A Core Activity Associated with the N Terminus of the Yeast RAD52 Protein Is Revealed by RAD51 Overexpression Suppression of C-Terminal rad52 Truncation Alleles

Corresponding author: Dennis M. Livingston, Department of Biochemistry, Molecular Biology and Biophysics, 4-225 Millard Hall, 435 Delaware St. SE, University of Minnesota, Minneapolis, MN 55455-0347., livin001{at}maroon.tc.umn.edu (E-mail)

Communicating editor: L. S. SYMINGTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

C-terminal rad52 truncation and internal deletion mutants were characterized for their ability to repair MMS-induced double-strand breaks and to produce viable spores during meiosis. The rad52-251 allele, encoding the N-terminal 251 amino acids of the predicted 504-amino-acid polypeptide, supports partial activity for both functions. Furthermore, RAD51 overexpression completely suppresses the MMS sensitivity of a rad52-251 mutant. The absence of the C terminus in the truncated protein makes it likely that suppression occurs by bypassing the C-terminal functions of Rad52p. RAD51 overexpression does not suppress the low level of spore viability that the rad52-251 allele causes and only partially suppresses the defect in rad52 alleles encoding the N-terminal 292 or 327 amino acids. The results of this study also show that intragenic complementation between rad52 alleles is governed by a complex relationship that depends heavily on the two alleles involved and their relative dosage. In heteroallelic rad52 diploids, the rad52-251 allele does not complement rad52 missense mutations altering residues 61 or 64 in the N terminus. However, complementation is achieved with each of these missense alleles when the rad52-251 allele is overexpressed. Complementation also occurs between rad52-327 and an internal deletion allele missing residues 210 through 327. We suggest that the first 251 amino acids of Rad52p constitute a core domain that provides critical RAD52 activities.

RAD52 plays a critical role in recombinational repair of double-strand breaks in yeast. Previous work on RAD52 suggests that it may have multiple functions in this process. Yeast Rad52p has been found to bind DNA and to promote annealing of complementary single-stranded DNA (MORTENSEN et al. 1996 ). Other work has detected binding between Rad52p and the strand exchange protein, Rad51p (MILNE and WEAVER 1993 ; SUNG 1997 ), as well as the heterotrimeric single-stranded DNA binding protein, replication protein A (RPA; FIRMENICH et al. 1995 ; HAYS et al. 1998 ). Studies using the human Rad52p homolog, hRad52p, show that it binds to itself (SHEN et al. 1996 ) suggesting that the yeast protein may also self-associate and function as a dimer or as a larger aggregate. Besides the identification of these interactions, recent in vitro studies have shown that Rad52p works with RPA to promote Rad51p-mediated strand exchange (SUNG 1997 ; NEW et al. 1998 ; SHINOHARA and OGAWA 1998 ), and similar results have been obtained with the human Rad52p homolog (BENSON et al. 1998 ). More recent studies on human Rad52p demonstrate that the protein forms a multimeric structure (SHINOHARA et al. 1998 ; VAN DYCK et al. 1998 ; PASSY et al. 1999 ) that binds to the ends of a DNA double-strand break (VAN DYCK et al. 1999 ). This indicates that Rad52p may be required to mediate access of recombinational repair proteins to the site of double-strand breaks. Still unknown however are the roles that the domains in Rad52p have in integrating the functions of Rad51p and RPA during double-strand break repair.

Our laboratory has created rad52 mutations and has characterized their effects on mitosis and meiosis. One study from our laboratory examined two C-terminal truncation alleles, rad52-327 and rad52-169 (BOUNDY- MILLS and LIVINGSTON 1993 ). The rad52-327 allele encodes the first 327 amino acids of the predicted 504-amino-acid Rad52p polypeptide, and it supports partial double-strand break repair activity. The rad52-169 allele contains the first 167 amino acids of Rad52p, and mutants carrying this allele behave in a manner similar to that of a mutant with a deletion of the coding region. Characterization of the rad52-327 allele also found that the intact N-terminal end of Rad52p provided by this truncation allele could be fully complemented by providing a second allele with a wild-type C-terminal end like the N-terminal missense mutation rad52-2 (P64L). Because a rad52-169/rad52-2 diploid did not exhibit intragenic complementation, this study defined a region between residues 169 and 327 needed for complementation. Further examination demonstrated that the rad52-327 allele could be suppressed by overexpression of RAD51 while the rad52-169 allele could not (MILNE and WEAVER 1993 ; KAYTOR and LIVINGSTON 1996 ). The results of the intragenic complementation and RAD51 suppression studies are indicative of the partial activity of the rad52-327 peptide that is no longer present in the rad52-169 peptide. This suggests that a region between amino acids 169 and 327 is vital to Rad52p function.

To better understand these phenomena, a set of rad52 alleles with C-terminal truncations and in-frame deletions that eliminate portions of the region between residues 169 and 327 have been created to answer three questions. First, how much of the N terminus is required for partial activity? Second, are there domains in Rad52p that must be intact for RAD51 overexpression to effectively suppress mutant rad52 alleles? Third, what domains of Rad52p must be on the same polypeptide to achieve intragenic complementation between two mutant rad52 alleles?

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains and growth conditions:
All yeast strains in this study are isogenic with SSL204 (DORNFELD and LIVINGSTON 1991 ) and are listed in Table 1. Cells were grown on YPD agar plates and in YM-1 liquid medium or, for strains carrying plasmids, on Ura omission plates and in SC-Ura liquid medium.

MMS survival:
Sensitivity to methyl methanesulfonate (MMS)-induced double-strand breaks was measured by the method of PRAKASH and PRAKASH 1977 . Cells were grown at 30° to saturation and treated with 0.5% MMS at room temperature for varying lengths of time. Viability was determined by plating aliquots of cells onto YPD or Ura omission plates and incubating them at either 30° or 33°. The Ura omission medium was used for RAD51 overexpression studies when strains were transformed with the 2µ vector YEp24 or YEp24-RAD51, or for intragenic complementation studies when rad52 alleles were overexpressed from the 2µ vector pRS426 (CHRISTIANSON et al. 1992 ).

Spore viability:
Diploid cells were patched onto YPD plates and then replica plated to 1% potassium acetate plates supplemented with required amino acids and nucleic acid bases and incubated at 30°. After 3–5 days tetrads were dissected on YPD plates and allowed to germinate at 30°. Spore colonies were counted after 3 days.

Construction of rad52 alleles:
C-terminal truncation alleles were constructed by using PCR to introduce a stop codon and a BamHI site at specific positions in RAD52. Fragments of RAD52 were amplified using primer Rad52-2 (5'-ggctttggtgtgttgttgat-3') with primer Rad52-1597s (5'-aggatcctaatccgttggcctaaacaaa-3'), Rad52-1721s (5'-aggatcctacaggttcttcgtcgagtcg-3'), or Rad52-1840s (5'-aggatcctaatcctgtttggaagcatcg-3'). Conditions for PCR were 94° for 4 min followed by 30 cycles of 1 min at 94°, 1 min at 57°, and 2.5 min at 72°. This was followed by a 4-min incubation at 72°. The resulting fragments were subcloned into YIp5-RAD52 using restriction endonuclease digestion by BstEII and BamHI. New rad52 alleles were introduced into strains SSL204 and SSL204A (DORNFELD and LIVINGSTON 1991 ) or SSL231 (KAYTOR and LIVINGSTON 1994 ) by the two-step gene replacement method (SCHERER and DAVIS 1979 ). The presence of the deletions was confirmed by PCR using primer pairs Rad52-2 and Rad52-10 (5'-gaatatgcttggacgtagtctg-3') or Rad52-1 (5'-cccgttagtgattctcgatg-3') and Rad52-8 (5'-tttcggccaggaagcgtttc-3').

The internal deletion rad52-i210B was constructed by using primers Rad52-1 (5'-cccgttagtgattctcgatg-3') and Rad52-1597 (5'-aggatccgttggcctaaacaaattg-3') to amplify a fragment that was subcloned into RAD52 using the BstEII and BamHI sites. This fused amino acid 210 to downstream sequences at the BamHI site. rad52-i251B and rad52-i292B were constructed in the same manner except that primers Rad52-1721 (5'-aggatccaggttcttcgtcgagtcgg-3') and Rad52-1840 (5'-aggatcctgtttggaagcatcgag-3') were used.

The rad52-2:i210B allele contains two lesions, the rad52-2 missense mutation and the in-frame deletion from the rad52-i210B allele. It was constructed by subcloning the BglII-SalI fragment from rad52-i210B into rad52-2 carried in the vector YCp50.

rad52-96A is a missense mutation that was recovered from a library of mutagenized rad52 plasmids (KAYTOR and LIVINGSTON 1994 ). The causative mutation in this allele was localized and found to be an amino acid substitution of glycine for arginine 235.

Several rad52 alleles were maintained on either the CEN plasmid, YCp50, or on the 2µ plasmid pRS426 (CHRISTIANSON et al. 1992 ). rad52 alleles were overexpressed by subcloning EcoRI-SalI fragments from YCp50 plasmids containing rad52-2, rad52-301A, rad52-327, rad52-292, and rad52-251 into pRS426.

Visualization of mutant rad52 proteins:
To place an epitope tag at the C terminus of Rad52p, the stop codon and adjacent sequence were mutagenized to extend the coding sequence to include a NotI restriction site. The sequence bearing three copies of the epitope for hemagglutinin (HA1; TYERS et al. 1992 , TYERS et al. 1993 ) was cloned into the NotI site. The extended gene confers MMS resistance to a strain deleted for RAD52, suggesting that the tag does not interfere drastically with RAD52 activity. The tag was also placed on mutant alleles, and all were cloned between the EcoRI and SalI sites of the vectors pRS316 and pRS426 (CHRISTIANSON et al. 1992 ). Protein extracts were prepared from log-phase yeast cultures using a glass bead breakage method (KATZ and SOLOMON 1988 ). The proteins were separated by electrophoresis in a SDS gel and then electroblotted to nitrocellulose (Schleicher & Schuell, Keene, NH). HA-tagged polypeptides were detected by incubating the filter with a mouse anti-HA monoclonal antibody (Roche Molecular Biochemicals, Indianapolis), followed by incubation with a goat anti-mouse IgG conjugated to horseradish peroxidase (HRP; Bio-Rad, Hercules, CA). Proteins were visualized using HRP substrate reagents followed by exposure to Hyperfilm ECL according to the manufacturer's instructions (Amersham Pharmacia, Buckingham, UK).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Partial activities conferred by rad52 truncation alleles:
We had shown previously that a strain harboring a rad52 allele (rad52-327) retaining the first 327 amino acids of the 504-amino-acid open reading frame maintains partial resistance to MMS exposure, while a strain with a shorter truncation allele (rad52-169) retaining only 167 amino acids is as sensitive as a strain with a deletion allele (BOUNDY-MILLS and LIVINGSTON 1993 ). To more precisely define how little of Rad52p is required for this activity, three additional truncation alleles were assessed for the ability to promote survival to MMS exposure. These new alleles retain the first 292, 251, and 210 amino acids and are designated rad52-292, rad52-251, and rad52-210, respectively (Fig 1).

View larger version (32K): In this window In a new window Download PPT slide   Figure 1. rad52 alleles used in this study. The RAD52 transcription unit is represented by the arrow, and the relative locations of eight missense mutations are indicated above the arrow. rad52 truncation and internal deletion alleles are shown below. The solid boxes denote the regions of RAD52 present in each allele, and dotted lines denote regions deleted.

The sensitivity of these strains to MMS exposure was determined by incubating stationary phase cells in a MMS solution and withdrawing aliquots at various times. Fig 2 depicts the time course of survival for the truncation mutants. The rad52-327 strain is ~200-fold more resistant to MMS exposure than the rad52-169 strain after 30 min of exposure to MMS. The three new truncation mutants are all more resistant than the rad52-169 strain after 30 min of exposure to MMS. The rad52-292 and the rad52-251 strains are 100- and 80-fold more resistant, respectively, than the rad52-169 strain, while the rad52-210 strain is at best 2-fold more resistant than the control strain. Thus, partial resistance to MMS exposure persists when only 251 amino acids are present and may even be found in a protein with 210 amino acids.

View larger version (35K): In this window In a new window Download PPT slide   Figure 2. MMS sensitivity of rad52 truncation alleles and the effect of RAD51 overexpression. (A) The log fraction of surviving cells after exposure to MMS from strains with the rad52 C-terminal truncation alleles rad52-327, rad52-292, and rad52-210, as well as from a wild-type strain (RAD52) and from a deletion strain (rad52-HS), was plotted vs. time of exposure. (B) The log fraction of surviving cells after exposure to MMS from strains with the rad52 C-terminal truncation alleles rad52-251, rad52-210, and rad52-169 was plotted vs. time of exposure. The solid symbols are the strains containing the vector control. The open symbols are the strains transformed with YEp24::RAD51. These studies were done at 33°.

As an independent measure of retention of function, we also assessed the ability of homoallelic truncation diploids to produce viable spores. We previously determined that homoallelic diploids of rad52-327 yield viable spores, the percentage being a function of strain background and of sporulation conditions (BOUNDY-MILLS and LIVINGSTON 1993 ; NGUYEN and LIVINGSTON 1997 ). In the current study the homoallelic rad52-327 diploid yielded ~40% spore viability in contrast to the >90% viability of the wild-type diploid (Table 2). The rad52-251 and rad52-292 homoallelic diploids also yielded nearly 40% viable spores, while the rad52-210 diploid did not sporulate. Thus, the intermediate spore viability correlates with the partial retention of resistance to MMS exposure. Together, these phenotypic assays suggest that a Rad52 polypeptide chain with as little as 251 amino acids can promote a low level of MMS damage repair and partially succeed at sporulation.

N-terminal truncation alleles that retain partial activities can also be suppressed by high-level expression of RAD51:
Work from our laboratory and from other laboratories has shown that high-level expression of RAD51 can suppress some rad52 alleles (MILNE and WEAVER 1993 ; SCHILD 1995 ; KAYTOR and LIVINGSTON 1996 ). In our studies we found that both the truncation allele rad52-327 and a ts allele, rad52-252 (D371G), containing a missense mutation in the C-terminal region can be suppressed to near wild-type levels in tests of resistance to MMS treatment (KAYTOR and LIVINGSTON 1996 ). Five other alleles [rad52-1 (A90V), rad52-2 (P64L), rad52-76A (N97T), rad52-23A (V128I), rad52-22B (V162A)], all missense mutations mapping to the N terminus of the protein, are not suppressed to nearly the same extent as rad52-327 or rad52-252.

To further these studies, we attempted to suppress the shorter truncation alleles by overexpression of RAD51. The results (Fig 2A and Fig B) show that, like the rad52-327 allele, both the rad52-292 and rad52-251 alleles are nearly completely suppressed by overexpression of RAD51. For example, the rad52-251 strain is 400-fold more resistant to MMS treatment (at 30 min) when RAD51 is overexpressed. This resistance level is similar to that of a RAD52 strain used as a control. The rad52-210 strain is partially suppressed by RAD51 overexpression and the rad52-169 strain is not suppressed.

In addition we retested the six missense alleles, rad52-1, rad52-2, rad52-22B, rad52-23A, rad52-76A, and rad52-252, that we had previously examined (KAYTOR and LIVINGSTON 1996 ) along with another missense allele, rad52-96A (R235G). These tests were all done using liquid treatment of a cell suspension and a complete time course. Only the rad52-252 and rad52-96A alleles are suppressed to near wild-type levels; none of the five missense mutations in the N-terminal portion are suppressed (data not shown). These results confirm the positional allele specificity for RAD51 suppression that we previously observed.

Suppression by RAD51 overexpression also works during meiosis. The results in Table 2 show that the partial spore viability of a homoallelic diploid of rad52-327 is raised from 45 to 80% and the value for a homoallelic diploid of rad52-292 is raised from 39 to 62% when RAD51 is overexpressed. Significant suppression does not occur for a homoallelic diploid of rad52-251, and the rad52-210 and rad52-169 homoallelic diploids do not sporulate even when RAD51 is overexpressed. The results show that while truncation alleles can be suppressed during meiosis, the level of suppression is less than that for MMS resistance in mitotic cells. In addition the suppression does not extend to alleles rad52-251 or rad52-210 in meiosis, whereas it does in measurements of MMS resistance in mitosis.

Taken together, these results define three characteristics about suppression by RAD51 overexpression. First, suppression by RAD51 overexpression requires an intact N-terminal sequence because none of the missense mutations mapping to the N terminus can be suppressed to the same level as the truncation alleles that have the wild-type sequence in this region. Second, these results show that RAD51 overexpression most likely suppresses MMS sensitivity by bypassing the need for the C terminus of Rad52p because truncations that retain only the first 251 residues and are missing the putative Rad51p binding site (MILNE and WEAVER 1993 ; MORTENSEN et al. 1996 ) are nearly completely suppressed. Third, RAD51 high copy suppression does not work as efficiently during meiosis.

rad52 alleles containing in-frame deletions:
To learn whether the properties of the truncation alleles could be altered by including the C-terminal third of Rad52p, we constructed internal deletion alleles. These alleles, rad52-i210B, rad52-i251B, and rad52-i292B, fuse amino acids 210, 251, and 292, respectively, of Rad52p in frame to residue 328. The truncation alleles were tested for their MMS resistance, with and without overexpression of RAD51, and for their ability to sporulate as homoallelic diploids. For the most part the internal deletion alleles have phenotypes much like those of the truncation alleles.

First, while the internal deletion mutants are sensitive to exposure to MMS, like the truncation alleles, they retain partial activity to resist MMS killing (Fig 3). After 30 min exposure, rad52-i210B is as sensitive to MMS exposure as rad52-327. rad52-i251B and rad52-i292B are more resistant than rad52-327 at this time. (After 45 min exposure all mutants lost viability.) Second, diploids homoallelic for the internal deletion alleles sporulate and yield approximately half the number of viable spores of a wild-type diploid (Table 2). In this case the phenotypes of rad52-i251B and rad52-i292B are similar to the phenotypes of the corresponding truncation alleles rad52-251 and rad52-292, respectively. A major difference is that a homoallelic diploid of rad52-i210B sporulates and yields viable spores. The corresponding truncation allele rad52-210 does not sporulate. A possible reason for this difference could be that residues extending from 328 to the C terminus are sufficient to support a reduced level of successful sporulation. Third, like the truncation alleles, the MMS sensitivity of the internal deletion mutants can be suppressed by high-level expression of RAD51 (Fig 3). Interestingly, while high level of expression of RAD51 suppresses rad52-251 and rad52-292 nearly to wild-type resistance (Fig 2), the three internal deletion alleles are not completely suppressed. The lack of complete suppression is evidenced by the result that after 45 min none of the three internal deletion mutants are as resistant as a suppressed rad52-327 mutant (Fig 3). We suspect that the inclusion of the C-terminal residues containing the Rad51p binding domain (MILNE and WEAVER 1993 ; MORTENSEN et al. 1996 ) interferes and inhibits the bypass suppression observed for the truncation alleles. The effect of high-level RAD51 expression on the sporulation capacity of the internal deletion mutants was not examined.

View larger version (28K): In this window In a new window Download PPT slide   Figure 3. MMS sensitivity of rad52 internal deletion alleles and the effect of RAD51 overexpression. The log fraction of surviving cells after exposure to MMS from strains with the rad52 internal deletion alleles rad52-i292B, rad52-i251B, and rad52-i210B were compared to the survival of a strain with the truncation allele rad52-327. The solid symbols are the strains containing the vector control. The open symbols are the strains transformed with YEp24::RAD51. These studies were done at 33°.

The phenotypes of the internal deletion alleles again show that partial activities are retained in the absence of the full primary sequence of Rad52p. While the retention of partial activities of the internal deletions most likely results from the preservation of the N-terminal domain, other possibilities are plausible. In the case of the internal deletion alleles, the partial activities might result not from the loss of a functional domain but rather from the inhibitory effect that the N- and C-terminal domains exert on each other by being brought closer together than they are in their normal arrangement.

Intragenic complementation between mutant forms of Rad52p appears to require a continuous wild-type sequence of the first 327 amino acids:
Earlier work in our laboratory found that N-terminal and C-terminal functions may be provided in trans by the rad52-327 and rad52-2 (L64P) alleles (BOUNDY-MILLS and LIVINGSTON 1993 ). To ascertain whether any of the shorter truncation alleles could also complement the rad52-2 mutation, heteroallelic diploids containing the shorter truncation alleles and the rad52-2 allele were made and subjected to MMS treatment. Surprisingly, none of the shorter truncation alleles, rad52-292, rad52-251, or rad52-210, complement the rad52-2 mutation under these conditions (Fig 4A). Consistent with this observation is the result that the rad52-2/rad52-327 heteroallelic diploid is also the only strain to demonstrate complementation in meiosis. This diploid yields a wild-type level of spore viability, while heteroallelic diploids containing rad52-2 with either the rad52-292 or the rad52-251 truncation allele yield the partial spore viability of the respective homoallelic truncation diploids (data not shown). As expected, heteroallelic diploids of rad52-210 or rad52-169 with rad52-2 do not sporulate.

View larger version (24K): In this window In a new window Download PPT slide   Figure 4. Summary of intragenic complementation and overexpression results. The figure represents a compilation of the relative resistance to MMS exposure of heteroallelic rad52 diploids or, in the case of overexpression, rad52 haploids carrying a second rad52 allele on a high copy plasmid. +, either wild-type levels of survival or a deficit of less than an order of magnitude relative to a wild-type control after a 45-min exposure. -, either a decrease of three to four orders of magnitude relative to wild-type or no significant increase compared to an appropriate negative control after a 45-min exposure. The positive control for complementation experiments was either a homoallelic wild-type diploid or a heteroallelic diploid known to demonstrate complementation, e.g., rad52-2/rad52-327. The negative control was either a homoallelic rad52 deletion diploid or a heteroallelic diploid known to be incapable of complementation, e.g., rad52-2/rad52-169. The positive control for the overexpression studies was a rad52-2 strain transformed with rad52-327 on a high copy plasmid. This strain has wild-type levels of resistance to MMS. The negative control was a rad52-327 strain with rad52-2 overexpressed. This strain is no more resistant than either haploid strain by itself. For the diploids, C/C refers to the chromosomal expression of both alleles. For the haploids, C/P and P/C refer to the expression of one of the two alleles from a 2µ vector, the first letter designating the allele depicted at the top of each section heading. (A) Heteroallelic diploids with rad52-2 crossed to the truncation alleles. (B) Heteroallelic diploids with rad52-301A crossed to the truncation alleles. (C) Heteroallelic diploids with rad52-2 crossed to the internal deletion alleles. (D) Heteroallelic diploids with rad52-327 crossed to the internal deletion alleles. In the case of the rad52-2:i210B allele, complementation was judged in a rad52-327 haploid transformed with the two-lesion allele expressed from a CEN vector. All experiments were done at 30°.

To assess whether the failure to achieve complementation was a peculiarity of the rad52-2 allele, a second N-terminal missense allele, rad52-301A (K61N), was also tested. Again, the rad52-301A allele can be complemented for MMS resistance by the rad52-327 allele, but not by any of the shorter truncation alleles (Fig 4B). These results suggest that complementation by mutant protein chains may require more than the core defined by the partial activity and RAD51 suppression results.

The set of three internal deletion alleles, rad52-i210B, rad52-i251B, and rad52-i292B, corroborated the need for more than the first 251 amino acid residues to achieve complementation in heteroallelic diploids. None of these internal deletion alleles are capable of complementing the MMS sensitivity of a rad52-2 allele in a heteroallelic diploid (Fig 4C). At this point we had yet to find complementation with any of the truncation or internal deletion alleles that deleted amino acid residues between 210 and 327.

Additional experiments revealed that truncation and internal deletion alleles can be complemented under certain conditions. First, all three internal deletion alleles, the most severe (rad52-i210B) removing residues 210 through 327, complement the rad52-327 allele (Fig 4D). This result dispelled the possibility that previous failures to achieve complementation between the internal deletion alleles and the rad52-2 allele resulted from instability of the internal deletion peptides. Rather, the results implied that to achieve complementation, one of the complementing peptides has to contain wild-type sequences that include residues between amino acids 251 and 327. Consistent with this interpretation is the failure of the rad52-292 allele to complement any of the internal deletion alleles (data not shown), but the success of the rad52-327 allele in complementing a complex allele containing both the rad52-2 missense mutation and the internal deletion of the rad52-i210B allele (Fig 4D). This latter success illustrates well that nearly wild-type RAD52 activity can be achieved between two peptides, one wild type for the first 327 residues and one wild type for the residues beyond position 327.

We could also achieve complementation under a second condition in which one of the two mutant alleles is expressed on a high copy plasmid. In these experiments haploids containing one rad52 allele were transformed with a 2µ plasmid-based vector containing a second allele. We found that when the truncation alleles rad52-292, rad52-251, and rad52-210 are expressed from high copy plasmids, they complement the missense alleles rad52-2(P64L) and rad52-301A(K61N) (Fig 4A and Fig B). The shortest truncation allele, rad52-169, also complements rad52-301A under these conditions. Control experiments (data not shown) revealed that high-level expression of any truncation allele in a mutant bearing a chromosomal copy of the same mutation does not enhance the mutant's MMS resistance. Thus, the truncated Rad52 polypeptides can be complemented by full-length missense mutations having a wild-type C-terminal sequence as long as the truncation alleles are expressed on a high copy plasmid.

We also performed studies to learn whether the reverse configuration of high copy expression of the missense alleles within mutants bearing the truncation alleles would lead to complementation. The rad52-2 allele does not complement any of the truncation alleles (Fig 4A). This result is consistent with our previous observation (repeated in this study) that the same relationship holds for the rad52-2 and rad52-327 alleles, two alleles that, unlike the shorter truncations, complement in heteroallelic diploids (BOUNDY-MILLS and LIVINGSTON 1993 ). While high copy expression of rad52-2 does not lead to complementation, high copy of rad52-301A, when expressed in any of the truncation alleles, does lead to complementation (Fig 4B). Again, we stress that control experiments show that MMS resistance cannot be achieved simply by overexpressing a mutant allele in a strain bearing a chromosomal copy of the same mutation, indicating that the complementation that is observed is dependent on the interaction or sequential function of both mutant alleles.

To investigate whether the differential expression levels of the mutant proteins might contribute to some of the differences in the complementation patterns, we analyzed relative protein levels by epitope-tagging mutant proteins at their C-terminal end. Blotting and probing revealed that all mutant forms could be detected when the alleles were expressed on a high copy plasmid except for the product of rad52-2 that was seen only in small amounts (Fig 5). The detection of all mutant protein forms, save for the low level of the rad52-2 product, shows that all mutant forms are expressed. We note that the products of the truncation alleles and internal deletion alleles appear to be present in greater amounts than the full-length, wild-type, and rad52-301A products (Fig 5). Whether or not the greater amounts represent greater stability of these truncated forms or simply greater survival during extraction or more efficient blotting is unknown. Nevertheless, the study suggests that the diminished capacity of the truncation and internal deletion alleles in functional tests is unlikely to result from unstable protein products. Furthermore, while we could detect the epitope-tagged proteins expressed from the high copy plasmids, the signal from the same tagged proteins expressed from low copy plasmids was significantly less detectable (data not shown). The diminished signal from proteins expressed from low copy plasmids suggests that each mutant form is present in lower amounts when expressed from a single chromosomal copy than from a high copy plasmid.

View larger version (61K): In this window In a new window Download PPT slide   Figure 5. Visualization of mutant proteins. The procedures used to produce this blot are described in MATERIALS AND METHODS. The film was exposed for 5 min. The full-length product of the rad52-2 allele was detectable after exposure for 2 hr.

Thus, these measurements indicate that most of our complementation data is not complicated by gross instability of mutant forms. The exception is the very low level of the rad52-2 product we can detect. If this failure is indicative of a severe cellular instability for this mutant form, it would partially explain why high-level expression of rad52-2 does not complement the truncation alleles. The possibility that the truncation products stabilize the rad52-2 product when the truncation products are expressed from a high copy plasmid is being investigated.

Our overall conclusion is that two mutant forms of Rad52p can complement each other under two conditions. The first is when both alleles are expressed from chromosomal copies in a heteroallelic diploid, and one allele encodes wild-type sequences containing the first 327 residues of the N terminus and the other allele encodes wild-type sequences for the remainder of the C terminus. The second is that the requirement for the first 327 residues can be relaxed to 210 residues, and even to 169 residues, in some cases, if one of the complementing proteins is expressed at high levels.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The RAD52 core domain:
Our study aimed to define the minimal requirements for functionality of truncated and internally deleted RAD52 products. The allele rad52-251, encoding a peptide that is less than half of the longest possible open reading frame, retains a remarkable amount of activity in tests of resistance to MMS exposure and in spore viability. In these tests that rely on expression of a single chromosomal copy, no shorter truncation allele could function as well, nor could the truncation alleles encoding 292 and 327 residues greatly exceed rad52-251 in these capacities. On the basis of functionality of single copy sequences we consider the first 251 residues to contain the core of RAD52.

In our studies we have also created situations where shorter protein products can exhibit core properties or where a longer protein product exceeds the ability of the rad52-251 product. These situations rely on the presence of complementing rad52 alleles, high copy expression of rad52 alleles, or high copy expression of RAD51. For example, an exception to the generalization that rad52-251 meets the functionality of longer protein products is the ability of rad52-327, but not rad52-251, to complement missense mutations in a heteroallelic diploid. In this situation, functionality is achieved by the interaction of two partially active mutant forms of RAD52. There are many examples of the shorter truncation alleles exhibiting substantial activity. One example is the rad52-210 product that is partially rescued by high-level expression of RAD51. However, we note that while the rad52-210 allele is partially rescued by high copy expression of RAD51, the rad52-251 allele is brought to near wild-type functionality in tests of MMS resistance by excess amounts of Rad51p. Another striking example is the ability of rad52-169 to complement rad52-301A when the truncated polypeptide is expressed in high copy. This is the first experimental evidence that the rad52-169 product retains any activity in vivo. Indeed, its ability to complement a missense mutation, in the absence of any other measurable activity, suggests that it may rescue the protein product of the missense mutation by binding to it. In turn, this would place the putative self-interaction domain within the first 169 residues. Thus, while rad52-251 neither outperforms all other alleles in every test nor performs as well as rad52-327 in some tests, its partial function in mitosis and meiosis and its ability to be brought to near wild-type MMS resistance by high copy expression of RAD51 demonstrate its retention of core function.

The size of this domain is interesting because it is approximately the size of RAD59, a RAD52 family member that is needed for RAD51-independent recombination events (BAI and SYMINGTON 1996 ). The first 251 residues of RAD52 are 28% identical and 50% similar in amino acid sequence to RAD59. Functionally, the two genes must have some common activity because overexpression of RAD52 can suppress the -ray sensitivity of a rad59 mutant. Curiously, the experiment does not work in reverse; high copy expression of RAD59 does not suppress a rad52 mutant. One explanation for this failure might be that Rad59p lacks the C-terminal functions of Rad52p. If this were so, then overexpression of RAD51 might be expected to suppress the rad59 mutant as it does the rad52-251 allele. This experiment failed as well (BAI and SYMINGTON 1996 ). The failure of RAD51 to suppress a rad59 mutant while being able to suppress the rad52-251 mutant implies that the truncated RAD52 protein product possesses some unique activity in addition to a Rad51p binding site that distinguishes it from Rad59p. Accordingly, we attempted to suppress a rad59 mutant by overexpressing rad52-251 either by itself or by simultaneous overexpression of RAD51 from the same plasmid. In both cases we failed to achieve suppression as judged by growth on agar plates containing MMS (data not shown). [Control experiments revealed that we could duplicate the original observation that overexpression of RAD52 suppresses the MMS sensitivity of a rad59 mutant (BAI and SYMINGTON 1996 ) and that we could also show that the plasmid simultaneously overexpressing both rad52-251 and RAD51 confers MMS resistance to a rad52 null strain.] The results suggest that the N termini of Rad52p and Rad59p are not interchangeable.

Suppression of rad52 mutations by RAD51 overexpression:
Our studies on suppression of rad52 mutations by overexpression of RAD51 show that this most likely occurs by bypassing the C-terminal function of Rad52p. A corollary of this conclusion is that the core of RAD52 must be intact in order for high-level expression of RAD51 to suppress the MMS sensitivity of a rad52 mutant allele. The conclusion that RAD51 overexpression works by bypassing the C-terminal half of Rad52p comes directly from its suppressive effects on truncation alleles as small as rad52-210. The protein products of these truncation alleles have eliminated the putative Rad51p binding site on Rad52p (MILNE and WEAVER 1993 ; MORTENSEN et al. 1996 ). Whether RAD51 overexpression suppresses the missense alleles rad52-96A (R235G) and rad52-252 (D371G) that have changes in the vicinity of the Rad51p binding domain by direct contact, rather than through a bypass mechanism, is not known. The observation of a strict positional specificity of RAD51 suppression of the MMS sensitivity of missense mutants, however, demonstrates the need for an intact core to achieve suppression. None of five N-terminal missense mutations that change residues in the region from position 61 to position 162 are suppressed, while the missense mutations with changes at positions 235 and 371 are. The suppressive effect on the rad52-96A allele is interesting because the amino acid change (R235G) lies in the residues that distinguish rad52-251 from rad52-210. As rad52-251 is nearly completely suppressible by overexpression and rad52-210 is not as completely suppressible, the suppressibility of a mutation at residue 235 may signal that core functions might be found in truncation alleles shorter than rad52-251.

RAD51 overexpression suppression of meiotic functions is interesting for two reasons. First, while the truncation alleles rad52-327 and rad52-291 are partially suppressed for the production of viable spores, they are not fully suppressed. In the case of mitotic suppression of MMS sensitivity, the suppression brings the mutants to near the resistance of a wild-type strain. One explanation for the difference in degree of suppression between mitotic and meiotic functions would be that Rad52p and Rad51p interact differently, or are present in different stoichiometry, during meiosis. Another possibility is that the truncation alleles remove not only the Rad51p binding domain of Rad52p but destroy a Rad52p function that is more important for meiotic division than for MMS resistance during mitotic growth. For example, the RPA interaction domain of yeast Rad52p has been inferred to lie in this region (HAYS et al. 1998 ), and the action of this protein may be different in meiosis than mitosis. The second interesting observation about the meiotic suppression results is that both the rad52-251 and rad52-210 alleles can be suppressed during mitosis, but neither can be suppressed in meiosis. While this might have been expected for the rad52-210 allele that does not sporulate, the failure to achieve meiotic suppression is somewhat surprising for the rad52-251 allele because its unsuppressed level of sporulation is not significantly different from the two longer alleles, rad52-291 and rad52-327, that can be partially suppressed. The drop-off in meiotic suppressibility again suggests that the region between residues 210 and 327 may hold a function that is more important for the meiotic function of Rad52p than for its mitotic function.

Intragenic complementation among rad52 alleles:
While the rad52-251 and the rad52-291 mutants expressing the core amino acids can be rendered nearly completely resistant to MMS exposure by RAD51 overexpression, these mutations cannot be complemented in a heteroallelic diploid with the rad52-2 allele. We have shown previously, and confirmed in this study, that a heteroallelic diploid of rad52-327 and rad52-2 has near wild-type resistance to MMS exposure (BOUNDY-MILLS and LIVINGSTON 1993 and this article). Further exploration of the intragenic complementation between mutant Rad52 protein chains showed that the rules governing complementation are complex. Nevertheless, we point out two features of the complementation that are clear from our results.

We first note that rad52-251 can complement rad52-2 but only if the expression of the truncation allele is altered by expressing it from a high copy plasmid in a strain with a chromosomal copy of the rad52-2 allele. The rad52-251 allele also fails to complement a second missense allele, rad52-301A, in a heteroallelic diploid but again does so when the quantity of the truncation protein is elevated above that of the rad52-301A protein. In both cases we suspect that the rad52-251 allele is providing the core function because rad52-2 and rad52-301A mutants have little activity by themselves or when RAD51 is overexpressed in them. Thus, while the rad52-251 allele does not behave like the rad52-327 allele when these alleles are expressed as chromosomal copies in heteroallelic diploids with a chromosomal copy of the rad52-2 or rad52-301A missense alleles, rad52-251 nonetheless can be brought to near wild-type levels of MMS resistance by expressing it from a high copy plasmid in a cell with either of these two missense rad52 alleles.

Second, we conclude that the rules of intragenic complementation show that while the N- and C-terminal portions of Rad52p can be supplied in trans, wild-type N-terminal sequences must extend to residue 327 for successful complementation in cells where both mutant alleles are expressed as single chromosomal copies. Three experiments argue for this conclusion. First, only the rad52-327 allele and none of the shorter truncation alleles or internal deletion alleles complement the rad52-2 allele when present in heteroallelic diploids. Second, the rad52-327 allele complements internal deletion alleles, including rad52-i210B, encoding a Rad52 protein missing residues 210 through 327. Finally, the rad52-327 allele complements an allele containing both the missense mutation of rad52-2 and the internal deletion removed from the rad52-i210B allele. Thus, complementation is achieved with the rad52-327 peptide and numerous other mutant peptides that have wild-type sequences extending from residue 328 to the C terminus. The complementation of the internal deletion alleles by the rad52-327 allele shows that the residues from 210 to 327 need not be present in both mutant protein molecules. This means either that a putative Rad52p self-association domain cannot reside in this region or that the complementation really works by the independent action of the two mutant proteins, one supplying the N-terminal functions and the other the C-terminal functions.

Conclusion:
These studies have shown how few of the amino acid residues of the open reading frame of RAD52 need to be present to achieve partial activity in tests of resistance to MMS exposure during mitotic growth and of spore viability after meiosis. By virtue of the differences in the degree to which the truncation alleles can be rescued by RAD51 overexpression and by intragenic complementation with rad52 missense alleles, the results also suggest that the portion of Rad52p between residues 210 and 327 contains domains necessary for fully successful sporulation and intragenic complementation.

	ACKNOWLEDGMENTS

We thank Lorraine Symington for providing the rad59 strain described in DISCUSSION. This work was supported by National Science Foundation grant MCB-9603515.

Manuscript received October 26, 1998; Accepted for publication July 1, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BAI, Y. and L. S. SYMINGTON, 1996  A Rad52 homolog is required for RAD51-independent mitotic recombination in Saccharomyces cerevisiae. Genes Dev. 10:2025-2037[Abstract/Free Full Text].

BENSON, F. E., P. BAUMANN, and S. C. WEST, 1998  Synergistic actions of Rad51 and Rad52 in recombination and DNA repair. Nature 391:401-404[Medline].

BOUNDY-MILLS, K. L. and D. M. LIVINGSTON, 1993  A Saccharomyces cerevisiae RAD52 allele expressing a C-terminal truncation protein: activities and intragenic complementation of missense mutations. Genetics 133:39-49[Abstract].

CHRISTIANSON, T. W., R. S. SIKORSKI, M. DANTE, J. H. SHERO, and P. HIETER, 1992  Multifunctional yeast high-copy-number shuttle vectors. Gene 110:119-122[Medline].

DORNFELD, K. J. and D. M. LIVINGSTON, 1991  Effects of controlled RAD52 expression on repair and recombination in Saccharomyces cerevisiae.. Mol. Cell. Biol. 11:2013-2017[Abstract/Free Full Text].

FIRMENICH, A. E., M. ELIAS-ARNANZ, and P. BERG, 1995  A novel allele of Saccharomyces cerevisiae RFA1 that is deficient in recombination and repair and suppressible by RAD52.. Mol. Cell. Biol. 15:1620-1631[Abstract].

HAYS, S. L., A. A. FIRMENICH, P. MASSEY, R. BANERJEE, and P. BERG, 1998  Studies of the interaction between Rad52 protein and the yeast single-stranded DNA binding protein RPA. Mol. Cell. Biol. 18:4400-4406[Abstract/Free Full Text].

KATZ, W. and F. SOLOMON, 1988  Diversity among beta-tubulins: a carboxy-terminal domain of yeast beta-tubulin is not essential in vivo.. Mol. Cell. Biol. 8:2730-2736[Abstract/Free Full Text].

KAYTOR, M. D. and D. M. LIVINGSTON, 1994  Saccharomyces cerevisiae RAD52 alleles temperature-sensitive for the repair of double-strand breaks. Genetics 137:933-944[Abstract].

KAYTOR, M. D. and D. M. LIVINGSTON, 1996  Allele-specific suppression of temperature-sensitive mutations of the Saccharomyces cerevisiae RAD52 gene. Curr. Genet. 29:203-210[Medline].

MILNE, G. T. and D. T. WEAVER, 1993  Dominant negative alleles of RAD52 reveal a DNA repair/recombination complex including Rad51 and Rad52. Genes Dev. 7:1755-1765[Abstract/Free Full Text].

MORTENSEN, U. H., C. BENDIXEN, I. SUNJEVARIC, and R. ROTHSTEIN, 1996  DNA strand annealing is promoted by the yeast Rad52 protein. Proc. Natl. Acad. Sci. USA 93:10729-10734[Abstract/Free Full Text].

NEW, J. H., T. SUGIYAMA, E. ZAITSEVA, and S. C. KOWALCZYKOWSKI, 1998  Rad52 protein stimulates DNA strand exchange by Rad51 and replication protein A. Nature 391:407-410[Medline].

NGUYEN, M. M. and D. M. LIVINGSTON, 1997  Cold-sensitive rad52 alleles of yeast. Curr. Genet. 32:100-107[Medline].

PASSY, S. I., S. YU, Z. LI, C. M. RADDING, and E. H. EGELMAN, 1999  Rings and filaments of beta protein from bacteriophage lambda suggest a superfamily of recombination proteins. Proc. Natl. Acad. Sci. USA 96:4279-4284[Abstract/Free Full Text].

PRAKASH, S. and L. PRAKASH, 1977  Isolation and characterization of MMS-sensitive mutants of Saccharomyces cerevisiae.. Genetics 87:33-55[Abstract/Free Full Text].

SCHERER, S. and R. W. DAVIS, 1979  Replacement of chromosome segments with altered DNA sequences constructed in vitro.. Proc. Natl. Acad. Sci. USA 76:4951-4955[Abstract/Free Full Text].

SCHILD, D., 1995  Suppression of a new allele of the yeast RAD52 gene by overexpression of RAD51, mutations in srs2 and ccr4, or mating type heterozygosity. Genetics 140:115-127[Abstract].

SHEN, Z., S. R. PETERSON, J. C. COMEAUX, D. ZASTROW, and R. K. MOYZIS et al., 1996  Self-association of human RAD52 protein. Mutat. Res. 364:81-89[Medline].

SHINOHARA, A. and T. OGAWA, 1998  Stimulation by Rad52 of yeast Rad51-mediated recombination. Nature 391:404-407[Medline].

SHINOHARA, A., M. SHINOHARA, T. OHTA, S. MATSUDA, and T. OGAWA, 1998  Rad52 forms ring structures and co-operates with RPA in single-strand DNA annealing. Genes Cells 3:145-156[Abstract].

SUNG, P., 1997  Function of yeast Rad52 protein as a mediator between replication protein A and the Rad51 recombinase. J. Biol. Chem. 272:28194-28197[Abstract/Free Full Text].

TYERS, M., G. TOKIWA, R. NASH, and B. FUTCHER, 1992  The Cln3-Cdc28 kinase complex of S. cerevisiae is regulated by proteolysis and phosphorylation. EMBO J. 11:1773-1784[Medline].

TYERS, M., G. TOKIWA, and B. FUTCHER, 1993  Comparison of the Saccharomyces cerevisiae G1 cyclins: Cln3 may be an upstream activator of Cln1, Cln2 and other cyclins. EMBO J. 12:1955-1968[Medline].

VAN DYCK, E., A. Z. STASIAK, A. STASIAK, and S. C. WEST, 1999  Binding of double-strand breaks in DNA by human Rad52 protein. Nature 398:728-731[Medline].

VAN DYCK, W., N. M. A. HAJIBAGHERI, A. STASIAK, and S. C. WEST, 1998  Visualization of human Rad52 protein and its complexes with hRad51 and DNA. J. Mol. Biol. 284:1027-1038[Medline].

This article has been cited by other articles:

				I. Plate, S. C. L. Hallwyl, I. Shi, L. Krejci, C. Muller, L. Albertsen, P. Sung, and U. H. Mortensen Interaction with RPA Is Necessary for Rad52 Repair Center Formation and for Its Mediator Activity J. Biol. Chem., October 24, 2008; 283(43): 29077 - 29085. [Abstract] [Full Text] [PDF]

				C. Seong, M. G. Sehorn, I. Plate, I. Shi, B. Song, P. Chi, U. Mortensen, P. Sung, and L. Krejci Molecular Anatomy of the Recombination Mediator Function of Saccharomyces cerevisiae Rad52 J. Biol. Chem., May 2, 2008; 283(18): 12166 - 12174. [Abstract] [Full Text] [PDF]

				G. Octobre, A. Lorenz, J. Loidl, and J. Kohli The Rad52 Homologs Rad22 and Rti1 of Schizosaccharomyces pombe Are Not Essential for Meiotic Interhomolog Recombination, but Are Required for Meiotic Intrachromosomal Recombination and Mating-Type-Related DNA Repair Genetics, April 1, 2008; 178(4): 2399 - 2412. [Abstract] [Full Text] [PDF]

				F. Storici, J. R. Snipe, G. K. Chan, D. A. Gordenin, and M. A. Resnick Conservative Repair of a Chromosomal Double-Strand Break by Single-Strand DNA through Two Steps of Annealing Mol. Cell. Biol., October 15, 2006; 26(20): 7645 - 7657. [Abstract] [Full Text] [PDF]

				M. Valencia-Burton, M. Oki, J. Johnson, T. A. Seier, R. Kamakaka, and J. E. Haber Different Mating-Type-Regulated Genes Affect the DNA Repair Defects of Saccharomyces RAD51, RAD52 and RAD55 Mutants Genetics, September 1, 2006; 174(1): 41 - 55. [Abstract] [Full Text] [PDF]

				P. H. Thorpe, V. A. Marrero, M. H. Savitzky, I. Sunjevaric, T. C. Freeman, and R. Rothstein Cells Expressing Murine RAD52 Splice Variants Favor Sister Chromatid Repair Mol. Cell. Biol., May 15, 2006; 26(10): 3752 - 3763. [Abstract] [Full Text] [PDF]

				A. A. de Mayolo, M. Lisby, N. Erdeniz, T. Thybo, U. H. Mortensen, and R. Rothstein Multiple start codons and phosphorylation result in discrete Rad52 protein species. Nucleic Acids Res., January 1, 2006; 34(9): 2587 - 2597. [Abstract] [Full Text] [PDF]

				M. Tsukamoto, K. Yamashita, T. Miyazaki, M. Shinohara, and A. Shinohara The N-Terminal DNA-Binding Domain of Rad52 Promotes RAD51-Independent Recombination in Saccharomyces cerevisiae Genetics, December 1, 2003; 165(4): 1703 - 1715. [Abstract] [Full Text] [PDF]

				E. N. Asleson and D. M. Livingston Investigation of the Stability of Yeast rad52 Mutant Proteins Uncovers Post-translational and Transcriptional Regulation of Rad52p Genetics, January 1, 2003; 163(1): 91 - 101. [Abstract] [Full Text] [PDF]