Regulation of the Yeast Transcriptional Factor PHO2 Activity by Phosphorylation*

The induction of yeast Saccharomyces cerevisiae gene PHO5 expression is mediated by transcriptional factors PHO2 and PHO4. PHO4 protein has been reported to be phosphorylated and inactivated by a cyclin-CDK (cyclin-dependent kinase) complex, PHO80-PHO85. We report here that PHO2 can also be phosphorylated. A Ser-230 to Ala mutation in the consensus sequence (SPIK) recognized by cdc2/CDC28-related kinase in PHO2 protein led to complete loss of its ability to activate the transcription of PHO5 gene. Further investigation showed that the Pro-231 to Ser mutation inactivated PHO2 protein as well, whereas the Ser-230 to Asp mutation did not affect PHO2 activity. Since the PHO2 Asp-230 mutant mimics Ser-230-phosphorylated PHO2, we postulate that only phosphorylated PHO2 protein could activate the transcription of PHO5 gene. Two hybrid assays showed that yeast CDC28 could interact with PHO2. CDC28 immunoprecipitate derived from the YPH499 strain grown under low phosphate conditions phosphorylated GST-PHO2 in vitro. A phosphate switch regulates the transcriptional activation activity of PHO2, and mutations of the (SPIK) site affect the transcriptional activation activity of PHO2 and the interaction between PHO2 and PHO4. BIAcore® analysis indicated that the negative charge in residue 230 of PHO2 was sufficient to help PHO2 interact with PHO4 in vitro.

activates the transcriptional factor PHO4 (11). When the concentration of phosphate in the medium is sufficiently low, PHO81 protein, which is a cyclin-dependent kinase inhibitor, inhibits the kinase activity of the PHO80⅐PHO85 complex (12), thus allowing the hypophosphorylated PHO4 protein, together with PHO2 protein, to activate the transcription of PHO5 gene.
PHO2 has four conspicuous regions (see Fig. 1). The first one is a glutamine-rich region (amino acids 23-52), the deletion of which was showed to have no effect on PHO2 function (13). The other three regions include the homeodomain (amino acids 77-136) (14), the acidic region (amino acids 294 -329), and the region (amino acids 241-258) with homology to PHO80 protein.
Homeo boxes were found in many genes of higher eukaryotes involved in development (e.g. Drosophila, mouse and human). The PHO2 homeodomain is directly involved in DNA recognition. The acidic region of PHO2 contains continuous stretches of aspartic acid and asparagine. The PHO80 homologous region may be an association domain of PHO2 with PHO4, the deletion of which inactivated the PHO2 protein (13). The deletions at the C-terminal end of PHO2 protein may go up to residue 408 without strongly affecting the derepression of PHO5 (15).
The results of the two-hybrid assay (16) argued that DNA binding by PHO4 is dependent on the phosphate-sensitive interaction with PHO2. It was also suggested that interaction with PHO2 increases the accessibility of the activation domain of PHO4 (17). Recently, immunoprecipitation experiments and protein binding assays showed that PHO2 and PHO4 form a complex with a DNA fragment and interacted with each other directly in vivo (18).
Protein phosphorylation has been found to play an important role in the control of diverse cellular processes, especially that of transcriptional factors in the regulation of gene expression. PHO80-PHO85 cyclin-CDK (CDK, cyclin-dependent kinase) complex can regulate the expression of PHO5 gene by phosphorylating PHO4 (11). Cyclic AMP-dependent protein kinase (protein kinase A) has also been observed to exert its function in the synthesis of repressible acid phosphatase (19). As well, cdc2/CDC28 type kinase can affect gene expression by phosphorylation. The consensus motif of sites phosphorylated by cdc2/CDC28-type kinase is (Ser/Thr)-Pro-X-(Lys/Arg) (20). This sequence is found more frequently in nuclear proteins involved in transcriptional regulation than in proteins generally. Many of these proteins also contain DNA binding motif such as zinc fingers or helix-turn-helix structures (21)(22)(23)(24).
In this report we describe the identification of a potential phosphorylation site near the PHO80 homologous region of PHO2 protein. The mutation of the site can affect the function of PHO2 protein. Moreover, the introduction of a negative charge to this residue seems to be necessary for PHO2 functions. We also demonstrate that PHO2 can be phosphorylated by CDC28 kinase in vitro.
Yeast culture were grown at 30°C in either YPD (1% yeast extract, 2% glucose, 2% peptone) or synthetic medium (0.67% yeast nitrogen base, 2% glucose) supplemented with the appropriate amino acids as required. All yeast transformations were done using the high efficiency lithium acetate method of Gietz and Schiestl (24). For the experiments involving high or low phosphate, Burkholder medium (25) was used, but the content of KH 2 PO 4 was changed as described previously (26).
Polymerase chain reaction was used with yeast genomic DNA to amplify a 0.9-kb fragment including the entire coding region of CDC28 gene. The 5Ј oligonucleotide primer was ATCC GGAT CCTG ATGA GCGG TGAA TTAG CA, and the 3Ј oligonucleotide primer was GGAG CTGC AGTT ATGA TTCT TGGA AGTA GG. Polymerase chain reaction was also used with yeast genomic DNA to tag the CDC28 protein by adding a C-terminal extension of GAYPYDVPDYASLG, which includes a hemagglutinin antigen (HA) epitope. The same 5Ј oligonucleotide primer was used in conjunction with the 3Ј oligonucleotide primer: AGTC TGCA GTCA TCCC AAGC TAGC GTAG TCAG GAAC GTCA TATG GATA GGCG CCTG ATTC TTGG AAGT AGGG GTGG. Amplified fragments were digested with BamHI and PstI, which cleaved within the primers and, respectively, cloned into the BamHI/PstI sites of pGBT9 and the yeast expression vector pVTU102 (27) to create plasmids pGBT9-CDC28 and pVTU102-CDC28HA. The absence of polymerase chain reaction-introduced mutations was verified by DNA sequencing.
To construct plasmid pBL-PHO4, a 1.0-kb NcoI/HindIII fragment was prepared from PHO4 gene and inserted into the NcoI/HindIII sites of plasmid pBL. This plasmid produces the entire PHO4 protein (312 amino acids).
Site-directed Mutagenesis of the PHO2 Gene-Five synthetic oligonucleotides were used for site-directed mutagenesis with a U-DNA mutagenesis kit (Roche Molecular Biochemicals): 5Ј-ATCAGATCTGC-TATTTTCTTT-3Ј; 5Ј-GAGAAACCTAGCGCCAATAAA-3Ј; 5Ј-GTTGAG-AAACCTAGACCCAATAAAGAT-3Ј; 5Ј-AAACCTATCGTCAATAAAGA-T-3Ј; 5Ј-ATCGCCAATACAGATTAATAAC-3Ј. A 1.0-kb PvuI/HindII fragment of the PHO2 gene was cloned into the M13mp18 vector to produce single-stranded DNA. The mutagenesis procedures were then conducted according to the kit protocols. The mutants were isolated and sequenced to ensure that the intended base changes were present. The fragment from the replication form of M13mp18, containing the designed mutation, was then cloned back to pRS2.
Activity Assay of Acid Phosphatase-The strains to be tested were grown in 3 ml of synthetic medium lacking Leu overnight. The cells were harvested, washed twice, and used to inoculate into 3 ml of Burkholder high or low phosphate medium to an A 600 of 0.05. Incubation was carried out at 30°C for 16 h with shaking. The cells were harvested, washed with 0.05 M acetate buffer (pH 4.0), and resuspended in 1 ml of 0.05 M acetate buffer. Acid phosphatase activity was assayed according to the method described previously (16,26).
Preparation of Cell Extracts-Cell cultures were first grown to saturation in synthetic medium lacking uracil at 30°C. The cells were harvested by centrifugation, washed twice with sterilized water, and then incubated into the Burkholder low phosphate medium at an A 600 of 0.05. These cultures were grown at 30°C with shaking for 8 to 16 h to a final A 600 of 0.5-1.5 before analysis. 50 ml of these cultures were washed with 10 ml of sterilized water. All subsequent steps were carried out at 4°C. Cells were resuspended in 450 l of HSB buffer (45 mM HEPES-KOH (pH 7.5), 400 mM NaCl, 10% glycerol, 1 mM EDTA, 0.5% Nonidet P-40, 2 mM dithiothreitol, 2 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml pepstatin A, 2 g/ml leupeptin, 4 g/ml antipain, 10 mM NaF, 80 mM ␤Ϫglycerol phosphate), and 0.8 mg of acid-washed glass beads (425-600 m, Sigma) were added. Cells were lysed on a Vortex (Vortex-Genie2) for 30 s at the maximum setting, followed by cooling in an ice bath for 30 s. This procedure was repeated 10 times. Lysates were clarified by centrifugation at 4°C for 15 min at 14,000 rpm twice.
Expression in Escherichia coli and Purification of GST-PHO2 Fusion Protein and PHO4 Protein-Purification of recombinant GST-PHO2 protein and its mutants was performed essentially as described (28). BL21(DE3)-plysS E. coli cells (29) harboring the appropriate expression vector were grown in 50 ml of LB containing ampicillin (50 g/ml) overnight at 37°C. This culture was diluted into 500 ml of LB containing ampicillin (50 g/ml) and grown at 30°C until the A 600 was approximately 0.5. Isopropyl-␤-D-thiogalactopyranoside (Roche Molecular Biochemicals) was added to a final concentration of 1 mM, and the incubation was continued for 3 h at 30°C. The following procedures were performed at 4°C with ice-cold buffers. After the cells were centrifuged, the cell pellet was washed with 20 ml of phosphate-buffered saline (PBS) (150 mM NaCl, 16 mM Na 2 HPO 4 , 4 mM NaH 2 PO 4 , 10% glycerol (pH 7.3)). Cells were resuspended in 10 ml of PBS ϩ1% Triton X-100. Cell suspension was frozen in liquid nitrogen for 15 min and then frozen in Ϫ70°C for 1 h. Frozen cells were thawed as quickly as possible at room temperature. Cell lysate was clarified by centrifugation at 14,000 rpm for 15 min 3 times. Then the cell lysate was loaded on a 1-ml Glutathione-Sepharose 4B (Amersham Pharmacia Biotech) column equilibrated in PBS ϩ 1% Triton X-100. After loading, the column was washed with 2ϫ 5 ml of PBS ϩ 1% Triton X-100, then with 3ϫ 5 ml of PBS. The bound fusion protein was eluted with 5 ml of elution buffer (10 mM glutathione in 50 mM Tris⅐HCl (pH 8.0), and 0.5-ml fractions were collected and analyzed by SDS-polyacrylamide gel electrophoresis. BL21(DE3)-plysS cells harboring the PHO4 expression vector pBL-PHO4 were grown in 4 ml of LB containing ampicillin (50 g/ml) overnight at 30°C. The culture was diluted into 400 ml of LB containing ampicillin (50 g/ml) and grown to A 600 of 0.5 at 30°C, then induced for 3 h at 42°C. The following procedures were performed as described (11).
In Vitro Phospho-labeling Assays-For kinase reaction, 1 l of YPH499 cell extract was added to a 19 l of kinase mixture to form a reaction mixture (20 mM Tris⅐HCl (pH 7.5), 10 mM MgCl 2 , 4 g GST-PHO2 (wild type or Pro-231 to Ser mutant), 50 M ATP, 10 Ci of [␥-32 P] ATP (Amersham Pharmacia Biotech)). The reaction mixtures were incubated at 30°C for 20 min. The reaction was stopped by the addition of 20 l 2ϫ SDS sample buffer (30) and heating to 95°C for 5 min. Denatured samples were electrophoresed on SDS, 10% polyacrylamide gel, vacuum dried, and autoradiographed.
␤ϪGalactosidase Filter Assays-The yeast two-hybrid system was detailed in MatchMaker two-hybrid system protocol (CLONTECH (PT1265-1)). pGBT9-CDC28 and pGAD424-PHO2 as well as combinations of vectors and the two-hybrid constructions were cotransformed into SFY526. Transformants were grown on synthetic medium agar plates lacking Trp and Leu at 30°C for 2-4 days. Some clones were spotted onto nitrocellulose filters placed on synthetic medium (-Trp, -Leu) agar plate. The plate was then placed at 30°C for 2 days. The filter was subsequently removed, snap-frozen on an aluminum foil case in a liquid nitrogen bath, and placed on Whatman 3MM paper saturated with Z buffer (31) containing 0.33 mg/ml X-gal (5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside). The filter was incubated at 30°C for the appearance of blue clones.
Immunoprecipitation and Protein Kinase Assay-Yeast cell extracts (in any given experiment all samples were normalized to contain the same amount of total protein (32)) were immunoprecipitated with anti-HA monoclonal antibody 12CA5 (Roche Molecular Biochemicals) (2 g in each reaction) for 1 h on ice. After centrifugation at 12,000 rpm for 2 min, the supernatant was added to 30 l of slurry of protein A-agarose (Roche Molecular Biochemicals) equilibrated in HSB buffer followed by a 2-h rotation at 4°C. Immunoprecipitates were washed with HSB buffer three times. Histone H1 kinase assays were performed as described previously (33). For phosphorylation of GST-PHO2, immunoprecipitates were washed twice with kinase assay buffer (20 mM Tris⅐HCl (pH 7.5), 10 mM MgCl 2 ) and resuspended in 20 l of kinase assay mixture (20 mM Tris⅐HCl (pH 7.5), 10 mM MgCl 2 , 4 g of GST-PHO2, 50 M ATP, 10 Ci of [␥-32 P]ATP (Amersham Pharmacia Biotech)). These kinase reaction mixtures were incubated at 30°C for 20 min. Reactions were stopped by addition of 20 l of 2ϫ SDS sample buffer and heating to 95°C for 5 min. Denatured samples were electrophoresed on SDS, 10% polyacrylamide gel, vacuum-dried, and autoradiographed.
Liquid Culture ␤-Galactosidase Assay-Cell cultures were first grown to saturation in selective synthetic medium at 30°C. The cells were harvested and washed twice with sterilized water and then incubated into 3 ml of Burkholder low (or high) phosphate medium at an A 600 of 0.2. Cultures were grown at 30°C with shaking to a final A 600 of 1.5. The cells were harvested and washed twice with Z buffer. Cells were resuspended in 0.6 ml of Z buffer. 100 l of cell suspension was removed to a fresh microcentrifuge tube. Assays for ␤-galactosidase activity were then performed as described in the protocol described in the MATCHMAKER two-hybrid system (PT1265-1) (CLONTECH).

Mutations in the Consensus Sequence (SPIK) Recognized by cdc2/CDC28-type Kinase in PHO2 Led to Complete Loss of Its
Activity-A survey of potential recognition sites in PHO2 protein for cdc2/CDC28-type kinase or cyclic AMP-dependent protein kinase (protein kinase A) revealed a SPIK (230 -233) site near the PHO80 homologous region and a RKKIS (107-111) site in the second helix of the PHO2 homeodomain (Fig. 1). The amino acid sequence surrounding serine residue 230 resembles the consensus sequence of sites phosphorylated by cdc2/CDC28 protein kinase (20 -23). The consensus sequence R/KR/KXS is the optimal protein kinase A recognition site (34). The serines in the two consensus sequences were considered to be theoretical phospho-acceptors. We assumed that protein kinase A phosphorylated Ser-111 and cdc2/CDC28-type kinase phosphorylated Ser-230, thereby affecting PHO2 function. To verify our hypothesis, we mutated Ser-111 and Ser-230 to Ala and trans-formed the plasmids pRS2 (YCp containing a wild type PHO2 gene), pRS2 (Ser-111 3 Ala), and pRS2 (Ser-230 3 Ala) into PHO2-disrupted strain YJ1, then determined the acid phosphatase activity of the cell suspensions as described under "Experimental Procedures." As can be seen in Table I, the substitution of Ala for Ser-111 had no significant effect on the acid phosphatase activity under high or low phosphate conditions, whereas the mutation of PHO2 Ser-230 to Ala resulted in the loss of its ability to activate PHO5 expression. The results suggested that Ser-230 might be a critical residue for PHO2 to activate PHO5 expression. Since Ser-230 is located in the consensus sequence SPIK, which may be recognized and phosphorylated by cdc2/CDC28 type kinase in vivo, its phosphorylation may be necessary for PHO2 function. This possibility is supported by the finding that another PHO2 mutant (Pro-231 3 Ser), in which the consensus sequence has been changed, also showed no ability to confer repressible acid phosphatase activity. Interestingly, the PHO2 variant containing an acidic residue in place of Ser-230 (Ser-230 3 Asp) could activate PHO5 expression under low phosphate conditions. It seems feasible that the hydroxyl group may serve as a phospho-acceptor in vivo, and the introduction of a negative charge at residue 230 may be necessary for PHO2 function. Intriguingly, the alteration of Lys-233 to Ala showed no effect on PHO2 function to activate PHO5 expression.
In Vitro Phosphorylation of GST-PHO2 Fusion Protein by YPH499 Cell Extract-To ascertain whether Ser-230 can be phosphorylated, the E. coli-expressed GST-PHO2 and GST-PHO2 (Pro-231 3 Ser) were subjected to in vitro phospholabeling experiments. The labeled proteins were separated by SDS-polyacrylamide gel electrophoresis and visualized by autoradiography. When the cell extract from a wild type YPH499 strain grown under low phosphate conditions was used as the enzyme with the GST-PHO2 fusion protein as the substrate, an approximate 90-kDa protein was observed to be phosphorylated (Fig. 2B, lane 2), but no phosphorylated band was found at the corresponding position when GST-PHO2 (Pro-231 3 Ser) was used as the substrate under identical labeling conditions (Fig. 2B, lane 3) nor when fusion protein was absent (Fig. 2B,  lane 1). These results inicated that the PHO2 portion (wild type but not PHO2 (Pro-231 3 Ser)) of the fusion protein is the target of the protein kinase in cell extract from YPH499 strain grown under low phosphate conditions, and the Pro-231 to Ser mutation diminished or prevented the phosphorylation of PHO2 protein. Furthermore, we also found that GST-PHO2 could be phosphorylated by cell extract from YPH499 cells grown under high phosphate conditions (data not shown). These data allowed us to suggest that PHO2 protein can be phosphorylated in vitro by a protein kinase(s) from YPH499 cells, and the Ser-230 may be the major phosphorylation site.
Protein Kinase CDC28 Can Interact with PHO2 in the Twohybrid System-Because the potential phosphorylation site of PHO2 resembles the consensus motif of sites phosphorylated by cdc2/CDC28-type protein kinase, yeast CDC28 protein was fused to the GAL4 binding domain (BD) (pGBT9-CDC28) and tested directly for interaction with GAL4 activation domain (AD)-PHO2 (expressed by pGAD424-PHO2) in the two-hybrid assay. Interaction of the two proteins was determined for their ability to activate transcription of a GAL1::lacZ reporter gene in SFY526 by ␤-galactosidase filter assays (Fig. 3). Clones containing pGBT9-CDC28 and pGAD424 or pGBT9 and pGAD424-PHO2 were white, whereas clones containing pGBT9-CDC28 and pGAD424-PHO2 displayed a distinct blue color (Fig. 3). These results showed that CDC28 protein and PHO2 protein can interact with each other in the two-hybrid system.
Immunoprecipitate of CDC28HA from Low Phosphate Cell Extract Can Phosphorylate GST-PHO2-The interaction between CDC28 and PHO2 in the two-hybrid system suggested that CDC28 may be the protein kinase of the transcriptional factor PHO2. We fused an HA tag to the C terminus of CDC28 by polymerase chain reaction for the facility of immunoprecipitation. To ascertain whether CDC28 can phosphorylate PHO2 in vitro, we used anti-HA to immunoprecipitate CDC28 from the strain YPH499 bearing CDC28HA on a multicopy expression plasmid (pVTU102-CDC28HA) grown under low phosphate conditions. The immunoprecipitated CDC28HA could phosphorylate GST-PHO2 (Fig. 4A, lane 1), whereas the immunoprecipitate from control strain YPH499 containing vector pVTU102 had very low kinase activity (Fig. 4A, lane 2). The immunoprecipitate derived from cell extract containing CDC28HA phosphorylated histone H1 (Fig. 4B, lane 1), whereas the control had only background kinase activity (Fig.  4B, lane 2). These results showed that PHO2 could be phosphorylated in vitro by CDC28 kinase complex from YPH499 (pVTU102-CDC28HA) grown under low phosphate conditions.
Phosphate Switch Regulates the Transcriptional Activation Activity of PHO2 and Mutations of the (SPIK) Site Affect the Transcriptional Activation Activity of PHO2 and the Interaction between PHO2 and PHO4 -PHO2 was fused to the DNA BD of yeast transcriptional factor GAL4. The plasmid expressing GAL4 BD-PHO2 chimera, pGBT9-PHO2, was transformed into a yeast strain, SFY526, containing an integrated GAL1::lacZ reporter gene. By measuring the expression of the reporter gene (lacZ) under high and low phosphate conditions (Table II), it was found that the chimera alone could activate the expression of lacZ, and the ␤-galactosidase activity was regulated by the phosphate switch. The activity under low phosphate condition was 8-fold more than that under high phosphate conditions. In contrast, when pVA3 (containing the fusion of murine p53 and GAL4 DNA binding domain) and pTD1 (containing a fusion of SV40 large T-antigen and GAL4 activation domain) were cotransformed into SFY526, the expression of lacZ was also activated, but the ␤-galactosidase activity was not affected by the phosphate switch. These results demonstrated that the transcriptional activation activity of PHO2 was regulated by the phosphate switch, which was repressed under high phosphate condition and was derepressed under low phosphate conditions.
The SPIK (230 -233) site near PHO80 homologous region is a potential phosphorylation site of PHO2, and the phosphorylation of Ser-230 is necessary to activate PHO5 transcription. The mutation of Ser-230 to Ala resulted in the loss of ability to be phosphorylated, so GAL4BD-PHO2 (Ser-230 3 Ala) chimera could not activate the expression of lacZ irrespective of phosphate concentration (Table II). The substitution of Asp for Ser-230, which mimics the phosphorylation state by providing a negative charge at residue 230, had no significant effect on the activation activity of PHO2. These results showed that the   3. Interaction between CDC28 and PHO2 in yeast twohybrid system. pGBT9 contains a GAL4 DNA binding domain. pGAD424 contains a GAL4 activation domain. pGBT9-CDC28 contains the fusion of CDC28 and GAL4 DNA binding domain. pGAD424-PHO2 expresses GAL4AD-PHO2 chimera. pVA3, murine p53 in pGBT9. pTD1 contains the fusion of SV40 large T-antigen and GAL4 activation domain. The cotransformant of pTD1 and pVA3 was a positive control. ␤-Galactosidase activity of SFY526 transformants was assayed as described under "Experimental Procedures." Each was repeated twice as shown.
phosphorylation of PHO2 Ser-230 is necessary for the activation activity of PHO2. The phosphorylation site-containing negative charge may resemble an acidic activation domain. However, the activity of GAL4BD-PHO2 (S230D) chimera was also regulated by the phosphate switch. The activity under low phosphate condition was 5-fold more than that under high phosphate condition.
The expression of lacZ could also be activated by GAL4BD-PHO2 chimera in another strain, SFY526 (pho4::HIS3). The activity under low phosphate conditions was 3-fold more than that under high phosphate condition (Table III), although it was less than that in SFY526. When pGBT9-PHO2 was cotransformed with pRS4 (YCp containing a PHO4 gene), it had no significant effect on the ␤-galactosidase activity under high phosphate condition, but the activity under low phosphate conditions was 1-fold more than that using the GAL4BD-PHO2 chimera alone (Table III). The results demonstrated that these two proteins may interact with each other. When pGBT9-PHO4 (for expression of GAL4BD-PHO4 chimera) was cotransformed into SFY526 (pho4::HIS3) with pGAD424-PHO2 (for expression of GAL4 AD-PHO2 chimera), the ␤-galactosidase activity was 40% more than that using pGBT9-PHO4 and pGAD424. The result also indicated that PHO4 and PHO2 may interact in vivo.
Another result showed that the phosphorylation of PHO2 is necessary for the interaction between PHO2 and PHO4. When GAL4BD-PHO2 (S230A) was coexpressed with GAL4AD-PHO4 chimera in SFY526 (pho4::HIS3), the galactosidase activity could not be detected irrespective of the phosphate switch (Table III). The result indicated that the phosphorylation site may be essential not only for the transcriptional activation activity but also for the interaction between PHO2 and PHO4, which may play a key role in maintaining the transcriptional activity of PHO2.
BIAcore® Analysis of the Interaction between PHO2 and PHO4 in Vitro-The data above showed that the phosphorylation of PHO2 Ser-230 is necessary for the interaction between PHO2 and PHO4. To investigate whether the phosphorylation of Ser-230 is sufficient to mediate interaction between PHO2 and PHO4, we used purified GST-PHO2, GST-PHO2SA (Ser230 3 Ala), GST-PHO2SD (Ser-230 3 Asp), and PHO4 for BIAcore® analysis.
PHO4 was covalently linked to a sensor chip (CM5). GST-PHO2, GST-PHO2SA, and GST-PHO2SD, at the same concentration of 5 M, were injected over the surface. As shown by the sensorgrams obtained (Fig. 5), a strong signal was detected when GST-PHO2SD was injected (Fig. 5B), whereas signals were very weak when GST-PHO2 (Fig. 5A) or GST-PHO2SA (Fig. 5C) was used. The sensorgram of GST-PHO2SD showed a very typical interaction. The effect of impurities and the fused GST was eliminated for the results of GST-PHO2 and GST-PHO2SA. The results indicated the negative charge of PHO2 (Ser-230 3 Asp), which mimics the phosphorylation of Ser-230, may help the interaction between PHO4 and GST-PHO2SD. DISCUSSION The functions of many transcriptional factors are regulated by phosphorylation and dephosphorylation, which may act in several aspects. First, phosphorylation can change the cellular location of some transcriptional factors. An S. cerevisiae transcriptional factor, SWI5, which is localized in the cytoplasm at the G 2 phase of cell cycle due to its phosphorylation by the cyclin-CDC28 protein kinase, enters the nucleus at Start in the G 1 phase and activates HO gene expression (35). Similarly, PHO4 is localized in the cytoplasm under high phosphate conditions (36) due to hyperphosphorylation by the PHO80-PHO85 cyclin-CDK (CDK, cyclin-dependent kinase) (11). Second, some transcriptional activation factors, such as GAL4 and GCN4 (37,38), which have acidic transcriptional activation domains, may have a higher transcriptional activation activity upon point mutations to increase negative charges. It is conjectured that phosphorylation and dephosphorylation may regulate the transcriptional activation activity by the increase or decrease of negative charges in response to the change of the circumstance.  4. CDC28 phosphorylates PHO2 in vitro. A, phosphorylation of GST-PHO2 assay. The phosphorylation of GST-PHO2 was assayed by immunoprecipitates from the strain YPH499, bearing CDC28HA on a multicopy expression plasmid (pVTU102-CDC28HA) (lane1), and a control strain YPH499 containing the empty plasmid (pVTU102) (lane 2). B, histone H1 kinase assay. The immunoprecipitates from the strain YPH499 containing CDC28HA (lane 1) and the control strain (lane 2) was used in histone H1 kinase assay.

TABLE II
Analysis of transcriptional activation activity of yeast PHO2 protein PHO2 and its mutants, PHO2 (S230A) and PHO2 (S230D), were fused to the GAL4 DNA binding domain (in pGBT9) and expressed in yeast strain SFY526. ␤-Galactosidase activity derived from activation of the GAL1-lacZ reporter were measured as described under "Experimental Procedures." Values are the means for triplate determinations. pVA3, murine p53 in pGBT9. pTD1 contains the fusion of SV40 large T-antigen and GAL4 activation domain.  Third, phosphorylation may affect the interaction between transcriptional factors. Here we found that phosphorylation of PHO2 can affect the transcriptional activation activity. Only when PHO2 is phosphorylated, the secreted acid phosphatase, PHO5, can be expressed under low phosphate conditions. Further investigations indicated that the phosphorylation of PHO2 not only improve the transcriptional activation activity but facilitate the interaction between PHO2 and PHO4. These results suggested the pleiotropy of PHO2 phosphorylation. How the critical phosphorylation contributes to transcriptional activation activity and interaction is unclear. Because the previous deletion analysis showed that the possible PHO4 binding domain of PHO2 does not include the SPIK site (13,17) and our results indicated that the negative charge of PHO2 (Ser230 3 Asp) is sufficient to help the interaction between PHO2 and PHO4 in vitro suggests that the phosphorylation of the Ser-230 residue in PHO2 may alter the conformation of PHO2, making the PHO4 binding domain of PHO2 suitable for interaction with PHO4 but not participating in the interaction directly. The phosphorylation may also increase the ability of PHO2 to activate transcription by altering the three-dimensional conformation of PHO2. Furthermore, the phosphate group may be directly involved in transcriptional activation by increasing the negative charges of activation domain of PHO2. PHO2 has a potential phosphorylation site, SPIK (230 -233), that resembles the consensus sequence of site recognized and phosphorylated by cdc2/CDC28-type kinase (20 -23). We found that GST-PHO2 fusion protein expressed in E. coli could be phosphorylated by yeast cell extract from YPH499 strain grown under low phosphate conditions. Further investigation showed that CDC28HA immunoprecipitate derived from yeast grown under low phosphate conditions could phosphorylate GST-PHO2 in vitro. These results indicate that CDC28 may be the kinase of PHO2. CDC28 is an essential regulator of cell cycle progression, whereas PHO2 is required for several metabolism processes. Another yeast cyclin-dependent kinase, PHO85, has provided a link between cell cycle regulation and phosphate metabolism (11,39,40). By association with different cyclin subunits, PHO85 may be directed to distinct biological functions. If CDC28 could be proved to be the physiological kinase of PHO2, another link between cell cycle progression and metabolism would be found. Perhaps a new cyclin would be found to help CDC28 participate in metabolism. In addition to PHO85, CDC28 might also serve to coordinate nutritional state and cell cycle progression.
Homeo box-encoding genes appear to be associated with developmental control or cell type regulation. Previous investigations carried out by Gilliquet and Berben (41) reveals the involvement of PHO2 that contains a homeo domain in the life cycle. Homozygous PHO2 null diploids show a deficiency to progress through meiosis, which can be corrected when the cells are transformed with PHO2-bearing plasmid. Many nuclear factors involved in the cell cycle have been observed to be phosphorylated and regulated by cdc2/CDC28-type protein kinase, which is activated only at specific cell cycle stages (42). Our present findings show that PHO2 may be phosphorylated by CDC28. These evidences also suggest that there is some association between metabolism and cell cycle machinery.