, and PI-3,4,5P3) appear to be resistant to cleavage by phospholipase C (6,7).
In Saccharomyces cerevisiae, PI is the third major phospholipid in membranes (8) and is essential for cell growth (9,10). The synthesis of phosphatidylinositol 4,5-bisphosphate (PIP2) * 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  in yeast cells is regulated by glucose (11,12) and by sterol (13). PI kinase, phosphatidylinositol 4-phosphate (PIP) kinase and protein kinase C (PKC) have been characterized biochemically in yeast (14- 19), and the former two were shown to be regulated by the RASlcAMP cascade (20). A gene homologous to mammalian PKCa, PKCl, has been isolated in the yeast S. cerevisiae and was shown to be essential for cell division cycle with cells arresting mostly in the G2/M phase (21). VPS34, a gene whose product is homologous to the catalytic subunit of mammalian PI3-kinase (22), has been identified and shown to function in the vacuolar protein sorting pathway (23).
Biochemically, 45-kDa (p45) and 55-kDa (p55) isoforms of PI4-kinase have been purified from the membrane fraction of S. cerevisiae (14, 16). The two isoforms have different physicochemical and enzymological properties. Another isoform of PI4kinase (125 kDa) (~1 2 5 ) has been purified from the cytosol (17). On the other hand, PI3-kinase has not been purified, although its activity has been detected (15). In vitro assay of PI kinase showed that activity of PI4-kinase is approximately 5-fold higher than that of PI3-kinase, and in vivo labeling assay with [3H]inositol showed that PI-3P is as abundant as P1-4P, although PI-3P constitutes only 3-10% of the PIP in mammalian cells (15).
Staurosporine, isolated as an antifungal activity derived from Streptomyces sp., is one of the most specific inhibitors of PKC (24). I t inhibits PKC, probably by direct binding to the enzyme (25,26). To understand the physiological roles of the PKC pathway, we isolated staurosporine-sensitive mutants which were also temperature-sensitive (stt ). One of them, sttl, has been shown to be allelic to pkcl (27). Together with the finding that the PKCl gene confers staurosporine resistance to yeast cells in a dose-dependent manner (27), we have suggested that mutations in the PKCl regulatory pathway result in the staurosporine-sensitive phenotype.
In this paper, we report a potential PI4-kinase gene, STT4. Genetic interaction between pkcllsttl and sttl strongly suggests involvement of P14-kinase in regulation of PKC activity.
EXPERIMENTAL PROCEDURES Strains, Growth Conditions, and Dansformations-All strains used in this study were derivatives of wild-type strains YS3-3D or YS3-6D except those used for mapping of the location of STT4. Yeast cells were grown in YPD (1% yeast extract, 2% polypeptone, 2% glucose) (28). Synthetic minimal medium, SD, (28) supplemented with appropriate nutrients was used to select for plasmid maintenance and gene replacement. Transformation was performed using the lithium acetate method (29) and genetic manipulation was carried out as described (28). plasmids and phagemids. Phagemids BLUESCRIFTII (KS' and KS-) Bacterial strains XL1-Blue (30) were used for the propagation of all and the helper phage M13K07 (31) were used to generate singlestranded template DNA for sequence determination. Escherichia coli cells were cultured in Luria broth and transformed or infected with phagemids by standard methods (32).
The S. cerevisiae genomic DNA libraries were constructed on the shuttle vectors, YEpl3 (33) and YCp50 (34), respectively. DNA Manipulations-DNA was prepared from yeast strains by the method of Winston et al. (35). It was subjected to electrophoresis through agarose gels aRer digestion with restriction endonucleases and transferred to nitrocellulose filters by standard procedures (36). Plasmid DNA was prepared from E. coli using alkaline lysis (32) or from S. cereuisiae using glass beads (28). DNA sequence analysis was camed out by dideoxy chain termination (37) with a 370A DNA sequencer (Applied Biosystem) according to the supplier's instructions. Both strands of the 8.6 kb BamHIIPstI fragment were sequenced (Fig. 2 B ) .
STT4 Gene Integration, Replacement, and Plasmids-The 3.4-kb BamHIIHindIII fragment of plasmid pSTS4-14 was subcloned into the integrating vector pRS306 (38). This plasmid was digested with XhoI, and the linearized DNA was used to transform a wild-type strain RA1-1B. Restriction mapping and hybridization analysis of genomic DNA from the resulting transformants were carried out to confirm that integration had occurred a t the STT4 locus. One such transformant was mated with the stt4-1 mutant strain (SYT41-22C). The resulting diploid was subjected to sporulation and tetrad analysis.
A deletion mutant allele of STT4 was constructed by the method of Rothstein (39). The 4.1-kb XhoIISpeI fragment of STT4 ( Fig. 2.4) was replaced on pUC119 by the 1.3-kb XhoIIXbaI fragment of HIS3 from pJJ217 (40) in reverse transcriptional orientation. The resulting plasmid was digested with EcoRI to confirm that the 2.2-kb EcoRI fragment contained the entire s t t 4 3 I S 3 allele. This 2.2-kb EcoRI fragment was used to transform a wild-type diploid strain, YS3-6D-HO, by selection for histidine prototrophy. Restriction mapping and hybridization analyses of genomic DNA from the resulting transformants were carried out to confirm that transplacement had occurred a t the STT4 locus.
Mapping of the STT4 Gene-A strain, YP148, designed for electrophoretic separation of chromosomes (28) by orthogonal-field alternation gel electrophoresis (OFAGE) was used for chromosomal mapping (41). The 1.1-kb BamHYHindIII fragment of pES4-1, derived from the STT4 gene, was used as a probe for hybridization.
Assay of PI Kinase-Cell homogenates for assay of PI kinase activity were prepared as follows; yeast cells were grown in 50 ml of YPD medium containing 1 M sorbitol (1-2 x lo6 celldml) overnight a t 25 "C, pCi/tube), and enzyme protein (50 pg). The reaction was started by the addition of [32P]ATP, incubated for 10 min a t 30 "C, and then terminated by the addition of 2 ml of chlorofodmethanol (2:l. v/v). The mixture further received 500 pl of 1 N HCl and was extracted, vigorously vortexed, and centrifuged a t 1000 x g for 5 min at room temperature. The chloroform phase was washed once with synthetic upper phase.
Anti-Stt4p Antibody and Immunoblotting-The 1.4-kb BamHIISpeI fragment from pSTS4-14 was inserted into pUR288 (43) to make an in-frame gene fusion of lac2 and STT4. The resultant plasmid was introduced into the E. coli strain XL1-Blue and the transformant was induced by isopropyl-P-D-thiogalactopyranoside to express the l a d -STT4 fusion gene product. This hybrid protein was purified as described by Nakano et al. (44) and used to immunize rabbits.

RESULTS
Genetic Znteraction between pkcl lsttl and sttPTo identify mutations in the PKCl pathway, we focused on the stt mutations that were suppressed by overexpression of the PKClI STTl gene with respect to staurosporine sensitivity. Among nine stt mutations, the stt4 mutation was well suppressed (data not shown). The sttl stt4 double mutant was more sensitive to staurosporine than either the sttl or stt4 single mutant (Fig. 1). Further, like sttl mutants, all three stt4 mutant alleles isolated showed temperature-sensitive phenotype at 37 "C on YPD plates, but could grow not only on plates containing 1 M sorbitol but also on those containing 250 m M CaC12, MgC12, or inositol (27). However, temperature sensitivity of the stt4 mutants was not suppressed by the addition of 100 m sorbitol, CaC12, MgCl,, or inositol (data not shown).
Cloning and Nucleotide Sequence of the STT4 Gene-The STT4 gene was cloned by complementation of the stt4-1 mutation. The stt4-1 strain SYT44 was transformed with a yeast genomic library constructed on yEpl3 with the LEU2 marker (33) or with a yeast YCp50 genomic library with the URA3 marker (34). Among 20,000 transformants screened, two clones (pES4-1 and pES4-2) and one clone (pCS4-1) that complemented both temperature and staurosporine sensitivity of the stt4-1 mutation were obtained from the YEpl3 and the YCp50 libraries, respectively. The restriction map of the inserts in these clones showed that they contained an overlapping region ( Fig. 2A ). Subcloning and complementation analyses indicated that the complementing activity resided in the 8.6-kb BamHI1 PstI fragment (pSTS4-14, Fig. 2 A ) .
To confirm that this complementing ability is not due to extragenic suppression, integration mapping was carried out. The 3.4-kb BamHIIHindIII fragment was cloned from pSTS4-14 into the integrating vector pRS306 with the URA3 marker. The resulting plasmid was linearized at the internal unique XhoI site. A wild-type strain RA1-1B was transformed with this integrating plasmid. One of the Ura+ transformants, in which the integration of URA3 at the cloned gene locus was confirmed by Southern analysis (data not shown), was crossed with SYT41-22C carrying the stt4-1 mutation. In 19 dissected tetrads, both Ura' and Stt" phenotypes segregated 2:2 and always cosegregated with each other; 19 asci all showed parental ditype. From this result, we concluded that pSTS4-14 contained the authentic STT4 gene.
Mapping of the Chromosomal Location of S T T P T h e 1.1-kb BamHIIHindIII fragment of pSTS4-14 (STT4) was used for Southern hybridization of chromosomes from a strain of S. cerevisiae designed to allow electrophoretic resolution of all 16 chromosomal DNAs. The probe exclusively hybridized to chromosome XI1 (data not shown). To map STT4 more precisely on chromosome MI, standard meiotic linkage analyses were carried out using strains marked at the ASP5, URA4, CDC42, CDC25, and CDC3 loci. Tetrad analysis (Table I) revealed that the location of the STT4 gene was 3.9 centromere distal to CDC25. The DNA sequence of the 8.6-kb BamHUPstI fragment of pES4-1 showed that MET25 and ACOl genes were located  a2 0 aa proximal to the STT4 gene (Fig. 2 A ) . Disruption of the STT4 Gene-The STT4 gene was disrupted by using a single-step procedure (39). The 4.1-kb XhoIISpeI fragment was replaced with the HIS3 gene. The plasmid containing the disrupted STT4 gene was digested with EcoRI and introduced into a wild-type diploid strain YS3-6D-HO carrying multiple auxotrophic markers. His' transformants were isolated and confirmed for the disruption of one STT4 copy by Southern hybridization (data not shown). A heterozygous diploid (stt4:5ZZS3/STT4) was subjected to sporulation and tetrads were dissected. Only two spores in each tetrad gave rise to colonies on a YPD plate. All haploid colonies obtained were shown to be His-, indicating that deletion of the STT4 gene resulted in the loss of spore viability.
Next, the tetrad dissection was performed on a YPD plate containing 1 M sorbitol. Two spores gave rise to large colonies and two yielded small ones. Southern hybridization analysis and marker tests showed that all small colonies were disrupted for STT4. Thus, the STT4 gene is not essential for growth on YPD plates containing 1 M sorbitol. The doubling time of the stt4 deletion mutant was 300 min, which was much longer than that of wild-type (120 min) in YPD medium containing 1 M sorbitol at 25 "C. However, the doubling time of the stt4-1 mutant (110 min) was only a little longer than that of wild-type (100 min) in the YPD medium at 25 "C. This osmolarity-dependent phenotype is very similar to that of the pkcl disrupted strain (48).
STT4 Encodes a Protein Homologous to Vps34p and the Catalytic Subunit ofMammaEian PI3-Kinase, pll&In order to gain insight into possible functions of the STT4 gene product, a homology search of the Stt4p amino acid sequence in the protein database (NBRF) was performed. This search showed that the Stt4p sequence had noticeable similarity to Vps34p, a hydrophilic protein of 875 amino acids from S. cereuisiae (231, and p110, the catalytic subunit of mammalian PI3-kinase consisting of 1,068 amino acids (22) (Fig. 4). The amino acid sequence identity with Vps34p and pll0 is approximately 26 and 27% in 387 and 470 overlapping amino acid sequences, respectively.
Vps34p is required for vacuolar protein sorting and vacuole segregation (23). The ups34 mutant cells form pink colonies with the genetic background of the ade2 mutation because they cannot accumulate red pigment in the vacuoles (49). However, the stt4 mutant cells harboring the d e 2 mutation formed red colonies, suggesting that stt4 mutant is not defective in vacuolar functions.
Assay of PI Kinase ActivityAs Stt4p has homology to the catalytic subunit of mammalian PIS-kinase, the possibility was examined that Stt4p is one of the yeast PI kinases. PI kinase activity of yeast cell homogenates was assayed as described under "Experimental Procedures." In the stt4 null mutant, P I 4 P production was greatly reduced relative to that of wildtype, although PI-3P production was equivalent to that of wildtype (Fig. 5).  When expressed on a multicopy plasmid, STT4 had no effect on the PI kinase activity (Fig. 6A ). Western blot analysis using an anti-Stt4p antibody showed that there was no difference in amount of Stt4p between wild-type cells and the stt4 null mutant cells harboring an STT4 multicopy plasmid (stt4:flZS3/ pSTI'4). Rather, overexpression of STT4 appeared to cause degradation of excess Stt4p, since lower molecular weight crossreacting material was detected in these extracts (Fig. 6B). DISCUSSION We have isolated the STT4 gene by complementation of the temperature-sensitive growth of the staurosporine-sensitive stt4 mutant. The STT4 gene product predicts a hydrophilic protein of 1,900 amino acids. The C-terminal region of Stt4p has homology to the C-terminal regions of the yeast Vps34p (26% identity in 387 amino acids) and the catalytic subunit of Wild-type cells and the stt4 deletion mutant cella harboring an STT4 multicopy plasmid (stt4:.JIIS3/pSl"4) were incubated in YPD medium a t 25 "C overnight. A, separation of PIP by borate thin layer chromatography. P14-kinase activities of wild-type and the stt4 null mutant cells harboring the STT4 plasmid were 2.5 k 1.0 x 10" and 2.0 f 1.1 x 10" pmol/midmg protein, respectively, in three independent experiments. B, Western blot analysis of Stt4p. Their homogenate samples for PI kinase assay (50 pg of protein) were subjected to SDS-polyacrylamide gel electrophoresis (7.5% (w/v) gel). Stt4p was detected using an anti-Stt4p polyclonal antibody. mammalian P13-kinase (p110) (27% identity in 470 amino acids). P14-kinase activity of the stt4 deletion mutant cells is about 15% of wild-type cells (Fig. 5). Therefore, Stt4p is likely one of several yeast P14-kinases. Furthermore, previous biochemical analyses have shown that multiple PI kinases are present in yeast, i.e. p45, p55, and p125 P14-kinases and PI3kinase (14-17). Thus, it is not surprising that the sttl null mutant cells have residual PI4-kinase activity.
Vps34p is a hydrophilic protein of 875 amino acids and is required for vacuolar protein sorting and vacuole segregation (23). PIS-kinase activity is not detected in yeast cells depleted for the VPS34 gene (50). This observation excludes the possibility that the Stt4p PI kinase has some residual PIS-kinase activity. Overexpression of STT4 had no effect on PICkinase activity (Fig. 6A). It is probable that an other factods) is required for the stabilization or activation of Stt4p PI4-kinase like mammalian PI3-kinase, a complex of the p85 regulatory subunit and the p l l 0 catalytic subunit (22).
STT4 is essential for cell growth in YPD medium but is dispensable for growth in the presence of an osmotic stabilizer. Overexpression of PKClISTTl in the stt4-1 mutants suppresses the staurosporine sensitivity, but not the temperature sensitivity (data not shown). This observation suggests that Stt4p either directly interacts with the Pkcl protein kinase or is functionally involved in the PKCl pathway. On the other hand, overexpression of VPS34 partially suppresses both the growth defects and vacuolar protein missorting defects observed with Vpsl5 protein kinase-deficient mutants (51). The disruption of the VPS34 gene results in a temperature-sensitive growth defect. We have observed that the sttl mutants are not defective in vacuolar morphology, suggesting PKClISTTl-STT4 and VPS15-VPS34 signal transduction systems are operating quite independently. The existence of these two signal transduction pathways in yeast provides the interesting common feature of intimate PI kinase interaction with another protein kinase.