A New Recombinational DNA Repair Gene From Schizosaccharomyces pombe With Homology to Escherichia coli RecA

Corresponding author: Wolf-Dietrich Heyer, Section of Microbiology, University of California, Davis, 1 Shields Ave., Davis, CA 95616., wdheyer{at}ucdavis.edu (E-mail)

Communicating editor: L. S. SYMINGTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A new DNA repair gene from Schizosaccharomyces pombe with homology to RecA was identified and characterized. Comparative analysis showed highest similarity to Saccharomyces cerevisiae Rad55p. rhp55⁺ (rad homologue pombe 55) encodes a predicted 350-amino-acid protein with an M_r of 38,000. The rhp55 mutant was highly sensitive to methyl methanesulfonate (MMS), ionizing radiation (IR), and, to a lesser degree, UV. These phenotypes were enhanced at low temperatures, similar to deletions in the S. cerevisiae RAD55 and RAD57 genes. Many rhp55 cells were elongated with aberrant nuclei and an increased DNA content. The rhp55 mutant showed minor deficiencies in meiotic intra- and intergenic recombination. Sporulation efficiency and spore viability were significantly reduced. Double-mutant analysis showed that rhp55⁺ acts in one DNA repair pathway with rhp51⁺ and rhp54⁺, homologs of the budding yeast RAD51 and RAD54 genes, respectively. However, rhp55⁺ is in a different epistasis group for repair of UV-, MMS-, or -ray-induced DNA damage than is rad22⁺, a putative RAD52 homolog of fission yeast. The structural and functional similarity suggests that rhp55⁺ is a homolog of the S. cerevisiae RAD55 gene and we propose that the functional diversification of RecA-like genes in budding yeast is evolutionarily conserved.

DOUBLE-strand breaks (DSBs) are an important genotoxic lesion caused by free radicals during cellular metabolism, by various chemical agents, and by ionizing radiation (IR). Moreover, DSBs appear as an intermediate during meiotic recombination in Saccharomyces cerevisiae (FRIEDBERG et al. 1995 ). These lesions can be healed by several DNA repair pathways that evolved in eukaryotes and were found in yeasts and mammals. Among these, homologous recombination, single-strand annealing (SSA), and nonhomologous DNA end-joining provide the main mechanisms for DSB repair (HABER 1992 ; KANAAR et al. 1998 ). Homologous recombination uses the information from the undamaged sister chromatid or homolog to repair the DSB. It constitutes the only intrinsically error-free mechanism for DSB repair (RESNICK 1976 ; SZOSTAK et al. 1983 ). In contrast, SSA and DNA end-joining repair DNA ends in an intrinsically error-prone manner, inducing mutations during the process (LIN et al. 1984 ; ROTH and WILSON 1988 ). The impact of the different mechanisms in DNA repair could differ between yeasts and mammals (KANAAR et al. 1998 ). Yeasts such as Schizosaccharomyces pombe or S. cerevisiae preferentially use homologous recombination for DSB repair to achieve high fidelity (HABER 1992 ). One reason could be that these organisms have genomes densely packed with genes that would result in a direct negative impact of error-prone repair (KANAAR et al. 1998 ).

The study of DSB repair in the budding yeast S. cerevisiae established that the RAD52 epistasis group of genes is involved in DNA repair by homologous recombination. The products of the RAD52 group genes are concomitantly involved in DSB repair as well as in mitotic and meiotic recombination (PETES et al. 1991 ; GAME 1993 ). At present 10 genes have been identified in this group: RAD50, MRE11, XRS2, RAD51, RAD52, RAD54, RAD55, RAD57, RAD59, and RFA1 (PETES et al. 1991 ; GAME 1993 ; FRIEDBERG et al. 1995 ; BAI and SYMINGTON 1996 ; HAYS et al. 1998 ). Among them, RAD50, MRE11, and XRS2 were shown to be involved both in recombinational repair and DNA end-joining (SCHIESTL et al. 1994 ; MILNE et al. 1996 ; MOORE and HABER 1996 ). RAD51, RAD52, and RAD54 are the most important genes of the group as judged by the extreme sensitivity to IR of the respective mutants (GAME 1993 ). Genetic and biochemical experiments indicated that Rad51p can interact with both proteins (SHINOHARA et al. 1992 ; MILNE and WEAVER 1993 ; DONOVAN et al. 1994 ; SCHILD 1995 ; JIANG et al. 1996 ; CLEVER et al. 1997 ).

The RAD51, RAD55, and RAD57 genes encode proteins with sequence similarity to the E. coli key recombination protein RecA (KANS and MORTIMER 1991 ; ABOUSSEKHRA et al. 1992 ; BASILE et al. 1992 ; SHINOHARA et al. 1992 ; LOVETT 1994 ). However, the three proteins do not perform redundant functions, because the individual mutation in any of the three genes causes overlapping but distinct phenotypes (GAME 1993 ). Rad51p has been shown to form a nucleoprotein filament similar to that formed by RecA protein (OGAWA et al. 1993 ) and to catalyze homologous pairing and DNA strand exchange in vitro (SUNG 1994 ). This strongly suggests that it is a functional budding yeast homolog of bacterial RecA. Null mutants of both RAD55 and RAD57 exhibit a curious enhancement of DNA damage sensitivity at lower temperatures (LOVETT and MORTIMER 1987 ; HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ). Both proteins interact (HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ) and form a stable heterodimer (SUNG 1997 ), which is consistent with the largely identical phenotypes of the respective mutants. An interaction between Rad55p and Rad51p was identified in the two-hybrid system (HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ). Biochemical experiments revealed that the Rad55p:Rad57p heterodimer stimulates Rad51p in the strand exchange reaction in vitro (SUNG 1997 ). Unlike Rad51p, the Rad55p:Rad57p dimer cannot promote DNA pairing and strand exchange in the presence of RPA on its own (SUNG 1997 ). This suggests a unique function of the Rad55p:Rad57p heterodimer as an accessory protein to Rad51p.

Recent studies established the evolutionary conservation of some components of the budding yeast recombinational DNA repair machinery. Homologs of the RAD52 group genes, RAD51, RAD52, RAD54, RAD50, and MRE11 were identified in several eukaryotic organisms including humans (KANAAR et al. 1998 ). The isolation of some of these genes was greatly aided by the isolation of recombinational repair genes from the fission yeast S. pombe. As fission yeast and budding yeast show great primary sequence divergence (RUSSELL and NURSE 1986 ), the existence of homologs in both organisms makes it highly likely that the respective gene is conserved in eukaryotes. Homologs for RAD51 and RAD54 have been isolated from S. pombe and been named rhp51 and rhp54, respectively (MURIS et al. 1993 , MURIS et al. 1996 ; SHINOHARA et al. 1993 ; JANG et al. 1994 ). Mutations in both genes confer extreme sensitivity to IR in fission yeast similar to the situation in budding yeast. However, the S. pombe mutants confer a degree of genomic instability not noted for the respective mutations in S. cerevisiae. The fission yeast rad22 gene product has been proposed to be a Rad52p homolog (OSTERMANN et al. 1993 ). However, the sequence homology is more limited than that of rhp51 or rhp54 with the S. cerevisiae counterpart, and the rad22 phenotypes do not match the severity of the S. cerevisiae rad52 mutant. In addition, the fission yeast rad32 gene was recognized to be a homolog of MRE11 (TAVASSOLI et al. 1995 ).

In Escherichia coli, a single RecA protein performs the central recombinational repair function (KOWALCZYKOWSKI et al. 1994 ). In the bacterium Myxococcus xanthus, two RecA genes have been identified, suggesting functional diversification (NORIOKA et al. 1995 ). This diversification is extended in the budding yeast with three RecAlike proteins for DNA repair (Rad51p, Rad55p, Rad57p) and the meiosis-specific Dmc1 protein (BISHOP et al. 1992 ). Not unexpectedly the situation in mammals is even more complex. Seven genes encoding proteins with homology to RecA have already been identified. Besides the rather strong homologies to the yeast Rad51 and Dmc1 proteins by the mammalian homologs (SHINOHARA et al. 1993 ; HABU et al. 1996 ), it is difficult to classify the remaining RecA-like proteins, R51H3, Rad51C, Xrcc2, Xrcc3, and hRec2 (THOMPSON 1996 ; RICE et al. 1997 ; CARTWRIGHT et al. 1998 ; DOSANJH et al. 1998 ) in the absence of information other than the primary sequence. Thus it is not known whether the mammalian genes represent homologs of the Rad51 protein or some of them might represent mammalian homologs to the budding yeast Rad55 and Rad57 proteins. Therefore, it remains unclear whether the functional diversification of RecA-like proteins seen in budding yeast is unique to this organism or a basic feature of the mechanisms of recombinational DNA repair in eukaryotes.

We have taken the fission yeast S. pombe as a model system to identify additional RecA-like proteins in this organism. The functional analysis possible in this organism provides additional criteria to the primary sequence information to classify new genes. Here we report the identification of a new RecA-like DSB DNA repair gene whose protein product has only slightly higher homology to the Rad55 protein of S. cerevisiae than to other RecA-like proteins. However, the analysis of the gene deletion provided strong evidence that this gene is a RAD55 homolog. Therefore, it was named rhp55⁺ for rad homolog S. pombe 55. Rhp55 protein acts in one DNA repair pathway together with Rhp51 and Rhp54 proteins, but in a different pathway than the putative Rad52p homolog Rad22p of fission yeast.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains, media, and growth conditions:
The S. pombe strains used in this study are listed in Table 1. The S. pombe strains used for the characterization of the rhp55⁺ gene had the mutations smt-0 or mat1P17::LEU2 representing deletions of the DSB-related site in the mat1 locus. The reason for this is discussed in the text. E. coli strain DH5 was used for recombinant DNA procedures in E. coli. S. pombe media malt extract agar (MEA), yeast extract agar (YEA), yeast extract liquid (YEL), and minimal medium (MMA), as well as the general genetic manipulations, have been described elsewhere (GUTZ et al. 1974 ; ALFA et al. 1993 ). When necessary, 0.01% (w/v) of supplements (amino acids and nucleosides) were added to the media. For meiotic time-course experiments the synthetic minimal medium S. pombe minimal (PM; BEACH et al. 1985 ) and PM without NH₄Cl (PM-N; WATANABE et al. 1988 ) were used. Methyl methanesulfonate (MMS) was added as a liquid to solid media cooled to 50° before pouring, and plates were used at the same day. S. pombe cells were transformed by the lithium acetate procedure of SCHIESTL et al. 1993 . S. pombe strains were grown at 30°, except when 18° or 20° was used as specified. All meiosis experiments were performed at 25°, the standard temperature for S. pombe meiosis.

DNA libraries and plasmids:
The S. pombe cDNA library in pDB20 (FIKES et al. 1990 ) and a S. pombe genomic DNA library in phage -ZAP (kind gift from T. Enoch) were used to clone the rhp55⁺ cDNA and genomic DNA, respectively. Several recombinant clones were isolated and the plasmids with DNA inserts were excised from phage DNA using ExAssist helper phage (Stratagene, La Jolla, CA). The oligonucleotides used for screening the cDNA library were oligo.rhp55#1 5'-AGCACCTGGGATGGAAAAAC-3' and oligo.rhp55#2 5'-GAATAGGCATTGATAGGTTGTC-3'. Two plasmids, pZAP-8 and pZAP-11 were found to contain ~3.9-kbp inserts encompassing the rhp55 open reading frame (ORF) as judged by DNA sequencing. The S. pombe complementation vector pIBG81 carrying the rhp55⁺ promoter and coding sequence was constructed by cloning the 1836-bp SalGI-EcoRV from pZAP-8 in the SalGI- and SmaI-cleaved LEU2 vector pIRT2 (HINDLEY et al. 1987 ). Plasmid pIBG48 with the rhp55 cDNA under the control of the adh promoter was constructed by recloning of the BamHI-KpnI cDNA fragment from pIBG40 into the pART1 plasmid. pART1 is a derivative of pIRT2, having the adh promoter. pIBG40 is the pBlueScript KS with HindIII insert of rhp55⁺ cDNA.

Construction of the rhp55 deletion:
Construction of the rhp55::ura4⁺ deletion mutant was performed by replacement of the complete ORF by the S. pombe ura4⁺ gene. Two PCR fragments representing the adjacent sequences flanking the rhp55⁺ ORF were generated. Two sets of primers were used to amplify 314- and 337-bp fragments flanking the rhp55⁺ ORF: 5'-GGCTCTAGAGAGCTTACTCTTGCATTCCTG-3' and 5'-GGCGGATCCTCCAGCACAGTGCGACCTAC-3'; and 5'-GGCATCGATTTTGGAATGTGAGTCCTAGG-3' and 5'-GGCGTCGACGTCCGTACACTGAATGCGAC-3'. Restriction digestion resulted in two fragments with cohesive ends: XbaI-BamHI and ClaI-SalI. Recloning of these fragments into the plasmid pBS (±) carrying a 1764-bp HindIII insert of ura4⁺ gene resulted in a plasmid pIBG62-8, where the ura4⁺ gene is flanked by regions of homology to sequences surrounding the rhp55⁺ ORF in the genome. The gene replacement module was excised with SalI/XbaI and used to transform the haploid strain h^- smt-0 ura4-D18 ade7-152 by selection for ura⁺ transformants. To construct the rhp55::arg3⁺ deletion mutant the new gene replacement plasmid pIBG62-9 was constructed as follows. The ura4⁺ gene was excised from pIBG62-8 by HindIII digestion and replaced by the 1783-bp SmaI-PstI fragment carrying the arg3⁺ gene from plasmid paR3 (WADDELL and JENKINS 1995 ). The gene replacement module was excised by KpnI/SacI digestion and transformed in S. pombe IBGY18 selecting for arg⁺ transformants. The resultant ura⁺ and arg⁺ transformant were analyzed by Southern hybridization to confirm the deletion of the entire rhp55⁺ gene.

DNA manipulation:
Standard procedures were used for DNA sequencing, DNA hybridization, screening of DNA libraries, and mapping of transcription initiation sites by primer extension. DNA primer 5'-TCCAGCACAGTGCGACCTAC-3', complementary to base pairs -21 to -40 of the rhp55⁺ cDNA, was used for primer extension analysis. Chromosomal DNA from S. pombe was isolated as described (WRIGHT et al. 1986 ).

MMS-, UV-, and -ray-sensitivity tests:
MMS and UV sensitivity were tested by drop assays. Sequential 10-fold dilutions of exponentially growing cells were spotted on the appropriate plates with or without MMS. To test for UV sensitivity, cells were irradiated on plates after spotting. Plates were incubated at 30° or 18°. To examine IR survival, exponentially growing cells were washed, resuspended in saline, and irradiated with -rays using a ⁶⁰Co source with a dose rate of 70 Gy/min. Appropriate dilutions were plated on complete media to determine the survival at 30° or 20°. All irradiation experiments were repeated at least twice.

Sporulation efficiency and spore viability:
To evaluate sporulation efficiency the number of spores, asci, and vegetative cells was scored microscopically. The sporulation efficiency (percentage sporulation) was calculated as (0.25 S + A)/(0.25 S + A + 0.5 C), where S is the number of spores, A the number of asci, and C the number of vegetative cells. To determine the spore viability, tetrads were dissected on YEA plates.

Meiotic time course and Northern analysis:
Synchronized meiosis was induced by shifting vegetatively growing cells of the diploid strains PA39 from PM to PM-N medium (time point, 0 hr). Meiotic prophase in this time course started ~5 hr after induction, as determined by the amount of "horse tail" nuclei after staining with DAPI (BAHLER et al. 1993 ). The first meiotic division started ~9 hr after induction as determined by the amount of cells with more than one nucleus (BAHLER et al. 1993 ). Total RNA was isolated from aliquots withdrawn at various time points as described elsewhere (GRIMM et al. 1991 ). A total of 20 µg of RNA was transferred to Zeta-Probe membrane (BioRad, Richmond, CA) after agarose gel electrophoresis denaturing and used for hybridization with ³²P-labeled rhp55⁺, byr1⁺ (NADIN-DAVIS and NASIM 1990 ), and ura4⁺ DNA probes under conditions recommended by the membrane supplier. The 314-bp PCR-generated fragment of rhp55⁺ cDNA and a 0.4-kb byr1⁺ PCR fragment were used as hybridization probes. Quantitation was performed using ImageQuant software after exposure in a phosphoimager.

Meiotic recombination:
Frequencies of meiotic intra- and intergenic recombination were determined by random spore analysis. For each interval crosses were performed in triplicate on MEA plates. For intragenic recombination analysis the spores were microscopically counted and plated on MMA and YEA with appropriate supplements. The amount of recombinant spores on MMA was counted and normalized to the amount of viable spores. For intergenic recombination analysis spore clones were randomly picked, grown on YEA master plates, and then replicated on MMA with appropriate supplements to determine the amount of recombinant spores. Spontaneously generated diploids were identified by replica plating on MEA plates and excluded from the quantitation.

Cytology and flow cytometry:
DAPI and Calcofluor staining of the exponentially growing cells was performed as described (ALFA et al. 1993 ). Slides with stained cells were analyzed using an epifluorescence microscope (Zeiss Axiovert) and the images were recorded with a monochrome cooled CCD camera (Kappa). To determine the cellular DNA content, propidium-iodide-stained cells were analyzed using a Becton-Dickinson (Franklin Lakes, NJ) FACScan as described (ALFA et al. 1993 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

rhp55⁺ encodes a new S. pombe protein with homology to RecA:
By DNA data bank analysis we identified on S. pombe cosmids 3C7 and 25A8 an ORF encoding a polypeptide with significant homology to bacterial RecA proteins. These cosmids had been mapped on chromosome I of S. pombe near the ras1⁺ gene (HOHEISEL et al. 1993 ). The greatest homology was detected between the predicted new protein and Rad55p from S. cerevisiae (Figure 1). The gene was named rhp55⁺, for rad homolog pombe 55, not only because of the sequence homology, but also because of the functional analysis discussed below. Analysis of this ORF revealed the presence of NTP-binding motifs (Walker boxes A and B) in the predicted protein typical for RecA-like proteins. To identify the complete coding sequence of the new gene we screened a S. pombe cDNA library using as a probe the 314-bp PCR fragment encompassing both Walker boxes, which was generated from cDNA. The identified recombinant plasmid contained a cDNA insert of ~1.4 kb. The insert was recloned as a HindIII fragment in pBlueScriptKS(±) and sequenced. The cDNA sequence data are available from GenBank under accession no. AF053410. The cloned cDNA contained the complete ORF for rhp55⁺ as well as 5'- and 3'-untranslated regions. A run of (T)₂₄, likely to represent the poly(A) tail of the mRNA, was found to be attached to the 3' end of cDNA (position 1303; the numbering system refers to the DNA sequence starting at the first ATG as +1). The 1443-bp cDNA contained an ORF of 1050 bp, which can be translated into a protein of 350 amino acids (aa) with a predicted molecular mass of 38,900. Because the ATG initiation codon is preceded by two in-frame stop codons, and the cloned cDNA exhibited biological activity (Figure 4B), the use of this ATG for initiation of translation is highly likely. By comparison of the cDNA sequence and the genomic DNA sequence from cosmid 3C7, one 44-bp intron close to the ATG start codon was identified (Figure 1A). The first exon comprises nucleotides 1–16 of the coding sequence and the intron/exon junctions perfectly match the S. pombe consensus sequence for 3' and 5' splice and branch sites: GTA(A/T)GT (5' intron site) ... CTAAPy (branch site) ... PyAG (3' intron site), where Py is pyrimidines (PRABHALA et al. 1992 ). In the 5' region of the gene, three putative transcription starts were mapped to positions -100, -153, and -182 by primer extension (data not shown). Upstream of these transcription initiation sites, two potential TATA-boxes, TAAAATAA (position -235) and TATAAA (position -283), can be identified. The spacing of the TATA boxes in relation to the transcription starts is in agreement with the reported 35–120 nucleotide (nt) distance for known S. pombe genes (RUSSELL 1989 ). In the 3'-untranslated region a possible polyadenylation signal was found at position 1116 (ATTAAT), as well as possible termination signals for mRNA synthesis (TTTTTA) at positions 1181 and 1253.

View larger version (81K): In this window In a new window Download PPT slide   Figure 1. Structure of the rhp55+ gene and the relationship of Rhp55p to other RecA-like proteins. (A) Exon-intron structure of the fission yeast rhp55+ gene. The open reading frame of the rhp55+ gene is shown as an open arrow. The black box represents the intron. The numbering refers to the nucleotides of the rhp55+ protein-coding sequence. (B) Amino acid (aa) alignment of S. pombe Rhp55p with other RecA-like proteins from S. cerevisiae and S. pombe. The alignment was generated using Pile-Up (GCG software package; DEVEREUX et al. 1984 ) and viewed with Boxshade (Bioinformatics Group, ISREC). Identical aa are highlighted in black, similar aa are shaded as follows: P, A, G, S, and T; E, D, N, and Q; V, I, L, and M; F, W, and Y; K, R, and H. Dots indicate highly conserved residues in the RecA sequence important for protein structure and function. (C) Dendrogram of clustering relationship between eukaryotic RecA-like proteins. The dendrogram was generated by Pile-Up (GCG software; DEVEREUX et al. 1984 ). Sc, S. cerevisiae; Sp, S. pombe; Hs, Homo sapiens.

View larger version (21K): In this window In a new window Download PPT slide   Figure 2. Induction of rhp55+ expression during meiosis. Total RNA was analyzed by Northern hybridization at the indicated time points after induction of meiosis. The signals were quantified on a phosphoimager using ImageQuant. The level of rhp55 transcript was normalized to byr1 as an internal control.

View larger version (60K): In this window In a new window Download PPT slide   Figure 3. Fluorescence microscopy and FACS analysis of the rhp55 mutant. (A) DAPI and Calcofluor double staining of wild-type and rhp55 cells. (B) Results from FACS analysis. Left, plots of cell counts over propidium iodide staining; right, distributions of cells plotting the peak of intensity of propidium iodide fluorescence and the total intensity. This gives an indication of the heterogeneous nature of the cells and the nuclear organization. Control wild-type cells (h- smt-0 ura4-D18 strain) grown in minimal media with limiting nitrogen source were used to demonstrate the DNA contents of cells in G1 and G2 (top row). Cells of wild-type strain (middle row) and rhp55 mutant strain IBGY19 (bottom row) were used during exponential growth in full medium (YEL medium).

View larger version (113K): In this window In a new window Download PPT slide   Figure 4. MMS and UV sensitivity of the rhp55 mutant. (A) MMS and UV survival test of the rhp55 mutant. MMS and UV sensitivity was tested by drop assay (see MATERIALS AND METHODS). The following isogenic strains were used: h- smt-0 ura4-D18 (wild type), IBGY19 (rhp55), IBGY20 (rhp51), IBGY21 (rhp54), and IBGY22 (rad22). (B) Complementation of MMS and UV sensitivity of the rhp55 mutant by cloned cDNA and cognate gene. Complementation was tested by drop-assays (see MATERIALS AND METHODS). The strains were grown at the indicated temperatures and plates were photographed after incubation for the indicated times. rhp+, wild-type strain h- smt-0 ura4-D18; rhp552, rhp55 strain IBGY84; +pIRT2 and +pIRT2rhp55+, IBGY84 transformed with either empty vector or pIBG81; +pART1 and +pART1rhp55+, IBGY84 transformed with an empty vector and pIBG48.

Amino acid comparison of Rhp55p with E. coli RecA and related proteins from budding and fission yeast revealed a significant level of identity (Figure 1B). The highest level of homology among the five proteins exists in the central region encompassing the NTP-binding motifs A and B (from ~aa 24–157 in the Rhp55p sequence). The N-terminal regions of Rhp55p and Rad55p are shorter than the N termini of ScRad51p, SpDmc1p, and ScRad57p. Pairwise alignment using ClustalW (version 1.74) showed the highest overall homology between Rhp55p and S. cerevisiae Rad55p (27.9% of identical and 52.5% of similar aa). The homologies to the S. cerevisiae Rad57 and Rad51 proteins were 26.5% identity and 52.1% similarity as well as 21.2% identity and 40.8% similarity, respectively. Despite the moderate overall homology between the two proteins, the majority of the critical aa positions responsible for ATP-binding and conformational changes in the RecA protein have been conserved (see Figure 1B). The dendrogram shows closer similarity of Rhp55p to S. cerevisiae Rad55p than to Rad57p, Dmc1p, or Rad51p (Figure 1C). The sequence homology of Rhp55p to RecA and RecA-like proteins from the RAD52 group implicated the protein in recombinational repair of DNA damage.

rhp55⁺ is transcribed in vegetatively growing cells and is induced during meiosis:
Expression of rhp55⁺ was examined in vegetatively growing cells and during meiosis. In a synchronized meiotic time course, total RNA was analyzed for the presence of rhp55-specific mRNAs. The result of this experiment (Figure 2) revealed that rhp55⁺ was expressed both in mitotically dividing cells before entering meiosis (time point, 0 hr) and throughout meiosis (time, 2–10 hr). Only one mRNA species of ~1.4 kb was detected on Northern blots, in accordance with the size of the isolated cDNA. During meiosis the expression of the gene gradually increased with a maximum occurring at 8 hr, which corresponded in this time course to meiosis I prophase. The level of rhp55⁺ transcript was normalized at each time point using byr1⁺ mRNA as a control. The maximal level of meiotic induction of rhp55⁺ transcript levels (8 hr) was found to be fourfold above its expression in mitotically dividing cells (0 hr). A similar increase was obtained when the ura4⁺ gene transcript was used for normalization (data not shown). Both controls, ura4⁺ and byr1⁺, exhibit essentially unchanged transcription levels during meiosis (PARISI et al. 1999 ). The expression pattern of the rhp55⁺ gene suggested that it has a role during vegetative growth and during meiosis.

rhp55 cells contain aberrant nuclei and exhibit an increased DNA content:
Mutants of the S. pombe recombinational repair genes rhp51⁺ and rhp54⁺ were shown to have slow growth and cell elongation phenotypes (MURIS et al. 1996 ). The rhp55 mutant did not exhibit a clear slow-growth phenotype: at 30° the mutant cells grew normally, whereas at 18° slightly slower growth was observed (see Figure 4A and data not shown). However, the previous studies with rhp51 and rhp54 are difficult to interpret because the strains analyzed were not deficient for the DSB related to mating-type switching in fission yeast. The DSB associated with mating-type switching in S. pombe (GUTZ and SCHMIDT 1985 ; EGEL 1989 ; ARCANGIOLI 1998 ) causes problems in experiments with DSB repair mutants that have not been fully appreciated in previous studies with S. pombe. The resulting bias against loss-of-function mutants in DSB repair genes is reflected in the results of screens for DNA repair mutants (LEHMANN 1996 ). In S. cerevisiae, this problem usually does not arise because standard laboratory strains contain a mutation in the HO gene, encoding the endonuclease delivering the DSB at MAT. Although a similar mutation is not available in S. pombe, the use of cis mutations (smt-0 or mat1P17::LEU2) affecting DSB formation at mat1 fulfill the same purpose.

Throughout this study we used strains containing these mutations to avoid compound effect between the DSB repair mutant and the DSB at mat1. The double DAPI-Calcofluor staining of rhp55 mutant cells in exponential growth at 30° revealed that ~17% of the cells were elongated at least more than twofold compared to wild type and that these cells contained aberrant nuclei (Figure 3A). This observation was further substantiated by fluorescence analysis of rhp55 mutants (Figure 3B). Wild-type cells growing under conditions of limited nitrogen accumulate in the G1 phase (Figure 3B, top). This served to visualize the two peaks, corresponding to G1 with a 1n DNA content and to G2 cells with a 2n content. Exponentially growing fission yeast wild-type cells in full medium contain mostly a 2n or higher DNA content. This is because fission yeast spends most of the time in the G2 phase of the cycle and, because after the completion of DNA synthesis (S phase), cells are still unseparated (Figure 3B, middle). This is in accordance with previous observations and typical for fission yeast (FORSBURG and NURSE 1991 ). However, in rhp55 mutants the peak of intensity of propidium iodide fluorescence shifted and broadened in the direction of greater DNA content. The distribution of cells (Figure 3B, bottom) also indicated a greater heterogeneity of cell size and DNA content in the mutant compared to wild type. Approximately 48% of mutant cells contained DNA exceeding 2n, in contrast to 26% in the wild-type cells. From these data we conclude that, even in the complete absence of a mating-type switch-associated DSB, rhp55 cells exhibit defects in nuclear morphology and DNA content. This may indicate a role of Rhp55p in processes other than DNA repair of exogenously induced DNA damage.

rhp55⁺ is a DNA-damage-repair gene:
To determine whether the rhp55⁺ gene is involved in DNA damage repair, we constructed a null allele by homologous recombination. The entire ORF of the rhp55 was substituted by the ura4⁺ or arg3⁺ gene. The deletion mutant was viable in haploid cells, indicating that rhp55⁺ is not essential for mitotic growth. We tested the sensitivity of the deletion strain rhp55::ura4⁺ to the alkylating agent MMS and to UV light in drop assays. For comparison we included in this test other mutants of S. pombe-DNA-repair genes: rhp51, rhp54, and rad22. The first two are homologs of the S. cerevisiae RAD51 and RAD54 genes (MURIS et al. 1993 , MURIS et al. 1996 ; SHINOHARA et al. 1993 ); the latter has been proposed to be a homolog of the RAD52 gene (OSTERMANN et al. 1993 ). As shown in Figure 4A, the rhp55 mutant was sensitive to MMS and to a lesser extent to UV at 30°. The sensitivity to MMS was less extreme than in rhp51 and rhp54 mutants. The sensitivity of the rhp55 mutant to UV light at 30° was more pronounced at higher UV doses (see also Figure 4B). In contrast, the rad22 mutant was almost fully resistant to these DNA-damaging agents. Interestingly, at 18° the rhp55 mutant exhibited a much higher sensitivity to MMS and UV. Under these conditions the survival of rhp55 mutants was indistinguishable from that of rhp51 and rhp54 mutants. This unusual property of cold-enhanced sensitivity to DNA-damaging agents is a unique feature of the S. cerevisiae rad55 and rad57 mutants (LOVETT and MORTIMER 1987 ; HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ). Because the damage to cells caused by MMS is rather complex and DSBs constitute only a part of the lesions induced by this drug (FRIEDBERG et al. 1995 ), we examined the effect of IR. In Figure 5 the -ray survival curves of the rhp55 mutant at two temperatures (30° and 20°) are shown along with those of the rhp51, rad22, and rhp54 single mutants. These data demonstrate that the rhp55⁺ gene is involved in the repair of DNA damage induced by -rays. Moreover, the -ray sensitivity is also enhanced at lower temperatures as was found for MMS and UV damage (Figure 5A). At 20° the rhp55 mutant is as sensitive to IR as the single rhp51 or rhp54 mutants. Again, the rad22 mutant was only slightly sensitive to -irradiation (Figure 5A). From this data we conclude that rhp55⁺ is a new DNA-damage-repair gene in fission yeast.

View larger version (18K): In this window In a new window Download PPT slide   Figure 5. -ray survival curves of single and double mutants of the S. pombe DSB repair genes. (A) Cell survival curves of rad22, rhp55, and rad22 rhp55 double mutants. (B) Survival curves of rhp51, rhp55, and rhp51 rhp55 double mutants. (C) Survival curves of rhp54, rhp55, and rhp54 rhp55 double mutants. The following strains were used: x, wild type (h- smt-0 ura4-D18); , rad22 (IBGY22); , rhp55 (IBGY19); , rhp55 at 20°; , rad22 rhp55 (IBGY43); , rhp51 (IBGY20); , rhp51 rhp55 (IBGY44); •, rhp54 (IBGY21); , rhp54 rhp55 (IBGY45). All survival curves were determined at 30° except when indicated at 20° for rhp55.

The rhp55⁺ cDNA complements the DNA-damage-repair defect caused by the rhp55 mutation:
To further ascertain that we have identified the full extent of the rhp55⁺ gene and not missed an intron, we constructed two complementation plasmids. Previous analysis of genes that control meiotic recombination in fission was complicated by the occurrence of multiple introns, e.g., rec12 (LIN and SMITH 1994 ; BERGERAT et al. 1997 ; FOX and SMITH 1998 ). In pIBG48 the rhp55⁺ cDNA was placed under the control of S. pombe adh promoter, and pIBG81 contained the rhp55⁺ gene with its authentic promoter. These constructs were transformed into the rhp55 strain IBGY84 and the resulting clones were analyzed for the DNA repair phenotype of the mutant. The data presented in Figure 4B demonstrate that both the rhp55⁺ cDNA and the cognate gene complement the MMS and UV sensitivity of the mutant equally efficiently, restoring the wild-type level of damage resistance. This indicated that the cloned cDNA encodes a functional protein.

To explore the functional relation of rhp55⁺ to known S. cerevisiae homologs we also examined the ability of the rhp55⁺ cDNA to complement the MMS sensitivity of S. cerevisiae rad51, rad55, and rad57 mutants by placing the cDNA onto centromeric plasmids under the control of a set of S. cerevisiae promoters of different strengths (MUMBERG et al. 1995 ). All constructs, when transformed in S. cerevisiae, failed to complement the MMS sensitivity of rad51, rad55, or rad57 mutants (data not shown). The absence of interspecies complementation between budding and fission yeasts is quite common and reflects the evolutionary divergence between both organisms (RUSSELL and NURSE 1986 ).

Rhp55p acts in one pathway with Rhp51 and Rhp54 proteins, but in a different pathway than the putative Rad52p homolog, Rad22p:
Double-mutant analysis allows the assignment of genes to one or different epistasis groups in DNA repair, which are believed to represent different pathways for the repair of a specific DNA lesion (GAME 1993 ; FRIEDBERG et al. 1995 ). In S. cerevisiae RAD51, RAD52, RAD54, and RAD55 have been assigned to one epistasis group, the recombinational repair group. The double mutants rhp55 rhp51 and rhp55 rhp54 showed the same -ray sensitivity at 30° as the rhp51 and rhp54 single mutants (Figure 5B and Figure C). This indicates that rhp55 is epistatic to rhp51 and rhp54 and functions in the same DNA repair pathway. However, the rhp55 rad22 double mutant exerted a synergistic sensitivity to -rays at 30° if compared to the single mutants (Figure 5A). This indicates that these genes act in different DNA repair pathways competing for the same substrate. We extended the epistatic analysis of putative recombinational repair genes by testing rad22 rhp55, rad22 rhp51, rad22 rhp54, rhp51 rhp55, and rhp54 rhp55 double mutants for sensitivity to MMS and UV. The data presented in Figure 6 demonstrated that rad22 rhp51, rad22 rhp54, and rad22 rhp55 double mutants were significantly more sensitive to low doses of DNA-damaging agents (0.0001% MMS and 40 J m^-2 UV) than the respective single-mutant strains. This synergistic behavior suggests that Rad22p acts in a different pathway than Rhp51p, Rhp54p, and Rhp55 for the repair of this type of DNA damage. Moreover, the rhp51 rhp55 and rhp54 rhp55 double mutants are as resistant to these doses as single rhp51 and rhp54 mutants, confirming the epistatic interaction identified previously for -ray sensitivity (Figure 5). Therefore, the double-mutant analysis shows the existence of an epistasis group for the repair of -ray, MMS, and UV damage in fission yeast similar to the RAD52 group in budding yeast consisting presently of rhp51⁺, rhp54⁺, and rhp55⁺. However, rad22⁺, a gene proposed to be the fission yeast RAD52 homolog, acts in a different pathway than the Rhp51, Rhp54, and Rhp55 proteins.

View larger version (59K): In this window In a new window Download PPT slide   Figure 6. Epistatic analysis of rad22+, rhp51+, rhp54+ and rhp55+ genes for repair of MMS and UV damage. MMS and UV sensitivity were tested by drop assay as described in MATERIALS AND METHODS. Isogenic strains used were wild type (h- smt-0 ura4-D18), rhp51 (IBGY20), rhp54 (IBGY21), rhp55 (IBGY19), rad22 (IBGY22), rhp51 rad22 (IBGY200), rhp54 rad22 (IBGY201), rhp55 rad22 (IBGY43), rhp51 rhp55 (IBGY44), and rhp54 rhp55 (IBGY45). Plates were incubated at 30° for 3 days and photographed.

rhp55 affects meiosis and causes a small reduction in meiotic recombination:
The S. cerevisiae rad51, rad55, and rad57 mutants are characterized by strong meiotic defects such as reduced sporulation efficiency and gross spore inviability (PETES et al. 1991 ; GAME 1993 ). Therefore, we examined the sporulation of rhp55 crosses in comparison with wild-type crosses at 25°, the standard temperature for meiosis experiments in fission yeast. We found that the sporulation efficiency was reduced 2.6-fold in the mutant (Table 2). The sporulation efficiency of wild-type crosses was consistent with previous determinations (BAHLER et al. 1993 ). Moreover, only about half of the mutant spores (54.4%) were found to be viable, whereas wild-type crosses resulted in 92.4% viable spores (Table 3). This difference was statistically significant with a probability of <0.01 (² analysis). Analyzing the distribution of inviable spores among tetrads revealed that more tetrads with 0 viable spores were observed than anticipated from a random process in the mutant (9 found vs. 3.73 expected). This difference was significant with a probability of <0.01. The reason for the increase in this particular class might be related to premeiotic events in the mutant. Vegetatively growing cells were found to contain aberrant nuclei and aberrant DNA content (see Figure 3). Mating of such cells may lead to the complete meiotic inviability observed in this class. However, mitotic events cannot account alone for the low-spore viability in rhp55 cells, because the classes with 4 and 3 viable spores are also significantly reduced in the mutant compared to wild type. Because fission yeast contains only three chromosomes, random chromosome segregation results in considerable viability among the meiotic products. The assumption of random MI segregation and equational MII segregation predicts 22% of the spores to be viable, including haploid, diploid, and chromosome III disomic spores (G. R. SMITH, personal communication). Fission yeast chromosome I and II disomes are not viable (NIWA and YANAGIDA 1985 ). The rhp55 data suggest that this protein is important for full spore viability, but not essential, because the spore viability is significantly above the expected level for random segregation. The reduction in spore viability found in the rhp55 mutant is less than found previously with rhp51, for which a spore viability of 1.7% has been reported (MURIS et al. 1997 ). However, that value was not determined by tetrad analysis but by determining the colony-forming units of a spore suspension. Moreover, it relates to a wild-type level that was artificially set to 100%. As shown in Table 3, the spore viability of the rhp51 mutant was 7.8% as determined by tetrad analysis, which is significantly below the expected 22% (P < 5%). As rhp55⁺ and rad22⁺ were found to act in different DNA repair pathways (see above), it was of interest to determine if the double mutant will show synergism for meiosis. Table 2 and Table 3 show the sporulation efficiencies and spore viability of rad22 and rad22 rhp55 mutant crosses. The rad22 mutant showed the reduction in sporulation efficiency similar to rhp55 mutant (2.4- and 2.6-fold, respectively); however, spore viability was two times lower (26.1 and 54.4%, respectively), which was significant (P < 0.01). The rhp55 rad22 double mutant showed less reduction in sporulation efficiency and spore viability in comparison with the single mutants. However, the difference with the rhp55 single mutant was statistically not significant. This indicates the absence of a synergistic effect of the two mutations in meiosis in fission yeast.

To determine to what extent rhp55 mutants were affected in meiotic recombination, we analyzed meiotic intra- and intergenic recombination. Intragenic recombination was examined in three intervals, two in the ade6 locus on chromosome III (ade6-469 x ade6-M26 and ade6-469 x ade6-M375 crosses) and one in the ade7 locus on chromosome II (ade7-150 x ade7-152 cross; Table 4). In the ade6 locus we observed a small reduction (1.6- and 2.9-fold) in intragenic recombination. This reduction was significant as indicated by the nonoverlapping standard deviations. No reduction of intragenic recombination was found in the ade7 locus. Intergenic recombination (Table 5) was analyzed in three intervals placed on the three fission yeast chromosomes. In the his1-lys7 (chromosome I), ade7-arg6 (chromosome II), and ade6-arg1 (chromosome III) intervals, meiotic intergenic recombination was reduced 1.8-, 1.7-, and 1.4-fold, respectively, in the mutant when compared to wild-type crosses. As indicated by the nonoverlapping standard deviations, these reductions were significant. Similarly, only small reductions in meiotic intergenic recombination were found in rhp51 and rhp54 mutants of S. pombe (MURIS et al. 1997 ). No region specificity was observed in intergenic recombination, as the reduction in the mutant was similar in all intervals tested. From this data we conclude that Rhp55p is required for full meiotic recombination in fission yeast.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study we report the isolation of rhp55⁺, a new recombinational DNA repair gene, from S. pombe encoding the likely fission yeast homolog of the S. cerevisiae Rad55 protein. Rhp55p acts in one DSB repair pathway with Rhp51 and Rhp54 proteins, which is different from the pathway in which the putative Rad52p homolog, Rad22p, acts. Rhp55 protein is important for recombinational DNA repair of exogenously induced DNA damage by MMS, UV, and IR. Moreover, Rhp55p is required for genomic stability, efficient sporulation, and full meiotic recombination in fission yeast. Throughout this study we used strains deficient for the mating-type switching-related DSB to avoid possible compound effects with the mutations in DSB repair.

Rhp55p is a new recombinational repair protein in S. pombe:
Recombinational repair is a major pathway for the repair of MMS-induced and -ray-induced DNA lesions in yeasts (PETES et al. 1991 ; GAME 1993 ; FRIEDBERG et al. 1995 ; LEHMANN 1996 ). Moreover, it appears that in S. pombe, an organism lacking photolyase for the repair of UV damage, recombinational repair constitutes a major pathway also for the repair of UV-induced DNA damage (LEHMANN 1996 ). Based on the following evidence, Rhp55p is a new recombinational repair protein in fission yeast. First, the severity and the spectrum of the sensitivities of rhp55 mutants toward MMS, -ray, and UV are consistent with a role of Rhp55p in recombinational repair. Second, the phenotypes of the rhp55 mutant are similar and, at lower temperatures, almost identical to that of two other recombinational repair mutants in fission yeast, rhp51 and rhp54 (MURIS et al. 1993 , MURIS et al. 1996 ; SHINOHARA et al. 1993 ). Third, double-mutant analysis established that Rhp55, Rhp51, and Rhp54 proteins act in the same DNA repair pathway. Fourth, analysis of meiotic recombination in rhp55 mutants demonstrated a mild but significant defect in intra- and intergenic homologous recombination.

The cold-enhanced sensitivity of the rhp55 deletion mutant to UV, MMS, and -ray is unique among the recombinational repair mutants in fission yeast, but similar to the phenotype of deletions in the S. cerevisiae RAD55 and RAD57 genes (LOVETT and MORTIMER 1987 ; HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ; see below). Protein:protein interaction leads to a net gain of entropy, and temperature sensitivity is often a consequence of impaired hydrophobic interactions within a protein (SHORTLE 1989 ) or between subunits of a complex (CANTOR and SCHIMMEL 1980 ). An example for an intrinsically cold-sensitive heteromultimeric protein complex is the microtubule (MANDELKOW and MANDELKOW 1994 ). Also the assembly of functional - and -tubulin is a cold-sensitive process in deletion mutants of the GIM genes (GEISSLER et al. 1998 ). Cold-sensitive assembly of a protein complex can also be found in mutants affecting ribosome assembly (GUTHRIE et al. 1969 ; LHOEST and COLSON 1981 ). The cold-sensitive phenotype of the rhp55 deletion suggests that it either stabilizes a heteromultimeric complex or is involved in the assembly of a protein complex important for the repair of DNA damage. The biochemical properties of the Rad55p:Rad57p heterodimer of budding yeast (SUNG 1997 ) are compatible with both models.

Rhp55p is the homolog of the S. cerevisiae Rad55 protein:
On the basis of the specific evidence discussed below, we suggest that Rhp55p is a homolog of the S. cerevisiae Rad55 protein. First, pairwise sequence comparisons and the clustering relationship show higher homology of Rhp55p to Rad55 protein than to any other RecA-like protein in budding yeast. Second, the overall protein sequence of Rhp55 protein resembles that of Rad55p more than that of any other RecA-like protein in S. cerevisiae. All RecA-like proteins are characterized by a central core containing the highest homology and the two consensus sequences for nucleotide binding. In addition, they have N- and C-terminal extensions of different lengths (HEYER 1994 ). Both Rhp55 and Rad55 proteins are characterized by rather short, almost identical length N-terminal extensions (see Figure 1B). Taking the conserved glycine at position 26 (G24 of Rhp55p equivalent to G44 in RecA; see Figure 1B) as a reference point, the N-terminal region of Rhp55p and Rad55p is significantly smaller than that of any other RecA-like protein in the two yeasts (Rhp55p, 25 aa; Rad55p, 23 aa). This is clearly distinct from the much longer N-terminal regions of Rad51-like proteins (126 aa for Rhp51p and 162 aa for Rad51p), Dmc1-like proteins (101 aa for SpDmc1p, 98 aa for ScDmc1p), or Rad57p-like proteins (102 aa). The C-terminal extension of Rhp55p is clearly different from Rad51p and Dmc1p in length and sequence, but similar to Rad55p and Rad57p (see Figure 1B and HEYER 1994 ). Third, the phenotypes of the rhp55 mutant in DNA damage repair are comparable to those of rad55 and rad57 mutants. Fourth, rhp51⁺ has been isolated as a high-copy suppressor of the rhp55 deletion mutation (G. V. SAVCHENKO, unpublished observations); similar high-copy suppression of the rad55 and the rad57 deletion by RAD51 has been identified in budding yeast (HAYS et al. 1995 ; JOHNSON and SYMINGTON 1995 ). Fifth, the cold-enhanced DNA-damage sensitivity of rhp55 mutants is the most compelling argument for its similarity to RAD55. The cold-enhanced phenotypes of the rad55 and rad57 deletion mutants are a unique feature for these DNA-damage-repair genes (see above; PETES et al. 1991 ; GAME 1993 ). The absence of interspecies complementation is not an argument against this interpretation as this is quite commonly found because of the evolutionary divergence between both organisms (RUSSELL and NURSE 1986 ). Failure of interspecies complementation between the two yeasts has been reported even when the mammalian homolog showed complementation in one of the yeast species. One example of this effect is the Rad54 protein (KANAAR et al. 1996 ; MURIS et al. 1997 ). Other examples can be found with components of the transcriptional apparatus, another evolutionarily conserved process (SHPAKOVSKI et al. 1995 ).

Therefore, the occurrence of a Rad55p homolog in fission yeast suggests that the functional diversification of RecA-like genes in S. cerevisiae with one homolog, Rad51p, performing a highly similar function to RecA (SUNG 1994 ) and two additional proteins, Rad55 and Rad57, performing a different accessory role (SUNG 1997 ), may not be a peculiarity of budding yeast but rather a general feature of the molecular mechanisms of recombinational repair in eukaryotes. Future biochemical work with the S. pombe protein will have to substantiate this notion. This leads to the prediction of the occurrence of a Rad57-like protein in fission yeast and a strong candidate for this has been recently identified (H. SHINAGAWA, personal communication). Thus, some of the mammalian RecA-like proteins (see Introduction) may turn out to perform a accessory role to the human Rad51 protein similar to that of the Rad55:Rad57 heterodimer to Rad51p in S. cerevisiae.

Rhp51, Rhp54, and Rhp55 proteins form a recombinational repair epistasis group that does not include the putative Rad52p homolog Rad22p:
Double-mutant analysis can be used to define epistasis groups that are likely to reflect an organization in pathways (FRIEDBERG et al. 1995 ). If a double mutant exhibits a sensitivity no greater than the most sensitive single mutant, the two genes in question are considered epistatic, i.e., being in one pathway. If, however, the double mutant exhibits an additive sensitivity, the two genes are likely to act in different pathways. Synergistic enhancement of the sensitivity in the double mutant not only suggests that both genes act in different pathways, but also that these pathways compete for the same substrate. DNA-damage-repair epistasis groups have been extensively defined in S. cerevisiae (FRIEDBERG 1988 ; GAME 1993 ; FRIEDBERG et al. 1995 ) but only to a much smaller extent in S. pombe. The mutant phenotypes of rad32 (TAVASSOLI et al. 1995 ), rad21 (BIRKENBIHL and SUBRAMANI 1992 ), rad22 (OSTERMANN et al. 1993 ), rhp55 (this article), rhp51, and rhp54 (MURIS et al. 1993 , MURIS et al. 1996 , MURIS et al. 1997 ) had suggested an involvement of the respective genes in a recombinational repair pathway similar to the RAD52 pathway of budding yeast. Here we greatly extend the limited epistasis analysis in S. pombe, which showed that rad32⁺ and rhp51⁺ act in one pathway for the repair of UV damage (TAVASSOLI et al. 1995 ), by demonstrating that this epistasis group also includes rhp55⁺ and rhp54⁺, but not rad22⁺, for the repair of -ray-, MMS-, and UV-induced damage (Figure 6). This strongly suggests that rad22⁺ is not a functional homolog of the S. cerevisiae RAD52 gene as previously suggested (OSTERMANN et al. 1993 ). This interpretation is also more consistent with the much weaker phenotypes of the rad22 mutant for UV, MMS, and -ray survival (this article; OSTERMANN et al. 1993 ; MURIS et al. 1997 ) than those reported for the S. cerevisiae rad52 mutant (PETES et al. 1991 ; GAME 1993 ; FRIEDBERG et al. 1995 ). In S. cerevisiae, Rad52p is central for homologous recombination and SSA in contrast to Rad51p, Rad55p, and Rad57p, which have no role in SSA (IVANOV et al. 1996 ).

The existence of Rad59p in S. cerevisiae, a Rad52p-related protein that functions in RAD51-independent recombination (BAI and SYMINGTON 1996 ), poses the question of whether rad22⁺ may be the fission yeast homolog of RAD59. This appears unlikely because, in S. cerevisiae, RAD59 and RAD51 are epistatic for repair of DNA damage (L. SYMINGTON, personal communication), unlike rhp51 and rad22 in fission yeast.

The question of whether there is a functional RAD52 and/or RAD59 homolog in fission yeast can at present not be answered. Database searches revealed the presence of a second ORF besides rad22 in the fission yeast genome, which may encode a protein with homology to Rad52p (V. I. BASHKIROV, unpublished observations). The function of this putative gene is unknown. Similar to the S. pombe rad22 mutant, homozygous mutations in genes with homology to RAD52 in mouse ES cells or chicken DT40 cells were found to lack obvious DNA repair defects. Moreover, such cells exhibited only minor deficiency in targeted recombination (RIJKERS et al. 1998 ; YAMAGUCHI-IWAI et al. 1998 ). Thus, also in chicken and mouse these proteins do not have the same important function Rad52p has in budding yeast.

Rhp55 protein is involved in maintaining genomic stability and in meiosis:
Previous work in the rhp51 and rhp54 mutants suggested a high degree of genomic instability in these mutants reflected by an accumulation of aberrantly long cells with a higher than normal DNA content (MURIS et al. 1996 ). However, it is difficult to fully interpret these results because these studies have been conducted in strains where the mat1-related DSB occurs, which may lead to compound effects in DSB repair mutants. In this study, we have used DSB-negative strains that harbor cis-acting mutation (smt-0, mat1P17::LEU2), preventing the formation of a mating-type associated DSB. Thus, the observed increase in elongated cells with greater than normal DNA content found in rhp55 mutants was not caused by a compound effect with the mat1-related DSB. Similar effects of mutations in recombinational repair genes in budding yeast have not been noticed. Therefore, DSB repair proteins exhibit a more visible role in genomic stability in fission yeast than in budding yeast. For a full discussion of this point and the reasoning for a possible involvement of DSB repair proteins in DNA replication see MURIS et al. 1996 . In conclusion, rhp55⁺ and the recombinational DNA repair pathway are required for genomic stability in fission yeast in the absence of exogenously added genotoxic stress. Thus, fission yeast might present an informative model system to understand the inviability of vertebrate cells lacking Rad51 protein (LIM and HASTY 1996 ; TSUZUKI et al. 1996 ; SONODA et al. 1998 ).

The Rhp55 protein is suggested to have a role in meiosis based on the following evidence. First, the mutant exhibited reduced sporulation efficiency and spore viability. Some of the observed meiotic lethality is probably a result of the accumulation of genomic aberrations before meiosis as suggested by the spore viability pattern and the cytological analysis of vegetatively growing cells. Aneuploid cells are exceedingly unstable in fission yeast (NIWA et al. 1986 ) and a majority of the rhp55 mutant cells appeared to have a normal DNA content, strongly suggesting that not all of the meiotic lethality is due to premeiotic events. This is supported by the spore viability pattern in the mutant. The spore viability of rhp55 mutants in fission yeast is significantly higher than that of rad55 mutants in budding yeast, which was reported to be <2.5% (L