DNA Polymerase ζ without the C-Terminus of Catalytic Subunit Rev3 Retains Characteristic Activity, but Alters Mutation Specificity of Ultraviolet Radiation in Yeast

DNA polymerase ζ (pol ζ) plays a central role in replicating damaged genomic DNA. When DNA synthesis stalls at a lesion, it participates in translesion DNA synthesis (TLS), which helps replication proceed. TLS prevents cell death at the expense of new mutations. The current model indicates that pol ζ-dependent TLS events are mediated by Pol31/Pol32 pol ζ subunits, which are shared with replicative polymerase pol δ. Surprisingly, we found that the mutant rev3-ΔC in yeast, which lacks the C-terminal domain (CTD) of the catalytic subunit of pol ζ and, thus, the platform for interaction with Pol31/Pol32, retains most pol ζ functions. To understand the underlying mechanisms, we studied TLS in normal templates or templates with abasic sites in vitro in primer extension reactions with purified four-subunit pol ζ versus pol ζ with Rev3-ΔC. We also examined the specificity of ultraviolet radiation (UVR)-induced mutagenesis in the rev3-ΔC strains. We found that the absence of Rev3 CTD reduces activity levels, but does not alter the basic biochemical properties of pol ζ, and alters the mutation spectrum only at high doses of UVR, alluding to the existence of mechanisms of recruitment of pol ζ to UVR-damaged sites independent of the interaction of Pol31/Pol32 with the CTD of Rev3.


Introduction
UV radiation (UVR) is a ubiquitous DNA-damaging mutagenic agent, and living cells possess multiple mechanisms of protection from its consequences [1][2][3]. UVR in the solar spectrum is a major cause of skin cancer [4][5][6]. UVR is divided into three wavelength groups, with varying degrees of effects on DNA [7,8]. This study utilized UVC (<280 nm) radiation, which is typically blocked by ozone; however, the ozone layer is being depleted in some areas and, thus, this deadly radiation can contribute to DNA damage [9]. UVC is often used to study UVR-induced mutagenesis because the major DNA lesions are the same as those resulting from UVA/UVB radiation, but are induced more efficiently [10]. The two most abundant DNA lesions induced by UVR are cyclobutane-pyrimidine dimers (CPDs) and 6,4 photoproducts (6-4PPs) [11]. Both lesions distort the DNA helix, impeding replication and transcription. CPDs comprise nearly 75% of UV lesions and are the most cytotoxic. 6-4PPs present a difficult template for DNA polymerases and have been shown to correlate with mutational hotspots better than CPDs [12].
When DNA repair does not occur before UVR-induced damage is encountered by the replication machinery, CPDs and 6-4PPs can cause replication fork arrest, double-stranded Given this documented role of the Rev3-CTD, it was surprising to discover that yeast strains with the rev3-∆C deletion mutation encoding for a variant of Rev3 without the CTD-thus lacking the whole platform for the assembly of four-subunit pol ζ ( Figure 1B), in contrast to point mutations affecting Fe-S binding, exhibit normal UVR mutagenesis at low doses and moderate defects in mutagenesis at higher doses [47]. Extending these studies, here we asked two additional questions, expanding our genetic work and introducing supporting biochemical analysis. First, we explored whether pol ζ with the Rev3-∆C catalytic subunit (named pol ζ Rev3∆C ) is catalytically active and retains the capability to bypass damaged DNA in vitro. Second, we investigated whether the specificity of UVR mutagenesis is changed in the rev3-∆C mutant to help further define the mechanism of mutagenesis by crippled pol ζ. We found that pol ζ Rev3∆C retains most properties of pol ζ, suggesting that it is recruited to TLS via a pathway independent of the Rev3-CTD. UVR mutation spectra in rev3-∆C strains differed at different doses, and generally showed severe defects in mutagenesis pathways leading to transversions, while being quite proficient in mutagenesis leading to transitions. The majority of mutations occurred in dipyrimidine sites, consistent with the dependence of mutagenesis on the two major types of UV-induced damage. Complex mutations, a hallmark result of the activity of the complete pol ζ complex, were not abrogated. Taken together, biochemical data and in vivo mutationspecificity analysis imply that the main properties of pol ζ are not be altered by the absence of the CTD (and, as a consequence, the absence of the 4Fe-4S cluster) of Rev3 and suggest that there are yet-to-be determined mechanisms compensating for the loss.

Yeast Strains, Media, and Plasmids
DNA pol δ was overproduced as described previously [48]. For pol ζ purification, we used a similar system to that designed by P. Burgers: protease-deficient Saccharomyces cerevisiae yeast strain BJ2168 was transformed by plasmids pBL813 (URA3 GAL1-10 REV3opt REV7opt) and pBL347 (LEU2 GAL1-10 POL31 7xHIS_POL32) [41]. This strain overproduces all four subunits of pol ζ after induction by galactose. A variant of pBL813 encoding for CTD-less Rev3 (Rev3∆C) was created in our laboratory by site-directed mutagenesis using the GenEdit kit (First Biotech, Kyoto, Japan). First, a fragment of codon-optimized REV3opt was amplified by PCR with primers cdel-SP1 5 TGTAAAACGACGGCCAGTGA and cdel-SP2 5 GACAAGATCCCAGACCCAGC, and the 3.1 kb HindIII-XhoI part of the PCR was cloned into a helper plasmid included in the kit. Site-directed mutagenesis was done using the primers cdel-OP1 5 TAAGCGAATTTCTTATGATTTATGA and cdel-OP2 5 AGAAGCTCTCTTAGACTTAACGA, which amplify the whole plasmid without the portion of REV3opt coding for the Rev3 CTD. After verification that the deletion was generated, the fragment was inserted back into pBL813. The plasmid encodes Rev7 and GST-Rev3-∆CTD, lacking amino acids 1381-1504 of Rev3 ( Figure 1).
We used standard yeast media with modifications and appropriate carbon sources, as described previously [50]. Agar, yeast nitrogen base, yeast extract, peptone, glucose are from Formedium, Norfolk, UK. Other medium supplements are from Thermo Fisher, Waltham, MA, USA.

DNA Substrates
For DNA polymerase activity assays, a synthetic 50-mer oligodeoxynucleotide with the sequence 5 -TTACTTTTCTCATCTTTTAATGGTATTGACCCACGTCTGTGGTGTGTTTG-3 , complementary to nucleotides 91-140 of the CAN1 gene, was used as the template. The same 49-mer oligodeoxynucleotide containing an abasic site (bold and underlined) was used as a damaged template. A 16-mer labeled with fluorescent dye Cy5 (5 -Cy5-CAAACACACCACAGAC-3 ) served as a primer. Oligos were synthesized by IDT (Coralville, IA, USA) Cy5-labeled primer and template strands (500 nM each) were annealed in 25 mM Tris-HCl, pH 7.5, and 150 mM NaAc by heating to 95 • C for 5 min and then slowly cooled to room temperature. Annealed DNA substrates were stored at −20 • C for up to 2 weeks.
DNA polymerase activity assays with fluorescent primers were performed as described in [51,52]. In brief, assays were performed at 30 • C in 10 µL reaction mixtures containing 50 mM Tris-HCl, pH 7.5; 8 mM MgCl 2 ; 100 µg/mL BSA; 1 mM DTT; 100 µM each of dATP, dCTP, dGTP, and dTTP; and 50 nM annealed DNA substrate. Polymerase concentration and reaction time are indicated in the legend off Figure 2. After incubation, reactions were terminated with 10 µL denaturing gel loading buffer containing 95% formamide, 10 mM ethylenediaminetetraacetic acid (EDTA), and 0.125% Orange G. After incubation at 95 • C for 5 min, the reaction mixtures were loaded onto a 12.5% denaturing polyacrylamide gel (7 M urea). Gel electrophoresis was performed in ThermoFisher Owl S3S box in buffer with 85 mM Tris, 85 mM boric acid, and 2 mM EDTA at 50 W for 2 h and scanned with a Typhoon™ FLA 9500 biomolecular imager (Cytiva). The pattern of DNA synthesis was determined by analyzing each extended band intensity in comparison to the intensity of the unextended primer in ImageQuant TL 8.2 analysis software (Cytiva). This activity was distinct from that of pol ∆, which could not elongate primer termini after the abasic site in the template (marked "θ" to the right of the gel). The pol ζ ratio to the substrate was 0.5, that of pol ζ Rev3∆C was 2, and that of pol ∆ was 0.1. Reaction times were 5, 10, 25 min. "no", no polymerase added to reaction. (C) Quantitative comparison of products of synthesis by pol ζ Rev3∆C (red) with pol ζ 4 (blue) on normal template. *, p < 0.05, one-way ANOVA.

Determination of Survival and Induced Mutation Frequencies Using Can r Assay
Experiments were performed essentially as in [47]. Yeast strains were grown for two days at 30 • C in 5 mL of yeast extract peptone dextrose medium with 60 mg/L adenine and uracil (YPDAU) with shaking. Cells were pelleted at 1000 g in a Beckman Model TJ-6 centrifuge for 2 min and re-suspended in 1 mL of sterile water. Cells were appropriately diluted, and 100 µL aliquots were plated on synthetic complete (SC) medium for estimation of viability; 50-100 µL of undiluted cells were plated on SC-arg medium containing 60 mg/L of L-canavanine (Can). Plates were irradiated with UVC within 10 min of plating. After three days of growth at 30 • C, colonies on SC plates were counted, and survival was calculated by dividing the number of colonies at each UV treatment by the number of colonies without exposure. After five days of growth, colonies on SC + Can plates were counted, and the mutation frequency was calculated by dividing the number of colonies on SC + Can plates at each UV dose by the number of colonies on the SC plate at the same dose (the SC colony count was first multiplied by the dilution factor). The induced Can r mutant frequency was calculated by subtracting the spontaneous frequency (without treatment) from the mutant frequency for each UV dose [20]. All data points are averages of at least four independent trials, with duplicates of each sample in each trial.

Analysis of Mutational Spectra
Cells were plated and irradiated as described above for the Can r assay. Then, independent colonies were collected from each strain (wild-type and rev3-∆C) at each dose and were colony-purified by streaking them out on SC + Can plates (one colony per quarter-plate). One colony was picked from each quarter, genomic DNA was isolated, and the CAN1 gene was amplified with PCR using DreamTaq Green DNA Polymerase (Thermo Scientific) and the primers CAN1ext-F2 and CAN1ext-R2 (F2-TCTTCAGACTTCTTAACTCC, R2-ATAGTAAGCTCATTGATCCC) or CAN1ampF 5 TTCTGTGTGGTTTCCGGGTGAG and CAN1ampR 5 TGCTTCTACTCCGTCTGCTTTCTTTTC using SimpliAmp Thermal Cycler (Applied Biosystems/ThermoFisher, Waltham, MA, USA). The PCR products were purified using PEG 8000 precipitation [53] and sent to Genescript (Piscataway, NJ, USA) or McLab (San Francisco, CA, USA) for sequencing using the primers CAN1seq520F: TG-GTTTTCTTGGGCAATCAC, CAN1seq699R: GATGGAAGCGACCCAGAAC, and CAN1 seq1026F: AGTTCCATACAATGACCCTA. DNA sequences were then analyzed using Geneious Pro software v.5-v.R10 (Biomatters, Aukland, New Zealand).

Statistical Analyses
Survival and mutation frequencies are represented by means ± standard deviations. Unpaired two-tailed Student's t-tests were used to test the difference in survival and induced mutant frequency between wild-type and rev3-∆C. The χ 2 test of independence was applied to determine the differences in the mutation spectra between wild-type and rev3-∆C, as well as the mutation spectra in dipyrimidine (DP) sites. Fisher's exact test was used to compare the mutation spectra when the sample sizes were small. All analyses were performed using the open source software R. P < 0.05 was considered statistically significant.

Pol ζ Rev3∆C Shows Hallmark Characteristics of the TLS Polymerase pol ζ 4 , but Has Lower Activity In Vitro
The results of the purification of four-subunit pol ζ and a variant that lacks the Cterminal domain of the catalytic subunit, Rev3-∆C, are shown in Figure 2A. The catalytic subunit, Rev3, was N-terminally tagged by GST in the overexpression constructs, which allowed affinity purification of protein complexes tightly bound to Rev3. GST was subsequently cleaved from Rev3 (see the Section 2). Preparations of pol ζ 4 showed all four expected subunits ( Figures 1A and 2A), while preparations of pol ζ Rev3∆C did not show Pol31/Pol32 subunits, confirming that their interaction was severely impaired by the CTD deletion preventing co-purification ( Figures 1B and 2B). Despite the lack of the subunits thought essential for pol ζ activity [39][40][41], preparations of pol ζ Rev3∆C were catalytically active in vitro, though higher concentrations were required to get synthesis to the end of the template ( Figure 2B). This was consistent with our previous in vivo data [47]. Importantly, pol ζ Rev3∆C was able to bypass and extend past an abasic site similar to pol ζ 4, in stark contrast to replicative pol δ. The distribution of extended products showed an elevated tendency of pol ζ Rev3∆C to terminate synthesis prematurely, especially before (position +3) the abasic site ( Figure 2C).

Rev3-∆C ιs Most Proficient at Participating in the Formation of Transition Mutations In Vivo
Additional tests confirmed our previous observations [47] on the increased sensitivity of rev3-∆C strains to UVR-induced killing and proficiency in UVR-induced mutagenesis. The rev3-∆C strain was up to 50% more sensitive to UV than the wild-type strain ( Figure 3A). UVR dramatically increased (two orders of magnitude) can1 mutant frequencies in both strains, but, at higher doses, the frequency of induced mutants in rev3-∆C was profoundly lower compared to the wild-type strain, though still far above the level of spontaneous mutations ( Figure 3B). The magnitudes of the effects were consistent with a previous study [47].
To explore further than previously, we determined can1 mutation spectra at three different doses to accurately reflect the dose-dependent changes in the rev3-∆C strain (Tables 1, S1 and S2 with more details on the statistical analysis). Mutations are shown mapped to the CAN1 gene in Figure S1A-C. For the analysis, mutations were divided into the major classes of base substitutions plus insertions/deletions (indels) and complex mutations (Tables 1 and S1). •; rev3∆C, red . Data were collected and analyzed as described in the Section 2. The significance of the difference between wild-type and rev3-∆C strains was evaluated by a two-tailed unpaired Student's t-test.

Dose of 20 J/m 2
At the lowest dose used, rev3-∆C showed small increases over the wild-type in the frequency of all types of mutations, except for complex mutations; however, differences in the overall proportions of mutations of all types were not statistically significantly different from the wild-type (Table 1, Figure S2A). The proportions of transition and transversion mutations were the same (Figure 4, two left bars). Approximately one-fifth of all mutations were indels. The spectra in the wild-type and rev3-∆C strains were similar, with congruency between several hot spots: single base substitutions at positions 452, 527, 938, 1471, and 1477 and coincidence of complex mutations at position 723 ( Figure S1A). The results suggest that, at this dose, pol ζ Rev3∆C was fully functional in its transactions in UVR-induced mutagenesis.

Dose of 40 J/m 2
The intermediate dose caused a 3.5-fold further increase in mutation frequency in the wild-type strain compared to the previous dose, but rev3-∆C strains showed only a subtle increase, which led to a twofold difference in total induced mutant frequency compared to the wild-type ( Figure 3, Table 1, last row). At this dose, the difference between the two strains' mutational spectra started to appear (Figure 4, two middle bars). The rev3-∆C spectrum displayed 3.5-fold lower frequencies of transversions and a 1.9-fold decrease in frameshift mutations (Table 1, Figure S2B). Interestingly, the frequency of transitions was unaffected compared to the wild-type strain. As a result, the proportions of different classes of mutations were different between the strains (Figure 4). Mutations are shown mapped to the CAN1 gene in Figure S1B. The overlap in the distributions of mutations between wild-type and rev3-∆C was less obvious than at 20 J/m 2 . We could see cases of coincidences in mutations at positions 527, 938, and 1234. Positions 930-931 showed similar complex mutations in both strains. Another interesting similarity was the occurrence of similar mutations at position 1477 in both spectra and doses so far discussed. We inferred from the results at this dose that pol ζ Rev3∆C has limitations in the processes, leading to all types of mutations except transitions.

Dose of 60 J/m 2
At the highest dose, rev3-∆C exhibited an 11-fold reduction in overall induced mutant frequency (the frequency of UVR-induced mutations still exceeded spontaneous mutation frequency by 50-fold ( Figure 3); the strain was UVR-mutable) and exhibited much lower frequencies for all types of mutations compared to the wild-type (Table 1, Figure S2). The de-fect in transitions was less severe at 6.4-fold. Transversions decreased by 22-fold, frameshifts decreased by 25-fold, and complex mutations decreased by 89-fold (Tables 1 and S1). At this dose, CG-to-TA transition mutations dominated the spectrum in rev3-∆C (71% of all mutations, Figure 4). Mutations are shown mapped to the CAN1 gene in Figure S1C. We observed a C-to-T transition mutation hotspot at position 527 in both strains, which was the same as in the 20 J/m 2 spectra. Furthermore, positions in the vicinity of 1478 also had several mutations, even complex mutations, in the rev3-∆C strain. Overall, the rev3-∆C strain was most proficient in UVR-induced transitions. We inferred that, at this dose, pol ζ Rev3∆C was partially impaired in transactions leading to UVR-induced mutations, but the ability to extend mismatches leading to transitions was less affected.

Pol ζ Rev3∆C Participates in Creating pol ζ-specific Complex Mutations with Reduced Efficiency
Nearly all UVR-induced mutations are dependent on pol ζ, and its mutational signature has been studied in vitro and in vivo (see the Introduction). One of the prominent mutagenesis features dependent on pol ζ is the formation of complex mutations, defined here as multiple mutations in the same clone within ten base pairs. Pol ζ Rev3∆C was capable of supporting pathways leading to complex mutations, but these mutations were found at lower frequencies than with the wild-type for all doses studied (see Table S1 and Figure S1).

The Majority of Mutations in Wild-type and rev3-∆C Strains Occur in Dipyrimidine (DP) Sites
UVR primarily induces CPDs and 6-4PPs, both of which occur in DP sequences. Correspondingly, the majority (up to 90%) of mutations induced by UVR occur in these DP sites. The percentages of mutations found in DP sites versus non-DP sites were compared between wild-type and rev3-∆C strains, excluding complex mutations. In both strains, mutations predominantly occurred in DP sites, and there was no statistical difference between the two strains ( Table 2). This result suggests that pol ζ Rev3∆C can participate in transactions leading to mutations during di-pyrimidine lesion bypass similarly to pol ζ4. The rev3-∆C strain showed a significant decrease in the proportion of TT sites mutated and an increase in the proportion of CC sites mutated at 60 J/m 2 compared to the wild-type ( Figure 5). It is possible that the activity of pol ζ Rev3∆C , when lesions are abundant, is not equally sufficient for different types of UVR lesions. * Complex mutations excluded. # -the only pair that was significantly different (p = 0.0177) between the wild-type and rev3-∆C strains is in bold, Fisher's exact test. OVL-overlapping sites, e.g., TCT, where mutated base is "C".

Discussion
Given the multi-faceted role of pol ζ in mutagenesis and carcinogenesis, it is essential to understand the mechanism, and the precise roles and functions, of all subunits of pol ζ in transactions on damaged DNA. The protein complex performing TLS is called the mutasome. The recruitment of pol ζ may occur through several partially overlapping mechanisms dependent on protein-protein interactions and regulated by Fe-S clusters in pols. We centered our study on a yeast mutant that lacked the C-terminal domain (CTD) of Rev3 and, thus, the crucial interaction platform for Pol31/Pol32, which were thought to be required for pol ζ function (Figure 1). Breaking the current paradigm in TLS, the strain with crippled pol ζ had almost normal UV-induced mutagenesis [47] (Figure 3), pointing to the void in our understanding of pol ζ and the existence of a backup mechanism for its recruitment. Due to the similarities in the genetic processes and conservation of genes encoding for pols in yeast and human cells, it is likely that analysis of yeast mutants would provide new information on the mechanisms of pol ζ-assisted replication relevant to the onset and cure of human disease. This knowledge could help lead to the development of new dual-target approaches to the inhibition of TLS (in addition to the inhibition of REV3L/REV7/REV1 interactions), a promising anticancer target [54], where not one but two pathways of recruitment of pol ζ would be inhibited simultaneously.
In this study, we asked whether (a) Pol ζ Rev3∆C possesses the same basic biochemical properties as wild-type pol ζ, and (b) the Pol ζ variant present in the rev3-∆C mutant strain is deficient in participation in all types, or only certain types, of mutations in response to UVR. Comparing the biochemical properties of the two forms of pol ζ and the UV-induced mutational spectra between wild-type and rev3-∆C strains further defined the role of the CTD of Rev3 during TLS. We found that Pol ζ Rev3∆C behavior in vitro in normal and dam-aged templates was similar to that of pol ζ, but the overall activity was reduced. The spectra of UVR-induced mutations in the rev3-∆C strain showed a bias toward transition mutations. There was a milder defect with respect to transitions than transversions/frameshifts, and a higher UVR dose was needed to elicit this effect.

Biochemical Properties of pol ζ Rev3∆C.
We purified the complex from yeast strains overexpressing GST-rev3-∆C, REV7, POL31, and POL32, the system described for wild-type pol ζ 4 [41] (Figure 2A). The main feature of the system is the use of the GST tag for affinity chromatography in one of the first steps of purification (note that the GST tag was cleaved after the initial pulldown of the GST-tagged catalytic subunit; see the Section 2). The protocol resulted in robust purification of the four-subunit pol ζ 4 (Figure 2A). In the case of pol ζ Rev3∆C , only two bands, corresponding to Rev3-∆C and Rev7, were seen on the gel, confirming that the deletion of the CTD of Rev3 led to an inability to support strong interactions with Pol31/Pol32 subunits that could withstand conditions of purification. Despite the deficit of Pol31/32, Pol ζ Rev3∆C retained substantial DNA polymerase activity and had the ability to synthesize DNA on a template with an abasic site ( Figure 2B,C), which was quite different from the inability to bypass abasic sites in the genuine two-subunit pol ζ 2 , purified from a strain with a deletion of the POL32 gene [41].These results suggest that the characteristic conformation of the active site of Rev3 [55] was unaltered and functional. We hypothesize that the Rev3 CTD in pol ζ 2 without Pol31/32 either adopted a conformation that was inhibitory for pol reaction or became purified as an inactive complex, whereas the Rev3 CTD without proper partners was misfolded; this constraint was bypassed in Pol ζ Rev3∆C . Another possibility is that the preparation of pol ζ Rev3∆C contained various unknown minor accessory proteins (not visible in the Coomassie gel) that were sufficient to support the activity of the complex. The complex was purified from a strain that overexpressed components of pol ζ 4 but had natural levels of all other replication proteins. A candidate protein was Rev1, which can interact with Rev7 and Pol32 [44,[56][57][58] and, thus, could have brought the necessary components to the pol ζ Rev3∆C complex. In this scenario, we should assume different interactions of Rev1 within the pol ζ Rev3∆C and pol ζ 5 complexes due to the presence of Rev1 in inhibiting the pol ζ 4 -Rev1 five-subunit complex [26]. The near-normal (at least at low or moderate levels of DNA damage) participation of Pol ζ Rev3∆C in UVR-induced mutagenesis indicated that the natural cell environment effectively compensated for the absence of the Rev3 CTD. It is possible that the high level of UVR mutagenesis in rev3-∆C was a result of a combination of the mechanisms discussed above.

UVR Mutation Spectra in rev3-∆C Strains
Studies of mutational spectra can provide a wealth of information that cannot be discerned by simply measuring rates of mutagenesis in cells [19,59,60]. Mutational spectra have been examined in the p53 gene in human skin cancers, which are primarily caused by UVR ( [61]; reviewed in [60,62]). Most mutations in p53 in skin cancers are CG-to-TA transitions (mainly C-to-T), and 90-96% of all base substitutions occur in dipyrimidine sites. This well-known UV signature is consistent with the fact that UVR primarily induces CPDs and 6-4PPs, which occur at these sites.
It has been shown that mutational spectra from skin tumors are very similar to the spectra observed in UVR targets in model systems, such as yeast, but differ from spectra described in other cancers (discussed in [63]). The UVR spectrum in yeast transformed by an irradiated vector containing human p53 cDNA was not significantly different from that found in patient non-melanoma skin tumors [63], highlighting the fact that yeast is a suitable model to study cancer-relevant UVR mutagenesis.
In yeast, the CAN1 gene is often used to study such mutational spectra (discussed in [19,20]). This system is able to detect all types of base-pair substitutions, deletions, insertions, and complex mutations. UVC-induced spectra in CAN1 consist predominantly of base-pair substitutions and tandem and non-tandem complex mutations ( [20]; discussed in [19]). These complex mutations are a signature of pol ζ [27,29,64,65]. Mutation spectra presented here for both strains share many features with the yeast CAN1 spectra previously reported in the literature (discussed in [19]).
UV spectra were examined at three doses to accurately characterize the rev3-∆C strains (Table 1, Figure S1). Compared to the wild-type strain, this strain showed no statistically significant changes in the levels or distribution of mutations at the lowest UVR dose. At an intermediate dose, it showed a decreased ability to form transversion and indel mutations. At the highest dose, rev3-∆C exhibited a reduction in overall induced mutant frequency and lowered all types of mutations compared to the wild-type; however, the level of induced mutagenesis remained high, exceeding the level of spontaneous mutations by almost two orders of magnitude ( Figure 3B). The decrease seen in the overall induced mutant frequencies with the increase in UVR doses in rev3-∆C strains could have been due to several reasons.
One reason could be the increased number of lesions, which a polymerase with lower activity would not be able to process. It is also plausible that lesions or sequence contexts where transversions usually form may not be efficiently bypassed by Pol ζ Rev3∆C . Crippled pol ζ might have difficulties inserting bases resulting in transversions compared to the wild-type and/or it may be less efficient in extending transversions inserted by other polymerases, such as Pol η. Pol ζ Rev3∆C may be recruited less efficiently to lesions where transversions usually arise. It is conceivable that Pol ζ Rev3∆C may have less difficulty with transitions, as these involve the substitution of structurally similar bases (same numbers of rings) [55]. Transversions would distort the double helix of DNA, posing difficulties for both insertion and extension. It is formally possible that pol ζ Rev3∆C is an inferior enzyme, so only the mutations that are easy to produce/extend remain (transitions). Another possibility is that Pol ζ Rev3∆C is more proficient at bypassing one of the two major types of UV lesions. It will be essential in the future to determine whether the regulatory role of the CTD is in recruitment to DNA or whether it affects catalysis. Without a solved structure showing Rev3 with the CTD in the full Pol ζ 5 complex, it is not clear where the domain lies relative to the catalytic residues or DNA binding site. Deamination of C residues in photoproducts [66] could also play a role in this process. If it takes more time for mutations to be addressed by pol ζ Rev3∆C , this could allow time for the C residues involved to be deaminated to U, resulting in an A incorporation.
A prominent signature of pol ζ in UVR mutagenesis is the formation of complex mutations [20]. These can occur through a template-switching mechanism [33] or when pol ζ produces errors/mutations during the extension of sequences downstream of the lesion site. They are defined here as multiple mutations in the same clone within 10 base pairs, although it has been shown that stretches synthesized by pol ζ can be up to several hundred nucleotides in length [32]. In our strains, pol ζ Rev3∆C was able to participate in the formation of these complex mutations (Table S1, Figure S1), providing additional evidence for the functional proficiency of the enzyme in vivo.
Most UVR-induced mutations in wild-type and rev3-∆C strains were found in DP sites, consistent with the fact that most lesions caused by UVR are CPDs or 6-4PPs (Table 2). 6-4PPs are usually formed preferentially at CC and CT sites, whereas CPDs are formed at TT sites [61]. We observed that a higher proportion of total mutations were found in CC sites, and a higher proportion of total CC sites and a lower proportion of total TT sites were mutated in rev3-∆C at 60 J/m 2 compared to the wild-type ( Figure 5). Additionally, there was an increase in the proportion of mutations found at CC and CT sites and in the proportion of total CC/CT sites mutated at 60 J/m 2 in rev3-∆C compared to lower doses. These data may reflect a stronger role for Pol ζ Rev3∆C in bypassing 6-4PPs (and deficiency in CPD bypass), especially at high doses of UVR.
Finally, we would like to emphasize that the complete absence of the domain of pol ζ necessary for binding the Fe-S cluster did not abolish the polymerase transactions necessary for TLS and induced mutagenesis, arguing against the involvement of electron transfer signaling in the regulation of pol ζ-assisted events hypothesized for the other members of B-family, primase-pol α, pol δ, and pol ε [67][68][69].
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/genes13091576/s1, Figure S1: Mutational spectra of wild-type and rev3-∆C mapped to CAN1 gene, Figure S2: UV dose-dependent changes in frequencies of transition, transversion, and indel mutations in the wild-type and rev3-∆C strains, Table S1: Differences in classes of mutations in wild-type and rev3-∆C strains. Table S2: Differences in mutation spectra in wild-type and rev3-∆C strains.