recombination in Saccharomyces cerevisiae. A

Downloaded from genesdev.cshlp.org on July 14, 2011 - Published by Cold Spring Harbor Laboratory Press

A Rad52 homolog is required for RAD51-independent mitotic recombination in Saccharomyces cerevisiae. Y Bai and L S Symington Genes Dev. 1996 10: 2025-2037 Access the most recent version at doi:10.1101/gad.10.16.2025

References

This article cites 61 articles, 36 of which can be accessed free at: http://genesdev.cshlp.org/content/10/16/2025.refs.html Article cited in: http://genesdev.cshlp.org/content/10/16/2025#related-urls

Email alerting service

Receive free email alerts when new articles cite this article - sign up in the box at the top right corner of the article or click here

To subscribe to Genes & Development go to: http://genesdev.cshlp.org/subscriptions

Copyright © Cold Spring Harbor Laboratory Press


A Rad52 homolog is required for RAD51-independent mitouc recombination in Saccharom yces cerevisiae Y u n Bai and Lorraine S. S y m i n g t o n ~ Columbia University College of Physicians and Surgeons, Department of Microbiology and Institute of Cancer Research, New York, New York 10032 USA

With the use of an intrachromosomal inverted-repeat as a recombination reporter we have previously shown that mitotic recombination is dependent on the RAD52 gene. However, recombination was found to be reduced only 4-fold by mutation of RAD51, which encodes a homolog of bacterial RecA proteins. A rad51, which strain containing the recombination reporter was mutagenized to identify components of the RAD51-independent pathway. One mutation identified, rad59, reduced recombination 1200-fold in the presence of a rad51 mutation, but only 4- to 5-fold in a wild-type background. Thus the rad51 and rad59 mutations reduce recombination synergistically. The rad59 mutation reduced both spontaneous and double-strand-break-induced recombination between inverted repeats. However, the rate of interchromosomal recombination was increased in a rad59 homozygous diploid. These observations suggest that RAD59 functions specifically in intrachromosomal recombination. The rad59 mutant strain was sensitive to ionizing radiation, and this phenotype was used to clone the RAD59 gene by complementation. The gene encodes a protein of 238 amino acids with significant homology to members of the Rad52 family. Overexpression of RAD52 was found to suppress the DNA repair and recombination defects conferred by the rad59 mutation, suggesting that these proteins have overlapping roles or function as a complex.

[Key Words: Recombination; gene conversion; RAD51; RAD52; Saccharomyces cerevisiae] Received April 26, 1996; revised version accepted July 8, 1996.

Genes in the RAD52 epistasis group (RAD50-57, XRS2, and MRE11) were identified initially as essential for the repair of ionizing radiation-induced DNA damage or for meiosis (Game and Mortimer 1974; Ajimura et al. 1993). Although the members of this family share the property of conferring resistance to ionizing radiation, they show considerable heterogeneity in other assays for doublestrand break (DSB) repair and recombination. The RAD50, XRS2, and MRE11 genes form a subgroup with similar properties, and interactions between these gene products have been detected by the two-hybrid system (Johzuka and Ogawa 1995). Mutation of other members of this group (RAD51, RAD52, RAD54, RAD55, and RAD5 7) results in the inability to repair DSBs, reduction in spontaneous and induced mitotic recombination, and defects in mating-type switching and sporulation. Of these mutants, rad52 confers the most severe defects in DSB repair and recombination, suggesting that RAD52 plays a crucial role in these processes (Game 1993).

~Correspondingauthor.

Rad52 is not homologous to any other known protein of Saccharomyces cerevisiae and has no apparent sequence motifs indicative of its function (Adzuma et al. 1984). Homologs of Rad52 have been identified in other eukaryotic organisms and show high conservation of the amino-terminal region (Bezzubova et al. 1993; Milne and Weaver 1993; Ostermann et al. 1993; Bendixen et al. 1994; Muris et al. 1994; Shen et al. 1995). The yeast RAD51, RAD55, and RAD57 gene products have sequence homology to prokaryotic RecA proteins, which are essential for homologous recombination and the SOS response in bacteria (Kans and Mortimer 1991; Aboussekhra et al. 1992; Basile et al. 1992; Shinohara et al. 1992; Lovett 1994). Electron microscopic analysis shows that Rad51 forms helical filaments on both ssDNA and dsDNA, but only the Rad51-ssDNA nucleoprotein filament is functionally relevant for strand exchange (Ogawa et al. 1993b; Sung and Robberson 1995). Rad51, Rad55, and Rad57 contain conserved putative nucleotide-binding motifs called the Walker A-type and B-type boxes (Walker et al. 1982). Mutation of the conserved lysine residue within the A-motif results in mutant phenotypes in DNA repair and meiosis for Rad51

GENES & DEVELOPMENT10:2025-2037 © 1996 by Cold SpringHarborLaboratoryPress ISSN 0890-9369/96 $5.00

2025


Bai and Symington

and Rad55 but not for Rad57 (Shinohara et al. 1992; Johnson and Symington 1995). Several lines of evidence suggest that the Rad proteins function in a complex. The Rad51 and Rad52 proteins have been shown to interact with each other by affinity chromatography (Shinohara et al. 1992) and the two-hybrid system {Milne and Weaver 1993; Hays et al. 1995). The amino-terminal region of Rad51 interacts with the carboxy-terminal region of Rad52 (Milne and Weaver 1993; Donovan et al. 1994). In addition, the two-hybrid system has revealed interactions between Rad51 and Rad54 (T. Kodadek, pets. comm.}, RadS1 and Rad55, Rad55 and Rad57, and between Rad51 and itself (Hays et al. 1995; Johnson and Symington 1995). Interactions among Rad proteins are also suggested by the finding that mutations in many of these genes can be suppressed by overexpressing some others. For example, RAD51 or RAD52 on a high-copy-number or CEN plasmid suppresses the DNA repair and recombination defects caused by rad55 and rad57 mutations (Hays et al. 1995; Johnson and Symington 1995). Overexpression of RAD51 also suppresses certain alleles of RAD52 (Milne and Weaver 1993; Schild 1995). The large subunit of yeast ssDNA-binding protein, encoded by the RFA1 gene, has been found to be involved in recombination (Firmenich et al. 1995; Smith and Rothstein 1995). An rfal missense mutation that reduces plasmid-to-chromosome gene conversion and results in defects in DNA repair is suppressed by RAD52 in a dosage-dependent manner (Firmenich et al. 1995). Finally, both the DNA repair and mitotic recombination defects of rad55 and rad57 null mutants are more severe at 23°C than at higher temperatures {Lovett and Mortimer 1987; Johnson and Symington 1995; Rattray and Symington 1995). This cold-sensitive phenotype suggests that Rad55 and Rad57 are part of, or act to stabilize, a protein complex. Spontaneous mitotic recombination is generally measured by the rate of prototroph formation between two different mutant alleles {heteroalleles) of a given gene. The heteroalleles may be present on homologous chromosomes in diploids, or as repeated sequences in haploids. Using an intrachromosomal inverted-repeat system we have shown previously that spontaneous mitotic recombination is reduced more than 3000-fold in rad52 mutants but only 4-fold in tad51 mutants (Rattray and Symington 1994). This result was surprising because Rad51 is thought to be the structural and functional RecA homolog in yeast, and therefore a more severe phenotype was expected. A rad51 rad52 double mutant showed the same rate of recombination as the rad52 single mutant, indicating that RAD52 is epistatic to RAD51. Similarly, RAD51 was found to be epistatic to RAD54, RAD55, and RAD57 {Rattray and Symington 1994, 1995). These results indicate that spontaneous mitotic recombination between inverted repeats is mostly dependent on RAD52, but only partially dependent on RAD51. Although mating-type switching, a doublestrand-break-induced gene conversion event, is dependent on RAD51, RAD52, RAD54, RAD55, and RAD57,

2026

GENES & DEVELOPMENT

other HO-induced conservative recombination events can occur in the absence of RAD51, RAD54, RAD55, and RAD57 {Sugawara et al. 1995; Ivanov et al. 19961. These results suggest that an alternate pathway exists for both spontaneous and DSB-induced mitotic recombination events in the absence of the RAD51 pathway. To elucidate the mechanism of RAD51-independent recombination, we searched for genes whose inactivation would further reduce recombination in the rad51 mutant background. In this paper we describe the identification of a novel gene that encodes a Rad52 homolog and show that mutation of this gene in combination with a rad51 mutation synergistically reduces recombination.

Results

Isolation of mutants in the RAD51-independent pathway A previously described ade2 inverted-repeat substrate was used to measure recombination (Rattray and Syo mington 1994)(Fig. 1). Using this substrate the recombination frequency can be visually assessed by a red/ white colony sectoring assay, and quantitated by the number of Ade + prototrophs within a population. The average rate of recombination was determined to be 1.8 x 10 -4 events/cell per generation in wild-type strains and 3.5 x 10 -s in rad51 mutants. A screen for recombination mutants was carried out in the rad51 background. A tad51 strain {B356-13D) was mutagenized by N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), and mutants displaying reduced sectoring were isolated. Of -15,000 colonies screened, five of the mutants isolated were shown by complementation tests to be RAD52 alleles. One other mutant, #25, displayed a low-sectoring phenotype similar to that of a rad52 strain. When this mutant was crossed to a rad52 strain not carrying the ade2 inverted-repeat substrate (LSY255), the resulting diploids were resistant to ~/-ray irradiation and sectored at wildtype levels. This result indicated: First, the low-sectoring phenotype of mutant #25 was not due to loss or mutation of the inverted-repeat substrate; second, the unidentified mutation in #25 was not allelic to RAD52; and third, the mutation was recessive. Mutant #25 was then backcrossed to an isogenic rad51 strain (B356-11A). The resulting diploids were homozygous for the rad51 mutation, so a RAD51-containing plasmid was introduced into the diploids to complement the rad51 sporulation defect. The diploids were sporulated, and, after tetrad dissection, plasmid-free haploid progeny were obtained by counterselecting against the plasmid maker gene URA3 on 5-fluoro-orotic acid (5-FOA)-containing medium. The low-sectoring phenotype segregated 2:2 in this backcross, indicating that the unidentified mutation in mutant #25 was a single gene trait. Mutant #25 was derived from a rad51 strain and was thus extremely sensitive to -y-ray irradiation. To test whether the unidentified mutation in #25 would confer ~/-ray sensitivity by itself, a strain was made that carried the unidentified mutation but had a wild-type RAD51 gene (B357-1D).


RAD51-independentrecombination ~ ade2-5'A>

Figure 1. ade2 inverted-repeat substrate and

recombination events. The substrate contains an intrachromosomal inverted duplication of alleles of the ade2 gene integrated at the HIS3 J ~/ I ADEZ f locus on chromosome XV. One of the ade2 alleles contains a deletion on the 5' side (ade2-5'Ll), and the other one has a frame-shift mutation at a NdeI site on the 3' side (ade2ade2.5'A > i n). Both ade2 alleles are nonfunctional and ~sover n the initial strain {Ade-) forms red colonies on nonselective media due to accumulation of a red pigment in the adenine biosynthetic pathI ADE2 f I ade2-n 1~ way. If recombination between the two alleles produces a wild-type ADE2 gene, a white sector will be formed within the red colony. Thus the recombination level of a strain can be visually assessed by colony sectoring and quantitatively determined by measuring the frequency of Ade + prototrophs within a popADE2 f ulation of cells. Ade + recombinants can arise by gene conversion and/or intrachromosomal Ade- (red) Ade+(white) crossover events. Gene conversion restores the wild-type NdeI sequence within the full-length repeat, and crossover results in inversion of the TRP1 gene between the two ade2 alleles. The types of recombination events can be distinguished by restriction endonuclease digestion and Southern hybridization.

|ade2-s'a,In~

i

X !

0

"

B357-1D displayed intermediate ~t-ray sensitivity. Based on this property and epistasis to rad52 (see below), the m u t a t i o n in #25 was designated rad59-1 (radiation sensitive mutation).

Cloning of the RAD59 gene The RAD59 gene was cloned by c o m p l e m e n t a t i o n of the ~-ray sensitivity of the rad59-1 mutant. Following transformation of strain B357-1D (rad59-1) with a S. cerevisiae genomic D N A library (Carlson and Botstein 1982) and selection for ~-ray resistance, three independent clones were isolated. Plasmids recovered from these transformants carried different sizes of inserts, but D N A sequence analysis revealed that all three inserts contained a c o m m o n D N A sequence. A subclone containing a 1.2-kb ScaI-BglII fragment was capable of rescuing the ~/-ray sensitivity of strain B357-1D (rad59-1) and the lowsectoring phenotype of m u t a n t #25 (rad51 rad59-1). A D N A fragment containing the 1.2-kb ScaI-BglII fragment was sequenced and an ORF encoding a predicted polypeptide of 238 a m i n o acids was identified. The determ i n e d RAD59 nucleotide sequence has been deposited into the GenBank data base under accession n u m b e r U53668. This sequence is completely identical to the recently released sequence of chromosome IV ORF D2548 in the Saccharomyces G e n o m e Database (SGD). A n u l l allele of RAD59 (rad59::LEU2) was made by replacing the entire coding region of RAD59 with a LEU2 gene. A rad59::LEU2 strain (B361-4C) was then crossed to the rad59-1 strain (B357-1D), and diploids were found to be sporulation proficient. For tetrads with four viable spores, all spores were ~-ray sensitive, indicating that the cloned RAD59 gene was allelic to the locus m u t a t e d in the rad59-1 mutant.

Rad59 is homologous to proteins of the Rad52 family A search of the protein data bases revealed that the predicted Rad59 protein is homologous to proteins of the Rad52 family, including the Rad52 proteins from S. cerevisiae, Kluyveromyces lactis, h u m a n , mouse, chicken, and the Rad22 protein from Schizosaccharomyces pombe (Bezzubova et al. 1993; Milne and Weaver 1993; Ostermann et al. 1993; Bendixen et al. 1994; Muris et al. 1994; Shen et al. 1995). The Rad52 f a m i l y m e m b e r s are most highly conserved at the a m i n o terminus. Rad59 is about half the length of the Rad52 family proteins and is also homologous to the a m i n o - t e r m i n a l conserved region (Fig. 2). However, Rad59 is the least conserved of the Rad52-1ike proteins. The Rad59 and Rad52 proteins from S. cerevisiae have 28% residue identity and 50% similarity. W i t h i n the most conserved region, residue identity and similarity are 37% and 55%, respectively.

rad59 and rad51 synergistically reduce recombination The rad59-1 mutant, like the rad51 mutant, showed only slightly reduced colony sectoring compared with the wild type. However, m u t a n t #25 (rad51 rad59-1) was isolated from a rad51 m u t a n t background by the strong reduction in colony sectoring (Fig. 3). This synergistic reduction in recombination was verified by quantitatively determining the rates of recombination for each strain (Table 1). Compared w i t h wild-type strains, recombination rates in rad51 and rad59-1 single m u t a n t s were reduced only 4- to 5-fold, whereas the rates in rad51 rad59-1 double m u t a n t s were reduced about 1200-fold. The rad59::LEU2 null allele gave the same results as the rad59-1 allele in recombination and ~-ray sensitivity assays. Thus rad59-1 is likely to be a null mutation. For

GENES& DEVELOPMENT

2027


Bai and Symington

48 I G L L ~ K I E R Y T Y N I ~ H N N K Y B K H N L S K - ~ A L ~ Q F ~ r ~ T ~ D ~ R

51 S E D ~ T K ~ D K ~ K ~ F ~ T S R I A I I I N N W R V ~ ~ Q ~ S ~ 29 37 38 38

K D ol~KBo K ~ ~ n N K U ~ F ~ S S RBA~WUA~Rim"JliNDNlS - - - L~SBS R~V~RBS~PIBII~F S B S R R I m S WBABEmm~wmmmmmFm~mm~S -i Ki R~ MAl l | ~ l l | , i | ' l l l l l - ' l R~ M i K ~ R ~ M A B m q D B m m " I " a ; I R ~ | M -~ HIln:~lhNl'illLlm Q Ailllonnll'lnlm:n~vm I S lm,_lmn 14~ F ILqiIVm~l

o.,,,,c

P

101 T ENK s N v I n - - 61 TE~JK NT ~ . ~ 76 S m l R S IIINIE 86 I q -

Figure 2. Sequence alignment of the conserved amino-terminal region of Rad52-1ike proteins. Sc: S. cerevisiae; KI: K. lactis; Sp: S. pombe; Hs: H. sapiens {human); Mm: M. musculus (mouse); Gd: G. domesticus (chicken).

147 c N ~ B R

i T -NsNR

99

VlRN E ~ R ~ S B F

I ~ T

"

;r!d

~ E c Y N R smnni~

K

S

simplification the rad59::LEU2 allele will be referred to as rad59 in subsequent analyses. Recombination rates in rad51 rad52 and rad52 rad59 double m u t a n t s and in rad51 rad52 rad59 triple m u t a n t s were similar to that in rad52 single mutants, indicating that R A D 5 2 is epistatic to RAD51 and R A D 5 9 w i t h respect to the overall rate of recombination. With the use of the ade2 inverted-repeat recombination assay, in c o n j u n c t i o n w i t h Southern hybridization analysis, it was s h o w n that 50% of the Ade + prototrophs recovered from wild-type strains were noncrossover events (simple gene conversions), and the other 50% were crossover events (simple crossovers and gene conversions w i t h crossing-over)(Table 1). In rad51 m u t a n t s noncrossover events were reduced 20-fold, while crossover events were reduced only 3-fold. In rad52 m u t a n t s overall r e c o m b i n a t i o n rates were greatly reduced, but the distribution of r e c o m b i n a t i o n events was similar to wild-type strains. In rad51 rad52 double m u t a n t s the overall r e c o m b i n a t i o n rates were close to that of rad52 single m u t a n t s . However, noncrossover events in rad51 rad52 double m u t a n t s were preferentially reduced compared w i t h crossovers, an outcome similar to that observed in rad51 single m u t a n t s . The distribution of the classes of events in rad59 m u t a n t s is similar to that in wild-type strains but significantly different from that in radS1 strains (Table 1; P < 0.05). The Ade + prototrophs examined from rad51 rad59 double m u t a n t s show a bias toward the crossover class, but not as extreme as observed in rad51 and rad51 rad52 double mutants. This distribution of events is significantly different from wild-type and rad59 strains (P < 0.05), but not significantly different from the distribution observed in rad51 strains (P > 0.05).

L S

125 125

G

SpRad22 HsRad52 MmRad52 ~ GdRad52 ScRad59 ScRad52 KIRad52 SpRad22 HsRad52 MmRad52 GdRad52

VNNGENTNTSEV RQ - - -

R

ScRad59 ScRad52 KIRad52

~

K

.........

"

6

C

~

EBI

L

IB-

Yilm~l~

~ -

,

soR~es9 KIRadS2

MmRad52 q GdRad52

repeat substrate, a single m u t a t i o n in RAD51 or R A D 5 9 slightly reduced r e c o m b i n a t i o n on a Tn903 inverted-repeat substrate, whereas s i m u l t a n e o u s m u t a t i o n of both RAD51 and R A D 5 9 synergistically reduced recombination (Fig. 4A). On a leu2 inverted-repeat substrate, the rad59-1 m u t a t i o n reduced r e c o m b i n a t i o n 6-fold, but the rad51 m u t a t i o n had no effect. The rate of r e c o m b i n a t i o n in the rad51 rad59-1 double m u t a n t was the same as that observed for the rad59-1 single m u t a n t , indicating no synergism between the rad51 and rad59-1 m u t a t i o n s on this recombination reporter (Fig. 4B). Next we tested the r e c o m b i n a t i o n a l repair of DSBs induced by H O endonuclease in a l a c Z inverted-repeat plasmid IFig. 51. A D N A fragment w i t h an HO cut site, embedded in one copy of lacZ, serves as a site for DSB formation when H O endonuclease is expressed. The DSB can be repaired using wild-type sequences located on a second, promoterless, copy of l a c Z on the same plasmid. Unrepaired plasmids are lost from the population. The

RAD

rad51

4

!

-g

rad59-1

~, ?

*;J

rad51 rad59-1

RAD59 is required for intrachromosomal but not in terchrom os om al recom bin a tion To determine if the R A D 5 9 gene has a general role in recombination, we tested the effect of the rad59 mutation on other i n t r a c h r o m o s o m a l inverted repeats, on mitotic i n t e r c h r o m o s o m a l recombination, and on sporulation. Similar to their effects on the ade2 inverted-

2028

GENES & DEVELOPMENT

Figure 3. Colony sectoring for strains with the ade2 invertedrepeat substrate. During the growth of the colony recombination events that generate a wild-type ADE2 gene are visualized as white sectors within the red colony. Strains were grown at 30°C for 7 days on YEPD medium.


RAD51-independent recombination Figure 4. Recombination rates of the Tn903 and leu2 inverted-repeat substrates. (A) The Tn903 inverted-repeat substrate. A fragment of the bacterial transposon Tn903 was inteleu2-5'A~ " / ~ IS903 grated into yeast chromosome 111 between the HIS4 and HML loci. The fragment conpRS416-SU Tn903 inverted repeats tains the kanamycin resistance gene (kanq of Tn903, flanked by two inverted copies of the IS903 sequences. The inverted IS903 repeats Straingenotype Rateof Kanrcells Relativerate Straingenotype Rateof Leu÷cells Relativerate can undergo an intrachromosomal reciprocal RAD 2.3 X 10-7 100 RAD 3.3 (+1.4)X 104 100 exchange resulting in inversion of the kanr rad51 5.3X 10.8 23 rad51 5.5 (_+0.5)X104 169 gene. A yeast strain containing a single copy rad59-1 5.1 X 108 22 rad59-1 5.1 (+2.0)X 10-7 16 of the kanr gene in its original orientation is sensitive to 0.5 m g / m l of G418. After inverrad51rad59-1 < 1.0 X 10-s