Introduction

A largely conserved network of proteins is in place to prevent lethal and mutagenic consequences of DNA damage and replication stress in eukaryotes (Friedberg et al., 2006). In this context, the regulation of dNTP levels has attracted considerable attention and many details surrounding the regulation of the heterotetrameric ribonucleotide reductase (RNR) are known in Saccharomyces cerevisiae. RNR activity and level is increased in S-phase and in response to DNA damage or replicational stress (e.g. by hydroxyurea [HU]) through various mechanisms: by transcriptional regulation (inactivation of the Crt1 co-repressor), by subunit localization, by inhibitor interaction and degradation, and by altered feedback inhibition (Chabes et al., 2003; Elledge et al., 1993; Huang et al., 1998; Lee and Elledge, 2006; Wu and Huang, 2008; Yao et al., 2003; Zhao et al., 2001).

In budding yeast, an elevated dNTP level can prevent the lethality of checkpoint kinase (Mec1 or Rad53) deletions. These checkpoint kinases are necessary for cell cycle arrest and facilitation of DNA repair (Friedberg et al., 2006; Nyberg et al., 2002). Elevated dNTP levels can be achieved by deletion of negative regulators such as Sml1, Ctr1 or Dif1 or by overexpression of RNR subunit Rnr1 (Desany et al., 1998; Lee et al., 2008; Zhao et al., 1998).

Additional genes of unknown function have been identified through a selection for genes that, if overexpressed, suppress the lethality of a RAD53 and/or MEC1 deletion (srl , for suppressor of Rad53 lethality) (Desany et al., 1998). Some of these genes have remained uncharacterized. The present study concerns SRL3 whose deletion in a commercially available strain was identified by us in a screen for spontaneous mutator mutants. However, its high spontaneous mutability and DNA damage sensitivity phenotype were found to be associated with an independent mutation which was identified as a truncation of MMS2. MMS2 is involved in a replicative DNA damage tolerance pathway (Broomfield et al., 1998; Broomfield et al., 2001; Friedberg et al., 2006; Gangavarapu et al., 2006; Hofmann and Pickart, 1999; Torres-Ramos et al., 2002; VanDemark et al., 2001; Xiao et al., 1998). While we found no direct connection between SRL3 and MMS2, important information pertaining to both gene products was uncovered.

Materials and Methods

Yeast strains and strain construction

Determination of spontaneous mutation rates

Survival assays

Checkpoint assays following UV irradiation

Results

The commercially available collection of deletion mutants of non-essential genes of a haploid yeast strain (BY4741) was screened for strains with higher spontaneous mutation frequencies, using canavanine resistance as a forward mutation marker (Gong and Siede, 2009). While this screen was similar to the one described earlier by others (Huang et al., 2003), two novel genes were identified: SRL3 and TOP3.

We focused on srl3∆ and confirmed quantitatively that spontaneous mutation rates were approximately 10 fold higher in this strain than in the isogenic wild-type (Table 1). (In order to avoid confusion, we have designated the genotype of this strain as srl3∆ uvs since subsequent analysis indicated the existence of a second mutation.) SLR3 has been described as a suppressor of lethality of a Rad53 deletion if overexpressed and in general, higher dNTP pools have been identified as a mechanism for this effect (Desany et al., 1998). Therefore, using the CAN1 system, we determined spontaneous mutation rates of mutants that have presumably lower or higher dNTP levels due to deletion of RNR subunits (rnr1∆, rnr3∆, rnr4∆) or of a negative regulator of RNR (sml1∆) (Zhao et al., 1998), respectively. In no instance, we found an influence on spontaneous mutation rates that is quantitatively comparable to that of srl3∆ uvs (Table 1). At the most, an approximately 3 fold enhanced rate was found for RNR1 deletion compared to wild type.

Figure 1

Table 1: Spontaneous rates of canavanine-resistance mutations, expressed as fold difference over wild type (= 1).

Data shown are from representative single experiments, except for strain srl3∆ uvs which represents the average of three measurements including standard deviation.

Also, determination of UV sensitivity in srl3∆ uvs indicated a phenotype different from other mutants with possibly lower dNTP levels. The srl3∆ uvs strain showed notably increased UV sensitivity while none of the RNR mutants did (Fig. 1 A). Using cells synchronized in G1, S or M, UV sensitivity of srl3∆ uvs was demonstrated to be cell cycle stage independent (data not shown). Sensitivity was also shown towards the alkylating agent MMS and the double-strand break-inducing agent streptonigrin (Fig. 1 B, C).

Figure 2

Figure 1: Survival of BY4741 following exposure to UV radiation (A), MMS (B) and streptonigrin (C). Additionally, the UV survival of strains BY4741 , and are shown (A). Logarithmic-phase cells were treated, as described in Material and Methods, and the fraction of colony forming cells was plotted as a function of dose

UV sensitivity was not due to a checkpoint defect. When cells were synchronized in G1 with alpha-factor, in S with HU and in M with nocodazole before treatment with UV, a more extended arrest was found in all tested cell cycle stages (Fig. 2 A, B, C).

Figure 3

Figure 2: Checkpoint arrest duration in in comparison to wild type. Exit from G1 (A), S (B) and M phase (C) following UV radiation treatment (80 J/m) was analyzed by determining the fraction of small budded cells in liquid medium (A) or by the emergence of microcolonies (with more than two cells) on plates (B, C)

Prompted by the fact that the second commercially available srl3∆ strain of opposite mating type did not exhibit DNA damage sensitivity (not shown), we hypothesized that there might be an unlinked mutation present in the used strain. Therefore, we crossed the strain against wild type, sporulated the diploid strain and isolated individual spores by micromanipulation. Indeed, the mutagen sensitivity did not co-segregate with the SRL3 deletion which can be identified through the kanMX4 marker gene conferring G418 resistance (Fig. 3 A). The unknown mutation was termed uvs. Additionally, elevated spontaneous mutation rates were caused by the uvs mutation but not the SRL3 deletion (Table 1).

In order to identify possible candidate genes whose mutation may result in the observed phenotypes, we performed epistasis analyses with well-characterized single gene deletions conferring similar phenotypes. Briefly, in epistasis analysis, an uncharacterized mutant gene affecting DNA damage sensitivity is assigned to the same DNA repair or tolerance pathway as a characterized mutant gene if the double mutant’s phenotype is not more severe than that of the most sensitive single mutant (Brendel and Haynes, 1973). Given the sensitivity of the uvs mutant to both UV radiation and the alkylating agent MMS, a role in either homologous recombination (Rad51 pathway) or replicative DNA tolerance (Rad6 pathway) is likely (Friedberg et al., 2006). If combined with a RAD52 deletion (representing the Rad51 pathway), non-epistatic interaction was found for both UV and MMS (not shown). If combined with a RAD5 deletion, representing an error-free sub-pathway of the Rad6 pathway, epistatic interaction was found for both agents (Fig. 3 B, C). We also investigated interaction with RAD50, encoding a protein of the MRN complex, involved in several pathways depending on double-strand break recognition and processing (D'Amours and Jackson, 2002; Friedberg et al., 2006). We found an epistatic interaction with uvs for MMS and for low doses of UV which became additive at higher doses (Fig. 3 B, C).

Next, we identified candidate genes of the error-free subpathway of the Rad6 epistasis group or the MRN complex whose mutation may account for the uvs phenotype. A haploid uvs strain was mated with deletion mutants of RAD50, MRE11, SRL3, RAD5, UBC13 and MMS2 of opposite mating type and the generated diploid strains were tested for complementation of UV and MMS sensitivity. Absence of complementation as indicated by increased sensitivity was found for mms2∆ (Fig. 3 D).

Figure 4

Figure 3: UV sensitive phenotype of is caused by a mutation in (= ). A. Segregation of UV sensitivity. Following crossing of to wild-type, spores were isolated containing deletion as indicated by the transplacement marker , conferring G418 resistance. As shown, deletion and UV sensitivity did not cosegregate. B, C. Epistasis analysis with uvs. The unknown mutation () was combined with deletions of and . In logarithmic-phase cells, UV (B) and MMS (C) sensitivities of double mutants were determined and compared to single mutant sensitivities. D. A strain was mated with strains carrying deletions of , , , , or . UV and MMS sensitivities of the diploid strains were characterized in gradient assays. Enhanced sensitivity was only found for the combination.

Consequently, the MMS2 gene of the uvs strain was isolated by PCR and subjected to sequencing using appropriate primers. Inspection of the DNA sequence revealed a G –> A transition at residue 393, resulting in a change of TAG (encoding tryptophan) to TGA (stop codon) and thus a truncation of the C-terminal 35 amino acids of Mms2 (Fig. 4).

Figure 5

Figure 4: DNA and protein sequence. The mutation identified in is indicated.

Discussion

In a screen of the commercially available deletion collection in the haploid yeast strain BY4741, deletions of SRL3 and TOP3 were among previously unidentified mutants enhancing spontaneous mutability. SRL3 has been isolated before as a gene that can suppress in high copy number the lethality of a deletion of checkpoint kinase Rad53 and, more poorly, Mec1 (Desany et al., 1998). Since no additional phenotype was known, we proceeded with a more detailed characterization and, in addition to approximately 10 fold higher mutation rates in the CAN1 system, we found sensitivity to UV radiation, MMS and streptonigrin. Extended checkpoint arrest responses correlated well with the likely DNA repair deficiency in this mutant.

Further analysis, however, revealed that the described phenotype in this particular strain originated from another unlinked mutation that results in truncation of the damage tolerance protein Mms2. An SRL3 deletion alone does in fact neither influence spontaneous mutability nor mutagen sensitivity. Elevated dNTP levels are the only known mechanism to overcome the lethality of deletions of the checkpoint kinases Mec1 or Rad53, possibly by facilitating replication through chromosomal regions that are difficult to replicate and comprise fragile sites (Cha and Kleckner, 2002). Absence of any large effect on spontaneous mutability or UV sensitivity is in agreement with the phenotype of other yeast mutants affecting dNTP pools, negatively (RNR subunits) or positively (Sml1 mutant). However, it should be noted that RNR mutants very well affect damage-induced mutation frequencies when dNTP levels appear to become more limiting than in undamaged cells. Reduced UV-mutation frequencies were found for rnr1 and rnr4 deletions (Lis et al., 2008; Strauss et al., 2007)(Gong and Siede, unpublished). Similarly, elevated dNTP levels result in higher induced mutation frequencies in response to certain types of DNA damage (Chabes et al., 2003).

Mms2 functions in a sub-pathway of the DNA-damage tolerance pathway (Rad6 group) that facilitates replication of a damaged template (Andersen et al., 2008). Mms2 is a ubiquitin-conjugating (E2) variant enzyme that forms a heterodimer with the E2 enzyme Ubc13 (VanDemark et al., 2001) and interacts with the ubiquitin-ligase Rad5 (Broomfield et al., 1998; Gangavarapu et al., 2006; Hofmann and Pickart, 1999; Torres-Ramos et al., 2002; Xiao et al., 1998). Their primary target during lesion bypass appears to be PCNA that is subject to polyubiquitination (Hoege et al., 2002). Error-free template switching is postulated as the used mechanism of lesion bypass (Branzei et al., 2008). There is one earlier report of higher spontaneous mutation rates in mms2∆ measured in a point mutant reversion system (trp1-289) (Broomfield et al., 1998). Our studies confirm the phenotype in a forward mutation system which will detect a much broader range of sequence alterations. It has been speculated that Rad5-Mms2-Ubc13 is required to restrain the action of the error-prone bypass polymerase zeta which is responsible for most spontaneous mutations; the exact mechanism, however, is not known. Our data also attest to the importance of the C-terminus of Mms2. The observed truncation does also delete Trp 393, a residue which is well conserved in various E2 enzymes (Ptak et al., 2001).

Given the role of Mms2 in the replicative bypass of UV lesion, extended checkpoint arrest of its mutant in S-phase can easily be rationalized. This is not the case with the observed extension of G1 and M phase arrest where the assumed bypass mechanism cannot easily contribute to UV resistance. We hypothesize that there may be an unknown role of Rad5-Mms2-Ubc13 during nucleotide excision repair. The observed extension of checkpoint arrest was independent of srl3∆ (data not shown).

We have no indication for any functional connection between Srl3 and Mms2 and assume that the appearance of both mutations together is a coincidence. For example, continued propagation of srl3∆ did not result in the accumulation of UV-sensitive cells (data not shown). This issue, however, will require more detailed studies in the future.