Description

The RecQ helicases play a crucial role in the maintenance of genome stability. The importance of these helicases is most apparent in genetic diseases associated with RecQ helicases. In human, defects in BLM, RECQL4 or WRN result in Bloom’s syndrome, Rothmund–Thomson syndrome and Werner’s syndrome, respectively (Ellis et al. 1995; Yu et al. 1996; Kitao et al. 1999). A common characteristic of all these syndromes is a predisposition to genome stability and cancer (Croteau et al. 2014). In Saccharomyces cerevisiae, Sgs1p is a RecQ helicase homolog. Sgs1p functions as part of the STR complex (Sgs1p-Top3p-Rmi1p complex) and regulates recombination to prevent illegitimate structures (Croteau et al. 2014). For instance, cells deficient in SGS1 accumulated Rad51p-dependent aberrant DNA structures at replication forks when cells were treated with methylmethane sulfonate (MMS) (Liberi et al. 2005). How Sgs1p regulates recombination, in part, depends on sumoylation, a process in which a SUMO (small ubiquitin-like modifier) peptide is covalently conjugated to target proteins. In response to DNA damage, Sgs1p becomes sumoylated (Branzei et al. 2006; Bermúdez-López et al. 2016; Bonner et al.. 2016). Sumoylation of Sgs1p is mediated by an E3 SUMO ligase, Mms21p and this modification supports the ability of Sgs1p to resolve aberrant DNA structures at replication forks (Branzei et al. 2006; Bermúdez-López et al. 2016; Bonner et al. 2016).

Although recombination events help to overcome the impediments at replication forks, failure to regulate these processes can become the source of genome instability. Thus, it is not surprising that more than one pathway function at replication forks to prevent illegitimate recombination. Srs2p is another helicase responsible for inhibiting aberrant DNA structures in S. cerevisiae. Although Sgs1p and Srs2p are both 3’-to-5’ helicases and generally considered to be anti-recombinogenic, almost synthetic lethal phenotype of sgs1 srs2 mutants implies that they function in independent pathways (Gangloff et al. 2000; Marini and Krejci 2010). Consistent with this idea, a recent study detailing the homologous recombination (HR) process involving Sgs1p and Srs2p uncovered that these helicases target different types of D-loops during HR (Piazza et al. 2019). However, genetic data suggest that Sgs1p and Srs2p have some overlapping functions. Sgs1p can suppress Srs2p deficiency in a dose-dependent manner suggesting that Sgs1p may partially compensate for the loss of Srs2p (Mankouri et al. 2002). Likewise, overexpression of Srs2p in sgs1Δ reduces MMS-induced accumulation of X-molecules at replication forks (Liberi et al. 2005).

Based on these data, we asked if SGS1 SUMO mutants, similar to sgs1 knock out cells, would exhibit genetic interaction with srs2 mutants. To this end, we genetically modified SGS1 at the endogenous locus to create a condition in which Sgs1p cannot be sumoylated (SGS1-SuOff). In this strain, SGS1 gene is fused to the domain of Ulp1p, harboring the region where SUMO contacts with Ulp1p and the isopeptidase enzymatic activity. The enzymatic activity of this domain deconjugates SUMO from the target protein, preventing Sgs1p from being sumoylated (Almedawar et al. 2012; Bermúdez-López et al. 2016; Wei and Zhao 2016). This tag indeed reduces the sumoylation of Sgs1p (Bermúdez-López et al. 2016). We also generated a strain to mimic constant sumoylation of Sgs1p (SGS1-SuOn). This strain carries a transgene encoding for Sgs1p fused to the Ulp1p isopeptidase domain that retains its affinity for SUMO but harbors a mutation that renders it catalytically dead (Almedawar et al. 2012; Wei and Zhao 2016). This genetic modification is known to promote sumoylation of the protein product of the target gene likely due to an increase in local concentration of SUMO (Wei and Zhao 2016). Increase in sumoylation of the target when fused to the SuOn tag has been demonstrated with Scc1p and Mcm6p (Almedawar et al. 2012; Wei and Zhao 2016). As a control, a similar tag with point mutations that disrupt the respective function of the SUMO-contact region and the isopeptidase activity was used (SGS1-SuC) (Almedawar et al. 2012; Wei and Zhao 2016). The sumoylation status of SGS1-SuOn and SGS1-SuC has not been examined in our study.

Wild-type SGS1 and SGS1-SuC strains grew at a comparable rate and showed similar sensitivity to 0.05% MMS, suggesting that the SUMO tag does not significantly impair the growth of wild-type cells (Figure 1A). Compared to the SGS1-SuC strains, the SUMO mutants (SGS1-SuOn and SGS1-SuOff) did not exhibit appreciable MMS sensitivity (Figure 1B). Consistent with our data, Bermúdez-López et al. observed a lack of MMS sensitivity for sgs1 mutants in which three lysine residues where Sgs1p is sumoylated are converted to arginines (Bermúdez-López et al. 2016). Lack of MMS sensitivity does not preclude the possibility that Sgs1p sumoylation contributes to its function. In fact, sumoylation of Sgs1p mediates its function in the context of the STR complex and discourages crossover events during double-stranded break repair (Bermúdez-López et al. 2016). To further understand the contribution of Sgs1p sumoylation, we took a genetic approach and studied genetic interaction between SGS1 SUMO mutants and srs2. Previous studies suggest that srs2Δ is synthetically lethal with sgs1Δ (Gangloff et al. 2000). To understand if a similar genetic interaction exists between srs2 mutants and SGS1 SUMO mutants, we knocked out SRS2 in strains with wild-type SGS1 or SGS1 fused with different SUMO tags. SGS1-SuC srs2Δ cl.1 exhibits MMS sensitivity similar to that of srs2Δ clones (cl.1 and 2) (Figure 1A). However, this phenotypic similarity was not observed for SGS1-SuC srs2Δ cl.2 and therefore, this clone was excluded from later experiments (Figure 1A). We generated independent clones of SGS1-SuOn srs2Δ and SGS1-SuOff srs2Δ and chose the clones that showed consistent phenotypes for downstream experiments (Figure 1C). SGS1-SuOn srs2Δ double mutants exhibited similar MMS sensitivity compared to SGS1-SuC srs2Δ cells (Figure 1B). However, SGS1-SuOff allele suppressed the weak MMS sensitivity that srs2Δ mutants exhibited (Figures 1B, C and D).

We reason that previously reported genes that exhibit a positive genetic interaction with srs2Δ may provide insights into the function of Sgs1p sumoylation. For example, disabling the function of Rad51p rescues the sensitivity of srs2Δ mutants to MMS, UV and γ-radiation, suggesting that srs2Δ cells accumulate Rad51p-dependent recombination structures (Aboussekhra et al. 1992). Srs2p can indeed disrupt Rad51p presynaptic filaments to regulate unscheduled recombination (Krejci et al. 2003; Veaute et al. 2003). Since SGS1-SuOff mutation also rescues the weak sensitivity of srs2Δ observed in this strain, one simple explanation is that deficiency to sumoylate Sgs1p prevents or attenuates unregulated Rad51p-mediated recombination in srs2Δ cells. The phenotype of SGS1-SuOff srs2Δ can be confirmed by using another SUMO mutant of Sgs1p (sgs1-K612R or sgs1-3KR). Furthermore, knocking out RAD51 in SGS1-SuOff srs2Δ can also elucidate if Rad51p is responsible for the observed phenotype.