Dissection of the Functions of the Saccharomyces cerevisiae RAD6 Postreplicative Repair Group in Mutagenesis and UV Sensitivity

Corresponding author: Zuzana Storchová, Institute of Medical Radiobiology, University of Zurich, August Forel Str. 7, 8008 Zurich, Switzerland., zuzana{at}imr.unizh.ch (E-mail)

Communicating editor: L. S. SYMINGTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

The RAD6 postreplicative repair group participates in various processes of DNA metabolism. To elucidate the contribution of RAD6 to starvation-associated mutagenesis, which occurs in nongrowing cells cultivated under selective conditions, we analyzed the phenotype of strains expressing various alleles of the RAD6 gene and single and multiple mutants of the RAD6, RAD5, RAD18, REV3, and MMS2 genes from the RAD6 repair group. Our results show that the RAD6 repair pathway is also active in starving cells and its contribution to starvation-associated mutagenesis is similar to that of spontaneous mutagenesis. Epistatic analysis based on both spontaneous and starvation-associated mutagenesis and UV sensitivity showed that the RAD6 repair group consists of distinct repair pathways of different relative importance requiring, besides the presence of Rad6, also either Rad18 or Rad5 or both. We postulate the existence of four pathways: (1) nonmutagenic Rad5/Rad6/Rad18, (2) mutagenic Rad5/Rad6 /Rev3, (3) mutagenic Rad6/Rad18/Rev3, and (4) Rad6/Rad18/Rad30. Furthermore, we show that the high mutation rate observed in rad6 mutants is caused by a mutator different from Rev3. From our data and data previously published, we suggest a role for Rad6 in DNA repair and mutagenesis and propose a model for the RAD6 postreplicative repair group.

MUTATIONS play a fundamental role in evolution and contribute to aging, carcinogenesis, and genetic diseases. Spontaneous mutations occur during DNA replication by incorrect nucleotide incorporation, followed by skipping the proofreading activity of replicative polymerases and the activity of the mismatch repair pathway; arise as DNA repair errors; or are introduced by some mutagenic system. Mutations resulting after mutagenic treatment are called induced mutations and appear to result from mutagenic repair (FRIEDBERG et al. 1995 ). In the past decade, it was shown that mutations also occur during prolonged nonlethal starvation in nongrowing cells in both bacteria and unicellular eukaryotes. These mutations are called adaptive or starvation associated. Several mechanisms have been suggested for adaptive mutations in bacteria (e.g., FOSTER 2000 ). They are suggested to be mainly a result of incorrect DNA repair of endogenous lesions arising in starving cells (e.g., BRIDGES 1996 ); however, the existence of such lesions has not yet been substantiated. The ability of cells to generate mutations in even a quiescent state appears to be a general phenomenon, at least among unicellular organisms. However, we do not understand the exact mechanism of starvation-associated mutagenesis (SAM) and we do not know its contribution to survival and evolution of microorganisms.

One of the DNA repair pathways suggested to be involved in SAM of the unicellular eukaryotic organism Saccharomyces cerevisiae is the RAD6 postreplicative repair pathway (STORCHOVA et al. 1998 ). This pathway is involved in the repair of UV-induced DNA lesions and other bulky lesions that block DNA replication and in mutagenesis. The RAD6 epistatic group can be dissected into various repair subpathways, but they are poorly understood (LIEFSHITZ et al. 1998 ; XIAO et al. 2000 ).

The pivotal gene of this group is RAD6. Its product functions in various cellular processes including DNA repair and mutagenesis, gene silencing, protein degradation, sporulation, and histone H2B ubiquitination. Null rad6 mutants exhibit a pleiotropic phenotype—they possess a defect in all of the above-listed functions—that contributes to their extreme sensitivity to various DNA-damaging agents, enhanced spontaneous and impaired induced mutagenesis, lower growth rate, decreased viability under stress conditions, and so on (LAWRENCE 1994 ).

The protein Rad6 consists of 172 amino acid residues, from which the last 23 form an almost entirely acidic C-terminal tail (MORRISON et al. 1988 ). It is a ubiquitin-conjugating enzyme (E2) catalyzing the ubiquitination of proteins in cooperation with other proteins. Rad6 ubiquitinates either by forming a Lys-48 polyubiquitin chain, which serves as a signal for proteosomal degradation of the ubiquitylated proteins (DOHMEN et al. 1991 ; WATKINS et al. 1993 ), or by monoubiquitination, which is not a signal for degradation and is known, for example, for histones (ROBZYK et al. 2000 ). The protein was shown to interact tightly with either Ubr1 for ubiquitin-mediated N-end rule protein degradation (WATKINS et al. 1993 ) or Rad18 for the DNA repair functions (BAILLY et al. 1997A , BAILLY et al. 1997B ).

According to mutational analysis, the functions of the Rad6 protein can be attributed to its distinct domains. The first nine amino acids are required for interaction with Ubr1 and thus for N-end rule-mediated protein degradation (WATKINS et al. 1993 ). They are probably also involved in error-free repair (BROOMFIELD et al. 1998 ). The amino acids 142–149 and, less importantly, residues 10–22 are responsible for the interaction with Rad18 and thus for the DNA repair functions of Rad6 (BAILLY et al. 1997A , BAILLY et al. 1997B ). Cysteine at position 88 is required for binding of a ubiquitin molecule (SUNG et al. 1990 ). The acidic C-terminal tail is involved in ubiquitination of histone H2B, sporulation, meiotic functions, and, less importantly, in nonspecific interaction with Ubr1 (SUNG et al. 1988 ; ROBZYK et al. 2000 ; ULRICH and JENTSCH 2000 ).

As mentioned above, Rad6 interacts with Rad18, the product of another member of the RAD6 group. This interaction is necessary for Rad6 DNA repair functions, because Rad18, unlike Rad6, shows an affinity to ssDNA and appears to target Rad6 toward damaged DNA (BAILLY et al. 1994 ). The Rad6/Rad18 complex has been assumed to be required for all recognized subpathways within the RAD6 group (see below), although rad18 mutants do not show a DNA repair deficiency as strong as rad6 mutants.

Epistasis studies have identified three subpathways within this group up to now. An error-free pathway acting by a yet-unknown mechanism involves Rad5, Mms2, Ubc13, Pol30, and pol (XIAO et al. 1999 , XIAO et al. 2000 ). One of the members, Rad5, has DNA helicase and zinc-binding domains, shows affinity to ssDNA, and transiently interacts with the ubiquitin-conjugating protein complex Mms2/Ubc13 (JOHNSON et al. 1994 ; ULRICH and JENTSCH 2000 ). The products of MMS2 and UBC13 have both been shown to be members of the RAD6 epistasis group and involved in error-free repair and form a tight heterodimer possessing the ability to conjugate a polyubiquitin chain in vitro via an unusual Lys-63 (HOFMANN and PICKART 1999 ). In vitro studies also showed interaction of Rad5 and Rad18. It was suggested that the two similar ubiquitin-conjugating enzymes Rad6 and Mms2/Ubc13 form complexes with either Rad18 or Rad5, respectively. These complexes are held together by interaction between Rad5 and Rad18 and bind DNA (ULRICH and JENTSCH 2000 ). Furthermore, POL30, encoding proliferating cell nuclear antigen (PCNA), also belongs to this group. It probably acts with pol to perform DNA synthesis after repair of lesions (TORRES-RAMOS et al. 1996 ).

The key role in a second mutagenic subpathway is played by polymerase , encoded by REV3 and REV7 (catalytic and regulatory subunits, respectively). This nonessential DNA polymerase has the capability to bypass thymine dimers and other replication-blocking lesions at the cost of an increased mutation frequency (translesion synthesis; NELSON et al. 1996 ; BAYNTON et al. 1999 ). The role of the Rad6/Rad18 heterodimer in this process is not very clear, but it was suggested that it allows the recruitment of pol to the stalled replication fork (BAILLY et al. 1997B ).

On the basis of recent epistatic studies, RAD30 belongs to neither of the two above-mentioned subpathways and probably represents a third subpathway (XIAO et al. 2000 ). It encodes translesion polymerase that exhibits low fidelity and tolerance to DNA damage (WASHINGTON et al. 1999 ). The significance of this pathway in DNA repair remains unclear.

We have shown previously that the rad6-1 mutation significantly enhanced SAM (STORCHOVA et al. 1998 ). To clarify the role of Rad6 in SAM in more detail, we constructed a rad6 null mutant and complemented the mutation by various rad6 alleles carried on plasmids. The analysis showed that, in particular, Rad6 DNA repair function is responsible for maintaining the low level of SAM in the wild-type strain. We therefore analyzed the effect of deletion of various RAD6 repair group genes on spontaneous and starvation-associated mutagenesis and sensitivity to UV light to elucidate the role of the error-free and error-prone subpathways in SAM. Our experimental system allowed us to compare the relative importance of various RAD6 group genes in SAM and spontaneous mutagenesis. The analysis of rev3, rad5, rad18, and mms2 single and multiple mutants allowed us to propose a model for RAD6-mediated subpathways acting in repair and mutagenic processes.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

General methods:
For nonselective growth, YPD medium consisting of 2% glucose, 1% bactopeptone, and 0.5% yeast extract was used (all Difco, Detroit). For selection of clones with replacement of the gene of interest by kanMX4 cassette, YPD with G418 (400 µg/ml; GIBCO BRL, Paisley, Scotland) was used. Synthetic dropout (SD) medium (0.75% yeast nitrogen base without amino acids, 2% glucose, and dropout solution; Difco) was used for selective growth when the essential supplement for selection was omitted. All the media were solidified in 2% agar (Difco). Yeast genetics methods were used essentially as described (AUSUBEL et al. 1994 ). All yeast strains were propagated under aerobic conditions at 30°.

Yeast strains:
The yeast strains used in this study for analysis of reversion of the suppressible amber ade2-101 allele were derivatives of YPH499/500 strains (SIKORSKI and HIETER 1989 ) and are listed in Table 1. The replacement of RAD6, REV3, RAD5, and MMS2 was performed using a kanMX4 replacement cassette obtained by PCR with specifically designed primers using pUG6 plasmid as a template (GULDENER et al. 1996 ). Sequences of primers used are available on request. The disruption of RAD18 with the functional LEU2 gene was performed via homologous recombination of the BamHI-HpaI fragment from YCp50-11 plasmid (FABRE et al. 1989 ), kindly provided by F. Fabre. Yeast transformation was performed as previously described (GIETZ et al. 1995 ). Transformants were streaked on appropriate omission medium and single colonies were picked for further analysis. All mutants were verified both genomically using PCR and/or Southern blotting and phenotypically using a UV-sensitivity test. Multiple mutants were prepared by crossing single mutant strains followed by tetrad dissection. We have always constructed and tested two to three independent isolates to reduce the chance of spontaneously occurring suppressors appearing in the genetic background. All details about strain construction are available on request.

Plasmid manipulation:
Plasmids carrying various alleles of the RAD6 gene (BAILLY et al. 1997A ) were kindly provided by S. Prakash. The control plasmids were prepared by excision of the RAD6 open reading frame from the plasmids. For the list of given plasmids see Table 2.

Reversion in ade2-1:
Spontaneous mutation rates and accumulation of starvation-associated Ade⁺ revertants were determined by conducting a modified fluctuation test. Three-day-old colonies (10–24 independent isolates) were transferred to liquid media and grown nonselectively up to a density of 2 x 10⁸ cells/ml, harvested, and washed with sterile water. Each culture was plated on SD plates lacking adenine (usually 5 x 10⁷ cells per plate) and after appropriate dilution also onto SD plates supplemented with adenine to determine the number of colony-forming units (cfu). Revertants to Ade⁺ prototrophy were counted daily from the 3rd to the 12th day of selective starvation. The growth-dependent mutation rate was calculated according to LEA and COULSON 1949 .

Cell survival analysis:
Yeast cultures cultivated nonselectively up to a density of 2 x 10⁸ cells/ml were harvested, washed, and diluted. Drops containing 5 x 10⁵ cells were spotted onto SD plates lacking adenine. The agar cubes with drops were cut out in regular intervals, washed, diluted if necessary, and plated onto SD plates supplemented with adenine to score total survivors. In the case of the strains containing plasmids, the appropriate supplement was omitted to reveal only the number of living cells with plasmid. Each result is an average of at least three independent experiments.

UV killing analysis:
The late exponential phase cultures were harvested, washed, and diluted. Proper dilutions were plated on YPD plates and irradiated (UV Stratalinker 1800). The number of colonies was counted after 4 days of cultivation in a dark chamber. Each result is an average of three to five independent experiments.

Evaluation of starvation-associated mutagenesis:
The data describing mutagenesis and viability in stationary phase are presented in three different graphics in Fig 1, Fig 3, and Fig 6. In Fig 1A, Fig 3A, and Fig 6A, the accumulation curve shows the cumulative average of the total number of revertant colonies that appeared during selective starvation. The data were normalized to 10⁸ cfu on the day of plating. Fig 1B, Fig 3B, and Fig 6B show the percentage of colony-forming units plotted against the time of starvation. And Fig 1C, Fig 3C, and Fig 6C show both accumulation and viability data combined. It shows the hypothetical number of starvation-associated revertant colonies per 10⁸ living cells during the whole period of starvation. Because the numbers are noncumulative, it shows the increase of revertant colonies in defined intervals. The basic formula used for calculating the transformed number of revertant colonies after t days of starvation is = n_t(), where n_t is the observed number of revertant colonies after t days of starvation and cfu_t-T is the percentage of colony-forming units after t-T days of starvation. T (days) accounts for the average delay between the mutation event and the observation of a visible colony. We have also established a factor D (days), which accounts for the interval when 95% of the cells form visible colonies under corresponding experimental conditions (presence of nongrowing background cells). Note the difference between T (formation of 50% of colonies) and D (95%). Fig 1C, Fig 3C, and Fig 6C show only revertants that appeared after D days of starvation—the starvation-associated revertants. The curve is noncumulative. Further details about data analysis are available on request.

View larger version (24K): In this window In a new window Download PPT slide   Figure 1. Accumulation of revertants (A), viability (B), and daily noncumulative increment of SAM (C) during prolonged starvation of strains overexpressing various RAD6 alleles. Wild type (x); rad6 (); RAD6 (•); rad6150-172 (); rad6-1 (); rad61-9 (); and rad6Ala88 (). For A, 95% confidence interval of the total number of revertants (day 12) represents the following: wild type (±11.6); rad6 (±20.2); RAD6 (±9.8); rad6150-172 (±5.0); rad6-1 (±21.0); rad61-9 (±22.0); rad6Ala88 (±24.0).

View larger version (18K): In this window In a new window Download PPT slide   Figure 2. UV sensitivity of strains overexpressing various RAD6 alleles. Wild type (x); rad6 (); RAD6 (•); rad6150-172 (); rad6-1 (); rad61-9 (); and rad6Ala88 ().

View larger version (26K): In this window In a new window Download PPT slide   Figure 3. Accumulation of revertants (A), viability (B), and daily noncumulative increment of SAM (C) of rad5, rad6, and rad18 single-, double-, and triple-mutant strains. Wild type (x); rad6 (); rad18 (); rad5 (); mms2 (); rad6 rad18 (); rad5 rad6 (); rad5 rad18 (•); and rad5 rad6 rad18 (). For A, 95% confidence interval of the total number of revertants (day 12) represents the following: wild type (±11.6); rad6 (±20.2); rad18 (±21.7); rad5 (±32.8); mms2 (±45.5); rad6 rad18 (±24.8); rad5 rad6 (±18.8); rad5 rad18 (±15.0); and rad5 rad6 rad18 (±37.7).

View larger version (21K): In this window In a new window Download PPT slide   Figure 4. UV sensitivity of rad5, rad6, and rad18 single-, double-, and triple-mutant strains. Wild type (x); rad6 (); rad18 (); rad5 (); mms2 (); rad6 rad18 (); rad5 rad6 (); rad5 rad18 (•); and rad5 rad6 rad18 ().

View larger version (22K): In this window In a new window Download PPT slide   Figure 5. The epistasis of rev3, rad5, rad6, and rad18 single and double mutants in UV sensitivity. Wild type (x); rad6 (); rad18 (); rad5 (); rev3 (+); rev3 rad18 (—); rev3 rad6 (); and rev3 rad5 (*).

View larger version (23K): In this window In a new window Download PPT slide   Figure 6. Accumulation of revertants (A), viability (B), and daily noncumulative increment of SAM (C) of rad5, rad6, and rad18 mutants and its dependency on rev3 mutation. Wild type (x); rad6 (); rad18 (); rad5 (); rev3 (+); rev3 rad18 (—); rev3 rad6 (); and rev3 rad5 (*). For A, 95% confidence interval of the total number of revertants (day 12) represents the following: wild type (±11.6); rad6 (±20.2); rad18 (±21.7); rad5 (±32.8); rev3 (±4.2); rev3 rad18 (±6.2); rev3 rad6 (±12); and rev3 rad5 (±6.1).

	RESULTS AND DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

The experimental system for analysis of spontaneous and SA mutagenesis:
Yeast auxotrophic strains can form two types of revertant colonies when plated on plates lacking an essential supplement. The first type of colony appears early, on average 3–4 days after plating. The corresponding mutations occurred very likely by replication errors during the nonselective growth before plating. The second type of revertant colony appears later and accumulates up to 20 days as a consequence of mutation events occurring during the starvation on selective plates (e.g., HALL 1992 ; STEELE and JINKS-ROBERTSON 1992 ; STORCHOVA et al. 1998 ).

The evaluation of late revertant accumulation is complicated by two facts:

1. The cells are losing their colony-forming ability during starvation, and individual mutant strains differ in their sensitivity to starvation. This effect strongly influences the number of arising revertants.

2. It is difficult to estimate the proportion of preexisting and starvation-associated (SA) revertants, because although the colonies of SA revertants arise later, there is a certain interval at the beginning of starvation when both types of revertant colonies arise.

We performed a reconstruction experiment, which was suggested to address this problem (STEELE and JINKS-ROBERTSON 1992 ). In this experiment the formation of visible colonies was followed on complete SD plates with a background of 5 x 10⁷ nongrowing nonrevertible cells, which slows down the colony appearance. We have ascertained that under these conditions the formation of a visible colony takes, on average, 3 days for the wild-type strain YPH499 (parameter T = 3; see MATERIALS AND METHODS), whereas the period required for the formation of 95% of the colonies was 3.5 days (D = 3.5). The various mutant strains derived from YPH499 grew similarly (T = 3; D = 3.5) except for the rad6 strain and all strains carrying the rad6 mutation in combination with any other mutation, as well as strains carrying the rad6Ala88 allele (T = 3.5; D = 4). Thus, we considered all the colonies arising before day D to be a result of spontaneous events occurring during nonselective growth before plating for selection, and the counted numbers were used for calculation of the spontaneous mutation rates. All colonies arising later are a result of starvation-associated mutagenesis.

To describe starvation-associated mutagenesis we followed the accumulation of revertants (Fig 1A, Fig 3A, and Fig 6A) and cell viability (Fig 1B, Fig 3B, and Fig 6B) during prolonged starvation. In Fig 1C, Fig 3C, and Fig 6C, both data are combined, and a daily increment of starvation-associated revertant colonies is plotted in a noncumulative manner. Thus, Fig 1C, Fig 3C, and Fig 6C show numbers not influenced by the loss of viability during starvation. This approach allows us to compare the strains even if they differ in their growth, viability, and in the number of preexisting revertants. For a detailed description of the evaluation of our experimental data, see MATERIALS AND METHODS.

The effect of RAD6 alleles on SAM suggests an involvement of Rad6 repair functions:
We have shown previously that a mutation in the RAD6 gene increases the level of starvation-associated mutagenesis (STORCHOVA et al. 1998 ). Because Rad6 acts in a number of various cellular processes, we used the defined alleles, whose products possess a well-known lack of function, to elucidate in more detail the role of Rad6 in SAM.

As in the experimental approach, we complemented the chromosomal RAD6 deletion with a plasmid carrying one of the various alleles of RAD6 (Table 2). In strains overexpressing mutant proteins we followed spontaneous mutagenesis and SAM. Because the analyzed alleles were carried on plasmids, we also performed all the necessary controls with rad6 and wild-type strains transformed with plasmids lacking the coding sequence for any of the Rad6 mutant proteins. There were no significant differences in mutation rates among transformed strains, not between rad6 strain and rad6 transformed with plasmids without the coding sequence for various RAD6 alleles or between wild-type strain and wild type transformed with the control plasmids (Table 3). We observed slight differences in cell viability during starvation that accounted for slight differences in the accumulation of starvation-associated revertants, which were caused by differential loss of plasmids under stress (data not shown).

The expression of Rad6 protein with deletions of the acidic C terminus, varying in length up to the amino acid 150, complemented the rad6 phenotype to the wild-type level in both starvation-associated (Fig 1A and Fig C) and spontaneous mutagenesis (Table 3). The viability of these strains during starvation is also comparable with the wild type (Fig 1B). The C-terminal part of Rad6 was shown to be involved in direct ubiquitination of H2B by Rad6 (ROBZYK et al. 2000 ), and mutated strains also had a reduced sporulation efficiency (MORRISON et al. 1988 ). The N-end rule protein degradation is decreased, because the C terminus of Rad6 is also involved in interaction with Ubr1 (WATKINS et al. 1993 ). Thus, all these functions are dispensable for maintaining a low level of spontaneous and SA mutagenesis.

The mutation frequency in the rad6 strain was not changed by the presence of the rad6-1 allele encoding a protein shortened from the C terminus to residue 142. The truncated protein is very likely able to ubiquitinate but it cannot interact tightly with Rad18, because this interaction occurs via the missing eight amino acids 142–149 (BAILLY et al. 1997A , BAILLY et al. 1997B ). Because Rad18, unlike Rad6, possesses an ssDNA-binding activity and probably targets Rad6p toward damaged DNA, the truncated Rad6 protein is deficient in a majority of its DNA repair functions (BAILLY et al. 1997A , BAILLY et al. 1997B ). The rad6-1 strain is only mildly sensitive to a starvation, and combination of the accumulation and viability curves reveals a high increase in the number of SA revertants (Fig 1).

The rad61-9 allele encodes the Rad6 protein lacking the first nine amino acids, which are necessary for the Rad6/Ubr1 interaction (WATKINS et al. 1993 ). The strain overproducing this allele behaves similarly to the rad6-1 strain and shows both increased spontaneous (Table 3) and SA mutagenesis together with moderate starvation sensitivity (Fig 1). Mutant rad61-9 was shown to have impaired Ubr1-dependent ubiquitination function, which labels certain proteins for proteosomal degradation (SUNG et al. 1990 ). Because the proteosomal mutants do not have increased UV sensitivity or mutagenicity (DOR et al. 1996 ), we can conclude that the observed increase of SA spontaneous mutagenesis in this mutant is not due to impaired proteosomal degradation. Mutant rad61-9 was also shown to be sensitive to UV, and UV-induced mutagenesis in this mutant was as high as in the wild type (WATKINS et al. 1993 ). This suggests that the first nine amino acids are essential for the error-free repair but dispensable for the error-prone repair. The increased level of spontaneous and SA mutagenesis reflects a preferential repair of DNA lesions by the error-prone repair pathway, because the error-free pathways in the rad61-9 strain are inactive.

The strain expressing rad6-Ala88 (mutation Cys88 Ala88) produces the Rad6 protein that is completely lacking ubiquitin-conjugating activities (SUNG et al. 1990 ). This strain and the rad6 mutant show a very similar phenotype, consistent with the generally accepted hypothesis that ubiquitin-conjugating activities are required for all the RAD6 functions (BROOMFIELD et al. 1998 ). Both rad6 and rad6-Ala88 strains are extremely sensitive to starvation, which in combination with a rather strong mutator phenotype, leads to fast revertant accumulation up to the sixth or seventh day of starvation, followed by considerable slowing down of revertant accumulation (Fig 1). Very likely the dramatic decrease of viability during starvation prevents further formation of revertant colonies. The rad6 and rad6-Ala88 mutants have the same phenotype in spontaneous and starvation-associated mutagenesis that rad6-1 and rad6 1-9 strains have (Fig 1C). However, the different sensitivity to starvation is notable: while rad6 and rad6-Ala88 are extremely sensitive, rad6-1 and rad61-9 show only moderate sensitivity. The complex phenotype of the rad6 and rad6-Ala88 strains is very likely a consequence of the simultaneous damage of many different biological processes. The increased mutagenesis in rad61-9, rad6-1, rad6-Ala88, and rad6 is probably due to inefficient RAD6-dependent error-free repair. Thus, the RAD6-dependent error-free functions are required for maintaining a low level of both starvation-associated mutations and spontaneous mutations.

The observed phenotypes could be a consequence of an overexpression of the mutant rad6 alleles. However, because Rad6 is rather abundant in yeast cells during the entire cell cycle, this effect appears to be highly unlikely (DOR et al. 1996 ).

UV sensitivity of strains overexpressing various RAD6 alleles:
The strains carrying various plasmid-borne RAD6 alleles were also analyzed for their UV sensitivity (Fig 2). The alleles coding for protein with the acidic tail deleted ensure the same UV resistance as wild-type Rad6 protein. Mutants rad61-9 and rad6-1 are markedly less sensitive to UV irradiation than rad6-Ala88 and rad6 mutants. We hypothesize that this difference is caused by remaining DNA repair activity of proteins coded by rad61-9 and rad6-1, because it was shown that both of them can at least partially ensure the repair functions (DOR et al. 1996 ). The absence of ubiquitination in both rad6-Ala88 and rad6 mutants results in a very strong phenotype independently of the presence or absence of the protein. These two mutants are similarly sensitive to starvation. Thus, the phenotypes observed with high sensitivity to stress are probably caused by loss of other cellular RAD6 functions, whereas in rad61-9 and rad6-1 mutants the observed phenotypes are very likely only the consequence of a deficiency in RAD6 repair functions.

Epistatic relationships between RAD6, RAD5, and RAD18:
To analyze the role of various genes from the RAD6 epistatic group, we created a set of strains with chromosomal deletions of genes of interest (Table 1) and analyzed spontaneous and starvation-associated mutagenesis in each strain. After analyzing single mutants, we observed a similar spontaneous mutation rate in rad5 mutant as in rad6 mutant (Table 4). The spontaneous mutation rate of rad18 strain was slightly, but not significantly, lower.

The rad5 rad18 double mutant has a very low mutation rate, comparable with wild type (Table 4). However, the triple rad6 rad18 rad5 mutant shows a similar spontaneous mutation rate as any double mutant in combination with the rad6 mutation. We assume that the mutagenic repair mediated by Rad6 requires the presence of either Rad18 or Rad5. In the rad5 rad18 double mutant and in the presence of Rad6, the repair is channeled into a different error-free repair pathway, which results in low mutagenesis. This assumption can be applied also to SA mutagenesis because the epistasis follows exactly the same pattern (Fig 3).

The UV-sensitivity analysis supports these data (Fig 4). The rad5 and rad18 single mutants are moderately sensitive; the rad18 strain is more sensitive than a rad5 mutant. The sensitivity of the rad6 strain is very high, and the rad6 rad5 and rad6 rad18 double-mutant strains show the same sensitivity. Also, the double-mutant rad5 rad18 is as sensitive as the rad6 strain as well as the rad5 rad6 rad18 triple mutant. Thus, the RAD6 is epistatic and the effects of RAD18 and RAD5 are synergistic. For repair, RAD6 and either RAD18 or RAD5 are required. The repair is mutagenic in the absence of RAD6 and/or in the absence of either RAD18 or RAD5 or nonmutagenic in the absence of both RAD5 and RAD18.

The rad5 and rad18, but not rad6, mutator phenotype is completely dependent on REV3:
The important mutator in the RAD6 repair group appears to be the DNA polymerase , which is responsible for 60% of spontaneous mutations (ROCHE et al. 1994 ). We analyzed the strain with a chromosomal deletion of REV3, which codes for the catalytic subunit of this mutagenic polymerase. Rev3 is expected to act downstream of Rad6 in a Rad6-dependent manner (LAWRENCE and CHRISTENSEN 1976 ). The rev3 strain shows a low level of spontaneous mutagenesis (Table 4) as previously described (ROCHE et al. 1995 ). The mutagenic effect of the rad18 deletion is completely dependent on Rev3 functions because the mutation rate of the rad18 rev3 double mutant is as low as that of the rev3 single mutant (Table 4). Similarly, the rad5 rev3 mutant shows the same mutation rate as rev3. This means that Rev3 can act either in a subpathway mediated by Rad18 or in a subpathway mediated by Rad5. However, the mutator phenotype of rad6 is only slightly decreased by the rev3 deletion and, according to the statistical analysis, it is statistically indistinguishable from rad6, rad5 rad6, or rad18 rad6 mutant strains. It is interesting that rad6 shows any mutator phenotype, because the pol activity has been considered to be completely dependent on Rad6 function (LAWRENCE and CHRISTENSEN 1976 ). Our results clearly show that the cause of a higher mutation rate in rad6 mutants is unrelated to polymerase . Another situation is observed in UV-induced mutagenesis, where the rad6 and rev3 strains show the same antimutator phenotype (KUNZ et al. 2000 ).

Which pathways are responsible for the high mutagenesis rate in rad6 and rad6 rev3 mutants? We propose the existence of another mutagenic repair pathway different from Rev3-mediated repair. This pathway is responsible for nearly all mutations in rad6 null mutants and all multiple mutants in combination with rad6 and does not occur if at least partially functional Rad6 protein is present. Recently, it was shown that POL32 might be responsible for the mutator phenotype of rad6 null mutant, because the double mutant rad6 pol32 shows nearly no mutations in the canavanine resistance forward assay (HUANG et al. 2000 ).

The analysis of UV sensitivity revealed the additive sensitivity in double mutants rad5 rev3 and rad18 rev3 (Fig 5), suggesting that RAD5 and REV3, and RAD18 and REV3, respectively, form two separate and independent DNA repair pathways. It also confirmed the parity of RAD5- and RAD18-mediated subpathways. We also observed that the relative sensitivity decreases with higher levels of irradiation in the rad18 rev3 mutant strain (Fig 5). This would suggest the existence of an unknown DNA repair activity induced by a high level of DNA damage.

In contrast to previously published results, our detailed analysis revealed higher UV sensitivity of rad6 rev3 mutants than a single rad6 mutant, suggesting that Rev3 could work in two different modes. The first is dependent on the RAD6 repair pathway and the second is independent of RAD6.

MMS2 belongs to the error-free pathway:
The highest mutation rate was observed in the mms2 mutant (Table 4). This confirms the engagement of MMS2 in the error free branch of RAD6 repair pathway (BROOMFIELD et al. 1998 ). This strain was considered as a control for the situation when only error-prone pathways of RAD6 repair pathways are functional. Also the SAM is very high in the mms2 strain, confirming the hypothesis that an error-free RAD6 repair pathway is responsible for maintaining the low level of SAM (Fig 3).

A single mms2 mutant shows very low sensitivity to UV and at low doses of UV it is almost indistinguishable from the wild-type strain (Fig 4). Thus, this protein has a more important function in avoiding mutagenic repair than in the repair of UV-induced lesions.

The mutants show a similar phenotype in both starvation-associated and spontaneous mutagenesis:
The effect in starvation-associated mutagenesis of the mutants analyzed follows basically the same principles as in spontaneous mutagenesis. Our findings might suggest that spontaneous and starvation-associated mutagenesis cannot be shown to be different processes. However, it was already shown that the type of reversion and mutation spectrum is different in spontaneous and SA mutagenesis (STORCHOVA et al. 1998 ; HEIDENREICH and WINTERSBERGER 2001 ). We suppose that the RAD6 postreplicative repair pathway is a nonspecific pathway ensuring tolerance to any blocking lesion in DNA and thus helping to maintain chromosomal stability. Such a function can be important in any stage of the yeast life cycle, which results in similar epistasis observed for both spontaneous and starvation-associated mutagenesis. The RAD6 repair pathway is believed to be activated as the replication machinery stalls when it is unable to bypass a lesion to further synthesize DNA (for review, see KUNZ et al. 2000 ). We assume that this happens not only during conventional DNA replication in growing cells, but also during repair synthesis in starving and nongrowing cells.

The strains can be classified into three groups according to the accumulation of starvation-associated revertants: (1) the mms2, rad6, rad5, and rad18 single mutants, the rad6 rad5, rad6 rad18, and rad6 rev3 double mutants, and the rad6 rad5 rad18 triple mutant show the highest level of SAM, similar to spontaneous mutagenesis; (2) relatively low levels of revertants, again similar to spontaneous mutagenesis, are seen in the rad5 rad18 double mutant; and (3) the lowest numbers of SA revertants were observed in wild-type, rev3, and rev3rad5 strains. Although the number of SA revertants in the rev3 strain is significantly lower in comparison to the wild type, the mutant strain is less viable during starvation, so the resulting SA mutagenesis of the rev3 strain is only partially lower than that of the wild type (Fig 6). Polymerase is synthesized at the same level during the entire cell cycle (SINGHAL et al. 1992 ), unlike major replicative DNA polymerases (MORRISON et al. 1988 ; JOHNSON et al. 1994 ), and thus its effect could be higher in the stationary phase. Hence, this polymerase was suggested to play a key role in starvation-associated mutagenesis as a result of mutagenic synthesis either on undamaged templates (HOLBECK and STRATHERN 1997 ) or during translesion synthesis (BAYNTON et al. 1998 ). However, our results surprisingly suggest that the role of error-prone polymerase in the mutagenesis of nongrowing cells is not higher than in spontaneous mutagenesis. The starvation-associated mutations can arise in the wild-type strain either because more mutations are formed during starvation or because they are less efficiently removed from DNA. Our data on the rev3 mutant support the latter hypothesis.

Overall, it appears that the genetic relationships in spontaneous and SA mutagenesis are similar. An important exception is the rev3 rad18 double mutant, which shows an unusual course in the accumulation of starvation-associated revertants. At the beginning, the mutagenesis is very low, as in wild-type and rev3 mutants. The numbers of revertants increase later with the period of starvation and finally reach levels as high as in rad5 or rad18 single mutants (Fig 3C). This suggests the possible induction of another DNA repair pathway acting in a mutagenic manner. The mutant also reveals very unusual survival after UV irradiation: it is very sensitive to low-dose UV irradiation, but relatively less sensitive to the higher doses. Induced activation of an alternative pathway as a result of severe DNA damage caused by UV irradiation or long-term starvation might explain this phenomenon. This pathway is very likely not mediated by another error-prone polymerase coded by RAD30 because its activity requires the Rad6/Rad18 heterodimer (WASHINGTON et al. 1999 ).

RAD6 repair pathway is active in starving cells:
Although stationary phase cells have 10-fold less RAD6 mRNA than exponential phase cells (MADURA et al. 1990 ), our results suggest that the RAD6 repair pathway is active in nondividing cells. The Rad6 protein is rather abundant in the exponential phase and its concentration is not limiting the DNA repair (unlike the concentration of Rad18; DOR et al. 1996 ). Therefore, even a substantial decrease of Rad6 concentration can sustain activity of the RAD6 repair pathway. Furthermore, we confirmed the presence of RAD6 mRNA in S. cerevisiae even after 5 days of starvation in liquid culture without a single supplement (Z. STORCHOVÁ, unpublished data).

Model for RAD6 postreplicative repair group:
We showed that the Rad6 protein acts in two different ways during DNA repair and mutagenesis: (1) it functions in both error-free and error-prone DNA repair pathways in cooperation with Rad18 and/or Rad5 and (2) it interferes with other DNA repair pathways. For example, the phenotype of the rad5 rad18 mutant is different in the presence or absence of RAD6. Both rad5 and rad18 single mutants exhibit a very high mutation rate, almost exclusively dependent on REV3. Deletion of both genes brings about a dramatic decrease in mutagenesis and a synergistic effect on UV sensitivity, which suggests partial redundancy of Rad5- and Rad18-mediated pathways and competition for the same substrate, which might be processed by REV3 if any of Rad5 or Rad18 are present. However, further deletion of rad6 increases the mutation rate of the rad5 rad18 double mutant again. The presence of the Rad6 protein might therefore block other repair pathways.

On the basis of our genetic data and data published by other groups, we suggest the following working model for the RAD6 postreplicative group. We hypothesize that the RAD6 repair group consists of three different subpathways (Fig 7). The first complex is responsible for major error-free repair and consists of Rad18/Rad6/Rad5/Mms2/Ubc13 proteins allowing the error-free bypass of lesions in DNA. All these proteins were shown to interact at least transiently (ULRICH and JENTSCH 2000 ). MMS2, UBC13, pol, and PCNA were clearly shown to belong only to the error-free pathway (TORRES-RAMOS et al. 1996 ; BROOMFIELD et al. 1998 ; HOFMANN and PICKART 1999 ).

View larger version (18K): In this window In a new window Download PPT slide   Figure 7. Model of interactions within the RAD6 repair group.

The second subpathway requires Rad18/Rad6 dimers and ensures that translesion synthesis is performed by either Rev3 or Rad30. It was shown that Rad30 can be involved in both error-free and error-prone repair, depending on lesion type (YUAN et al. 2000 ), and its activity definitely requires Rad18 and Rad6 (MCDONALD et al. 1997 ). This is not the case for Rev3, because the mutator phenotype of both rad5 and rad18 mutants is completely dependent on Rev3. Thus, pol is functional in the Rad18/Rad6-mediated subpathway as well as in the Rad5/Rad6 subpathway. The Rad6/Rad5/Rev3 subpathway is responsible for the majority of mutations in rad18 strains, whereas the Rad6/Rad18/Rev3 subpathway is responsible for the mutations in rad5 strains.

The model presented above is based almost exclusively on genetic analysis, which presupposes linear unbranched processes. To the best of our knowledge, the scarce biochemical data available are in agreement with our model. Our study underlines the need for more detailed biochemical studies, which will be necessary to perform to understand the exact role of the RAD6 repair pathway in both growing and starving cells.

	ACKNOWLEDGMENTS

We thank S. Prakash, F. Fabre, and L. Hlavatá for plasmids used in this study. Z.S. is grateful to P. Schär and M. Räschle for comprehensive suggestions about the manuscript and Ann E. Randolph for English grammar corrections. This work was supported partially by the Grant Agency of Charles University 1999/293 B BIO to Z.S. and partially by the Institute of Medical Radiobiology with kind support of J. Jiricny.

Manuscript received May 9, 2001; Accepted for publication August 17, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN, J. A. SMITH and K. STRUHL, 1994 Current Protocols in Molecular Biology. John Wiley & Sons, New York.

BAILLY, V., J. LAMB, P. SUNG, S. PRAKASH, and L. PRAKASH, 1994  Specific complex formation between yeast RAD6 and RAD18 proteins: a potential mechanism for targeting RAD6 ubiquitin-conjugating activity to DNA damage sites. Genes Dev. 8:811-820[Abstract/Free Full Text].

BAILLY, V., S. LAUDER, S. PRAKASH, and L. PRAKASH, 1997a  Yeast DNA repair proteins Rad6 and Rad18 form a heterodimer that has ubiquitin conjugating, DNA binding, and ATP hydrolytic activities. J. Biol. Chem. 272:23360-23365[Abstract/Free Full Text].

BAILLY, V., S. PRAKASH, and L. PRAKASH, 1997b  Domain required for dimerization of yeast Rad6 ubiquitin-conjugating enzyme and Rad18 DNA binding protein. Mol. Cell. Biol. 17:4536-4543[Abstract].

BAYNTON, K., A. BRESSON-ROY, and R. P. FUCHS, 1998  Analysis of damage tolerance pathways in Saccharomyces cerevisiae: a requirement for Rev3 DNA polymerase in translesion synthesis. Mol. Cell. Biol. 18:960-966[Abstract/Free Full Text].

BAYNTON, K., A. BRESSON-ROY, and R. FUCHS, 1999  Distinct roles for Rev1p and Rev7p during translesion synthesis in Saccharomyces cerevisiae. Mol. Microbiol. 34:124-133[Medline].

BRIDGES, B. A., 1996  Mutation in resting cells: the role of endogenous DNA damage. Cancer Surv. 28:155-167[Medline].

BROOMFIELD, S., B. L. CHOW, and W. XIAO, 1998  MMS2, encoding a ubiquitin-conjugating-enzyme-like protein, is a member of the yeast error-free postreplication repair pathway. Proc. Natl. Acad. Sci. USA 95:5678-5683[Abstract/Free Full Text].

DOHMEN, R. J., K. MADURA, B. BARTEL, and A. VARSHAVSKY, 1991  The N-end rule is mediated by the UBC2 (RAD6) ubiquitin-conjugating enzyme. Proc. Natl. Acad. Sci. USA 88:7351-7355[Abstract/Free Full Text].

DOR, Y., B. RABOY, and R. G. KULKA, 1996  Role of the conserved carboxy-terminal alpha-helix of Rad6p in ubiquitination and DNA repair. Mol. Microbiol. 21:1197-1206[Medline].

FABRE, F., N. MAGANA-SCHWENCKE, and R. CHANET, 1989  Isolation of the RAD18 gene of Saccharomyces cerevisiae and construction of rad18 deletion mutants. Mol. Gen. Genet. 215:425-430[Medline].

FOSTER, P. L., 2000  Adaptive mutation: implications for evolution. Bioessays 22:1067-1074[Medline].

FRIEDBERG, E. C., G. C. WALKER and W. WIEDE, 1995 DNA Repair and Mutagenesis. American Society for Microbiology, Washington, DC.

GIETZ, R. D., R. H. SCHIESTL, A. R. WILLEMS, and R. A. WOODS, 1995  Studies on the transformation of intact yeast cells by the LiAc/SS-DNA/PEG procedure. Yeast 11:355-360[Medline].

GULDENER, U., S. HECK, T. FIEDLER, J. BEINHAUER, and J. H. HEGEMANN, 1996  A new efficient gene disruption cassette for repeated use in budding yeast. Nucleic Acids Res. 24:2519-2524[Abstract/Free Full Text].

HALL, B. G., 1992  Selection-induced mutations occur in yeast. Proc. Natl. Acad. Sci. USA 89:4300-4303[Abstract/Free Full Text].

HEIDENREICH, E. and U. WINTERSBERGER, 2001  Adaptive reversions of a frameshift mutation in arrested Saccharomyces cerevisiae cells by simple deletion in mononucleotide repeats. Mutat. Res. 473:101-107[Medline].

HOFMANN, R. M. and C. M. PICKART, 1999  Noncanonical MMS2-encoded ubiquitin-conjugating enzyme functions in assembly of novel polyubiquitin chains for DNA repair. Cell 96:645-653[Medline].

HOLBECK, S. L. and J. N. STRATHERN, 1997  A role for REV3 in mutagenesis during double-strand break repair in Saccharomyces cerevisiae. Genetics 147:1017-1024[Abstract].

HUANG, M. E., A. DE CALIGNON, A. NICOLAS, and F. GALIBERT, 2000  POL32, a subunit of the Saccharomyces cerevisiae DNA polymerase delta, defines a link between DNA replication and the mutagenic bypass repair pathway. Curr. Genet. 38:178-187[Medline].

JOHNSON, R. E., S. PRAKASH, and L. PRAKASH, 1994  Yeast DNA repair protein RAD5 that promotes instability of simple repetitive sequences is a DNA-dependent ATPase. J. Biol. Chem. 269:28259-28262[Abstract/Free Full Text].

KUNZ, B. A., A. F. STRAFFON, and E. J. VONARX, 2000  DNA damage-induced mutation: tolerance via translesion synthesis. Mutat. Res. 451:169-185[Medline].

LAWRENCE, C., 1994  The RAD6 DNA repair pathway in Saccharomyces cerevisiae: What does it do, and how does it do it? Bioessays 16:253-258[Medline].

LAWRENCE, C. W. and R. CHRISTENSEN, 1976  UV mutagenesis in radiation-sensitive strains of yeast. Genetics 82:207-232[Abstract/Free Full Text].

LEA, D. E. and C. A. COULSON, 1949  The distribution of the numbers of mutants in bacterial population. J. Genet. 49:264-285.

LIEFSHITZ, B., R. STEINLAUF, A. FRIEDL, F. ECKARDT-SCHUPP, and M. KUPIEC, 1998  Genetic interactions between mutants of the ‘error-prone’ repair group of Saccharomyces cerevisiae and their effect on recombination and mutagenesis. Mutat. Res. 407:135-145[Medline].

MADURA, K., S. PRAKASH, and L. PRAKASH, 1990  Expression of the Saccharomyces cerevisiae DNA repair gene RAD6 that encodes a ubiquitin conjugating enzyme, increases in response to DNA damage and in meiosis but remains constant during the mitotic cell cycle. Nucleic Acids Res. 18:771-778[Abstract/Free Full Text].

MCDONALD, J. P., A. S. LEVINE, and R. WOODGATE, 1997  The Saccharomyces cerevisiae RAD30 gene, a homologue of Escherichia coli dinB and umuC, is DNA damage inducible and functions in a novel error-free postreplication repair mechanism. Genetics 147:1557-1568[Abstract].

MORRISON, A., E. J. MILLER, and L. PRAKASH, 1988  Domain structure and functional analysis of the carboxyl-terminal polyacidic sequence of the RAD6 protein of Saccharomyces cerevisiae. Mol. Cell. Biol. 8:1179-1185[Abstract/Free Full Text].

NELSON, J. R., C. W. LAWRENCE, and D. C. HINKLE, 1996  Thymine-thymine dimer bypass by yeast DNA polymerase zeta. Science 272:1646-1649[Abstract].

ROBZYK, K., J. RECHT, and M. A. OSLEY, 2000  Rad6-dependent ubiquitination of histone H2B in yeast. Science 287:501-504[Abstract/Free Full Text].

ROCHE, H., R. D. GIETZ, and B. A. KUNZ, 1994  Specificity of the yeast rev3 delta antimutator and REV3 dependency of the mutator resulting from a defect (rad1 delta) in nucleotide excision repair. Genetics 137:637-646[Abstract].

ROCHE, H., R. D. GIETZ, and B. A. KUNZ, 1995  Specificities of the Saccharomyces cerevisiae rad6, rad18, and rad52 mutators exhibit different degrees of dependence on the REV3 gene product, a putative nonessential DNA polymerase. Genetics 140:443-456[Abstract].

SIKORSKI, R. S. and P. HIETER, 1989  A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122:19-27[Abstract/Free Full Text].

SINGHAL, R. K., D. C. HINKLE, and C. W. LAWRENCE, 1992  The REV3 gene of Saccharomyces cerevisiae is transcriptionally regulated more like a repair gene than one encoding a DNA polymerase. Mol. Gen. Genet. 236:17-24[Medline].

STEELE, D. F. and S. JINKS-ROBERTSON, 1992  An examination of adaptive reversion in Saccharomyces cerevisiae. Genetics 132:9-21[Abstract].

STORCHOVÁ, Z., A. P. ROJAS GIL, B. JANDEROVA, and V. VONDREJS, 1998  The involvement of the RAD6 gene in starvation-induced reverse mutation in Saccharomyces cerevisiae. Mol. Gen. Genet. 258(5):546-552[Medline].

SUNG, P., S. PRAKASH, and L. PRAKASH, 1988  The RAD6 protein of Saccharomyces cerevisiae polyubiquitinates histones, and its acidic domain mediates this activity. Genes Dev. 2:1476-1485[Abstract/Free Full Text].

SUNG, P., S. PRAKASH, and L. PRAKASH, 1990  Mutation of cysteine-88 in the Saccharomyces cerevisiae RAD6 protein abolishes its ubiquitin-conjugating activity and its various biological functions. Proc. Natl. Acad. Sci. USA 87:2695-2699[Abstract/Free Full Text].

TORRES-RAMOS, C. A., B. L. YODER, P. M. BURGERS, S. PRAKASH, and L. PRAKASH, 1996  Requirement of proliferating cell nuclear antigen in RAD6-dependent postreplicational DNA repair. Proc. Natl. Acad. Sci. USA 93:9676-9681[Abstract/Free Full Text].

ULRICH, H. D. and S. JENTSCH, 2000  Two RING finger proteins mediate cooperation between ubiquitin-conjugating enzymes in DNA repair. EMBO J. 19:3388-3397[Medline].

WASHINGTON, M. T., R. E. JOHNSON, S. PRAKASH, and L. PRAKASH, 1999  Fidelity and processivity of Saccharomyces cerevisiae DNA polymerase eta. J. Biol. Chem. 274:36835-36838[Abstract/Free Full Text].

WATKINS, J. F., P. SUNG, S. PRAKASH, and L. PRAKASH, 1993  The extremely conserved amino terminus of RAD6 ubiquitin-conjugating enzyme is essential for amino-end rule-dependent protein degradation. Genes Dev. 7:250-261[Abstract/Free Full Text].

XIAO, W., B. L. CHOW, T. FONTANIE, L. MA, and S. BACCHETTI et al., 1999  Genetic interactions between error-prone and error-free postreplication repair pathways in Saccharomyces cerevisiae. Mutat. Res. 435:1-11[Medline].

XIAO, W., B. L. CHOW, S. BROOMFIELD, and M. HANNA, 2000  The Saccharomyces cerevisiae RAD6 group is composed of an error-prone and two error-free postreplication repair pathways. Genetics 155:1633-1641[Abstract/Free Full Text].

YUAN, F., Y. ZHANG, D. K. RAJPAL, X. WU, and D. GUO et al., 2000  Specificity of DNA lesion bypass by the yeast DNA polymerase eta. J. Biol. Chem. 275:8233-8239[Abstract/Free Full Text].

This article has been cited by other articles:

				E. Mito, J. V. Mokhnatkin, M. C. Steele, V. L. Buettner, S. S. Sommer, G. M. Manthey, and A. M. Bailis Mutagenic and Recombinagenic Responses to Defective DNA Polymerase {delta} Are Facilitated by the Rev1 Protein in pol3-t Mutants of Saccharomyces cerevisiae Genetics, August 1, 2008; 179(4): 1795 - 1806. [Abstract] [Full Text] [PDF]

				D. L. Daee, T. Mertz, and R. S. Lahue Postreplication Repair Inhibits CAG {middle dot} CTG Repeat Expansions in Saccharomyces cerevisiae Mol. Cell. Biol., January 1, 2007; 27(1): 102 - 110. [Abstract] [Full Text] [PDF]

K. Herzberg, V. I. Bashkirov, M. Rolfsmeier, E. Haghnazari, W. H. McDonald, S. Anderson, E. V. Bashkirova, J. R. Yates III, and W.-D. Heyer Phosphorylation of Rad55 on Serines 2, 8, and 14 Is Required for Efficient Homologous Recombination in the Recovery of Stalled Replication Forks Mol. Cell. Biol., November 15, 2006;