Introduction

A hallmark of low fidelity DNA polymerases, also known as DNA translesion synthesis polymerases (TLS polymerases), with no detectable proofreading activity, is their ability to catalyse DNA synthesis across a variety of bulky, helix-distorting DNA lesions (Prakash et al., 2005; Vaisman et al., 2017; Maiorano et al., 2021; Ling et al., 2022). As a result, the TLS polymerases play important roles in a plethora of cellular processes, including but not limited to epigenetics, immune signalling, viral mutagenesis and cancer development (Paniagua and Jacobs, 2023). Indeed, translesion synthesis is a source of mutagenesis, potentially contributing to the development of cancer and drug resistance (Lange et al., 2011; Baranovskiy et al., 2012; Pilzecker et al., 2019). The Saccharomyces cerevisiae TLS Polζ (hereafter referred to as ScPolζ), comprising catalytic subunit Rev3, two copies of Rev7 regulatory subunits and accessory subunits Pol31 and Pol32, plays a pivotal role in TLS (Johnson et al., 2012; Makarova et al., 2012). Several lines of evidence suggest that the error rate during ScPolζ-catalysed replication of undamaged DNA templates is much higher than that of DNA polymerases, as it lacks 3′-to-5′ proofreading exonuclease activity (Lawrence et al., 1985; Huang et al., 2002; Northam et al., 2006; Zhong et al., 2006; Kochenova et al., 2017). Consistent with this, S. cerevisiae rev3, rev7 or pol32 mutant strains show greatly reduced spontaneous mutation frequencies, confirming that ScPolζ is responsible for DNA damage-induced mutagenesis (Quah et al., 1980; Lawrence et al., 1985; Morrison et al., 1989; Nelson et al., 1996, Makarova et al., 2015).

The TLS Polζ is conserved in a wide range of unicellular and multicellular eukaryotes, including fungi, plants, and animals (Maiorano et al., 2021; Ling et al., 2022; Paniagua et al., 2023). While the catalytic subunit Rev3 alone is capable of replicating damaged DNA, Rev7 augments its catalytic efficiency by twenty- to thirtyfold (Quah et al., 1980; Morrison et al., 1989; Nelson et al., 1996, Makarova et al., 2015) and the accessory subunits Pol31 and Pol32 further raise it by three- to tenfold (Johnson et al., 2012; Makarova et al., 2012), suggesting that they enhance the processivity of translesion DNA synthesis by Polζ (Acharya et al., 2006; Bezalel-Buch et al., 2020). Biochemical and structural analyses have demonstrated that Rev7 (also known as MAD2B and MAD2L2) serves as a bridge between Rev3 and the scaffold protein Rev1 (Haracska et al., 2001; Kikuchi et al., 2012; et al., 2012; Pustovalova et al., 2012). Structural analysis of ScPolζ, using cryo-electron microscopy (cryo-EM), has revealed a remarkable level of organization, wherein its subunits interact with each other and align themselves to form a functional complex (Gómez-Llorente et al., 2013; Malik et al., 2020; Du Truong et al., 2021). Notably, structural studies on the ScPolζ holoenzyme suggest a model wherein subunits Rev3, Rev7, Pol31, and Pol32 form a pentameric ring-shaped structure, facilitated by an uninterrupted chain of protein-protein interaction networks (Gómez-Llorente et al., 2013; Malik et al., 2020; Du Truong et al., 2021). Furthermore, the cryo-EM structure of ScPolζ holoenzyme bound to an A:C mismatch show how the active site pocket accommodates mismatched duplex DNA (Malik et al., 2022).

A deluge of new findings have established that the tetrameric Shieldin complex - comprised of REV7, SHLD1, SHLD2 and SHLD3 – binds single-stranded DNA (ssDNA), blocks 5′ end resection and homologous recombination (HR), antagonizes the recruitment of BRCA1 to the DSB, while facilitating non-homologous end-joining (NHEJ) repair (Xu et al., 2015; Boersma et al., 2015; Mirman et al., 2018; Findlay et al., 2018; Gupta et el., 2018; Ghezraoui et al., 2018; Dev et al., 2018; Tomida et al., 2018; Gao et al., 2018; Noordermeer et al., 2018; Liang et al., 2020). Subsequent studies have demonstrated that the C-terminal “seat belt” region of human Rev7 binds to the N-terminus of SHLD3 (Gupta et el., 2018; Ghezraoui et al., 2018; Dev et al., 2018; Tomida et al., 2018; Gao et al., 2018; Noordermeer et al., 2018; Liang et al., 2020; Clairmont et al., 2020; Dai et al., 2020). Parenthetically, Rev7 deficiency has been linked to disease pathology: a notable example is the Fanconi anaemia in which biallelic REV7 mutations are associated with hypersensitivity to ICL-forming compounds, accumulation of chromosome breaks during S/G2 phase of the cell cycle and enhanced p53/p21 activation (Bluteau et al., 2016). While Rev7 is most well known as a component of TLS Polζ, it is also involved in the regulation of cell cycle progression by binding to CDH1, an activator of APC/C, modulate the metaphase-to-anaphase transition and facilitate mitotic spindle assembly and chromosome segregation (Listovsky and Sale, 2013; Vaisman et al., 2017; Ling et al., 2022). However, a mechanistic understanding of how Rev7 regulates cell cycle events and how such roles differ or relate to its role in TLS remains underexplored. Although the emphasis of the findings differs, the fact that Rev7 functions as an anti-resection factor has spurred a new wave of experiments on DNA repair pathways (Setiaputra and Durocher, 2019; Clairmont and D’Andrea, 2021).

At DSBs in S. cerevisiae, the Mre11–Rad50-Xrs2 (MRX) in conjunction with Sae2 first catalyses endonucleolytic cleavage of 5′-terminated DNA strands and then its 3’→5’ exonucleolytic activity produces a short 3’-ssDNA overhang; subsequently, Exo1 and Dna2-Sgs1 redundantly generate long tracks of ssDNA that are critically important for HR (Cejka and Symington, 2021). As such, the mechanism by which cells restrain over-resection of DSBs is unclear, although hyper-resection could potentially hinder optimal HR and trigger genomic instability. Surprisingly, however, Shieldin orthologs are absent in organisms such as yeast, fruit fly, nematode worm, zebrafish and frog (Setiaputra and Durocher, 2019). These data together prompted us to hypothesize that S. cerevisiae Rev7 (hereafter referred to as ScRev7) may recruit an unknown functional equivalent(s) of Shieldin orthologs to regulate the DSB repair pathway choice between HR and NHEJ. We thus reasoned that one specific alternative worthy of consideration is the MRX complex on the basis of current knowledge about its multiple roles in signalling, processing and repair of DSBs (Cejka and Symington, 2021). Here, we provide several lines of evidence, suggesting that ScRev7, through a 42-aminoacid C-terminal fragment in the “safety belt” region, directly interacts with Mre11, Rad50 or Xrs2 subunits, protects rev7Δ cells from G-quadruplex DNA/HU-induced toxicity and facilitates efficient NHEJ while antagonizing HR. Mechanistically, ScRev7 binds with low-nanomolar affinity and specificity to G-quadruplex DNA, inhibits Mre11 nuclease and Rad50’s ATPase activities, without affecting the ability of the latter to bind ATP. Collectively, our findings not only reveal unanticipated functional interactions between Rev7 and the MRX subunits, but also provide novel insights into how Rev7 regulates the pathway choice between HR and NHEJ in S. cerevisiae.

Results

ScRev7 directly interacts with the MRX subunits

As noted above, studies in cancer cell lines have documented that the Rev7-Shieldin effector complex facilitates NHEJ by blocking 5′ end resection and HR (Xu et al., 2015; Boersma et al., 2015; Mirman et al., 2018; Findlay et al., 2018; Gupta et el., 2018; Ghezraoui et al., 2018; Setiaputra and Durocher, 2019; Cejka and Symington, 2021). Given that Shieldin orthologues are absent in S. cerevisiae (Setiaputra and Durocher, 2019), we hypothesized that Rev7 may recruit alternative factors such as the MRX heterotrimer, to block HR and enable NHEJ. To test this premise, yeast two-hybrid assays (Y2H) were performed to identify potential protein–protein interactions between ScRev7 and the M/R/X subunits, whereas in follow-up studies we mapped the minimal region of ScRev7 required for its interaction with the MRX subunits. The yeast strain PJ69-4A was co-transformed with plasmids (prey vectors) encoding the Mre11, Rad50 or Xrs2 subunits and ScRev7 fused to the GAL4 DNA-binding domain (bait vector). The positive colonies were selected on SC/-Trp-Leu-His dropout medium containing 3-aminotriazole (3-AT) (Fields and Song, 1989; James et al., 1996). The results showed robust growth in the selection medium containing 3-aminotriazole (3-AT), indicating Y2H interactions between ScRev7 and the Mre11, Rad50 and Xrs2 subunits (Figure 1A), whereas cells bearing empty vector and a plasmid expressing either Mre11, Rad50 or Xrs2 subunits did not yield any Y2H signals. Conversely, yeast cells transformed with prey and bait vectors expressing Rev7 showed robust growth (bottom panel of Figure 1A), indicating that Rev7 forms homodimers (Rizzo et al., 2018), which served as a positive control.

Figure 1.

Given that Rev3 has been implicated in HR-mediated DSB repair (Sonoda et al., 2003), we asked whether the MRX subunits interact with ScRev7 in the rev3Δ mutant strain. To test this hypothesis, the S. cerevisiae rev3Δ strain was co-transformed with a combination of bait (pGBKT7 or pGBKT7-REV7) and prey vectors (pGADT7, pGADT7-REV7, pGADT7-MRE11, pGADT7-RAD50 or pGADT7-XRS2). Interestingly, we observed binary interactions between the subunits of MRX complex and ScRev7 in the rev3Δ mutant (Figure 1B), indicating that Rev3 is not necessary for the interaction between ScRev7 with the MRX subunits. Analogously ScRev7 showed binary interactions with the subunits of MRX complex in the mre11Δ rad50Δ xrs2Δ triple mutant strain (Figure 1C), further confirming that their association is independent of endogenous MRX subunits.

ScRev7 exhibits high-affinity binding to the MRX subunits

Microscale thermophoresis (MST) allows for quantitative analysis of protein-protein interactions in free solution (Wienken et al., 2010). Since the binding of ScRev7 to the MRX subunits was unexpected, we sought to validate their interaction by an orthogonal assay and determine their binding affinities using purified proteins. We assessed the binding of ScRev7-eGFP (Figure 1—figure supplement 1 and 2) to the Mre11, Rad50 and Xrs2 subunits using the MST assay. The fraction of the indicated protein bound to ScRev7-eGFP was plotted as a function of concentrations (Figure 1D). Normalized MST data were fitted to a logistic binding curve, resulting in an apparent dissociation constant (K_d) of 0.16 ± 0.07 and 0.23 ± 0.06 μM for Rad50 and Mre11 subunits, respectively, which is two- to threefold greater compared to Xrs2 (Figure 1F). To confirm these findings, we purified S. cerevisiae Rev1 (Figure 1— figure supplement 3) and assessed its interaction with ScRev7 as positive control. Consistent with previous studies (Rizzo et al., 2018; Guo et al., 2003), our results showed that Rev1 interacts with Rev7 with an affinity similar to that of Rev7 and Rad50 interaction (Figure 1D and F). On the other hand, our results demonstrated that the MRX subunits do not interact with the negative control eGFP reporter (Figure 1E). It is also noteworthy that the Hill coefficients (n_H) indicate larger than one, implying positive cooperativity (Figure 1F). The quantitative assessment of binding affinities together with Y2H data suggest that ScRev7 robustly interacts with the MRX subunits.

Models predicted by AlphaFold-Multimer reveal that Mre11 and Rad50 subunits independently associate with Rev7

To further characterize the interaction between the MRX subunits and ScRev7, we used a structure prediction algorithm, AlphaFold-Multimer (Evans et al., 2022), to generate structural models of ScRev7-Mre11 and ScRev7-Rad50 heterodimers. The models with the high pLDDT scores were considered for further analysis. Strikingly, the models indicated that Mre11 and Rad50 binding surfaces overlap with ScRev7 and thereby suggesting that the latter can only bind one of these proteins at a time (Figure 1). In the model of Rev7-Mre11 complex, residues Asp 206 and Ile 240 in the Rev7 C-terminal “safety belt region” (green) mediate dimerization with Mre11 through residues Asp 131 and Arg 181. Interestingly, His 127 in the Rev7 N-terminal HORMA domain also contribute its binding to Asp 131 of Mre11 (Figure 1—figure supplement 4A, Supplementary Table S1). Furthermore, inspection of Rev7-Rad50 complex model showed that the ScRev7 C-terminal residues Lys 168, Glu 184, Asn 189 and Asp 188 mediate dimerization with ScRad50 via residues Glu 577, Lys 596 and Arg 603 (Figure 1—figure supplement 4B, Supplementary Table S2). Curiously, AlphaFold-Multimer models also revealed that amino acid residues outside of the 42-residue fragment also contribute to pairwise interactions between Rev7 and the Mre11 and Rad50 subunits, although Y2H assays did not identify such interaction. Indeed, similar findings have been previously noted for several other interacting partners (You et al., 2006; Koegl and Uetz, 2007; Hoff et al., 2010). A structural model of Rev7-Xrs2 heterodimer was not generated because Xrs2 alone has no known biochemical activity and thus was not considered for further investigation.

A 42-aminoacid C-terminal segment of ScRev7 is critical for its interaction with the MRX subunits

As the data from Y2H and MST assays, as well as AlphaFold-Multimer-based structural modelling, implied direct interaction between the MRX subunits and ScRev7, we sought to identify the putative functional domains in the ScRev7 in regulating the function of MRX subunits. To this end, we generated three N-terminally truncated species and an equal number of C-terminally truncated variants. We refer to these variants as Rev7-N1; Rev7-N2; Rev7-N3, and Rev7-C1, Rev7-C2 and Rev7-C3, respectively (Figure 2A). We then examined for their ability to associate with the MRX subunits and thus survival of yeast cells on selective medium. Our experiments surprisingly revealed that cells expressing the ScRev7 N-terminally truncated variants showed robust growth in selection medium (bottom panel of Figure 2B) similar to the WT (Figure 1A-C), indicating that they associate with the subunits of MRX complex. Further, these results also revealed that the HORMA domain encompassing residues 1-149 at the N-terminus, an evolutionarily conserved protein–protein interaction module, is dispensable for interaction between the MRX subunits and ScRev7 (Figure 2B). Notwithstanding, we cannot exclude the possibility that it may play a function that is undetectable by the Y2H assay. We then performed Y2H experiments using the C-terminally truncated species (Figure 2C). Notably, deletion of C-terminal 42-amino acid residues (i.e., 203-245) completely abolished the viability of the yeast cells (bottom panel of Figure 2C). Similar abrogation of cell viability was also caused by deletion of the C-terminal 150 to 203 amino acid residues of ScRev7 (Figure 2C). These results indicate that the C-terminal 42-residue segment of ScRev7 is critical for interaction with the MRX subunits.

Figure 2.

However, it remained possible that the inability of cells expressing the C-terminally truncated variants of ScRev7 (with appropriate prey proteins) to grow in selection medium (Figure 2C, bottom panel) may be due to altered expression or decreased abundance of truncated species. To explore this possibility, whole cell lysates of cells expressing the N- and C-terminally truncated variants, tagged with c-Myc epitope at the N-terminus, were harvested and analysed by immunoblotting using an anti-c-Myc antibody. Reassuringly, similar amounts of N- and C-terminally truncated species of were detected in the whole cell lysates of strains that were employed for Y2H analyses (Figure 2—figure supplement 1).

The C-terminal 42-residue fragment of ScRev7 mitigates the G-quadruplex-HU induced toxic effects

Several studies have documented the genome-wide prevalence of G-quadruplex DNA structures in various organisms ranging from viruses to humans, which play regulatory roles in diverse cellular processes (Rhodes and Lipps, 2015; Spiegel et al., 2020; Yadav et al., 2021). For instance, it has been shown that Rev1-deficient chicken DT40 cells exhibit defects in replicating G quadruplex-forming motifs (Sarkies et al., 2010, 2012). Similarly, computational and genetic studies in S. cerevisiae have revealed that G-quadruplex-forming motifs cause slow growth in replication stressed Pif1 deficient cells and affect genome integrity (Capra et al., 2010; Paeschke et al., 2011, 2013). Inspired by these findings, we asked whether ScRev7 plays a role in genome maintenance using the assay developed by the V. Zakian lab (Paeschke et al., 2011). To investigate these questions, we first assessed the growth of WT and rev7Δ mutant cells under normal growth conditions and under conditions of HU-induced replication stress in the absence of G-quadruplex forming sequences. This analysis revealed that, like the WT, rev7Δ mutant cells grew robustly in the absence or presence of HU (Figure 3A).

Figure 3.

Next, we asked whether the G-quadruplex DNA motifs affect cell viability in the absence or presence of DNA damaging agents, such as HU. To this aim, we tested the effect of G-quadruplex forming sequences, positioned on either the leading or lagging template strands, on the growth of WT, rev7Δ mutant and cells expressing the ScRev7 variants (Rev7-C1 or 42-residue C-terminal segment) in the absence or presence of HU. We found that all the tested strains showed robust growth in the absence of HU (Figure 3B). On the other hand, while HU had no effect on the growth of WT cells, rev7Δ mutant was extremely sensitive to HU and this was also the case with cells expressing the variant Rev7-C1 (Figure 3C). Gratifyingly, however, cells expressing the 42-residue C-terminal segment of Rev7 exhibited robust growth as the WT in the presence of HU (Figure 3C). These results suggest that the 42-residue C-terminal segment, but not Rev7-C1 variant, can protect cells from the toxic effects of HU and G-quadruplex DNA forming sequences, regardless of whether they are in the leading or lagging template strands.

Rev7 exhibits high-affinity for G-quadruplex DNA structures

Despite extensive cellular and structural characterization of ScRev7 (Setiaputra and Durocher, 2019; Clairmont and D’Andrea, 2021), it is unknown whether it can recognize and bind to specific DNA sequences or has any enzymatic activity. In this regard, our results revealed that REV7 alleviates G-quadruplex DNA synergised, HU-induced toxicity, raising the possibility that ScRev7 may interact with G-quadruplex DNA structures. To test this premise, ScRev7 and ScRev7-C1 were expressed in and purified from E. coli whole cell lysates to homogeneity (Figure 4A and 4B). Electrophoretic mobility shift assays (EMSA) were performed under physiological conditions to test for their ability to bind a suite of ³²P-labeled DNA substrates. In a protein titration experiment, we found that ScRev7 bound effectively to the G-quadruplex DNA substrates (TP-G4, 6G3-G4 and 4G3-G4), yielding >90% protein-DNA complexes at the protein concentrations used (Figure 4C—figure supplement 1A-C). By contrast, it exhibited no detectable binding to mixed-sequence ssDNA or dsDNA (Figure 4— figure supplement 1D-E). Furthermore, the mobility of ScRev7-DNA complexes was progressively decreased with increasing ScRev7 concentrations, suggesting cooperative binding of ScRev7 to G-quadruplex structures. As expected, the binding was lost when the G4 DNA capable sequences were mutated (Figure 4C and 4D—figure supplement 1F). Based on the bound and the unbound DNA from multiple EMSA experiments, the apparent binding affinities (K_d) of ScRev7 for various G4 DNA substrates was determined and found to be in the low nanomolar range (Figures 4B and 4C). Under identical conditions, ScRev7-C1 variant showed no detectable binding to G4 DNA (Figure 4—figure supplement 1G). Interestingly, we observed that the affinity (K_d) of ScRev7 increases with increasing number of contiguous guanine residues (i.e., TP-G4 > 6G3-G4 > 4G3-G4; Figure 4D). The apparent K_d values obtained for ScRev7 are comparable to those reported for several G4 DNA-binding proteins using similar substrates (Rhodes and Lipps, 2015; Spiegel et al., 2020; Yadav et al., 2021). These results indicate that ScRev7 binds to G-quadruplex structures and thereby might counteract genome instability induced by these structures. Additionally, these data suggest that ScRev7 is a novel G4 DNA-binding protein.

Figure 4.

Rev7 inhibits Mre11 nucleolytic activities

From analysis of the history of trimeric MRX complex, it is clear that Mre11 is a bifunctional enzyme that binds dsDNA ends and displays both endo- and exonuclease activities, which are critical for DNA end resection (Ghosal and Muniyappa, 2007; Ghodke and Muniyappa, 2013; Paull and Gellert, 1998; Tsubouchi and Ogawa, 1998; Stracker and Petrini, 2011; Paull, 2018; Casari et al., 2019). Further, it has been demonstrated that Mre11 cleaves non-B DNA structures such as DNA hairpins, cruciforms and intra- and inter-molecular G-quadruplex structures (Ghosal and Muniyappa, 2007; Trujillo and Sung, 2001; Lobachev et al., 2002; Ghosal and Muniyappa, 2005). To explore the functional significance of Mre11-Rev7 complex formation, an assay was performed wherein 5’-³²P-labelled dsDNA and MRX subunits (in equimolar amounts) were incubated in the absence or presence of ScRev7. The reaction products were analyzed as previously described (Ghosal and Muniyappa, 2005). The results informed that: (a) ScRev7 has no nuclease activity (Figure 5A, lane 2) and (b) Mre11 cleaved ³²P-labelled dsDNA, resulting in a ladder-like pattern of DNA products (Figure 5A, lane 3). We then examined the effect of ScRev7 on MRX nuclease activity by co-incubating various concentrations of ScRev7 and fixed amounts of MRX subunits, prior to the addition of 5’-³²P-labelled dsDNA. The results revealed that ScRev7 inhibited the MRX cleavage activity in a dose-dependent manner and almost complete inhibition was observed at 2 μM (Figure 5A, lanes 4-12). Interestingly, somewhat similar results were obtained with ScRev7-C1, although its inhibitory effect was noticeably less efficient than that of full-length ScRev7. Based on these data, we surmise that ScRev7-C1 does not faithfully recapitulate the Y2H results. Although the reason for this remains unclear, one possible reason for these discrepant results might be due to the interaction of amino acid residues in the ScRev7-C1 with Mre11, consistent with AlphaFold-Multimer modeling of the ScRev7-Mre11 heterodimer structure.

Figure 5.

Many studies have demonstrated that Mre11 exhibits both endonuclease and exonuclease activities independently of Rad50 and Xrs2 subunits (Cejka and Symington, 2021). Thus, we next examined the effect of ScRev7 on the Mre11 ssDNA-specific endonuclease activity in reactions lacking Rad50 and Xrs2. Consistent with previous reports, we observed Mre11 digested all the input substrate into small fragments/nucleotides in a manner dependent on its concentration (Figure 5C). Interestingly, we found that the addition of increasing concentrations of ScRev7 coincided with a concomitant decrease in the Mre11 endonuclease activity (Figure 5D), although it retained ssDNA nicking activity generating liner ssDNA. Together, these results support the idea that inhibition of Mre11 endonuclease activity is due to its direct interaction with ScRev7, as the latter is devoid of ssDNA and dsDNA binding activities (Figure 4).

ScRev7 impedes Rad50’s ATPase activity without affecting its ATP binding ability

Several studies have demonstrated that the ATPase activity of Rad50 plays an important regulatory role in DNA recombination/repair functions (Cejka and Symington, 2021). Intrigued by the direct interaction of ScRev7 with ScMre11 and Rad50 subunits, we asked whether ScRev7 affects the ability of Rad50 to bind ATP and catalyze its hydrolysis. To test this possibility, different concentrations of purified Rad50 were incubated with [γ-³²P]ATP. Because this interaction is transient, the reaction mixtures were irradiated with UV light prior to subjecting the samples to SDS/PAGE. In this analysis, we found a single band that migrated as a 153 kDa species, which corresponds to the molecular mass of ScRad50 (Figure 6A). Quantitative analyses indicated that Rad50 binds [γ-³²P]ATP in a concentration dependent manner (Figure 6B). Parallel experiments uncovered no significant differences in the extent of [γ-³²P]ATP bound by Rad50 in the presence or absence of ScRev7 (Figure 6C). This finding suggests that ScRev7 does not affect the ability of Rad50 to bind ATP (Figure 6D).

Figure 6.

We next evaluated whether ScRev7 affects the ability of Rad50 to catalyze ATP hydrolysis using thin layer chromatography coupled with an image analysis. Consistent with previous work (Ghosal and Muniyappa, 2007), control reactions showed that Rad50 catalyzed [γ-³²P]ATP hydrolysis to ADP and ³²Pi, which increased in a manner dependent on its concentration (Figures 6E and F). Interestingly, while ScRev7 itself has no ATPase activity, it suppressed ATP hydrolysis of Rad50 in a dose-dependent manner (Figures 6G and I). Similar analysis showed that the ScRev7-C1 blocked ATP hydrolysis by Rad50, although three-fold less efficiently compared with full-length ScRev7 (Figures 6H and I). As noted above, a possible explanation for this observation is that amino acid residues in the ScRev7-C1 may interact with Rad50, consistent with AlphaFold-Multimer modeling of the ScRev7-Rad50 heterodimer structure. These results were validated using colorimetric molybdate/malachite green-based assay (Lanzetta et al., 1979). While ScRev7 blocked approximately 60% of total ATPase activity of Rad50 at the highest dose tested, Rev7-C1 exerted partial inhibitory effect in the same range of concentrations (Figure 6—figure supplement 1). Collectively, these data support a model wherein the effect of ScRev7 extends beyond being a passive binder of MRX subunits, but negatively affects their catalytic activities.

REV7 facilitates NHEJ in S. cerevisiae

In S. cerevisiae, the heterotrimeric MRX complex has been implicated in both Ku-dependent NHEJ and microhomology-mediated end joining repair (Zhang and Paull, 2005; Boulton and Jackson, 1998; Moore and Haber, 1996; Ma et al., 2003). Multiple studies in cancer cell lines have shown that Rev7 inhibits DNA end-resection and favors NHEJ over HR (Gupta et el., 2018; Ghezraoui et al., 2018; Dev et al., 2018; Gao et al., 2018; Liang et al., 1972). To our knowledge, it is unknown whether these findings have been extended to other species. However, Schizosaccharomyces pombe Rev7 has been shown to inhibit long-range resection at DSBs (Leland et al., 2018). Furthermore, it is unclear whether the S. cerevisiae Rev1, Rev3 and Rev7 subunits are required for NHEJ. To investigate this, the efficiency of plasmid-based NHEJ repair was analyzed by transforming BamHI-linearized plasmid pRS416, which has no homology with the genomic DNA, into the WT and isogenic mutant strains as described (Boulton and Jackson, 1998; Moreau et al., 1999) (Figure 7A). Consistent with a previous study (Boulton and Jackson, 1998), we observed that NHEJ was undetectable in the mre11Δ mutant (Figure 7B, Supplementary Table S3). Notably, while the rev1Δ and rev3Δ strains showed a modest decrease in the efficiency of NHEJ, rev7Δ single and rev1Δ rev3Δ double mutants exhibited approximately four- and twofold decrease, respectively, compared with the WT. Further analysis revealed that the NHEJ efficiency in double mutants - rev1Δ rev7Δ, rev3Δ rev7Δ and sae2Δ rev7Δ - was comparable to that of rev7Δ mutant (Figure 7B and 7C —Supplementary Table S3). Intriguingly, however, we found that ScRev7-42 aa peptide, but not the Rev7-C1 variant, fully restored NHEJ efficiency in rev7Δ cells to the WT levels. Together, these data indicate that Rev7 augments the efficiency to NHEJ-mediated DSB repair in S. cerevisiae.

Figure 7.

Next, we sought to verify our plasmid-based NHEJ results in a chromosomal context. To this aim, the NHEJ efficiency was determined using “suicide-deletion” assay in which the I-SceI induced DSB may be repaired preferentially via NHEJ (Karathanasis and Wilson, 2002). Briefly, it is based on an approach wherein galactose-induced I-SceI endonuclease creates a pair of site-specific DSBs, resulting in the deletion of its own coding region, thereby facilitating NHEJ-mediated repair (Figure 8A). Using this approach, we determined the frequency of Ade2+ recombinants in the WT and isogenic mutant strains. A critical NHEJ factor Ku70 was used as a control. As expected, while NHEJ was undetectable in strains lacking MRE11 and KU70, deletion of REV7, REV1 and REV3 led to a 9-fold, 2.3-fold- and 3.2-fold decrease, respectively, in the frequency of NHEJ compared with the WT. Of note, a fourteen-fold decrease in the frequency of Ade2+ recombinants were observed in the rev7-C1 cells, which could be restored to WT levels by expressing the C-terminal 42 amino acid peptide (Figure 8B, Supplementary Table S4). The results reinforce the notion Rev7 promotes NHEJ repair at DSBs. The PCR product derived from genomic DNA of Ade2+ recombinants showed a 1.3 kb amplicon, suggesting faithful repair of I-SceI induced DSB (Figure 8C).

Figure 8.

REV7 suppresses HR in S. cerevisiae

As noted above, the RIF1/REV7/Shieldin complex blocks DNA end resection and BRCA1-mediated HR, but promotes DSB repair through NHEJ in cancer cell lines (Gómez-Llorente et al., 2013; Malik et al., 2020; Du Truong et al., 2021; Xu et al., 2015; Boersma et al., 2015; Mirman et al., 2018; Findlay et al., 2018; Gupta et el., 2018; Ghezraoui et al., 2018; Dev et al., 2018; Tomida et al., 2018; Gao et al., 2018; Noordermeer et al., 2018; Liang et al., 2020); however, the generality of this finding remained largely uncharacterized. To this end, we leveraged a spot assay (Paeschke et al., 2013) to assess the putative role of S. cerevisiae REV7 in HR. In this assay, we examined the recombination frequency, under replication stress and standard growth conditions, between the ura3-1 allele located at its endogenous location on chromosome V and ura3-G4 allele on the pFAT10-G4 plasmid (ura3 allele interrupted by G4 motifs) (Figure 9A). While Ura3⁺ papillae formation was not observed in the ura3-1 strain carrying the empty vector in both selective and non-selective media, the same strain with a plasmid containing G-quadruplex capable sequences exhibited Ura3⁺ papillae formation, albeit fewer, only under non-selection conditions. Of note, we found that the ura3-1 rev7Δ strain carrying a plasmid with G-quadruplex motifs displayed a significant number of Ura3⁺ papillae on non-selective and selective media, indicating that REV7 suppresses HR frequency (Figure 9B). Consistent with a previous report (Cejka and Symington, 2021), the nuclease-deficient mutant mre11-D56N,H125N alone (used as a positive control) and in combination with rev7Δ had no discernible effect on Ura3+ papillae formation on selective medium (Figure 9B). As expected, neither ScRev7-42 residue peptide nor Rev7-C1 altered the frequency of Ura3+ papillae formation on selective medium (Figure 9B, bottom two rows). Quantification revealed that the ura3-1 rev7Δ double mutant showed a 2.4-fold increase in HR frequency as compared to ura3-1 and ura3-1 mre11-D56N,H125N mutant strains (Figure 9C). Since all tested strains formed Ura+ papillae in the absence of HU approximately at a similar frequency, they are likely to arise due to spontaneous DSBs and collapsed replication forks.

Figure 9.

To confirm whether the G-quadruplex motifs stimulate HR in the rev7Δ strain, the frequency of HR was determined using pFAT10-G4 -mut plasmid with mutated G4 sequences. The results revealed that the rate of HR frequency decreased by 5.8-fold compared to strain carrying plasmid pFAT10-G4 -mut with unmutated sequences, indicating that this decrease is due to mutation of G-quadruplex–forming sequences (Figure 9—figure supplement 1). To assess whether increased frequency of HR is due to the instability of G-quadruplex DNA in rev7Δ cells, we examined the length of G4 DNA inserts in the plasmids carrying sequences during HR assay. We found a DNA fragment of approximately 829 bp corresponding to the expected size in both WT and rev7Δ cells (Figure 9—figure supplement 2). It is therefore possible that increased frequency of HR in rev7Δ cells could be due to the loss of Rev7 function rather than instability G4 DNA capable sequences. Viewed together, these findings are consistent with the idea that REV7 completely abolishes DSB-induced HR in S. cerevisiae.

Discussion

In this study, genetic and biochemical analyses of the wild-type ScRev7 and its 42 residue variant have revealed unanticipated insights into the mechanism by which Rev7 regulates the DSB repair pathway choice between HR and NHEJ in S. cerevisiae. Specifically, we show that ScRev7 directly interacts with the individual subunits of the MRX complex via a 42-residue C-terminal fragment (residues 203-245), protects cells form G4 DNA-HU induced toxicity, facilitates NHEJ while blocking HR, thereby suggesting that this minimal region harbors structural principles required for Rev7 function in vivo. Additionally, we found that ScRev7 impedes Mre11 nuclease and Rad50’s ATPase activities, without affecting the latter’s ATP-binding capacity. We further show that ScRev7 binds with significantly high affinity (K_d in the range of 3.22 to 18.26 nM) to G-quadruplex structures, as opposed to no binding to mixed-sequence ssDNA and dsDNA. When seen from a teleological perspective, these findings are conceptually reminiscent of Shieldin complex-mediated suppression of 5′ end resection and repair of DSBs via NHEJ while antagonizing HR in cancer cell lines (Xu et al., 2015; Boersma et al., 2015). At this stage, we tentatively speculate that this alternative mechanism of regulation of DSB repair pathway choice in S. cerevisiae may be conserved across multiple species.

Historically, REV7 was discovered as playing an important role in DNA damage-induced mutagenesis in S. cerevisiae (Lemontt, 1971; Lawrence and Das, 1985). Further analyses showed that Rev7 associates with Rev3 and functions as a regulatory subunit of eukaryotic TLS DNA polymerase Polζ (Prakash et al., 2005; Maiorano et al., 2021; Ling et al., 2022; Paniagua and Jacobs, 2023). Of note, Rev7 has no known catalytic activity, but serves as a versatile scaffolding protein: the versatility of Rev7 is underpinned by its unique ability to function in diverse cellular processes, including TLS, DNA repair and cell cycle that were initially considered inconceivable (de Krijger et al., 2021; Decottignies, 2013). However, little is known about the nature of specific effector(s) that associate with Rev7 and regulate DSB repair in S. cerevisiae. As we discuss below, the Y2H assays showed unequivocally that ScRev7 interacts with the individual subunits of MRX complex. Consistent with this, MST analysis and AlphaFold-Multimer models revealed tight association between the MRX subunits and ScRev7. On the other hand, while co-immunoprecipitation experiments implicated that Mre11, Rad50 and Xrs2 subunits exist as a hetero-trimeric complex in vivo (Usui et al., 1998), in vitro experiments have shown the formation of dimeric Mre11-Rad50 and trimeric MRX complexes (Oh et al., 2016; Arora et al., 2017). Additionally, to our knowledge, the relative amounts of monomeric, dimeric and trimeric species of MRX subunits in vivo are unknown. Relatedly, current models suggest that Mre11 and Rad50 bind to DNA in Rad50/Mre11-dependent and -independent manner (Stracker and Petrini, 2011; Paull, 2018; Casari et al., 2019). This leads us to postulate that the monomeric, dimeric and trimeric species of MRX subunits may exist in a dynamic equilibrium under in vivo conditions.

Given that Rev7 and MRX subunits play essential roles in DNA repair pathways, our findings provide intriguing insights into how ScRev7 interacts with the MRX subunits. Structure-function analysis showed that deletion of 42-amino acid fragment (203-245) at the extreme C-terminus of ScRev7 abolished its binding to the MRX subunits, whereas loss of the N-terminal HORMA domain, an evolutionarily conserved protein-protein interaction module (de Krijger et al., 2021; Muniyappa et al., 2014; Rosenberg and Corbett, 2015) had no discernible effect on its interaction. Curiously, further analysis indicated that the 42-residue peptide alone was sufficient for its binding to the MRX subunits and regulation of DSB repair pathway choice between NHEJ and HR. Analogously, it will be of interest to determine the regions/domains of MRX subunits that interact with ScRev7 for a comprehensive understanding of the crosstalk between these components. Work is currently in progress to gather insights into these questions.

The data from MST-based assays informed that ScRev7 binds to MRX subunits with significant affinity, analogous to the interaction between HORMA-domain protein hMAD2 and hCDC20 (Piano et al., 2021). How might ScRev7 suppress the function of Mre11 and Rad50 subunits? Encouraged by the above-mentioned results, we tested the effect of ScRev7 on the biochemical activities of MRX subunits. Reassuringly, we found that nanomolar amounts of ScRev7 besides adversely affecting the Mre11 nuclease activity also impeded Rad50’s ATPase activity without obstructing its ability to bind ATP. Whilst the specific nuances of the interaction await further characterization, we surmise that direct association between ScRev7 and the Mre11/Rad50 subunits might contribute to the observed effects. Although we provide compelling in vivo evidence that the Rev7 C-terminal 42-residue peptide is critical for binding to the MRX subunits, rescue cells from G4 DNA-HU-induced toxicity and regulate the pathway choice between NHEJ and HR, we could not demonstrate its function in vitro because our multiple attempts to express and purify the peptide were unsuccessful. Unexpectedly, however, we found that the Rev7-C1 variant, which lacks the ability to interact with the MRX subunits in Y2H assays, partially blocked the catalytic activities of Mre11 and Rad50 subunits in vitro. Why this is the case is not clear, but one possibility is that amino acid residues outside the 42-residue fragment facilitate pairwise interactions between ScRev7 and the Mre11 and Rad50 subunits, as seen in AlphaFold-Multimer models. Conversely, the 42-amino acid fragment as a part of whole protein could potentially block the residues in the Rev7-C1 fragment, thereby enabling it to function effectively as a single site for binding to the MRX subunits in Y2H assays. Future studies are required to test this hypothesis.

Many lines of evidence indicate that G-quadruplex capable sequences are measurably enriched at certain functional regions in the genomes of all organisms, from humans to plants to microbes (Capra et al., 2010; Huppert and Balasubramanian, 2005; Lejault et al., 2021; Castillo Bosch et al., 2014). Furthermore, G-quadruplex structures modulate diverse cellular processes, including DNA replication, transcription, translation, and are associated with certain diseases, characterized by high rates of chromosomal instability (Rhodes and Lipps, 2015; Spiegel et al., 2020). Serendipitously, we found that deletion of REV7 resulted in G4 DNA-associated genomic instability during DNA replication, rendering cells hypersensitive to HU-induced genotoxic stress and cell death. This is consistent with accumulating evidence that DNA replication stress induced by G-quadruplex structures play a prominent role in triggering genomic instability, which is exacerbated in the presence of HU (Sato and Knipscheer, 2023). To this end, G-quadruplex DNA-specific binding properties of ScRev7 defined here enabled us to provide a possible explanation for its role in protecting cells against G-quadruplex-induced hypersensitivity to HU. Curiously, it was found that Rev7 binds with high affinity to G-quadruplex structures, as opposed to ssDNA or dsDNA containing random sequences. However, we cannot exclude the possibility that additional, less abundant, preferred DNA substrates may exist for Rev7 in S. cerevisiae cells. Based on published literature and our studies, it is tempting to speculate that ScRev7 may recruit G4-resolving helicase(s) such as Sgs1 and Pif1 to unwind G-quadruplex structures prior to or during DNA replication (Huber et al., 2002, Paeschke et al., 2013). Alternatively, or in addition, it may generate G4 DNA intermediates which may be processed by specific enzymes such as Mus81 or Sgs1/Top3. Regardless, our data align with the notion that G-quadruplex structures are endogenous sources of replication stress and their formation and persistence may lead to genomic instability and cell death.

Several studies have shown that Shieldin complex facilitates NHEJ-dependent DSB repair, while inhibiting DNA end resection and HR in cancer cell lines (Clairmont and D’Andrea, 2021; Paniagua and Jacobs, 2023). Since 5′ end resection is the primary step in HR-mediated DSB repair, we hypothesized that direct interaction between ScRev7 and MRX subunits might block resection of DNA termini, and then facilitate NHEJ-dependent DSB repair instead of HR. Consistent with this, we found that ScRev7 augments the efficiency of NHEJ while antagonizing HR. However, other regulatory components may also play a role in modulating the levels of NHEJ versus HR-mediated DSB repair. For instance, TRIP13 or p31^comet inhibits the interaction of Rev7 with the SHLD3 subunit and regulates DNA end resection at DSBs and promotes their repair by HR (Sarangi et al., 2020). Additionally, interaction between Rev7 and putative binding partners may be regulated by posttranslational modifications and chromatin accessibility under changing physiological conditions.

Although a comparison of the mechanism of Rev7-mediated regulation of DSB repair between the yeast and cancer cell lines indicate broad similarities, we observed some notable differences. A fundamental difference between the results from cancer cell lines and S. cerevisiae is that, while ScRev7 robustly interacts with and suppresses the biochemical activities of both Mre11 and Rad50 subunits to facilitate NHEJ, Rev7-Shieldin complex acts as a downstream effector of 53BP1-RIF1 in restraining DNA end resection to promote NHEJ (Clairmont and D’Andrea, 2021; Paniagua and Jacobs, 2023). Furthermore, hRev7 interacts with the SHLD3 subunit via the HORMA domain, which is entirely dispensable in the case of ScRev7. Interpreted broadly, the highly complex, Shieldin complex-mediated regulation of DSB repair pathway choice might be a source of evolutionary innovation as an additional layer of regulation. In summary, our data bridges a notable gap in the current understanding of the molecular mechanism underpinning ScRev7-mediated regulation of DSB repair pathway choice in S. cerevisiae.

Materials and methods

S. cerevisiae strains, DNA plasmids and oligonucleotides

All strains and primers used for the construction of DNA substrates used in this study are listed in Supplementary Table S5 and S6, respectively. The plasmids FT10/Chr.IV_G4_lg and FT10/Chr.IV_G4_le were a kind gift from Dr. Virginia Zakian.

Construction of strains used in the study

S. cerevisiae W1588a haploid strains carrying rev1Δ, rev3Δ, rev7Δ, rev1Δ rev3Δ, rev1Δ rev7Δ, rev3Δ, rev7Δ mre11Δ or sae2Δ rev7Δ single or double mutations were constructed using appropriate pairs of ODN primers as described (Sambrook and Russell, 2001; Janke et al., 2004). Briefly, REV7 gene was deleted using the KanMX4 (pFA6a-KanMX4) cassette with forward primer OSB11 and reverse primer OSB12. The loss of deleted gene was confirmed by PCR amplification using OSB13 and OSB14 primers. Similarly, other strains were constructed as follows: MRE11 was deleted using the hphNT1 cassette (pYM-hphNT1) along with OSB52 and OSB54 primers; the loss of deleted gene was verified using the OSB53 probe. REV1 was deleted using the hphNT1 cassette along with OSB55 and OSB56 primers; the loss of deleted gene was confirmed using the OSB57 probe. REV3 was deleted using KanMX4 cassette (derived from pFA6a-hphNT1, pFA6a-KanMX4 respectively) along with OSB58 and OSB59 primers; the loss of deleted gene was confirmed using the OSB60 probe. The hphNT1 (pYM-hphNT1) cassette was used in the generation of double mutants (listed in Supplementary Table S6). The rev7-C1 and rev7-42 mutant strains were constructed by incorporating the sequences encoding the truncated species at the endogenous locus by using rev7-C1-9MYC-hphNT1 and rev7-42-3MYC-KANMX4 cassettes through overlapping primer-based PCR. The rev7-C1-9MYC-hphNT1 cassette was generated using primer pairs OSB116, OSB117 and OSB118, OSB70. Likewise, rev7-42-3MYC-KANMX4 cassette was generated using primer pairs OSB122, OSB120 and OSB121, OSB70. The N- and C-terminal truncation variants were confirmed using the probes OSB13 and OSB14, respectively.

Similarly, mre11Δ, rev1Δ, rev3Δ, and rev7Δ single knockout mutants were generated in the strain YW714 using hphNT1 cassette (derived from pFA6a-hphNT1). The REV7 truncation variants - rev7-C1-9MYC and rev7-42 aa-3MYC - were generated as described above. REV3 was deleted using hphNT1 cassette in strain PJ694A using OSB58 and OSB59 primers, and deletion was confirmed using primer OSB60 as the probe. Likewise, mre11Δ rad50Δ xrs2Δ triple knockout mutant was generated in strain PJ694A as follows: mre11Δ::KANMX4 was generated by PCR-amplification of KANMX4 cassette (pFA6a-KANMX4) using OSB52 and OSB53 as forward and reverse primers, respectively, and deletion was confirmed using OSB54 as the probe. The rad50Δ::URA3 strain was constructed by PCR amplification of URA3 cassette (pAG60-URA3) using primers OSB133 and OSB75, and deletion was confirmed using OSB76 as the probe. Similarly, xrs2Δ::HphNT1 was generated by PCR amplification of hphNT1 cassette (pFA6a-hphNT1) using primers OSB134 and OSB78, and deletion was confirmed using OSB79 as the probe.

Construction of DNA plasmids for expression and recombination assays

The plasmids FT10/Chr.IV_G4_lg and FT10/Chr.IV_G4_le were a kind gift from Dr. Virginia Zakian. S. cerevisiae REV7 gene was amplified from genomic DNA by PCR using the primer pair (forward OSB01 and reverse OSB02) and Phusion DNA polymerase (NEB, Ipswich, MA) as described (Sambrook and Russell, 2001). This reaction yielded an amplicon of expected size, which was digested with BamHI/HindIII and ligated into a similarly digested pET28a(+) vector (Novagen) by utilizing the T4 DNA ligase. The resulting expression plasmid was designated pET-28a_REV7. In parallel, pET28a_ScREV7-C1 truncation was generated by PCR amplification of 609 bp (corresponding to the ScRev7-C1 truncation) using primers OSB01 and OSB125 using pET28a_ScREV7 plasmid DNA as template. The PCR product was digested with BamHI/HindIII and ligated into BamHI/HindIII digested pET28a vector. Likewise, S. cerevisiae RAD50 gene was PCR-amplified from a genomic DNA preparation using OSB33 forward and OSB34 reverse primers. The amplicon was digested with BamHI/Xho1 and ligated into a similarly digested pE-SUMO Kan vector (Life Sensors, Malvern, PA) by utilizing the T4 DNA ligase. The resulting expression plasmid was designated pE-SUMO_RAD50. The REV7-eGFP expression vector was constructed by PCR amplification of the DNA fragment encoding the S. cerevisiae REV7 gene in the pET-28a_REV7 plasmid using forward Rev7-eGFP primer and reverse Rev7-eGFP primer (Supplementary Table S6). The amplicon of expected size was digested with XbaI/XhoI and ligated into a similarly digested pPROEX vector, upstream of eGFP coding sequence. The resulting plasmid encodes ScRev7-eGFP fusion protein. The pPROEX::eGFP vector was a kind gift from Deepak Saini.

The ura3-G4 insert region was amplified by overlap extension PCR method using ODNs OSB80 and OSB82 as forward and OSB81 and OSB83 as reverse primers, respectively. The amplicon of expected size was digested with BamHI/SphI and ligated into a BamHI/Sph1 digested FAT10 vector. The ODN OSB82 corresponds to the DNA sequence 362751-362775 corresponding to the coding strand of S. cerevisiae Chr X and OSB81 is its complementary strand. Similarly, ura3-G4 mutant was generated using ODNs OSB80 and OSB130 as forward, and OSB129 and OSB83 as reverse primers. The full-length ura3-G4 mutant PCR product was digested with BamHI/Sph1 and cloned into FAT10 vector. The ODNs OSB129 DNA sequence 362751-362775 corresponds to mutant version of S. cerevisiae Chr X and OSB130 is its complementary strand.

Construction of DNA plasmids for yeast two-hybrid analysis

To construct the pGBKT7-REV7 the S. cerevisiae REV7 orf in the pET28a (+) REV7 construct was PCR-amplified by using Phusion DNA polymerase and forward OSB03 and reverse OSB04 primers, respectively. The amplicon of expected size was digested with Nde1/BamHI and ligated into a similarly digested pGBKT7 vector. The S. cerevisiae MRE11 gene from genomic DNA was PCR-amplified using forward OSB05 primer and reverse MRE11_RP primer. The amplicon of expected size was digested with BamHI/EcoRI and ligated into the pGADT7 prey expression vector at the BamHI/EcoRI site. The same procedure was leveraged for the construction of all other prey expression vectors: XRS2 was amplified from the genomic DNA using forward OSB36 and reverse OSB37 primers. The expected-size amplicon was digested with Nde1/BamHI and ligated into the pGADT7 prey expression vector at the Nde1/BamHI site; RAD50 was amplified from the pESUMO_RAD50 construct using forward OSB35 and reverse OSB34 primers. The expected-size amplicon was digested with EcoRI/Xho1 and ligated into the pGADT7 prey expression vector EcoRI/Xho1 site; REV7 was amplified from the pET-28a_REV7 construct using the same primers and ligated into pGADT7 prey expression vector as in the case of bait plasmid construction.

The N-terminally truncated ScREV7 variants (REV7-N1, REV7-N2, and REV7-N3) were constructed by PCR amplification of relevant portions of the REV7 gene using OSB61, OSB62, OSB63 as a forward primer and OSB4 as a reverse primer (common for all N-terminal deletions), respectively. The C-terminally truncated ScREV7 variants (REV7-C1, REV7-C2 and REV7-C3) were constructed via PCR amplification of relevant portions of the REV7 gene in the pET-28a_REV7 plasmid using ODN OSB3 as a forward primer (common for all C-terminal deletions) and ODNs OSB64, OSB65 and OSB66 as reverse primers, respectively. The amplicons of expected size, corresponding to the N- and C-terminal variants were digested with Nde1/BamHI and ligated into a similarly digested pGBKT7 vector. S. cerevisiae PJ694A strain was used in Y2H analyses.

Cell viability assay

S. cerevisiae wild type (WT) and isogenic mutant strains were grown in YPD or SC medium to an OD₆₀₀ of 0.5. At this stage, ten-fold serial dilutions were spotted onto YPD agar plates with or without various concentrations of HU. Similarly, cells were spotted onto SC selective medium plates lacking appropriate amino acids, but with or without inhibitors, as indicated in figures/figure legends. The plates were incubated at 30 °C for 3-4 days. The images were captured using epi-illumination at auto-exposure ChemiDoc MP imaging system (Bio-Rad).

Yeast two-hybrid interaction analysis

The Y2H assays were performed as previously described (Fields and Song, 1989; Thakur et al., 2020) Briefly, pairwise combination of plasmids expressing bait proteins, fused to the Gal4 DNA-binding domain (G4BD), and prey proteins, fused to the Gal4 activation domain (G4AD), were co-transformed into reporter WT strain PJ69-4A or its isogenic rev3Δ single or the mre11Δ rad50Δ xrs2Δ triple mutant strain. The empty prey vector and Rev7 expressing vector served as negative and positive controls respectively. The Y2H interactions were documented by spotting representative transformants onto SC/-Trp –Leu and SC/-Trp –Leu -His agar plates containing 15 mM 3-aminotrizole (3-AT), which were incubated for 3-5 days at 30°C. Growth of cells on SC-Leu-Trp-His +3-AT agar plates is indicative of moderate/strong protein-protein interactions between the bait and prey proteins.

Assay for G-quadruplex DNA/HU-induced toxicity

The assay was performed as previously described (Paeschke et al., 2011). The plasmid constructs used in this study were identical to the parent plasmid, pFAT10, except that its derivatives contain three tandem arrays of G sequences derived from chromosome IV, inserted in the lagging or leading templates. Briefly, S. cerevisiae wild type, rev7Δ mutant, and cells expressing ScRev7 truncation variants (ScRev7-C1, ScRev7-42 aa) harboring the pFAT10 plasmid were synchronized in the G1 phase using α-factor at 30 °C and then released from the pheromone block by washing the cells. Subsequently, the cultures were normalized to equivalent A₆₀₀ values, and tenfold serial dilutions of each culture was spotted onto solid SC/-Leu and SC/-Leu/+100 mM HU agar plates. The plates were incubated at 30 °C for 72 h.

Expression and purification of ScRev7

The S. cerevisiae MRX subunits were expressed and purified as previously described (Ghosal and Muniyappa, 2007). The S. cerevisiae REV7 gene was sub-cloned into pET28a(+) expression vector with an N-terminal His₆-tag. The resulting plasmid was designated pET28a(+)_REV7. E. coli BL21(DE3)pLysS cells harboring pET28a(+)_REV7 plasmid were cultured in Luria-Bertani broth (1% tryptone, 0.5% yeast extract, 1% NaCl, pH 7.0) containing 50 μg/ml kanamycin at 37 °C with shaking at 180 rpm to an OD₆₀₀ of 0.6. At that point, ScRev7 expression was induced by adding 1-thio-β-D-galactopyranoside (IPTG) to a final concentration of 0.1 mM. The IPTG-induced cultures were further incubated for 12 h at 25 °C. Cells were harvested by centrifugation at 6000 g the cell paste was resuspended in 50 ml of buffer A (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 10 % glycerol and 5 mM 2-mercaptoethanol containing 1 mM phenylmethylsulfonyl fluoride and 0.05 % Triton X-100). Cells were lysed by sonication on ice (7 x 1-min pulses) and quickly subjected to centrifugation at 30,000 rpm, 4°C for 20 min. Solid ammonium sulfate was added (0.472 gm/ml) to the supernatant with continuous stirring for 45 min at 24 °C. The precipitate was collected by centrifugation at 18000 rpm at 4°C for 1 h. The precipitate was dissolved in buffer A and dialyzed against the same buffer. The dialysate was loaded onto 5 ml Ni²⁺-NTA column (Qiagen, Valencia, CA). After washing the column with buffer A containing 20 mM imidazole, bound proteins were eluted with a gradient of 20→500 mM imidazole. The fractions containing ScRev7 were pooled and dialyzed against buffer B (20 mM Tris-HCl, pH 8.0, 1.2 M NaCl, 7 % glycerol, 5 mM 2-mercaptoethanol). It was further purified by chromatography on a Superdex S75 column attached to an AKTA Prime FPLC system, which had been equilibrated with and eluted using buffer B. The peak fractions containing ScRev7 were pooled and dialyzed against buffer C (20 mM Tris-HCl (pH 8.0), 30 mM NaCl, 30 % glycerol, 5 mM 2-mercaptoethanol), and loaded onto a heparin column (5 ml). The bound proteins were eluted with a gradient of 30→350 mM NaCl in buffer C. An aliquot from each eluted fraction was visualized by staining with Coomassie Brilliant Blue R250 after SDS-PAGE. The fractions that contained ScRev7 were pooled and dialyzed against buffer D (20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 10 % glycerol, 1 mM DTT), and stored at -80 °C.

Expression and purification of ScRev7-eGFP

The His₆-tagged ScRev7-eGFP fusion protein was expressed in and purified from whole cell lysates of E. coli BL21(DE3)pLysS host strain harboring the pPROEX/REV7-eGFP plasmid as described above. The His₆-tagged eGFP was purified from the cell lysates of E. coli BL21* (DE3) pLysS host strain harboring eGFP pPROEX expression plasmid by Ni^2+-NTA affinity chromatography. Briefly, the whole cell lysate was loaded onto the Ni^2+-NTA resin column (5 ml), which had been equilibrated with a buffer A (HEPES, pH 7.5, 50 mM NaCl and 10% glycerol). The column was washed with buffer A containing 70 mM imidazole. The bound proteins were eluted with a gradient of 70 → 800 mM imidazole in buffer A. An aliquot from each eluted fraction was analyzed by SDS-PAGE and visualized by staining with Coomassie Brilliant Blue R250. The protein concentrations were determined by the Bradford assay. The fractions that contained ScRev7-eGFP and eGFP were pooled, dialyzed against buffer A, and stored at -80 °C.

Expression and purification of S. cerevisiae Rev1

The GST-tagged Rev1 was purified as described previously with certain modifications (Johnson et al., 2006). Briefly, S. cerevisiae BJ5464 cells bearing plasmid pBJ842-Rev1-GST were selected on SC plates lacking leucine. Single colonies were inoculated into SC/-Leu liquid media (200 ml) containing 2% raffinose for 12-14 h at 30°C until they reached an optical density (OD) at 600 nm reached 0.1. The cultures were incubated at 250 rpm, 30°C, till the O.D_600nm is equal to 0.5. Following this, the cultures were centrifuged at 4000 rpm for 10 min and the pellet was washed thrice with MilliQ water. Cells were then resuspended into SC/-Leu media containing 2% galactose. The cultures were further incubated at 30 °C/250 rpm for 7 h. The cells were harvested by centrifugation at 4000 rpm for 10 min, resuspended in CBB buffer (50 mM Tris-HCl pH 7.5, 10% sucrose, 1 mM EDTA, 10 mM β-mercaptoethanol and 5 μg/mL of protease inhibitor cocktail). Cells were lysed using a FastPrep 24 homogenizer. Following lysis, the cell debris was separated by centrifugation at 10,000 rpm for 10 min. The supernatant was further cleared by centrifugation at 35000 rpm for 30 min. The GST-tagged protein was precipitated from the cell lysate using ammonium sulfate (0.208 g/ml), followed by centrifugation at 20,000 rpm for 45 min. The pellet was resuspended in the GST-binding buffer (GBB, 50 mM Tris-HCl pH 7.5, 10% glycerol, 1 mM EDTA, 10 mM β-mercaptoethanol and 5 μg/mL of protease inhibitor cocktail) and dialyzed extensively against the same buffer. The sample was then loaded onto the GST-column (GSTrap High performance column, 5 ml, Cytiva Life Sciences), loaded at the rate of 0.5 mL/min. The column was then washed with GBB buffer containing 500 mM NaCl, to remove non-specific interactors. Bound protein was eluted into 30 fractions (1 mL each) at a flow rate of 1 mL/min in GBB buffer containing 40 mM glutathione. The fractions were analysed on 10% SDS-PAGE, followed by Coomassie staining, and the fractions containing purified Rev1-GST protein were pooled and dialysed against MST buffer (HEPES, pH 7.5, 50 mM NaCl, 10% glycerol), and stored at -80°C. The concentration of ScRev7 and its derivatives was determined by the dye-binding assay.

SDS-PAGE and Immunoblot analysis

Conventional protocols were employed for SDS-PAGE and immunoblotting using a polyvinylidene difluoride (PVDF) membrane was performed as described (Mahmood and Yang, 2012). The S. cerevisiae PJ69-4A strains were co-transformed with empty vectors pGBKT7 and pGADT7, or a bait vector harboring Rev7 truncations in combination with prey vectors harboring REV7, MRE11, XRS2 or RAD50. The transformants were selected by plating onto SC/-Leu,-Trp medium, and single colonies were inoculated into 5 ml SC/ -Trp,-Leu media containing 2% dextrose. The cells were grown at 30°C for 12-14h at 250 rpm to OD_600nm = 0.15, which were then transferred onto 10 ml of SC/ -Trp,-Leu medium. Incubation was continued at 250 rpm/30°C to OD_600nm = 0.5 to 0.6. The cells were harvested by centrifugation at 4000 rpm for 5 min and the pellet was resuspended using lysis buffer (50 mM sodium-HEPES pH 7.5, 200 mM sodium acetate pH 7.5, 1 mM EDTA, 1 mM EGTA, 5 mM magnesium acetate and 5% glycerol). Cells were lysed with 2 min. bursts at 4.0 m/s on a FastPrep-24 homogenizer in the presence of 200 ml of acid-washed glass beads. The samples were centrifuged at 7000 rpm for 5 min to remove the beads and cell debris. The supernatant was further centrifuged at 13000 rpm for 10 min. Fifty micrograms of proteins were separated by10% SDS-PAGE. Then, proteins were transferred to 0.45 µm PVDF membranes, followed by blocking for 1 h at 25°C with 5% nonfat dry milk in TBST (20 mM Tris–HCl buffer, pH 7.5,150 mM NaCl, 0.1% Tween 20) buffer. The blot was rinsed with TBST twice and incubated with anti-myc antibodies (diution 1:3000 dilution) for 1 h at 25 °C. The blots were washed thrice with TBST buffer and incubated with HRP-conjugated anti-rabbit antibodies (dilution 1:20000; Sigma-Aldrich) at 25°C for 1 h. Following incubation, the blots were washed thrice with TBST buffer and developed using chemiluminescence substrates (Bio-Rad Laboratories, CA, USA) in the Bio-Rad ChemiDoc Imaging systems. Anti-PGK1 antibodies were obtained from Santa-Cruz Biotechnologies, CA, USA.

Structure prediction of heterodimer protein complexes

The sequences of full-length ScRev7, Mre11 and Rad50 were obtained from the Saccharomyces Genome Database (https://www.yeastgenome.org/). Structural models for the complexes between ScRev7-Mre11 and ScRev7-Rad50 were built using publicly available versions of AlphaFold-Multimer algorithm (Evans et al., 2022; Varadi et al., 2022) and ColabFold software (Mirdita et al., 2022) to run AlphaFold-Multimer (https://github.com/sokrypton/ColabFold). The models with the highest confidence were analysed using LigPLot software, Version 2.2, to detect the amino acid residues involved in the interactions between the Mre11/Rad50 subunits and ScRev7. Approximately, 44% and 43% of the residues across the binding interface of ScRev7 and Mre11 displayed confident pLDDT scores greater than 60. Similarly, 67% and 100% of the residues lining the ScRev7 and Rad50 interface exhibited pLDDT scores greater than 60. The PyMOL Molecular Graphics System, Version 2.5.5, was used for visualizing the structures. Of the 5 models generated by AlphaFold-Multimer, the top models (based on average pLDDT score) were chosen for display.

Microscale thermophoresis assay

MicroScale Thermophoresis (MST) assay was performed on a Monolith NT.115 instrument (NanoTemper Technologies GmbH) according to the manufacturer’s instructions. Samples were prepared in 20 μl MST buffer (HEPES, pH 7.5, 50 mM NaCl, 10% glycerol) and ScRev7-eGFP or eGFP and with varying concentrations of ligands (Mre11, Rad50 or Xrs2) in the following concentration range: Mre11 (0.00015 to 5 μM), Rad50 (0.00006 to 2 μM), Xrs2 (0.00015 to 5 μM) or Rev1 (0.00015 to 5 μM). After incubation at 37 °C for 15 min, samples were transferred into Monolith NT.115 glass capillaries and and MST measurements were performed using 40% MST power with laser on/off times of 30 s and 5 s. MST signals were normalised to fraction bound (X) by X= [Y(c)− Min]/(Max − Min), error bars (SD) were normalized by stdnorm = std(c)/(Max − Min). The Fnorm values were plotted against the ligand concentration, to obtain an estimate of binding affinity. The K_d values and Hill coefficients (n_H) were calculated using MO Affinity Analysis software (NanoTemper). All data were statistically analyzed and graphed using GraphPad Prism software (v5.0).

Preparation of radiolabeled DNA substrates

The sequences of and oligonucleotides ( ODNs) used for the preparation of substrates are listed in supplementary Table S7.The ODNs were labeled at the 5’ end using [γ-³²P]ATP and T4 polynucleotide kinase as described (Sambrook and Russell, 2001). The unincorporated [γ-³²P]ATP was removed using a Sephadex G-50 superfine mini-column. The ³²P-labeled G-quadruplexes were prepared by incubating the G-rich containing single-stranded ODNs (OSB28, OSB92 and ODN6G3) in annealing buffer [10 mM Tris-HCl (pH 8.0), 1 mM EDTA and 120 mM KCl] by heating at 95 °C for 2 min and then slowly cooling down to 24 °C. These ODNs contain arrays of G residues, which adopt thermally stable G4 topologies (Muniyappa et al., 2000). The dsDNA substrates were prepared as follows: 41-bp dsDNA by mixing aliquots of 5’-end ³²P-labeld OSB17 (upper strand) with a small excess of unlabeled complementary OSB20; 60-bp dsDNA by mixing OSB41 (upper strand) and OSB42 (lower strand) in 1X SSC buffer (0.3 M sodium citrate buffer, pH 7.0, containing 3 M NaCl), followed by heating at 95 °C for 5 min and then slowly cooling to 24 °C over a period of 90 min. The substrates were resolved by non-denaturing 8% PAGE for 4 h at 4 °C. Finally, the dsDNA substrates were eluted from the gel slices in TE buffer (10 mM Tris-HCl, pH 7.5, and 1 mM EDTA) and stored at 4°C for further use.

Electrophoretic mobility shift assay

The electrophoretic mobility shift assays (EMSA) were performed as previously described (Thakur et al., 2021). The reaction mixtures (20 μl), which contained 20 mM Tris-HCl (pH 7.5), 0.1 mM DTT, 0.2 μg/ml BSA and 5% glycerol), 0.5 nM of the specified ³²P-labeled DNA substrate and increasing concentrationsof ScRev7/ScRev7-C1 as indicated in the figure legends. After incubation at 37 °C for 90 min, 2 μl gel-loading solution (0.1% of bromophenol blue and xylene cyanol in 20% glycerol) was added to each sample. The samples were loaded onto a non-denaturing 4% polyacrylamide gel. Electrophoresis was performed in 45 mM Tris-borate buffer (pH 8.3) containing 10 mM KCl and 1 mM EDTA, initially at 150 V for 20 min, and then at 10 V/cm at 4°C for 3 h. The bands were visualized using a Fuji FLA-9000 phosphor imager analyzer and quantified using UVI-Band Map software (v. 97.04). The binding curves were submitted to a nonlinear regression curve analysis using the software GraphPad Prism software (version 5.0). The values were calculated by using the equation Y = Bmax X/(Kd + X).

AsCas12a exonuclease assay

The assay was performed as previously described (Arora et al., 2017) with slight modifications. Two steps are involved in the assay: first, assembly of the ScRev7-MRX complex and second, nuclease susceptibility assay. The MRX subunits (100 nM each) were mixed with increasing concentrations of ScRev7 (0.05 to 2.5 μM) or ScRev7-C1 (1.5 and 3 μM) and incubated on ice for 15 min, prior to the addition of 20 μl of reaction mixture that contained 25 mM MOPS (pH 7.0), 20 mM Tris-HCl (pH 7.5), 80 mM NaCl, 8% glycerol, 5 mM MnCl₂, 0.1 mM DTT, 200 μg/ml BSA and 10 nM of ³²P-labeled 60-bp dsDNA. Subsequently, the reaction mixtures were incubated at 37 °C for 1 h, and the reaction stopped by adding a 2 μl solution containing 2 mg/ml proteinase K, 50 mM EDTA and 1% SDS. After incubation at 37 °C for 30 min, 5 μl of a solution containing 10 μg/ml glycogen, 2 μl 3 M sodium acetate and 50 μl of absolute ethanol was added to each sample and were frozen at -80°C. The thawed samples were centrifuged at 15000 rpm at 4 °C for 30 min. The pellets were washed with 70% ethanol and centrifuged at 15000 rpm at 4 °C for 5 min. The dried pellets were resuspended in 10 μl gel-loading dye (80% formamide, 10 mM EDTA, 0.1% xylene cyanol and 0.1% bromophenol blue) and then incubated at 95 °C for 5 min. Aliquots were loaded onto an 8% denaturing polyacrylamide/7 M urea gel. Electrophoresis was performed in TBE buffer (89 mM Tris-borate (pH 8.3) and 1 mM EDTA) for 2 h at 40 W. The dried gel was visualized using a Fuji FLA 9000 phosphor imager.

AsCas12a endonuclease assay

The assay was carried out as previously described (Shibata et al., 2014) with slight modifications. Briefly, the reaction mixtures (10 μl) without ScRev7 contained 100 ng of M13 circular ssDNA, 30 mM Tris-HCl, pH 7.5, 25 mM KCl, 5 % glycerol, 1 mM DTT, 200 μg/ml BSA, 5 mM MnCl₂ and increasing concentrations of ScMre11 (0.1-2 μM). After incubation at 37 °C for 1 h, the reaction was quenched by adding 2 μl of stop solution (2 mg/ml proteinase K, 50 mM EDTA and 1 % SDS) and incubation was extended for 30 min. The reactions were quenched by adding 2 μl of gel-loading solution. The reaction products were separated by electrophoresis on a 0.8% native agarose gel using 44.5 mM Tris-borate buffer (pH 8.3) containing 0.5 mM EDTA at 10 V/ cm for 1 h. The gel was stained with ethidium bromide and the image was captured using the UVItec gel documentation system (UVItec, Cambridge, UK). To test the effect of ScRev7 on Mre11 endonuclease activity, the assay was carried out as described above, except that the ScRev7-Mre11 heterodimer was first formed by incubating with increasing concentration of ScRev7 on ice for 15 min, prior to the addition of 10 μl of reaction mixture. The reaction mixtures were incubated and processed as described above.

ATP crosslinking assay

The assay was carried out as previously described (Thakur et al., 2021). To test for the effect of ScRev7 on ATP binding by Rad50, we first ascertained its capacity to bind ATP. The assay was performed in a 20 μl reaction mixture containing 50 mM Tris-HCl (pH 5.0), 10 mM MgCl₂, 400 pmol [γ-³²P]ATP and increasing concentrations of Rad50 (0.1-1 μM). In parallel, a fixed concentration of Rad50 (1 μM) was incubated with increasing concentrations of ScRev7 (0.5 - 6 μM) at 4 °C for 15 min, prior to mixing with 20 μl buffer containing 50 mM Tris-HCl (pH 5.0), 10 mM MgCl₂ and 400 pmol [γ-³²P]ATP. After incubation at 4 °C for 25 min, samples were exposed to UV irradiation (1.2 × 10⁵ µJ/cm² in Hoefer UVC 500 ultraviolet crosslinker) at a distance of 2 cm. All cross-linking reactions were quenched by adding 5 μl of 5X Laemmli buffer (10 mM Tris-HCl, pH 6.8, 12.5% SDS, 40% glycerol, and 0.1% bromophenol blue). Samples were heated at 95 °C for 10 min and resolved by SDS/PAGE in 10% polyacrylamide gel at 35 mA for 2 h. The images were captured using a Fuji FLA-9000 phosphor imager and the band intensities of radiolabeled species was quantified using UVI-Band Map software (v. 97.04) and graphs were plotted using GraphPad Prism software (v5.0).

ATPase assay

The hydrolysis of ATP by ScRad50 was monitored by the release of ³²Pi from [γ-³²P]ATP as described (Thakur et al., 2021). Where indicated, a fixed concentration of Rad50 (0.5 μM) was incubated on ice for 30 min with increasing concentrations of ScRev7 or ScRev7-C1 to allow protein complex formation, prior to the reaction. Briefly, 200 μM [γ−³²P] ATP and Rad50 (0.5 μM) were incubated in 10 μl reaction mixture containing 20 mM Tris HCl (pH 7.5), 50 mM KCl, 0.2 mg/ ml BSA, 0.1 mM DTT, 1.0 mM MgCl₂ and 5 % glycerol, with indicated concentrations of Rad50. The reactions were carried out at 30 °C for 30 min and the reaction was quenched by adding 15 mM EDTA. Aliquots (2 μl) from each sample was spotted onto the PEI-cellulose thin-layer plate and was developed using a solution containing 0.5 M LiCl, 1 M formic acid and 1 mM EDTA as the mobile solvent. The TLC sheets were dried and the autoradiographs were visualized using Fuji FLA-9000 phosphor imager. The images were captured using a Fuji FLA-9000 phosphor imager and the band intensities of ³²Pi and [γ-³²P]ATP were quantified and plotted using UVI-Band map software. The isotherms were generated by non-linear regression analysis using GraphPad Prism (Version 5.0).

Malachite green phosphate assay

ATP hydrolysis was monitored by measuring the amount of inorganic phosphate released using acidic ammonium molybdate and malachite green assay (Lanzetta et al., 1979). The reaction mixtures (80 µl) contained 20 mM Tris-HCl, pH 7.5, 50 mM KCl, 5% glycerol, 0.1 mM DTT, 0.2 mg/ml BSA, 1 mM MgCl₂, 150 µM ATP, and increasing concentrations of ScRad50 (0.05 - 1 µM). After incubation at 30°C for 30 min, reaction was stopped by adding 20 µl of malachite green reagent (Sigma-Aldrich, MAK307). Incubation was continued for 15 min at 24°C to allow the formation of phosphomolybdate malachite green chromogenic complex. The absorbance values, as a measure of the extent of ATP hydrolysis, at 620 nm (y-axis) were plotted against increasing concentrations of ScRad50 (x-axis). To test the effect of ScRev7 or ScRev7-C1 on Rad50 ATPase activity, increasing concentrations of ScRev7 or ScRev7-C1 (0, 0.1, 0.3, 0.5, 1, 1.5, 2 µM) were incubated with 0.25 µM of ScRad50 on ice for 20 min, prior to mixing it with the reaction mixture. After incubation at 30 °C for 30 min, ATP hydrolysis was monitored as described above. The curves were fitted using non-linear regression analysis using GraphPad prism (v. 5.0).

Mass spectrometry analysis

Identity of Rev7-GFP protein was confirmed using Orbitrap mass spectrometry method (Zubarev and Makarov, 2013). In-gel trypsin digestion was performed as follows: 50 mM ammonium bicarbonate and acetonitrile (7:3) ratio was used for de-staining; following which, samples were reduced for 30 minutes by adding 10 mM DTT diluted in 50 mM ammonium bicarbonate. Following reduction, alkylation buffer (55 mM iodoacetamide in 50 mM ammonium bicarbonate) was added and samples were incubated at room temperature for further 30 min. Samples were treated with 10 ng trypsin enzyme (Mass spectrometry-grade, Sigma Aldrich, India), at 37°C for 12 h. Subsequently, peptides were extracted from the sample by desalting with 300 µL of 70% acetonitrile containing 0.1% trifluoracetic acid. The extracted peptides are dried in speed vacuum and resuspended in 40 µL resuspension buffer (2% acetonitrile in LC-MS grade Milli-Q). 10 µL of the sample was then injected into Orbitrap Fusion Tribid Mass spectrometer (Thermo Fisher Scientific Inc., USA). Samples were run for 110 minutes using Nanospray Ion source (NSI) and static spray voltage. Ions were detected by Orbitrap detector at 60000 resolution using quadrupole isolation. Daughter ions were detected by Ion Trap detector using quadrupole isolation mode. Following data acquisition, peptide spectrum matches (PSMs) and percentage peptide coverage were obtained for the samples. Collectively, these results confirm that the purified protein is GFP-tagged ScRev7.

Non-homologous end-joining assayed using linear plasmid substrate

The role of S. cerevisiae REV7 alone and in combination with various genes in NHEJ was assessed using a linear plasmid re-circularization assay as described (Zhang and Paull, 2005; Ghodke and Muniyappa, 2013). Briefly, 60 ng of BamHI-digested or undigested plasmid pRS416 was transformed into the WT and isogenic rev1Δ, rev3Δ, rev7Δ, mre11Δ, sae2Δ, rev1Δrev3Δ, rev1Δrev7Δ, rev3Δrev7Δ, mre11Δrev7Δ, sae2Δrev7Δ rev7-C1 or rev7-42 strains. In these experiments, the uncut plasmid pRS416 served as a control. The transformants arising from plasmid re-circularization were selected on SC/-Ura plates after growth at 30 °C for 3-5 days. The efficiency of transformation was calculated as a ratio of the number of transformants with digested plasmid DNA to that with undigested plasmid DNA. The graph was obtained using GraphPad Prism (Version 5.0) and statistical significance was calculated using one-way ANOVA Dunnett’s multiple comparisons test.

Non-homologous end-joining using “suicide deletion” reporter assay

The assay was performed as described (Karathanasis and Wilson, 2002). Briefly, single colonies of S. cerevisiae YW714 strains WT, rev1Δ, rev3Δ, rev7Δ, mre11Δ, ku70Δ, rev7-C1 and rev7-42 were inoculated into SC/-Ura liquid media for 30h, at 30°C, 250 rpm. The cells were harvested by centrifugation at 4000 rpm for 10 min. Cell pellets were washed thrice with MilliQ water and resuspended in SC media at O.D.₆₀₀ equal to 1. Following this, appropriate serial dilutions were plated parallelly onto SC/-Ura plates containing glucose and SC/-Ade plates containing galactose. The plates were incubated at 30°C for 5 days and the numbers of colonies were counted on both sets of plates for each strain. The rate of NHEJ, as a function of Ade⁺ colonies was calculated using the FALCOR software (Hall et al., 2009) for three independent experiments. The differences between experimental results and relevant controls were tested with Dunnett’s multiple comparison test. The graph was generated using GraphPad Prism (Version 5.0). Statistical tests and p-values are mentioned in each figure legend.

Plasmid-chromosome recombination assay

The assay was performed as previously described (Paeschke et al., 2013). The WT and isogenic ura3-1, ura3-1 rev7Δ, ura3-1 mre11-D56N,H125N ura3-1 rev7Δ mre11-D56N,H125N, ura3-1 rev7-C1 and ura3-1 rev7-42 strains were transformed with 60 ng of empty vector (pFAT10) or a plasmid bearing G4 DNA capable sequences (pFAT10-G4) (Paeschke et al., 2011). The synchronous cells were grown in SC/-Leu medium to OD₆₀₀ = 0.5. At this stage, equal number of cells were replica-spotted onto the SC/-Ura medium (left) and SC/-Ura medium supplemented with 0.1 M HU (right). Plates were incubated at 30 °C for 4-6 days. The recombination frequency was calculated as described (Hall et al., 2009).

Statistical and data analysis

Statistical comparison of two groups was evaluated using the one-way ANOVA Dunnett’s multiple comparison test or Tukey’s post hoc analysis. Statistical significance level was set as follows: *, P-value < 0.05; **, P-value < 0.01; ***, P-value < 0.001; ****, P-value < 0.0001. Unless otherwise specified, details of statistical methods used to analyse the data are provided in the Legends to the Figures. No data was excluded from analysis.

Acknowledgements

We thank Drs. Virginia Zakian, Maria Pia Longhese, Lorraine Symington and Thomas Wilson for kindly providing some of the strains used in this study, as well as to Dr. Narottam Acharya for the generous gift of ScRev7 and ScRev1 expression plasmids, S. cerevisiae BJ5464 strain and for his assistance in the purification of expressed proteins, and to Naren Chandran Shakthivel for his assistance with the generation of AlphaFold-Multimer models.

Funding

This work was supported by a grant (CRG/2021/000082) from the Science and Engineering Research Board, New Delhi to K. M., who was also the recipient of Bhatnagar Fellowship (SP/CSIR/425/2018) from the council of Scientific and Industrial Research, New Delhi.

Author contributions

Sugith Badugu, Kshitiza M. Dhyani, Manoj Thakur: Investigation, Methodology, Software, Validation, Data curation, Writing – original draft and editing; Kalappa Muniyappa, Conceptualization, Resources, Supervision, Funding acquisition, Formal analysis, Project administration, Writing – review and editing.

Conflict of interest statement

None declared

Data availability

All data associated with this study are in the article.

Supplemental figures and legends

Figure 1—figure supplement 1.

Figure 1—figure supplement 2.

Figure 1—figure supplement 3.

Figure 1—figure supplement 4.

Figure 2—figure supplement 1.

Figure 4—figure supplement 1.

Figure 6—figure supplement 1.

Figure 9—figure supplement 1.

Figure 9—figure supplement 2.

Table S1.

Table S2.

Table S3.

Table S4.

Table S5.

Table S6.

Table S7.