The Meiosis-specific Protein Kinase Ime2 Directs Phosphorylation of Replication Protein A*

In Saccharomyces cerevisiae, the cellular single-stranded DNA-binding protein replication protein A (RPA) becomes phosphorylated during meiosis in two discrete reactions. The primary reaction is first observed shortly after cells enter the meiotic program and leads to phosphorylation of nearly all the detectable RPA. The secondary reaction, which requires the ATM/ATR homologue Mec1, is induced upon initiation of recombination and only modifies a fraction of the total RPA. We now report that correct timing of both RPA phosphorylation reactions requires Ime2, a meiosis-specific protein kinase that is critical for proper initiation of meiotic progression. Expression of Ime2 in vegetative cells leads to an unscheduled RPA phosphorylation reaction that does not require other tested meiosis-specific kinases and is distinct from the RPA phosphorylation reaction that normally occurs during mitotic growth. In addition, immunoprecipitated Ime2 catalyzes phosphorylation of purified RPA. Our data strongly suggest that Ime2 is an RPA kinase in vivo. We propose that Ime2 directly catalyzes RPA phosphorylation in the primary reaction and indirectly promotes the Mec1-dependent secondary reaction by advancing cells through meiotic progression. Our studies have identified a novel meiosis-specific reaction that targets a key protein required for DNA replication, repair, and recombination. This pathway could be important in differentiating mitotic and meiotic DNA metabolism.

In Saccharomyces cerevisiae, the cellular singlestranded DNA-binding protein replication protein A (RPA) becomes phosphorylated during meiosis in two discrete reactions. The primary reaction is first observed shortly after cells enter the meiotic program and leads to phosphorylation of nearly all the detectable RPA. The secondary reaction, which requires the ATM/ ATR homologue Mec1, is induced upon initiation of recombination and only modifies a fraction of the total RPA. We now report that correct timing of both RPA phosphorylation reactions requires Ime2, a meiosis-specific protein kinase that is critical for proper initiation of meiotic progression. Expression of Ime2 in vegetative cells leads to an unscheduled RPA phosphorylation reaction that does not require other tested meiosis-specific kinases and is distinct from the RPA phosphorylation reaction that normally occurs during mitotic growth. In addition, immunoprecipitated Ime2 catalyzes phosphorylation of purified RPA. Our data strongly suggest that Ime2 is an RPA kinase in vivo. We propose that Ime2 directly catalyzes RPA phosphorylation in the primary reaction and indirectly promotes the Mec1-dependent secondary reaction by advancing cells through meiotic progression. Our studies have identified a novel meiosis-specific reaction that targets a key protein required for DNA replication, repair, and recombination. This pathway could be important in differentiating mitotic and meiotic DNA metabolism.
Meiosis is a specialized process in which a single diploid cell undergoes one round of DNA replication followed by two consecutive rounds of nuclear division to generate four haploid progeny. For most species examined, the first nuclear division is directly preceded by a programmed recombination phase that is required for proper chromosome segregation and also serves to enhance genetic variability. In multicellular eukaryotes, this remarkable capacity to both rearrange chromosomal content and halve the chromosome number is restricted to a subset of cells destined to form gametes. However, singlecelled eukaryotes such as the budding yeast Saccharomyces cerevisiae can also undergo meiosis, leading to the production of haploid spores upon starvation. In either case, the reduction in ploidy resulting from meiosis is a prerequisite to zygote formation, whereupon the cellular DNA content is returned to a diploid state. Studies in a variety of systems have revealed that regulation of chromosome dynamics and maintenance of genomic integrity during meiosis require the induction of numerous proteins upon commitment of cells to meiotic differentiation. Among these are the meiosis-specific protein kinases that initiate and/or propagate critical meiotic signaling pathways.
In S. cerevisiae, the protein kinase Ime2 is an important regulator of meiotic initiation (1)(2)(3)(4)(5). Ime2 is induced early in meiosis and is required for maximal induction of early, middle, and late meiosis-specific genes. In the absence of Ime2 function, cells do not progress normally into "pre-meiotic" DNA replication and are deficient in completing downstream events in the sporulation process. Recent studies have demonstrated that Ime2 isolated by immunoprecipitation catalyzes phosphorylation of Ndt80 (6, 7), a meiosis-specific transcription factor that is required for up-regulation of middle sporulation genes (8 -10). Generation of phosphorylated Ndt80 correlates with maximal Ndt80 activation, suggesting that Ime2 enhances Ndt80 activity through phosphorylation (6,7). Another study has shown that Ime2 immunoprecipitates catalyze phosphorylation of Ime1 (11), an early meiosis-specific transcription factor (12,13). In this case, Ime2-mediated phosphorylation is thought to negatively regulate Ime1 by promoting its degradation (11). There is also evidence that Ime2 regulates proteins that are more directly associated with controlling meiotic progression, such as the B-type cyclin inhibitor Sic1 (14) and the anaphase-promoting complex ubiquitin ligase activator Cdh1 (15). Further studies will be required to determine whether Ime2 catalyzes phosphorylation of these proteins.
Given that Ime2 is required for proper pre-meiotic DNA replication, other potential targets of Ime2 protein kinase activity include proteins that are involved in the initiation of DNA synthesis. One candidate is replication protein A (RPA), 1 an abundant single-stranded DNA (ssDNA)-binding protein that is required for DNA replication, repair, and recombination (for review see Ref. 16). RPA is an evolutionarily conserved heterotrimeric complex, and each subunit is essential for viability in yeast (17,18). A fundamental role of RPA is to stabilize the ssDNA that is generated during DNA transactions. RPA also interacts with several other proteins required for DNA replication, repair, and recombination. Consequently, RPA is likely to play an important regulatory role in DNA metabolism. Consistent with this hypothesis, RPA is phosphorylated peri-odically during the mitotic cell cycle and also in response to genotoxic stress in both human and yeast cells (19 -24).
Our recent studies have demonstrated that yeast RPA also becomes phosphorylated in at least two separate reactions during meiotic progression (25). The primary RPA phosphorylation reaction first occurs soon after cells enter meiosis and would appear to coincide temporally with Ime2 induction. The secondary reaction occurs upon initiation of meiotic recombination and requires the protein kinase Mec1, an important regulator of DNA metabolic checkpoints (26 -30) that is also required for RPA phosphorylation events in vegetative cells (23,24). Mec1 is homologous to the human protein kinase ATM, which is mutated in ataxia-telangiectasia (31), and is even more homologous to the human "ATM-and Rad3-related" protein kinase ATR (for review see Ref. 32). Despite the participation of Mec1, our data have suggested that the secondary RPA phosphorylation reaction is not involved in meiotic Mec1-dependent checkpoint delay processes. Instead, it is likely that Mec1-dependent RPA phosphorylation functions to regulate recombination itself. The possibility that phosphorylation modulates RPA function in DNA metabolism is supported by earlier studies employing the cell-free SV40 DNA replication system. These experiments have suggested that certain forms of phosphorylated human RPA are inefficient in supporting DNA replication and might preferentially function in DNA repair (22,33,34).
In order to further understand the function of RPA phosphorylation, we have investigated the mechanism underlying the sequential RPA phosphorylation reactions that occur during meiotic progression. We now demonstrate that proper timing of both primary and secondary meiotic RPA phosphorylation requires Ime2. Our genetic and biochemical evidence suggests that the initial meiosis-specific reaction is directly catalyzed by Ime2. The secondary Mec1-dependent reaction is indirectly activated by the Ime2-dependent promotion of meiotic progression. Nearly all of the detectable RPA becomes phosphorylated in the primary Ime2-catalyzed reaction, suggesting that this novel meiotic pathway plays a significant role in directing the meiosis-specific DNA metabolic functions of RPA.
Ime2 Plasmids-PCR-based cloning methods were employed to generate plasmids expressing C-terminally tagged versions of Ime2. A sequence encoding six histidine repeats followed by three hemagglutinin epitope (Ha) repeats was inserted at the NotI site of pRS426 (39) to generate pRS426-His/Ha. Oligonucleotide primers designed to engineer an XmaI site upstream of the IME2 promoter and a NotI site immediately prior to the IME2 stop codon were used to amplify the entire IME2 gene from pAM405 (1), generously provided by Craig Giroux (Wayne State University). The PCR fragment was inserted directly into pGEM-T (Promega) by "TA" cloning. DNA sequencing revealed two PCR-generated mutations that were corrected by replacement with a restriction fragment from pAM405. IME2 was then excised from the pGEM-T clone with XmaI and NotI and inserted between the XmaI and NotI sites of pRS426-His/Ha. The resulting plasmid, pDC005, encodes a C-terminally tagged version of Ime2 designated Ime2-His/Ha. The plasmid pDC006, which encodes a kinase-dead mutant of Ime2 (Ime2 kd -His/ Ha) resulting from a lysine-to-alanine mutation at residue 97 (11), was generated from pDC005 by site-directed mutagenesis (QuikChange, Stratagene). Yeast cells were transformed with plasmids using a lithium acetate-based procedure (40).
Media, Growth, and Sporulation-Complex medium containing glucose (YPD) or glycerol (YPG) and synthetic complete medium lacking uracil were prepared as described (41). Sporulation medium (SPM) consisted of 0.3% potassium acetate and 0.02% raffinose supplemented with leucine at 0.0025% and arginine, histidine, tryptophan, and uracil each at 0.0005%. All incubations were carried out at 30°C. SK-1 cells were synchronously sporulated as described (25). Briefly, single colonies were selected following growth on YPG and incubated in YPD overnight. The overnight cultures were used to inoculate 200 ml of YPA (1% yeast extract, 2% bactopeptone, 2% potassium acetate) at an A 600 of 0.2 for pre-sporulation growth. After 12 h of incubation, pre-sporulation cells were washed once with 20 ml of SPM, resuspended in 200 ml of SPM, and further incubated for the indicated times.
Western Blot Analysis-To prepare denatured extracts, cells that had been washed with water and stored at Ϫ80°C were resuspended in sample buffer (62.5 mM Tris, pH 6.8, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.001% bromphenol blue) and subjected to three cycles of freezing (dry ice for 2 min) followed by heating (95°C for 2 min). The suspensions were then centrifuged at 18,000 ϫ g for 5 min at room temperature to pellet cell debris, and aliquots of the supernatants were further analyzed as described below. Immunoprecipitated and purified proteins (see below) were dissolved in sample buffer, heated at 95°C for 10 min, and centrifuged at 18,000 ϫ g for 30 s. To allow for reproducible detection of immunoprecipitated Ime2 by Western blot analysis, 20 g of carrier protein (yeast crude extract devoid of Ime2-His/Ha) was added to the Ime2 immunoprecipitates. Samples to be analyzed for Rfa2 phosphorylation were subjected to electrophoresis through a 12% low cross-linking SDS-polyacrylamide gel (150:1 w/w acrylamide/bisacrylamide). For analyses not requiring such resolution, 10% SDS-polyacrylamide gels (37.5:1 w/w acrylamide/bisacrylamide) were employed. Separated proteins were transferred to nitrocellulose (0.2-m pore size, Schleicher & Schuell) in Western buffer (25 mM Tris, 192 mM glycine, 20% methanol). RPA subunits were detected after incubation with anti-Rfa1 (24), anti-Rfa2 (24), or affinity-purified anti-Rfa3 polyclonal antibodies followed by horseradish peroxidase-linked goat anti-rabbit antibody (Pierce). Antiserum directed against Rfa3 as well as the affinity-purified preparation and pre-immune serum were generously provided by Steven Brill (Rutgers University). Ime2-His/Ha was detected after incubation with monoclonal antibody HA.11 (Covance) followed by horseradish peroxidase-linked goat anti-mouse antibody (Pierce). Protein bands were visualized by autoradiography using chemiluminescence reagents (Pierce).
DNA Content Analysis-Cells were analyzed for DNA content by flow cytometry after staining with SYBR Green I (Molecular Probes) as described previously (25). DNA content histograms were generated using WinMDI software.
Immunoprecipitation-Stationary phase cells were harvested by centrifugation, resuspended in 0.1 M EDTA, 10 mM dithiothreitol (DTT) (2.5 ml/g cells), and incubated at 30°C for 10 min. The cells were harvested by centrifugation again, resuspended in YPS (1% yeast extract, 2% bactopeptone, 1 M sorbitol) containing 200 g/ml Zymolyase 100T (1 ml/g cells; U. S. Biochemical Corp.), and incubated at 30°C for 2.5 h. The resulting spheroplasts were collected, washed twice with YPS, and then resuspended in lysis buffer (50 mM Hepes, pH 7.4, 100 mM KCl, 0.2% Tween 20, 1 mM DTT, 0.1 mM EDTA, 25 mM NaF, 100 M sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 1 g/ml leupeptin). Resulting suspensions were incubated on ice for 30 min and then centrifuged at 12,000 ϫ g for 15 min at 4°C. Supernatants were collected and stored at Ϫ80°C. For immunoprecipitation, lysate was incubated on ice for 2 h with antibody (HA.11, anti-Rfa1, or anti-Rfa3) or the respective control (no antibody, Rfa1 pre-immune serum, or Rfa3 pre-immune serum). Immune complexes were collected on EZview Red Protein G Affinity Gel (Sigma) by incubation at 4°C for 2 h with gentle shaking and were subsequently washed once with lysis buffer containing 0.5 M NaCl and twice with lysis buffer alone.
RPA Purification-RPA was purified from exponentially growing W303-1A (wild type) and YGB23 (rfa2-S122D) cells and from synchronously sporulating DSY1089 cells (wild type; 6-h time point). Purification over three columns (Affi-Gel Blue, ssDNA cellulose, and Mono Q) was performed based on methods described previously (19,42). Protein concentration was determined by the method of Bradford (43) using bovine serum albumin as a standard. Relative levels of Rfa2 in the W303-1A and YGB23 RPA preparations were determined from a silverstained gel (Fig. 4C) using a Kodak Digital Science Image Station 440 CF equipped with 1D Image Analysis Software, version 3.5.3.
Protein Kinase Assay-HA.11 immunoprecipitates were washed two times with kinase buffer (10 mM Hepes, pH 7.4, 10 mM MgCl 2 , 1 mM DTT) and then assayed for protein kinase activity in kinase buffer containing 10 M ATP and 2.5 Ci of [␥Ϫ 32 P]ATP (25 l final volume). Purified RPA (5.3 pmol) and M13mp18 single-or double-stranded DNA (17 fmol; New England Biolabs) were added where indicated. Reaction mixtures were incubated at 30°C for 40 min and then subjected to 15% SDS-PAGE. Gels were dried either directly or after silver staining (Fig.  4C), and incorporation of radioactive inorganic phosphate into protein was detected by autoradiography and Storm PhosphorImager (Amersham Biosciences) analysis.

RESULTS
Meiotic RPA Phosphorylation Requires Ime2-Our previous studies (25) revealed that the middle subunit of yeast RPA (Rfa2) becomes phosphorylated in two separate reactions during meiotic progression. Because the primary reaction is induced early in meiosis when Ime2 is also known to be induced, we investigated Rfa2 status in wild type and ime2 mutant cells subjected to starvation conditions that induce entry into the meiotic program. As observed previously, two Rfa2 phosphoisomers were detected by Western blot analysis in wild type cells (Fig. 1A). A primary phosphoisomer was first observed early in meiosis prior to the bulk of pre-meiotic DNA replication (see Fig. 1B) and persisted throughout most of meiotic progression. A secondary hyperphosphorylated phosphoisomer, comprising only a small fraction of the total Rfa2, was induced later in sporulation. We demonstrated previously that generation of this phosphoisomer requires both Mec1 and initiation of recombination (25). Deletion of IME2 abolished nearly all meiotic Rfa2 phosphorylation as determined by Rfa2 electrophoretic mobility (Fig. 1A). Consistent with previous reports (1,3,5), the ime2 mutant cells did not progress normally through pre-meiotic DNA replication, exhibiting a significant delay in the onset of DNA synthesis (Fig. 1B). In addition, these cells were incapable of forming mature asci at later time points (data not shown).
Genetic studies suggest that Ime2 regulates pre-meiotic DNA replication by promoting phosphorylation and degradation of the B-type cyclin inhibitor Sic1 (14). To determine whether Ime2 indirectly activates meiotic Rfa2 phosphorylation by relieving Sic1-mediated inhibition of a cyclin-dependent kinase, we examined sic1 mutant and ime2 sic1 double mutant cells induced to enter meiosis. Deletion of SIC1 led to an accumulation of G 2 cells that were incapable of entering the sporulation program (Fig. 2B) (14). As a result, a residual mitotic Rfa2 phosphoisomer was detected in sic1 and ime2 sic1 cells throughout the time course ( Fig. 2A). In sic1 cells, two Rfa2 phosphoisomers were observed, whereas in ime2 sic1 cells, a single Rfa2 phosphoisomer was observed that did not appear to change in abundance relative to unphosphorylated Rfa2 during the time course. Therefore, deletion of SIC1 did not rescue the defect in Rfa2 phosphorylation conferred by deletion of IME2, arguing against the involvement of a B-type cyclin/cyclin-dependent kinase complex in catalyzing meiotic Rfa2 phosphorylation. It should be further noted that our ime2 sic1 double mutant did not appear to progress though meiosis as judged by flow cytometry (Fig. 2B) or formation of mature asci (data not shown). These results do not support the contention that a sic1 deletion can rescue the defect in ime2 cells that leads to defective meiotic progression. Ime2 Expression in Vegetative Cells Promotes RPA Phosphorylation-We further investigated the potential role of Ime2 as an RPA kinase by introducing IME2 into cells on a high copy 2-m vector to allow for inappropriate mitotic expression of the meiosis-specific protein kinase. Cells expressing a His/Hatagged version of Ime2 (Ime2-His/Ha) exhibited two Rfa2 phosphoisomers during exponential growth, whereas cells expressing the kinase-dead version of tagged Ime2 (Ime2 kd -His/Ha) exhibited the single Rfa2 phosphoisomer typical of mitotic cells (Fig. 3A, bottom panel, lanes 1 and 2). Expression of both forms of Ime2 was verified by Western blot analysis with antibody directed against Ha (Fig. 3A, top panel, lanes 1 and 2). The wild type Ime2 appeared to have slightly slower mobility than the kinase-dead version, possibly due to Ime2 autophosphorylation (see below).
Our previous studies (23) revealed that Rfa2 phosphorylation observed during normal proliferation requires Mec1, and recent mapping studies have localized the site of Mec1-dependent phosphorylation to serine 122. 2 To examine ectopic Rfa2 phosphorylation in the absence of Mec1-dependent Rfa2 phosphorylation, we employed a strain containing a serine-to-alanine mutation at Rfa2 residue 122 (S122A). Whereas expression of wild type Ime2-His/Ha in this mutant strain led to the  generation of a single Rfa2 phosphoisomer, no Rfa2 phosphorylation was observed when the kinase-deficient Ime2 kd -His/Ha was expressed (Fig. 3A, lower panel, lanes 3 and 4). These data revealed that a significant fraction of the Rfa2 had become phosphorylated in an Ime2-dependent pathway. Despite the induction of this unnatural Rfa2 phosphorylation reaction, no obvious effects on DNA replication or general cell cycle progression were evoked as measured by flow cytometry (Fig. 3B).
The Ime2-dependent induction of Rfa2 phosphorylation during mitotic growth suggests that Ime2 directly catalyzes the reaction. However, expression of Ime2 during mitosis might cause cells to enter a "meiosis-like" state, leading to induction of another meiosis-specific protein kinase that is responsible for catalyzing Rfa2 phosphorylation. In fact, it has been shown previously that Ime2 expression during mitosis induces Spo11mediated homologous recombination (1), which is normally observed only during meiosis. To investigate the possibility that ectopic Ime2 expression indirectly activates Rfa2 phosphorylation, we examined the effect of mitotic Ime2 expression on Rfa2 phosphorylation in a panel of strains deleted of genes encoding meiosis-specific protein kinases. Because stationary phase cells were largely devoid of Rfa2 phosphorylated in the Mec1-dependent reaction but exhibited Ime2-induced Rfa2 phosphorylation (Fig. 3C), we chose to examine the deletion mutants in stationary phase. We tested deletions of MEK1, RIM15, SPS1, and SMK1, which encode protein kinases that operate at various stages during meiotic progression, and we observed no obvious defects in Ime2-induced Rfa2 phosphorylation (Fig. 3D). Therefore, these protein kinases are not required for ectopic Ime2-dependent Rfa2 phosphorylation and, by extension, are not likely to be responsible for Rfa2 phosphorylation during normal meiotic progression.
RPA Is an Ime2 Substrate in Vitro-Our genetic studies clearly indicate that meiosis-specific RPA phosphorylation requires Ime2 protein kinase activity. To examine whether Ime2 can catalyze phosphorylation of RPA, we assayed the in vitro kinase activity of immunoprecipitated Ime2-His/Ha using purified RPA as a substrate. Ime2-His/Ha and Ime2 kd -His/Ha were expressed in vegetative ime2⌬ cells and immunoprecipitated with antibody directed against Ha. Western blot analysis confirmed that precipitation of the tagged versions of Ime2 required the presence of the Ha antibody (Fig. 4A). RPA kinase activity was not evident in mock immunoprecipitates incubated with or without substrate (Fig. 4B, lanes 1, 2, 5, and 6). In contrast, immunoprecipitates of Ime2 catalyzed phosphorylation of Rfa2 and a protein that was likely to be Ime2 based on its migration and previous reports of Ime2 autophosphorylation (4,6,7,11,44) (Fig. 4B, lanes 3 and 4). Replacement of Ime2-His/Ha with Ime2 kd -His/Ha dramatically reduced both Rfa2 phosphorylation and the putative Ime2 autophosphorylation (Fig. 4B, lanes 7 and 8), suggesting that Ime2 directly catalyzes Rfa2 phosphorylation in vitro.
Our use of the intact RPA heterotrimer allowed us to determine whether phosphorylation of the large subunit (Rfa1) or small subunit (Rfa3) is also catalyzed through an Ime2-dependent reaction in vitro. We have not detected phosphorylation of either subunit (Fig. 4, B-D), consistent with a specific targeting of Rfa2 by Ime2 and consistent with our studies in vivo, which have thus far revealed phosphorylation of Rfa2 but not Rfa1 during normal meiotic progression (25). To further assess the specificity of the in vitro reaction, we employed purified RPA containing a serine-to-aspartate mutation at Rfa2 residue 122 (S122D). As shown in Fig. 4C, this mutation had little effect on incorporation of phosphate into RPA, indicating that the Ime2dependent activity in vitro mainly targets a residue that is not normally phosphorylated in a Mec1-dependent manner in vivo. 2 Because RPA is an ssDNA-binding protein and addition of DNA stimulates RPA phosphorylation catalyzed by immunoprecipitated Mec1, 2 we tested the effect of DNA on Ime2-mediated RPA phosphorylation in vitro. We found that neither single-nor double-stranded DNA stimulated the reaction under conditions that would significantly stimulate Mec1-mediated RPA phosphorylation (Fig. 4D and data not shown). In fact, quantification of these data revealed a minor decrease in both Rfa2 phosphorylation and Ime2 autophosphorylation upon addition of either DNA. It is important to note that 0.03 mol of phosphate were incorporated per mol of Rfa2 in the reaction devoid of DNA (Fig. 4D, lane 2), indicating that stimulation hypothetically could have been observed under these conditions.
Phosphorylated RPA Binds ssDNA and Maintains a Heterotrimeric Structure-In order to determine whether Ime2-mediated phosphorylation modifies the general biochemical characteristics of RPA, we elected to purify RPA from wild type cells undergoing meiosis. We employed a protocol similar to the one that we use for purification of RPA from mitotic cells, which includes ssDNA-cellulose chromatography. We found that hyperphosphorylated Rfa2 was present after each stage in the purification, including the final Mono Q chromatography step (Fig. 5A, middle panel, fractions 13-15), indicating that Rfa2 phosphorylated in the Ime2-dependent reaction was capable of binding ssDNA either independently or as a part of the RPA complex.
Examination of the elution profiles of Rfa1 and Rfa2 revealed that full-length Rfa1 and the phosphorylated isomers of Rfa2 did not co-elute from the Mono Q column (Fig. 5A). Whereas the majority of full-length Rfa1 was contained in fraction 12, little if any phosphorylated Rfa2 was present in this fraction. In contrast, the majority of full-length phosphorylated Rfa2 was found in fraction 14, which contained only a small percentage of full-length Rfa1. These data might imply that Ime2-mediated phosphorylation disrupts the RPA heterotrimeric complex. However, further analysis indicated that both Rfa1 and Rfa2 became significantly degraded during the course of RPA purification (Fig. 5A, upper and lower panels), thereby confounding interpretation of the Mono Q data.
To investigate further whether Ime2-dependent Rfa2 phosphorylation affects the physical structure of the RPA complex, we immunoprecipitated the large and small subunits from wild type vegetative cells expressing the tagged versions of Ime2, and we determined the phosphorylation status of bound Rfa2 by Western blot analysis. Rfa1 immunoprecipitates from Ime2-His/Ha expressing cells contained phosphorylated Rfa2 (Fig.  5B, lane 2). As expected, Rfa1 immunoprecipitates from wild type cells expressing Ime2 kd -His/Ha contained unphosphorylated Rfa2 (Fig. 5B, lane 4). Examination of the supernatants from these immunoprecipitations indicated that nearly all of the Rfa2 was precipitated with Rfa1 ( Fig. 5B, lanes 6 and 8). A similar experiment revealed that Rfa3 immunoprecipitates contained Rfa2 phosphorylated specifically in the Ime2dependent reaction (Fig. 5C). These results indicate that Ime2mediated phosphorylation of Rfa2 does not interfere with the canonical heterotrimeric structure of RPA.

DISCUSSION
Proper meiotic progression and avoidance of germ line mutations relies on the induction of meiosis-specific proteins that coordinate DNA replication, recombination, and chromosome segregation. In budding yeast, the meiosis-specific protein kinase Ime2 is a key regulator of meiotic differentiation. Ime2 has been studied largely in the context of transcriptional induction; however, several lines of evidence now indicate that Ime2 function is not limited to its effect on transcription factors. The results presented here indicate that RPA, a critical protein in DNA replication, repair, and recombination, is directly targeted by Ime2. We suspect that this Ime2-mediated pathway plays an important regulatory role in meiotic DNA metabolism.
RPA phosphorylation during meiosis occurs in two stages (see Fig. 6). The primary reaction first occurs early in the meiotic program before extensive pre-meiotic DNA synthesis has occurred. The experiments reported here strongly suggest that Ime2 directly catalyzes this reaction. Addition of DNA does not have a sizable effect on Ime2-mediated RPA phospho-FIG. 3. Ime2 expression during mitosis leads to Rfa2 phosphorylation. A, Western blot analysis of exponentially growing wild type (wt) and Rfa2 phosphorylation site mutant (S122A) cells expressing Ime2-His/Ha (wt) or its kinase-dead counterpart Ime2 kd -His/Ha (kd). B, DNA content analysis of exponential cultures examined in A. C, Western blot analysis of ime2⌬ cells harboring vector alone (Ϫ), vector containing the gene encoding Ime2-His/Ha (wt), or vector containing the gene encoding Ime2 kd -His/Ha (kd) in exponentially growing and stationary phase cells. D, Western blot analysis of stationary phase ime2⌬, mek1⌬, rim15⌬, sps1⌬, and smk1⌬ cells (Invitrogen versions) harboring the gene encoding Ime2-His/Ha (wt) or Ime2 kd -His/Ha (kd). P-Rfa2, phosphorylated Rfa2. rylation in vitro, suggesting that the reaction does not require DNA-bound Ime2 or RPA. Although we cannot rule out the possibility that endogenous DNA in the Ime2 preparation supports the reaction, we propose that primary RPA phosphorylation is simply regulated by induction of Ime2 and does not depend on generation of specific DNA structures. In support of this conclusion, primary RPA phosphorylation proceeds to completion in clb5 clb6 cells (25), which are defective for premeiotic DNA replication (14,30). It is noteworthy that the Rfa2 phosphoisomer resulting from primary RPA phosphorylation is present throughout most of meiosis, suggesting either that this form of phosphorylated RPA is stable once generated or that Ime2-dependent RPA phosphorylation continues to occur during meiotic progression. Therefore, it is possible that Ime2 catalyzes phosphorylation of either free or DNA-bound RPA in vivo, as indicated in Fig. 6.
The second meiotic RPA phosphorylation event coincides with the programmed recombination phase that is initiated by generation of DNA double-strand breaks. As shown here, the correct timing of this reaction also requires Ime2, an indirect effect that results from the dependence of proper meiotic progression on Ime2 function. We have reported previously that the secondary reaction depends on Mec1 and Spo11 (25), the enzyme that catalyzes DNA double-strand break formation (45). Our recent studies have revealed that DNA significantly stimulates Mec1-mediated RPA phosphorylation in vitro, an effect that appears to be due, at least in part, to RPA-DNA binding. 2 Therefore, in contrast to the primary Ime2-catalyzed reaction, the secondary Mec1-catalyzed reaction is likely to be activated by DNA structures. There is evidence that the Mec1-interacting protein Ddc2 (46) (also known as Lcd1 (47) and Pie1 (48)) guides Mec1 to sites of DNA damage (49), possibly through a direct interaction between Ddc2 and RPA (50). Such a mechanism could lead to Mec1-dependent RPA phosphorylation during meiosis (see Fig. 6). Given that Mec1-mediated RPA phosphorylation occurs during mitosis in the absence of Ime2-mediated Rfa2 phosphorylation, it is unlikely that the secondary meiotic RPA phosphorylation reaction is dependent on the primary meiotic RPA phosphorylation reaction. In this context, it should be added that Ime2-catalyzed Rfa2 phosphorylation might continue to occur after Mec1-catalyzed Rfa2 phosphorylation, leading to rephosphorylation of RPA molecules that have become dephosphorylated at the primary site. Efficient Ime2-dependent phosphorylation of Rfa2 containing a phosphomimetic mutation at residue 122 (Fig. 4C) suggests that Mec1-dependent RPA phosphorylation does not preclude subsequent Ime2-dependent RPA phosphorylation.
Defining the exact functions of meiotic RPA phosphorylation will likely require identification and mutation of the Rfa2 residues that are phosphorylated in both the primary and secondary reactions. Nonetheless, insight can be gained by examining the characteristics of the two reactions. We noted in our previous study that nearly all of the detectable RPA (estimated in humans to be 10 4 -10 5 molecules per cell (16)) becomes phosphorylated in the meiosis-specific reaction (25). Based on the abundance and longevity of the primary RPA phosphoisomer and its specific generation during meiosis, we suggest that Ime2-catalyzed phosphorylation switches RPA from a mitotic to a meiotic protein. This modification could be important in directing RPA-associated activities that are peculiar to meiosis.

FIG. 4. RPA is a substrate of Ime2 in vitro.
Immunoprecipitates of Ime2-His/Ha or Ime2 kd -His/Ha were assayed for RPA kinase activity in vitro. A, Western blot analysis of Ha immunoprecipitates from ime2⌬ cells expressing Ime2-His/Ha (wt) or Ime2 kd -His/Ha (kd). B, autoradiograph exhibiting RPA kinase assay employing immunoprecipitates analyzed in A. C, protein kinase assay employing wild type (wt) and mutant (S122D) RPA. The relative levels of Rfa2 (arbitrary units) in the two RPA preparations (silver) and the amount of P i incorporated (autorad) were used to calculate the efficiency of Rfa2-S122D as an Ime2 substrate (table). D, autoradiograph exhibiting RPA kinase activity in the absence or presence of DNA. Equimolar concentrations of single-stranded (ss) and double-stranded (ds) M13mp18 DNA were employed. Ha Ab, antibody directed against the Ha epitope present in the tagged versions of Ime2.
Candidate events that could be affected by RPA phosphorylation include DNA replication, recombination, or a process such as chromosome segregation that is functionally linked to DNA replication (for review see Ref. 51). Each of these fundamental activities is regulated differently in mitosis and meiosis. For example, pre-meiotic DNA replication proceeds slower than mitotic DNA replication (52) and is tightly coupled to the subsequent programmed meiotic recombination phase (53)(54)(55). Homologous recombination during meiosis differs from homologous recombination during mitosis in several ways (56), including a preference for exchange between homologues rather than sister chromatids (57). Finally, meiotic division progresses through two stages, an initial reductional segregation of chromosomes that is specific to meiosis and a subsequent equational segregation, whereas mitotic division includes only an equational segregation. Although the current evidence would suggest that Ime2 regulates meiotic chromosome segregation through activation of Ndt80 (6, 7), it is possible that other effectors are also involved.
In contrast to the Ime2-catalyzed RPA phosphoisomer, the Mec1-catalyzed RPA phosphoisomer represents only a fraction of the total RPA and is relatively short lived. The dependence of the meiotic Mec1-mediated reaction on the initiation of recombination and the disappearance of the resulting RPA phosphoisomer upon completion of the recombination phase implies that Mec1-dependent RPA phosphorylation functions to regulate recombination. It is possible that the similar Mec1-dependent RPA phosphorylation reaction occurring during normal mitotic progression also regulates recombination. However, we emphasize that the Ime2-mediated RPA phosphorylation reaction first occurs prior to the Mec1-mediated reaction in meiosis and could influence the effect of the Mec1-generated RPA phosphoisomer. As a result, Mec1-mediated RPA phosphorylation might have significantly different effects on recombination depending on whether cells are in mitosis or meiosis.
The mechanisms by which the two meiotic RPA phosphorylation reactions might influence DNA metabolic processes include alterations in RPA-DNA binding, RPA-protein interac- FIG. 5. Phosphorylated RPA interacts with ssDNA and maintains intersubunit interactions. A, RPA was purified from meiotic cells by Affi-gel Blue, ssDNA-cellulose, and Mono Q chromatography. Western blot analysis of Mono Q eluate using antibodies directed against Rfa1 (upper panel) and Rfa2 (middle and lower panels). Proteins were separated by low cross-linking denaturing gel electrophoresis for resolution of RPA isomers (upper and middle panels) or by conventional denaturing gel electrophoresis (lower panel). B, Western blot analysis of Rfa1 and Rfa2 in the pellet (lanes 1-4) or supernatant (lanes 5-8) following immunoprecipitation with Rfa1 pre-immune (Ϫ) or immune (ϩ) serum. Immunoprecipitations were conducted using crude extract of ime2⌬ cells expressing Ime2-His/Ha (wt) or Ime2 kd -His/Ha (kd). C, Western blot analysis of Rfa2 and Rfa3 in the pellet following immunoprecipitation with Rfa3 pre-immune (Ϫ) or immune (ϩ) serum. Immunoprecipitations were conducted using crude extract of ime2⌬ cells expressing Ime2-His/Ha (wt) or Ime2 kd -His/Ha (kd). P-Rfa2, phosphorylated Rfa2; *, apparent RPA degradation products; Rfa1 Ab, antibody directed against Rfa1; Rfa3 Ab, antibody directed against Rfa3. tions, or both. Although we have found that the RPA phosphoisomers generated during meiosis are still capable of binding ssDNA, further detailed analysis will be required to determine whether Ime2-catalyzed or Mec1-catalyzed Rfa2 phosphorylation affects the affinity or specificity of RPA-DNA binding. A recent report (58) has indicated that a human RPA phosphoisomer isolated from the M phase of the normal mitotic cell cycle has lower affinity for double-stranded DNA and for certain proteins, most notably DNA polymerase ␣. In addition, a phosphomimetic mutant of human RPA exhibits a weakened interaction with double-stranded DNA (59). These studies provide evidence for modulation of RPA-macromolecule associations upon RPA phosphorylation, effects that could have a direct impact on DNA replication, repair, or recombination. Although we would offer that phosphorylation of RPA most likely affects DNA metabolism, there is evidence that a Mec1independent Rfa2 phosphorylation reaction leads to transcriptional induction of certain DNA repair genes (60). Therefore, it is conceivable that one or both meiotic Rfa2 phosphorylation reactions regulate DNA metabolism indirectly or even target a separate RPA-mediated process. Further studies will be aimed at distinguishing between these various possibilities and precisely defining the role of meiotic RPA phosphorylation.