The Yeast Immunophilin Fpr3 Is a Physiological Substrate of the Tyrosine-specific Phosphoprotein Phosphatase Ptp1*

The tyrosine-specific phosphoprotein phosphatase en- coded by the Saccharomyces cerevisiae PTP1 gene dephosphorylates artificial substrates in vitro , but little is known about its functions and substrates in vivo . The presence of Ptp1 resulted in dephosphorylation of mul- tiple tyrosine-phosphorylated proteins in yeast expressing a heterologous tyrosine-specific protein kinase, in- dicating that Ptp1 can dephosphorylate a broad range of substrates in vivo . Correspondingly, several proteins phosphorylated at tyrosine by endogenous protein kinases exhibited a marked increase in tyrosine phospho- rylation in ptp1 mutant cells. One of these phosphotyrosyl proteins (p70) was also dephosphorylated in vitro when incubated with recombinant Ptp1. p70 was purified to homogeneity; analysis of four tryptic peptides revealed that p70 is identical to the recently described FPR3 gene product, a nucleolarly localized proline rota- mase of the FK506- and rapamycin-binding family. The identity of p70 with Fpr3 was confirmed in the demon- stration that the abundance of tyrosine-phosphorylated p70 in ptp1 mutants was strictly correlated with the level of FPR3 expression; immobilized phosphotyrosyl Fpr3 was directly dephosphorylated by recombinant Ptp1. Site-directed

The tyrosine-specific phosphoprotein phosphatase encoded by the Saccharomyces cerevisiae PTP1 gene dephosphorylates artificial substrates in vitro, but little is known about its functions and substrates in vivo. The presence of Ptp1 resulted in dephosphorylation of multiple tyrosine-phosphorylated proteins in yeast expressing a heterologous tyrosine-specific protein kinase, indicating that Ptp1 can dephosphorylate a broad range of substrates in vivo. Correspondingly, several proteins phosphorylated at tyrosine by endogenous protein kinases exhibited a marked increase in tyrosine phosphorylation in ptp1 mutant cells. One of these phosphotyrosyl proteins (p70) was also dephosphorylated in vitro when incubated with recombinant Ptp1. p70 was purified to homogeneity; analysis of four tryptic peptides revealed that p70 is identical to the recently described FPR3 gene product, a nucleolarly localized proline rotamase of the FK506-and rapamycin-binding family. The identity of p70 with Fpr3 was confirmed in the demonstration that the abundance of tyrosine-phosphorylated p70 in ptp1 mutants was strictly correlated with the level of FPR3 expression; immobilized phosphotyrosyl Fpr3 was directly dephosphorylated by recombinant Ptp1. Site-directed mutagenesis demonstrated that the site of tyrosine phosphorylation is Tyr-184, which resides within the nucleolin-like amino-terminal domain of Fpr3. Protein kinase activities from yeast cell extracts can bind to and phosphorylate the immobilized aminoterminal domain of Fpr3 on serine, threonine, and tyrosine. Fpr3 represents the first phosphotyrosyl protein identified in S. cerevisiae that is not itself a protein kinase and is as yet the only known physiological substrate of Ptp1.
Phosphotyrosine-specific phosphoprotein phosphatases (PTPs) 1 have been identified in many evolutionarily divergent eukaryotes. These enzymes form a distinct superfamily and are unrelated in sequence to serine/threonine-specific phosphoprotein phosphatases (for reviews, see Refs. [1][2][3]. All PTPs possess stretches of sequence similarity within their catalytic domains, including the active site consensus sequence (I/V)HCX-AGXGR(S/T)G. This hallmark sequence contains an invariant Cys residue, which acts as the nucleophile during the dephosphorylation reaction, and a GXGXXG motif, which forms a phosphate-binding loop and is also found in nucleotide-binding proteins such as protein kinases and GTPases (4). The substrate-binding cleft of PTPs is surrounded by basic amino acids, which may explain the preference for acidic residues near the phosphorylated tyrosines in PTP substrates (5,6).
In the budding yeast, Saccharomyces cerevisiae, dedicated tyrosine-specific protein kinases have not been identified. However, a number of genes encoding PTPs have been reported. These PTPs include both phosphotyrosine-specific and dualspecific enzymes as seen in higher eukaryotes. Two of the S. cerevisiae PTPs appear to be MAP kinase phosphatases. The dual-specific PTP encoded by the MSG5 gene dephosphorylates Fus3 and thereby contributes to the reversal of pheromone arrest (7). The PTP2 gene product is thought to dephosphorylate Hog1, a MAP kinase involved in osmoregulation (8). At least two S. cerevisiae PTPs are involved in cell cycle control: the CDC14 gene product is required for progression through S phase (9), and the product of the MIH1 gene, the S. cerevisiae homolog of the fission yeast cdc25 ϩ , is thought to dephosphorylate the Cdc28 kinase (10). The YVH1 PTP gene is induced by nitrogen starvation and encodes a PTP that is required for maximal growth (11).
PTP1, the first PTP gene reported in budding yeast, was identified by the polymerase chain reaction using oligonucleotides corresponding to conserved PTP catalytic domain sequences as primers (12). Ptp1 appears to be phosphotyrosine specific and is comprised of a carboxyl-terminal catalytic domain and a unique 55-residue amino-terminal region of unknown function. Although Ptp1 is active in vitro against artificial substrates, the physiological role of Ptp1 is unknown; PTP1 disruption or overexpression does not overtly effect growth at extreme temperatures, sensitivity to different metal ions, osmotic stability, carbon source utilization, mating, or sporulation (12,13). 2 However, expression of PTP1 in fission yeast mimics cdc25 ϩ overexpression and leads to precocious mitosis (14). In addition, overexpression of PTP1 in S. cerevisiae rescues the synthetic lethality resulting from disruption of both PTP2 and PTC1, a gene encoding a putative Ser/Thrspecific phosphoprotein phosphatase of the PP2C class (15). These results suggest that when overproduced, Ptp1 may be capable of dephosphorylating Cdc2 and Hog1, but the relevance of these activities to normal Ptp1 function is unclear.
Here, we describe the identification of yeast phosphotyrosyl proteins that are dephosphorylated by Ptp1 in vivo and present evidence that one Ptp1 substrate is the nucleolar immunophilin, Fpr3.

MATERIALS AND METHODS
Yeast Strains and Culture Conditions-Yeast strains used in this work are described in Table I. Yeast transformation was carried out using electroporation (16). To generate the strain YBB200, a 3.1-kilobase ClaI fragment containing a HIS3-disrupted ptp1 allele (13) (kindly provided by P. James) was introduced into strain YPH499 (17) by DNA-mediated transformation. To generate strain YBB300, the same ClaI fragment was used to transform strain YBB100 (18). To generate the strain PJ55300, the fpr3-2::HIS3 insertion mutation was introduced into strain PJ55-16C, following a procedure previously described (18). To generate YLW200, a 1.25-kilobase PvuII-AseI fragment containing a URA3-disrupted ptp1 allele was excised from plasmid pGEM-ptp1::URA3 (12) (kindly provided by R. Deschenes) and used to transform strain BJ2168.
To induce transcription of genes driven by a GAL promoter, cultures were grown overnight to A 600 nm ϭ 1 in defined medium containing 2% raffinose. Galactose was then added to a final concentration of 2%, and the cells were grown for an additional 3 h prior to harvesting. For large scale purification of p70 FPR3 , strain PJ58 -2B (ptp1⌬ ptp2⌬ mih1⌬) was grown in YPD in a 200-liter fermenter with vigorous aeration to stationary phase (A 600 nm ϭ 3.5).
Immunoblot Analysis-For immunoblot analysis, cells were grown to late exponential phase (A 600 nm ϭ 1). Cells were harvested by centrifugation at 4°C at 1,000 ϫ g for 5 min, resuspended in 50 mM Tris-HCl (pH 7.2), 100 mM NaCl, 5 mM EDTA, and recentrifuged. Cell pellets were resuspended in an equal volume of ice-cold buffer A (50 mM Tris-HCl (pH 7.2), 100 mM NaCl, 5 mM EDTA, 0.1% Triton X-100, 12 M benzamidine, 5 M phenanthroline, 30 M phenylmethylsulfonyl fluoride, 100 M Na 3 VO 4 , and 0.5 g/ml of each of the following: antipain, leupeptin, chymostatin, aprotinin, and pepstatin). The cells were lysed by vigorous agitation with glass beads as described (18). Lysates were clarified by centrifugation at 10,000 ϫ g for 10 min, brought to 1 ϫ Laemmli sample buffer (19) by addition of an appropriate amount of a concentrated stock, and separated by SDS-PAGE. Alternatively, where indicated, lysates were prepared by an alkaline lysis method as described (20). Briefly, cell pellets were incubated at 4°C in 250 mM NaOH, 1% 2-mercaptoethanol for 10 min, and cellular proteins were precipitated in 8% trichloroacetic acid. Protein precipitates were rinsed twice in acetone, dried, and then resuspended prior to SDS-PAGE by boiling in Laemmli sample buffer.
Plasmids-The polymerase chain reaction (PCR) was used to introduce a mutation into PTP1 that resulted in a Cys-to-Ala substitution at amino acid residue 252 to generate a catalytically inactive variant. The upstream primer, 5Ј-TGAAATTCCCGCGGGAACCCCATTATCGTACA-CGCTTCCGCAGGC-3Ј, spanning nucleotides 724 -766 of PTP1 (counting from the translational start site), incorporated the TG-to-GC mutation (underlined) and a SacII restriction site (italicized). The downstream primer spanned nucleotides 865-886 just downstream of a BsmI restriction site at nucleotides 850 -856. YEp51-PTP1, a multicopy vector containing PTP1 under control of the GAL10 promoter (a gift from P. James), was used as the PCR template, and the reaction was carried out using standard procedures as described (27). The PCR product (162 bp) was treated with the Klenow fragment of Escherichia coli DNA polymerase I, digested with SacII and BsmI, and religated into YEp51-PTP1, which had been digested with SacII and BsmI.
The wild-type and mutant PTP1 genes were cloned into the glutathione S-transferase (GST) fusion vector PGEX-3X (Pharmacia Biotech Inc.). In preparation for these ligations, an adapter, 5Ј-CGGGATC-CAAATGCAGGCCTCTCGAGATCGATGAATTC-G3Ј, which contains a BamHI site (boldface type) followed by the first 7 translated nucleotides of PTP1 (underlined), and StuI, ClaI, and EcoRI sites (italicized) was first inserted between the BamHI and EcoRI sites in the vector. This strategy allowed the in-frame insertion of a PvuII-ClaI fragment containing the remainder of PTP1 (nucleotides 8 -3042 excised from YEp51-PTP1) into the modified vector between the StuI and ClaI sites. The resulting constructs encode GST-Ptp1 and GST-Ptp1(C252A) fusion proteins with a factor Xa cleavage site between the GST and Ptp1 coding segments. Plasmid YEp352GAL-v-src was described previously (28).
FPR3 expression plasmids YEp351-FPR3myc (pYB1010), YEp351GAL-FPR3myc (pYB124), YEp351GAL-FPR3 (pYB123), YEp351GAL-FPR3N (pYB126), YEp351GAL-FPR3C (pYB120), and pGXFPR3A (encoding GST-Fpr3N) are described in Ref. 18. Plasmids expressing mutant derivatives of FPR3 were generated by PCR using a pUC19-derived plasmid containing wild-type FPR3 (pNH2.2; described in Ref. 18) as template. A double mutant (Y184F,Y189F) was generated using two PCR primers, each of which contained both a change in codon 184 from TAT to TTT (nucleotides 550 -552) and in codon 189 from TAC to TTC (nucleotides 265-267). Primer 1 spanned nucleotides 546 -573 on the coding strand, and primer 2 spanned nucleotides 573-540 on the noncoding strand. In one PCR reaction, primer 1 was used with an additional downstream primer, spanning nucleotides 839 -857 (noncoding strand), to generate a 311-bp product. In a separate reaction, primer 2 was used with an upstream primer, spanning nucleotides 363-382 (coding strand), to generate a 210-bp product. In the final PCR reaction, an overlap extension, the 2 initial overlapping products were purified and used as template primers together with the upstream and downstream primers. This reaction generated a 494-bp product spanning nucleotides 363-857 with mutations at codons 184 and 189. The product was digested with BspE1 and EcoRI to generate a 455-bp fragment. This fragment was ligated to the 4.5-kilobase fragment of BspE1 and EcoRI-digested pNH2.2, thus replacing the corresponding region of wild-type FPR3. To generate versions of FPR3 containing each of the single mutations, Y184F and Y189F, a similar procedure was used, except that the primers "1" and "2" were changed to include a mutation only at codon 184 for Y184F and only at codon 189 for Y189F. To express the mutated FPR3 genes in yeast, they were excised from pNH2.2 with AflIII and HindIII, and the AflIII site was filled in with the Klenow fragment of DNA polymerase I. The resulting fragments were then ligated into the vector YEp351GAL (18), which had been opened with SalI, treated with Klenow, and then cut with HindIII. All constructs were verified by DNA sequencing.

Dephosphorylation of Phosphotyrosyl Proteins by GST-Ptp1 Fusion
Proteins-To purify GST-Ptp1 fusion proteins, E. coli transformed with plasmids expressing the desired fusion were grown to A 600 nm ϭ 1.0, induced with 100 M isopropyl ␤-D-thiogalactoside, and harvested by centrifugation. Approximately 1 g of wet cells were resuspended in 5 ml of buffer B (buffer A devoid of vanadate) and broken by sonic disruption. The lysates were clarified by centrifugation at 10,000 ϫ g for 10 min, diluted in buffer B to 5 mg of protein/ml, and gently rocked for 30 min at 0°C with 0.5 ml (drained volume) glutathione-Sepharose 4B (Pharmacia). The beads were then collected by centrifugation at 1,000 ϫ g for 20 s, resuspended in buffer B, loaded into a column, and washed with 20 volumes of buffer B containing 300 mM NaCl. The GST-Ptp1 beads were stored at 4°C until their use for p70 FPR3 dephosphorylation reactions. In these reactions, crude lysates or purified samples containing approximately 0.1 pmol of p70 FPR3 in buffer B were incubated with 0.1 l of GST-Ptp1 beads (0.5 pmol of GST-Ptp1) for 10 min at 30°C.
Dephosphorylation of Immobilized Fpr3-GST-Ptp1 beads (300 l containing 100 g of GST-Ptp1) were incubated with 2 g of factor Xa (New England Biolabs) in 1 mM CaCl 2 , 50 mM Tris (pH 7.5), 150 mM NaCl for 15 h at 4°C to release soluble Ptp1. The beads were removed by centrifugation at 1000 ϫ g for 20 s. The solution containing Ptp1 was diluted to a concentration of 30 g/ml (0.5 M) in 50 mM Tris-HCl (pH 6.9), 10 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol. Lysates of strain YBB300 overexpressing Fpr3 (50 g of protein/lane) were resolved by SDS-PAGE and electrophoretically transferred to PVDF membrane. Strips of membrane (approximately 1 cm 2 each) containing Fpr3 were rocked at 30°C with 1 ml of 0.5 M Ptp1. At intervals, strips were washed three times for 10 min each at 55°C in immunoblot blocking buffer and probed with anti-phosphotyrosine mAb as described above.
Purification of Fpr3 (p70)-Strain PJ58 -2B was grown in a 200-liter fermenter, harvested by centrifugation in an air-driven Sharples supercentrifuge rotor, and frozen at Ϫ80°C until used. Yeast (500 g, wet weight) were processed in five separate batches as follows. For each batch, a block of frozen cells (100 g) was broken into small pieces, mixed with 100 ml of buffer A, and agitated vigorously with 100 ml of glass beads for 4 min at 0°C in buffer A using a Biospec Bead Beater (Biospec Products, Bartlesville, OK). The lysate was clarified by centrifugation for 15 min at 10,000 ϫ g, and the resulting supernatant fraction was centrifuged at 100,000 ϫ g for 40 min. The p70 was eluted from the resulting pellet by stirring the resuspended particulate material for 1 h at 4°C in 50 ml of buffer A containing a final concentration of 1 M NaCl. Insoluble particulate matter was removed from the suspension by centrifugation at 100,000 ϫ g. The supernatant fraction was loaded onto a bed (150 ml) of cellulose gel (GH25, Amicon, Denvers, MA) in a column (60 ϫ 1.8 cm) and eluted with buffer C (40 mM HEPES (pH 6.9), 20 mM NaCl, 5 mM EDTA, 100 M Na 3 V0 4 , and the same protease inhibitors used in buffer A). The protein-containing fractions were pooled (50 -70 ml) and loaded onto a bed (60 ml) of S-Sepharose in a column (2 ϫ 20 cm) (Pharmacia). The column was washed with five column volumes of buffer C containing 220 mM NaCl and eluted with a gradient (550 ml total) from 220 to 500 mM NaCl. Fractions were analyzed for p70 by immunoblotting with anti-phosphotyrosine antibody and stored at Ϫ70°C.
S-Sepharose fractions enriched for p70 (300 ml total) collected from the five 100-g batches of yeast were brought to a final concentration of 1% Triton X-100, dialyzed twice for 2.5 h against 4 liters of buffer C containing 0.1% Triton X-100, and precleared by incubation with 0.5 ml of protein A-Sepharose CL-4B (Pharmacia) for 4 h at 4°C; the protein A-Sepharose was removed by centrifugation at 4000 ϫ g for 5 min. The resulting supernatant solution was mixed with 1.0 ml of protein A-Sepharose, to which had been coupled anti-phosphotyrosine mAb FB2 (23), and incubated with gentle rocking at 4°C for 6 h. The beads were washed with 10 ml of buffer C containing a final concentration of 120 mM NaCl, and p70 was eluted with 2 ml of 60 mM phenylphosphate. The eluted protein was precipitated with 10% trichloroacetic acid, treated with 10 mM 4-vinylpyridine to alkylate Cys residues, resolved by SDS-PAGE, and transferred to Immobilon-P. Ponceau S staining revealed a single band in the 70-kDa size range. This procedure yielded approximately 15 g of p70 from 500 g of yeast. The band was excised and digested with sequencing grade trypsin (Boehringer Mannheim). Tryptic peptides were separated by reverse phase chromatography (Brownlee C8 column, 1 ϫ 250 mm, Applied Biosystems) using a 172A microbore high pressure liquid chromatograph (Applied Biosystems) and subjected to microsequencing by Edman degradation in a 477A protein sequencer (Applied Biosystems, Inc.).
Two-dimensional Polyacrylamide Gel Electrophoresis-Two-dimensional gel electrophoresis was conducted as described elsewhere (29). A sample of the p70 preparation (ϳ0.1 g) was separated in the first dimension using nonequilibrium isoelectric focusing in a tube containing pH 3-10 ampholytes (Pharmalyte, Pharmacia) and in the second dimension by electrophoresis in a 7.5% SDS slab gel. The staining of separated proteins was carried out with colloidal gold (Aurodye TM , Amersham).
Metabolic Labeling with [ 32 P]Orthophosphate-Strains YPH499 and YBB200 transformed with the plasmid YEp351-FPR3myc were grown overnight to A 600 nm ϭ 0.4 in synthetic low phosphate medium (30,31), and then resuspended to A 600 nm ϭ 0.8 in 5 ml of prewarmed synthetic low phosphate medium containing 2 mCi of 32 P-labeled PO 4 3Ϫ (DuPont NEN) (final specific activity, 8 Ci/mmol). After 3 h in a 30°C gyratory water bath, the yeast were harvested by centrifugation, and cell proteins were prepared by alkaline lysis and trichloroacetic acid precipitation as described above. The dried precipitates were resuspended by boiling in 2% SDS for 5 min, clarified by centrifugation, diluted to 0.15% SDS with buffer A, and precleared by incubation with protein A-Sepharose for 1 h. Fpr3 was then immunoprecipitated with 0.4 g of affinitypurified rabbit antibody raised against a bacterially expressed GST-Fpr3 fusion protein (18). The immunoprecipitates were resolved by SDS-PAGE, transferred to Immobilon P, and exposed to x-ray film. For phosphoamino acid analysis, Immobilon strips containing the Fpr3 bands were excised and subjected to partial acid hydrolysis, and the resulting products were separated by two-dimensional electrophoresis (32) at pH 1.9 (first dimension) and pH 3.7 (second dimension) on thin layer cellulose plates (Merck). Phosphoamino acid standards were included with the samples and were visualized by ninhydrin staining. Autoradiographs were obtained using either a Phosphorimager TM (Molecular Dynamics) or x-ray film. Radioactivity incorporated into individual phosphoamino acids was quantitated by scintillation counting of the corresponding stained spots scraped from the thin layer plates.
Phosphorylation of Immobilized GST-Fpr3 by Yeast Kinases-Strain YLW200 was grown to mid-exponential phase in YPD and harvested by centrifugation. Cells were washed and resuspended in 0.5 ml of extraction buffer (50 mM HEPES (pH 7.8), 75 mM NaCl, 2.5 mM MgCl 2 , 0.1 mM EDTA, 0.05% Triton X-100, 0.5 mM dithiothreitol, 20 mM ␤-glycerophosphate, 0.1 mM Na 3 V0 4 , 2 g/ml leupeptin, 0.5 mM phenylmethylsulfonyl fluoride). An extract was prepared by vigorously agitating the resuspended pellet (approximately 50 A 600 nm units) with glass beads in 0.5 ml of extraction buffer for 5 min at 4°C. The resulting lysate was clarified by centrifugation at 12,000 ϫ g for 10 min and adjusted to a protein concentration of 7 mg/ml by dilution with extraction buffer. A GST fusion encoding the amino-terminal 279 residues of Fpr3 (Fpr3N) was purified from bacterial extracts by adsorption to glutathione-Sepharose beads (18), and a sample (1 l of beads containing 1 g of GST-Fpr3N) was mixed with ϳ200 l of the lysate. Following incubation (4°C for 120 min) on a rotating mixer, the beads were recovered by brief centrifugation and washed three times with 1 ml binding buffer

Ptp1
Has Broad Substrate Specificity-The product of the PTP1 gene (Ptp1) dephosphorylates several artificial substrates in vitro (12), suggesting that it may be capable of dephosphorylating a broad range of substrates in vivo. To explore this possibility, we compared the levels of tyrosine phosphorylation of yeast proteins in strain PJ55-16C, which harbors a disruption of the PTP1 gene (ptp1⌬), and the congenic PTP1 strain PJ55-16A. To enhance total cellular tyrosine phosphorylation, these strains were transformed with a multicopy plasmid expressing a known protein-tyrosine kinase, p60 v-src , under the control of the GAL1 promoter (28,33). Lysates prepared from these strains were examined by immunoblotting with anti-Src and anti-phosphotyrosine antibodies. The level of v-src expression in ptp1⌬ cells was similar to that of the control PTP1 cells (Fig. 1, bottom panel). In PTP1 strains, p60 v-src expression resulted in tyrosine phosphorylation of at least 10 different cellular proteins (Fig. 1, lane 1; see also Refs. 28 and 33). In the absence of Ptp1, both the number of phosphotyrosyl proteins and their extent of phosphorylation were greatly elevated (Fig. 1, lane 2). This result suggests that Ptp1 is able to dephosphorylate many different phosphotyrosyl proteins and has a broad substrate specificity in vivo. In contrast, disruption of PTP2 and MIH1 did not affect the level of tyrosine phosphorylation induced by v-src (data not shown).
To determine whether Ptp1 can dephosphorylate in vitro the proteins phosphorylated by p60 v-src , a bacterially expressed GST-Ptp1 fusion protein was incubated with the phosphotyrosyl proteins in lysates from either PTP1 or ptp1⌬ cells expressing v-src. The level of tyrosine phosphorylation was drastically reduced by incubation with GST-Ptp1 (Fig. 1, lanes 3  and 4). In contrast, incubation with catalytically inactive GST-Ptp1(C252A) did not significantly reduce the level of tyrosine phosphorylation (Fig. 1, lanes 5 and 6). This result confirms that Ptp1 has broad substrate specificity in vitro.
Disruption of PTP1 Enhances Detection of Endogenous Phosphotyrosyl Proteins-Physiological substrates of S. cerevisiae Ptp1 have not been reported. The phosphorylated proteins detected in ptp1⌬ yeast expressing v-src may not represent physiological substrates of Ptp1, as there is no evidence that these proteins are phosphorylated in the absence of v-src expression. However, the experiment described above suggested that disruption of PTP1 might lead to detectable increases in the level of phosphotyrosine in authentic Ptp1 substrates that are phosphorylated by endogenous kinases in vivo. We therefore sought to detect phosphotyrosyl proteins in Ptp1-deficient yeast that were not expressing any exogenous tyrosine kinase. Extracts of a ptp1⌬ ptp2⌬ mih1⌬ triple mutant and its congenic wild-type strain were examined by immunoblotting with several different anti-phosphotyrosine antibodies (Fig. 2). Many of the bands were common to the wild-type and triple mutant strains and appeared to correspond to abundant proteins because they were congruent with species visualized by staining with Coomassie Blue (data not shown). Reaction of these proteins with the anti-phosphotyrosine antibodies was variable and likely to be nonspecific because it was not reversed by competition with 50 mM phosphotyrosine (data not shown). In contrast, several proteins reactive with anti-phosphotyrosine antibody were detected in the triple PTP mutant that were not detectable in the wild-type strain (Fig. 2). Proteins of apparent molecular masses of 175 and 116 kDa were recognized by mAb FB2 (lanes  3 and 4), a protein of apparent molecular mass of 170 kDa was recognized by the rabbit polyclonal antibody (lane 7), and a protein of apparent molecular mass of 70 kDa was recognized by all of the antibodies with the exception of mAb 6G9 (lanes 1,  3, and 7). The latter protein, designated p70, was chosen for further study.
To determine whether the activity of other S. cerevisiae PTPs affected the tyrosine phosphorylation state of p70 (or any other protein) during normal growth, lysates of strains containing disruptions in PTP1, PTP2, and MIH1 were compared by antiphosphotyrosine antibody immunoblotting. Anti-phosphotyrosine antibody recognized p70 only in strains disrupted for PTP1 (Fig. 3A). Disruption of PTP2 or MIH1 did not result in a detectable increase in the level of tyrosine phosphorylation on p70 or in the appearance of additional phosphotyrosyl proteins (Fig. 3A, lanes 3 and 4). These results suggest that phosphotyrosyl p70 is dephosphorylated only by Ptp1. In addition, these results support the conclusion that Ptp2, which is thought to dephosphorylate the Hog1 kinase (8), and Mih1, which is thought to dephosphorylate the Cdc28 kinase (10), have more restricted substrate specificities than Ptp1.
ings described above suggested that p70 is dephosphorylated by Ptp1. Overexpression of PTP1 in a ptp1⌬ strain resulted in the disappearance of phosphotyrosyl p70 (Fig. 3B, lane 2), confirming that the appearance of phosphotyrosyl p70 in the ptp1⌬ strain was due to loss of PTP1 function.
To determine whether Ptp1 could dephosphorylate p70 in vitro, lysates from the ptp1⌬ ptp2⌬ mih1⌬ triple mutant strain were incubated with GST-Ptp1 or GST-Ptp1(C252A). Incubation with GST-Ptp1 resulted in the complete dephosphorylation of p70 (Fig. 4, lane 3), while GST-Ptp1(C252A) had no effect (Fig. 4, lane 4). The same results were obtained with soluble Ptp1 preparations generated by cleavage from the GST carrier by digestion with factor Xa (data not shown). This experiment suggests that p70 is a direct substrate of Ptp1 but does not exclude the possibility that Ptp1 activates another proteintyrosine phosphatase, which in turn dephosphorylates p70. The results of subsequent experiments (see below) provide evidence that p70 is a direct substrate of Ptp1.
Identification of p70 as the Yeast Immunophilin Fpr3-Phosphorylated p70 was purified from the ptp1⌬ ptp2⌬ mih1⌬ strain, PJ58 -2B, as described under "Materials and Methods," using anti-phosphotyrosine antibody to monitor the presence of phosphorylated p70 in different fractions (Fig. 5, panel A). A sample of purified p70 (Fig. 5A, lane 4) was subjected to twodimensional gel electrophoresis, electrophoretically transferred to Immobilon P, and gold stained. The most abundant protein species present in the preparation migrated with an apparent molecular mass of 70 kDa (Fig. 5B, left side) and was strongly reactive with anti-phosphotyrosine antibodies (Fig.  5B, right side). At each step in the purification, p70 was susceptible to dephosphorylation by purified recombinant GST-Ptp1; for example, phosphotyrosyl p70 in the peak fractions from S-Sepharose was efficiently dephosphorylated by GST-Ptp1 but not by GST-Ptp1(C252A) (Fig. 4, lanes 6 and 7). The same result was obtained with p70 purified by immunoaffinity chromatography with anti-phosphotyrosine mAb (data not shown). These results suggest that p70 is a direct substrate of Ptp1 rather than a substrate of another yeast phosphatase activated by Ptp1.
Tryptic peptides were generated from purified p70, and five were sequenced. Four of the five sequences matched precisely the amino acid sequence of S. cerevisiae Fpr3, a recently identified nucleolar FK506-binding protein (Fig. 6) (18,34,35). FK506-binding proteins (FKBPs) are immunophilins that bind the structurally related immunosuppressive drugs, FK506 and rapamycin. The formation of complexes between these drugs and the predominant cytosolic FKBP, FKBP-12, inhibits signal transduction pathways in both vertebrates and yeast, but the normal functions of these proteins are unknown (see "Discussion"). Genetic analysis confirmed that p70 is identical to Fpr3. When a ptp1⌬ strain was transformed with a high copy plasmid expressing galactose-inducible FPR3, p70 was greatly overproduced; conversely, p70 was completely absent in the ptp1⌬ fpr3⌬ strain (Fig. 7A). These findings, together with the sequence of the tryptic peptides derived from p70, verify that p70 is Fpr3. The anomalous electrophoretic mobility of Fpr3 (calculated molecular mass, 47 kDa) has been noted previously (18).
To determine whether phosphotyrosyl Fpr3 is a direct substrate of Ptp1, PVDF membrane strips containing Fpr3 were incubated either with buffer alone or with buffer containing 0.5 M soluble recombinant Ptp1. Incubation in buffer alone had no effect on the level of phosphotyrosyl Fpr3 (Fig. 7B, lane 1) nor did incubation with Ptp1(C252A) or with Ptp1 in the presence of vanadate (data not shown). However, incubation with Ptp1 for 1 h led to a ϳ70% reduction in the phosphotyrosine content of Fpr3 (Fig. 7B, lanes 2-4), as determined by densitometry. The amount of Fpr3 as detected by immunoblotting with anti-Fpr3 antibodies was equivalent in every lane after the incubation (data not shown). Because the substrate protein in this reaction was immobilized and the phosphatase was purified  1, 2, 5), GST- Ptp1 (lanes 3, 6), or GST-Ptp1(C252) (lanes 4, 7) and then boiled in SDS sample buffer and analyzed by immunoblotting with anti-phosphotyrosine mAb 4G10. The 65-kDa GST-Ptp1 fusion proteins (lanes 3-7) are stained nonspecifically by the 4G10 antibody. from bacteria (which lack PTPs), we conclude that Ptp1 can directly dephosphorylate Fpr3. Some phosphotyrosyl-Fpr3 remained phosphorylated following Ptp1 treatment; this may be because the phosphotyrosine residue was inaccessible in a fraction of the immobilized molecules.
Identification of the Tyrosine Phosphorylation Site in Fpr3-Fpr3 contains an amino-terminal nucleolin-like segment (amino acids 20 -290) and a COOH-terminal immunophilin domain (amino acids 291-413) (18). To determine which of these two regions of Fpr3 is subject to tyrosine phosphorylation, each region was expressed independently in fpr3⌬ and ptp1⌬ fpr3⌬ yeast strains. In the ptp1⌬ fpr3⌬ strain, phosphotyrosine was readily detected in both full-length Fpr3 and its amino-terminal domain (Fig. 8A, lanes 2 and 4), but phosphotyrosine was not detected in the COOH-terminal domain (Fig. 8A, lane 6).
These results indicate that the site of tyrosine phosphorylation in Fpr3 is one or more of the seven tyrosine residues within the nucleolin-like domain (Fig. 6).
Inspection of the sequence of the amino-terminal domain of Fpr3 revealed that two of the Tyr residues (Tyr-184 and Tyr-189) are immediately preceded by two or more acidic residues (Fig. 6). This sequence context is favored by many of the Tyrspecific protein kinases in higher eukaryotes (36). To determine if Tyr-184 and -189 were sites of tyrosine phosphorylation, these residues were changed to Phe by site-directed mutagenesis of FPR3. Neither the Y184F,Y189F double mutant nor the Y184F single mutant contained detectable phosphotyrosine (Fig. 8B, lanes 2 and 3). In contrast, the Y189F mutant possessed just as high a level of phosphotyrosine as wild-type Fpr3 (Fig. 8B, lanes 1 and 4). Thus, Tyr-184 appears to be the sole site of tyrosine phosphorylation of Fpr3.
Fpr3 Is Phosphorylated on Ser, Thr, and Tyr-To assess the phosphorylation state of Fpr3 in vivo, Fpr3 was immunoprecipitated from [ 32 P]orthophosphate-labeled cells and subjected to phosphoamino acid analysis (Fig. 9, panels 1 and 2). The relative phosphoamino acid content of Fpr3 from ptp1⌬ yeast The deduced amino acid sequence of the FPR3 gene product is shown in the one-letter code (18). The sequence of each of four peptides derived by digestion of purified phosphotyrosyl p70 with trypsin is underlined. The carboxyl-terminal catalytic domain of Fpr3, which possesses peptidylprolyl cis-and trans-isomerase activity and which is homologous to other FK506-and rapamycin-binding proteins, is overlined. All of the tyrosine residues are shown as white-on-black letters. Two tyrosine residues altered by site-directed mutagenesis (Tyr-184 and Tyr-189) are marked by the asterisks. The predicted 413-residue sequence of Fpr3 that we determined and have subsequently reconfirmed (18) (GenBank accession number L34569), differs from that reported for Fpr3/Npi46 by the laboratories of Mélèse and colleagues (34) (GenBank accession number X79379) and Movva and co-workers (35) by having two additional Glu residues (codons 241 and 242). was ϳ85% phosphoserine, ϳ11% phosphothreonine, and ϳ4% phosphotyrosine (values represent the mean of three independent experiments with S.E. of Յ1%). No phosphotyrosine was detectable in Fpr3 immunoprecipitated from PTP1 yeast. When the amount of radiolabel in Fpr3 was normalized to the amount of Fpr3 protein quantitated by staining, the stoichiometry of phosphate incorporated per mole of Fpr3 was ϳ3 mol of phosphoserine, 0.4 mol of phosphothreonine, and in the ptp1⌬ strain, 0.2 mol of phosphotyrosine.
A Protein-tyrosine Kinase Binds to and Phosphorylates Fpr3 in Vitro-To begin to characterize the protein kinase(s) responsible for phosphorylating Fpr3, a bacterially produced fusion protein consisting of GST and the amino-terminal 279 residues of Fpr3 (GST-Fpr3N) was adsorbed onto glutathione-Sepharose beads, washed, and incubated with yeast lysate. The beads were then washed exhaustively and incubated in a protein kinase reaction buffer containing [␥-32 P]ATP. Radiolabel was incorporated into GST-Fpr3N but not onto GST alone, suggesting that a protein kinase had bound to and phosphorylated the amino-terminal domain of the immobilized Fpr3 (data not shown). Phosphorylated GST-Fpr3N was subjected to phosphoamino acid analysis and was found to contain phosphotyrosine, phosphoserine, and phosphothreonine (Fig. 9, panel 3), indicating that one or more protein kinases capable of associating with and phosphorylating the amino-terminal domain of Fpr3 are present in yeast cell extracts. In addition, incubation of the phosphorylated GST-Fpr3N with Ptp1 prior to phosphoamino acid analysis led to the complete removal of phosphotyrosine but no significant change in levels of phosphoserine or phosphothreonine (data not shown), supporting previous findings (12) that Ptp1 is tyrosine specific.

DISCUSSION
The Substrate Specificity of Ptp1-We have shown here that in yeast cells lacking Ptp1, the nucleolar immunophilin Fpr3 and several unidentified proteins exhibit enhanced levels of tyrosine phosphorylation. In the case of Fpr3, we demonstrated that this protein is a direct substrate of Ptp1 in vitro. We also found that Ptp1 can substantially reverse phosphorylation of a  6, 12). The lysates were analyzed by immunoblotting with anti-phosphotyrosine mAb 4G10 (left panel) and anti-Fpr3 polyclonal antibodies (right panel). The samples in lanes 1-4 and 7-10 were separated on an 8.5% SDS-PAGE gel, and the samples in lanes 5, 6, 11, and 12 were separated on a 13% SDS-PAGE gel. Panel B, Tyr residue 184 is required for Fpr3 tyrosine phosphorylation. Protein extracts from strain YBB300 (ptp1⌬ fpr3⌬) expressing either wild-type (wt) FPR3 or FPR3 mutated at putative tyrosine phosphorylation sites were analyzed for relative levels of phosphotyrosyl Fpr3 by immunoblotting with anti-phosphotyrosine or with anti-Fpr3 antibodies. Wild-type or mutant FPR3 genes were expressed from the YEp351GAL plasmid. Lanes 1 and 5, wild-type FPR3; lanes 2 and 6, Y184F,Y189F double mutant; lanes 3 and 7, Y184F single mutant; lanes 4 and 8, Y189F single mutant. large number of yeast proteins phosphorylated at tyrosine by p60 v-src . These findings indicate that Ptp1 is a broad specificity PTP, similar to the mammalian enzyme PTP1B (37), which, when expressed in yeast (28), is also capable of dephosphorylating numerous proteins phosphorylated by p60 v-src . Additionally, when overexpressed in Schizosaccharomyces pombe, either S. cerevisiae PTP1 or mammalian PTP1B can complement a mutation in an endogenous PTP, Cdc25, and activate the cell cycle regulator, Cdc2 (14). In contrast to Ptp1, Ptp2 recognizes a very limited number of substrates. Disruption of S. cerevisiae PTP2 either in the presence or absence of v-src expression did not cause detectable increases in protein phosphotyrosine. These results are consistent with the previous observation that Ptp2 is unable to dephosphorylate artificial substrates in vitro (13).
The broad substrate specificity of Ptp1 suggests several possible functions for the enzyme. One extreme possibility is that Ptp1 may totally lack specificity for protein substrates, and function simply to reverse adventitious tyrosine phosphorylation by error-prone or promiscuous Tyr-specific or dual-specific protein kinases. Consistent with this idea, we observed that ptp1⌬ yeast were killed by mutants of the v-src tyrosine kinase that were only partially growth inhibitory in PTP1 strains. 3 However, other observations suggest that Ptp1 has some level of substrate specificity and thus that it may have a more specific role in yeast cell physiology. It is clear that Ptp1 is unable to fulfill the functional niches occupied by other PTPs in S. cerevisiae. The fact that cells carrying mutations in a PTPencoding gene, CDC14, undergo a cell cycle arrest (9) is evidence that, under normal conditions, Ptp1 cannot dephosphorylate the substrate(s) of Cdc14. Recent genetic evidence indicates that Ptp2 may function by dephosphorylating Hog1, the terminal MAP kinase of the osmosensory signaling pathway. The SLN1 gene encodes a histidine-protein kinase receptor that mediates this osmosensory pathway. Overexpression of PTP2 (but not of PTP1) rescues sln1 mutants, and normal expression of PTP2 (but not of PTP1) can compensate for a mutation in the functionally related phosphatase, Ptc1 (8,15). It will be of interest to determine whether the limitations on the activity of Ptp1 are a result of its subcellular localization (see below) or an inability to recognize and dephosphorylate certain phosphotyrosyl proteins.
Fpr3 Phosphorylation-The findings presented here indicate that the yeast nucleolar immunophilin Fpr3 is a substrate of Ptp1 in vivo. Phosphotyrosyl Fpr3 was detected in all of the ptp1⌬ strains and in none of the PTP1 strains deficient in Ptp2 or Mih1, indicating that Fpr3 is a physiological substrate of Ptp1 but not of other yeast PTPs.
At present, we do not know whether tyrosine phosphorylation affects Fpr3 function. Indeed, the precise cellular function of Fpr3 is unknown, but the properties of related mammalian and yeast immunophilins provide several clues. The peptidyl-prolyl isomerase activity of immunophilins suggests that they may catalyze protein folding (reviewed in Ref. 38). The immunosuppressant drugs, FK506 and rapamycin, which mimic the peptidyl-prolyl bond, bind to the FKBP class of immunophilins. In mammalian T-cells, the complex of drug and immunophilin blocks signal transduction. In yeast, exposure to FK506 inhibits calcineurin-mediated signal transduction and certain amino acid permeases (39), while exposure to rapamycin is lethal (40). Three FKBPs have been described in S. cerevisiae: Fpr1, Fpr2, and Fpr3. Fpr1, a homolog of the mammalian FKBP-12, is a cytosolic protein with high affinity for FK506 and rapamycin. FPR1-deficient yeast are resistant to these drugs, indicating that Fpr1 is largely responsible for mediating drug toxicity (41)(42)(43). S. cerevisiae FPR2, a homolog of mammalian FKBP-13, may be involved in the proper folding of proteins in the ER (44,45).
Fpr3 is an abundant nucleolar protein that is dispensable for growth. Yeast with disruptions in FPR3, including fpr1 fpr2 fpr3 strains, grow normally under a variety of growth conditions. The drug-binding and proline isomerase activities of Fpr3 are mediated by the conserved immunophilin domain, which represents the carboxyl-terminal third of Fpr3. When this domain is expressed independently, it is retained in the cytoplasm and restores FK506 and rapamycin sensitivity in fpr1⌬ strains. The amino-terminal two-thirds of Fpr3 contains striking regions of acidic and basic residues and is responsible for localization of Fpr3 in the nucleolus (18,34,35). This portion of Fpr3 exhibits some sequence similarity to nucleolin, a major nucleolar protein thought to be involved in ribosome assembly and shuttling of RNA or proteins through nuclear pores (for review, see Ref. 46).
Our findings also indicate that Fpr3 is tyrosine phosphorylated within the amino-terminal nucleolar localization domain. Preliminary immunofluorescence studies suggest that Ptp1 is localized primarily in the cytosol. 4 These observations raise the possibility that Fpr3 might be dephosphorylated by Ptp1 prior to its entry into the nucleus and that tyrosine phosphorylation and dephosphorylation of Fpr3 might regulate its subcellular localization. By dephosphorylating Fpr3, Ptp1 might also play a role in regulating the catalytic activity of Fpr3 and/or the ability of Fpr3 to associate with other proteins in the nucleolus. The Y184F mutant of Fpr3 does not undergo tyrosine phosphorylation and will serve as a useful reagent to study the possible effects of tyrosine phosphorylation/dephosphorylation on Fpr3 localization and catalytic activity.
The Fpr3 Tyrosine Kinase-Our results indicate that Fpr3 is only transiently tyrosine phosphorylated under normal growth conditions, since phosphotyrosyl-Fpr3 does not accumulate appreciably in PTP1 strains. Other S. cerevisiae proteins reported to contain phosphotyrosine include Cdc28 kinase, which is thought to be phosphorylated by Swe1, a S. cerevisiae homolog of S. pombe Wee1 (47), MAP kinases, which are phosphorylated by MAP kinase kinases (MEKs) (48,49), and dual-specific FIG. 9. Fpr3 is phosphorylated at Ser, Thr, and Tyr in vivo and in vitro. Panels 1 and 2, phosphoamino acid analysis of Fpr3 labeled in vivo. Strains YPH499 (PTP1) and YBB200 (ptp1⌬) expressing YEp351-FPR3myc were metabolically labeled with 32 P i for 3 h, harvested, and disrupted by alkaline lysis. Fpr3 was immunoprecipitated with rabbit anti-Fpr3 antibody and subjected to phosphoamino acid analysis. Autoradiography was carried out by exposure for 48 h in a Phosphorimager. Panel 3, phosphorylation of GST-Fpr3N in vitro. GST-Fpr3N adsorbed to glutathione-Sepharose beads was incubated with lysate from the protease-deficient ptp1⌬ strain YLW200, washed, and then incubated in the presence of [␥-32 P]ATP. The GST-Fpr3N was resolved by SDS-PAGE and subjected to phosphoamino acid analysis. Autoradiography was carried out by exposure for 24 h to x-ray film with an intensifying screen. In all three panels, position of phosphoamino acids detected by ninhydrin staining is marked by S (phosphoserine), T (phosphothreonine), or Y (phosphotyrosine). protein kinases such as Mck1 (50), Spk1 (51), and casein kinase I (52) that autophosphorylate on tyrosine. Thus, Fpr3 is the first tyrosine-phosphorylated protein identified in yeast that is not itself a protein kinase.
Because dedicated tyrosine-specific protein kinases have not been identified in unicellular eukaryotes, it is possible that a dual specificity kinase is responsible for the phosphorylation of Fpr3 at Tyr-184. Alternatively, the Fpr3 tyrosine kinase might be a novel yeast kinase that is tyrosine-specific. We have shown here that yeast extracts contain an activity or activities that phosphorylate Fpr3 at tyrosine, as well as at serine and threonine. Identification of the Fpr3 tyrosine kinase should allow us to distinguish between these possibilities.