Decision letter: Roles for mycobacterial DinB2 in frameshift and substitution mutagenesis

Full text Figures and data Side by side Abstract Editor's evaluation Introduction Results Discussion Methods Appendix 1 Data availability References Decision letter Author response Article and author information Metrics Abstract Translesion synthesis by translesion polymerases is a conserved mechanism of DNA damage tolerance. In bacteria, DinB enzymes are the widely distributed promutagenic translesion polymerases. The role of DinBs in mycobacterial mutagenesis was unclear until recent studies revealed a role for mycobacterial DinB1 in substitution and frameshift mutagenesis, overlapping with that of translesion polymerase DnaE2. Mycobacterium smegmatis encodes two additional DinBs (DinB2 and DinB3) and Mycobacterium tuberculosis encodes DinB2, but the roles of these polymerases in mycobacterial damage tolerance and mutagenesis is unknown. The biochemical properties of DinB2, including facile utilization of ribonucleotides and 8-oxo-guanine, suggest that DinB2 could be a promutagenic polymerase. Here, we examine the effects of DinB2 and DinB3 overexpression in mycobacterial cells. We demonstrate that DinB2 can drive diverse substitution mutations conferring antibiotic resistance. DinB2 induces frameshift mutations in homopolymeric sequences, both in vitro and in vivo. DinB2 switches from less to more mutagenic in the presence of manganese in vitro. This study indicates that DinB2 may contribute to mycobacterial mutagenesis and antibiotic resistance acquisition in combination with DinB1 and DnaE2. Editor's evaluation This important study uses a combination of compelling biochemical and genetic approaches to identify a highly mutagenic mycobacterial DNA polymerase, which drives a wide spectrum of mutations when overexpressed. The findings advance the understanding of mutagenesis in mycobacteria. The work will be of interest to bacteriologists working on mutation mechanisms and the emergence of drug resistance. https://doi.org/10.7554/eLife.83094.sa0 Decision letter Reviews on Sciety eLife's review process Introduction One of the primary mediators of chromosomal mutagenesis is the error-prone DNA damage tolerance pathway termed translesion synthesis (TLS) (Vaisman and Woodgate, 2017). TLS polymerases transiently replace the replicative DNA polymerase to traverse lesions that block replication. Because of the flexibility of their active site and lack of proofreading activity, these polymerases facilitate survival during DNA damage but are mutagenic. TLS has been extensively studied in Escherichia coli, which encodes two DNA damage-inducible polymerases: DinB (Pol IV) and UmuDC (Pol V) (Fuchs and Fujii, 2013; Fujii and Fuchs, 2020). DinB and UmuDC confer tolerance to multiple forms of DNA damage and mediate mutagenesis by incorporating both substitution and frameshift mutations. Whereas DinBs are ubiquitous in bacteria, the distribution of UmuDC is more restricted. Many bacteria, including mycobacteria, encode a second copy of the replicative DNA polymerase (Pol III), called DnaE2 (Cole et al., 1998; Erill et al., 2006), as a translesion polymerase. The Mycobacterium genus, which includes the causative agent of tuberculosis (TB) Mycobacterium tuberculosis (Mtb) and several others pathogens, encodes DnaE2 as well as several DinB paralogs (Cole et al., 1998; Timinskas and Venclovas, 2019). DnaE2 confers UV tolerance and antibiotic resistance in Mtb through its mutagenic activity and plays a role in pathogenicity (Boshoff et al., 2003). DnaE2 was initially thought to be the only active TLS polymerase in mycobacteria. The role of DinBs in DNA damage tolerance and mutagenesis was unclear because initial studies of dinB deletion strains failed to identify a function of Mtb DinBs (Kana et al., 2010). Mycobacterial DinBs exemplify three phylogenetic subfamilies found in many actinobacteria: DinB1/DinX, DinB2/DinP, and DinB3/msDinB3 (Cole et al., 1998; Timinskas and Venclovas, 2019). In a recent study, we revealed that Mycobacterium smegmatis and Mtb DinB1 contribute to alkylation damage tolerance, antibiotic resistance though a characteristic mutagenic spectrum, and chromosome diversification through frameshift mutations in homo-oligonucleotide runs (Dupuy et al., 2022). Some of these DinB1 activities, particularly homopolymeric run frameshift mutagenesis, are redundant with DnaE2 during exposure to DNA damage, suggesting that DinB1 and DnaE2, both of which interact with the β-clamp, exert their overlapping activities at the replication fork. M. smegmatis encodes two additional DinBs, DinB2 and DinB3, whereas Mtb encodes DinB2 but lacks DinB3. Biochemical studies of M. smegmatis DinB2 and DinB3 showed that both have polymerase activity (Ordonez et al., 2014). DinB2 is notable for its capacity to utilize ribonucleotides during templated DNA synthesis, an activity that is attributable to the absence of a polar filter (Johnson et al., 2019) and a permissive steric gate (Ordonez et al., 2014). Leucine 14 of DinB2 replaces the canonical aromatic amino acid which clashes with the 2’-OH of rNTPs and thereby confers selectivity to DNA polymerases. DinB2 also displays a metal-dependent mutagenic switch in which manganese supports efficient 8-oxoguanine utilization and nucleotide addition opposite 8-oxoguanine in the template (Ordonez and Shuman, 2014). These biochemical activities suggest that DinB2 is equipped to mediate a diverse range of mutation types in vivo. However, our prior data (Dupuy et al., 2022) indicate that DinB2 is expressed at very low levels in basal conditions and not induced by DNA damage, in contrast to DnaE2, which is induced ~100-fold by UV. Indeed, Patra et al., 2021 showed that DinB2 expression is actively repressed in M. smegmatis by the action of the TetR family repressor protein MSMEG_2294 encoded in an operon with the dinB2 gene. This explains why we did not detect any effects of deleting the dinB2 ORF on mutagenesis in vivo (Dupuy et al., 2022), because DinB2 is effectively absent under the conditions surveyed. Thus, an alternative approach is needed. In this study, we used inducible overexpression to gauge whether and how DinB2 (and DinB3) can affect genomic integrity. We show that DinB2 and DinB3 are both promutagenic in vivo, with distinct mutagenic signatures. In addition, we highlight the strong ability of DinB2, but not DinB3, to incorporate frameshift mutations in both short and long homo-oligonucleotide runs, an activity that is enhanced by manganese in vitro. Finally, we find that manganese enhances the growth inhibitory effects of DinB2 overexpression, suggesting that the metal switch operates in vivo. Results Overexpression of DinB2 causes cell death through its polymerase activity To examine the function of DinB2 and DinB3 in M. smegmatis, we expressed plasmid-borne copies of the dinB2 or dinB3 genes, encoding untagged or streptavidin-tagged (ST) versions of M. smegmatis DinB2 and DinB3, under the control of an anhydrotetracycline (ATc)-inducible promoter (tet promoter). DinB2 and DinB3 were detected by immunoblotting with anti-ST antibodies after inducer addition. Levels of DinB2 and DinB3 were similar 4 and 24 hr after ATc treatment (Figure 1—figure supplement 1A). Compared to the untreated condition, the levels of DinB2 and DinB3 were increased by between 5- and 15-fold, depending on ATc concentration (Figure 1A). Overexpression of tagged and untagged versions of DinB2, but not DinB3, caused a growth defect (Figure 1B and Figure 1—figure supplement 1B) and 10-fold loss of viability (Figure 1C) which was proportional to the concentration of inducer (Figure 1—figure supplement 1B–C). DinB2 overexpression also triggered the DNA damage response, as measured by RecA induction (Figure 1D). Despite no effect on growth, overexpression of DinB3 also induced the DNA damage response (Figure 1D). Figure 1 with 2 supplements see all Download asset Open asset Overexpression of DinB2 causes cell death through its polymerase activity. (A) Anti-streptavidin/RpoB immunoblots from indicated strains with indicated concentrations of inducer treatment (16 hr of treatment). Average and SEM of RpoB normalized band intensities (n=3, arbitrary units) are given below the image of a representative blot. (B) Growth and (C) viability of indicated strains in presence of 50 nM anhydrotetracycline (ATc). Note that the OD600 values in (B) are calculated values based on continuous dilution growth experiments (see Methods). (D) Anti-RecA/RpoB immunoblots from indicated strains with indicated times of inducer treatment (50 nM). Average of normalized band intensities, expressed relative to the empty vector strain, is given below the image of a representative blot. (E) Anti-streptavidin/RpoB immunoblots from indicated strains after 16 hr of inducer treatment (50 nM ATc). Average and SEM of normalized band intensities (n=3) are given below the image of a representative blot. (F) Growth of indicated strains in presence of 50 nM ATc. (G) Anti-RecA/RpoB immunoblots from indicated strains after 24 hr of inducer treatment. Average of normalized band intensities, expressed relatively to the empty vector strain, is given below the image of a representative blot. Empty = empty vector, tet = ATc-inducible promoter, DinB2=M. smegmatis DinB2, DinB3=M. smegmatis DinB3, ST=streptavidin tag, D107A=catalytically inactive M. smegmatis DinB2. Results shown are means (± SEM) of biological triplicates. Stars under the means mark a statistical difference compared to the empty vector reference strain (, p<0.01; *, p<0.001). Figure 1—source data 1 Uncropped immunoblots Figure 1A, D, E and G. https://cdn.elifesciences.org/articles/83094/elife-83094-fig1-data1-v1.zip Download elife-83094-fig1-data1-v1.zip Our prior work showed that overexpression of DinB1 in M. smegmatis also induces cell death, probably by interacting with the replicative machinery and competing with the replicative DNA polymerase (Dupuy et al., 2022), a phenotype that was exacerbated by a polymerase dead active site mutation in DinB1. Unlike DinB1, neither DinB2 nor DinB3 have a predicted β clamp binding motif (Kana et al., 2010). To investigate the cause of the cell death induced by DinB2 overexpression, we expressed DinB2-D107A, which lacks polymerase activity (Ordonez et al., 2014). Whereas DinB2 and DinB2D107A were expressed at similar levels (Figure 1E), DinB2D107A did not arrest bacterial growth (Figure 1F and Figure 1—figure supplement 1D) or induce the DNA damage response (Figure 1G). These results indicate that, unlike DinB1, the toxic effect of DinB2 is due to its polymerase activity. A study conducted in E. coli revealed that dinB overexpression toxicity is due to genomic incorporation and excision of 8-oxoguanines, leading to the formation of DNA double-strand breaks (Foti et al., 2012). This is reminiscent of the ability of DinB2 to utilize 8-oxoguanine for DNA synthesis in vitro (Ordonez and Shuman, 2014). To test if DinB2 overexpression causes cell death in vivo because of genomic incorporation of 8-oxoguanines, we measured the impact of mutTs and mutY/mutMs deletion on DinB2-dependant lethality. The mutT and mutY/mutM enzyme systems are involved in degrading free 8-oxoguanine nucleotides and excision of genomic 8-oxoguanines, respectively (Dupuy et al., 2020). We did not observe a significant impact of the absence of these anti-8-oxoguanine systems on the dinB2 overexpression growth defect (Figure 1—figure supplement 2A and C) or loss of viability (Figure 1—figure supplement 2B and D), indicating that DinB2-dependent lethality is not mainly due to genomic 8-oxoguanine incorporation. DinB2 and DinB3 confer antibiotic resistance through a distinct mutagenic profile Previous studies showed that mycobacterial DnaE2 and DinB1 confer antibiotic resistance by stimulating chromosomal substitution mutations with distinct mutation signatures (Boshoff et al., 2003; Dupuy et al., 2022). To investigate if DinB2 and DinB3 are similarly mutagenic, we measured the rifampicin resistance (rifR) frequency in M. smegmatis strains overexpressing DinB2 or DinB3. In the absence of inducer, strains carrying the empty vector, tet-dinB2 or tet-dinB3 had similar rifR frequencies. Although 16 hr of inducer treatment had no effect on the rifR frequency in the control strain, overexpression of DinB2 or DinB3 increased the rifR frequency by sixfold, compared to –ATc condition (Figure 2A). A similar effect was observed with an ST version of DinB2 but not DinB2D107A-ST(Figure 2B), showing that the polymerase activity of DinB2 is necessary to confer antibiotic resistance. DinB2 and DinB3 overexpression still enhanced rifR frequency in the ΔrecA and ΔdnaE2 strains (Figure 2—figure supplement 1), indicating that the antibiotic resistance conferred by DinB2 is not an indirect consequence of the activation of the DNA damage response or the induction of the DNA damage-inducible error-prone DnaE2 polymerase. Figure 2 with 1 supplement see all Download asset Open asset DinB2 and DinB3 overexpression confers antibiotic resistance through a distinct mutagenic profile. (A and B) Rifampicin resistance (rifR) frequency in indicated strains in absence (blue) or presence (red) of inducer (50 nM anhydrotetracycline [ATc]). Results shown are means (± SEM) of data obtained from biological replicates symbolized by gray dots. Stars above bars mark a statistical difference with the reference (same strain without inducer) (***, p<0.001). Pie charts and bar chart in (A) shows the relative and absolute frequencies of nucleotide changes, represented with colors, detected in rpoB of indicated strains rifR in presence of inducer (50 nM ATc). The number of sequenced rifR is given in the center of each pie chart. (C) Location and relative frequency in % of mutated nucleotides in rpoB found in empty (blue), tet-dinB2 (red), or tet-dinB3 (orange) rifR. (D) Absolute frequency of the main rpoB mutations found in indicated strains in presence of 50 nM ATc. Empty = empty vector, tet = ATc-inducible promoter, DinB2=M. smegmatis DinB2, DinB3=M. smegmatis DinB3, ST = streptavidin tag, D107A=catalytically inactive M. smegmatis DinB2. Rifampicin resistance is conferred by substitution mutations in the rifampin resistance determining region (RRDR) of the rpoB gene. To define the mutation spectrum stimulated by DinB2 and DinB3, we sequenced the RRDR of rifR colonies selected after overexpression of these polymerases (Figure 2A). In the strain carrying the empty vector, 38% of mutations conferring rifR were G>A or C>T and 24% were A>G or T>C, with a minority of other mutations, consistent with our prior report (Dupuy et al., 2022). DinB3 overexpression enhanced the relative frequency (expressed as %) of A>G or T>C mutations by 2.7-fold and the absolute frequency (expressed as mutants/108 CFU) by 11-fold with a minimal effect on other mutation types (Figure 2A). In contrast, DinB2 overexpression elicited a more diverse spectrum of mutations. Overexpressed DinB2 enhanced the relative frequency of A>C or T>G and G>T or C>A mutations by 2.8- and 2.3-fold, respectively. In addition, a previously undetected mutation type emerged: a trinucleotide mutation (G TC >T GT) spanning two codons and detected in 6.8% of sequenced rifR colonies. Overall, the absolute frequency of all mutation types was increased at least threefold after DinB2 overexpression, revealing the ability of the polymerase to stimulate diverse mutation types, in contrast to DinB1 (Dupuy et al., 2022). Mapping of the DinB2 or DinB3 stimulated rpoB mutations onto the RRDR sequence revealed that 57% of the mutations incorporated by DinB3 were localized in the second nucleotide of the His442 codon vs 23% at this position in the control (Figure 2C). The predominant missense change at this codon was CAC>CGC (His>Arg) (Figure 2D), the same mutation stimulated by DinB1 (Dupuy et al., 2022). The absolute frequency of this mutation was increased 12-fold after DinB3 overexpression (Figure 2D). In contrast, DinB2 associated mutations were more widely distributed in the RRDR, particularly at His442 (CAC>CGC, >CCC, or >TAC), Ser438 (TCG>TTG or >TGG), Leu437/Ser438 (CTG TCG>CTT GTG), Leu449 (CTG>CCG), Asn435 (AAC>AAA), and Asp432 (GAC>GGC) (Figure 2C and D). Overall, our results show that DinB2 and DinB3 can mediate rifampicin resistance but with different mutagenic profiles, with DinB2 driving a broader mutation spectrum. In comparison to our prior results with DinB1 and DnaE2, these data indicate that the mutagenic activities of DinB3 and DinB1 are relatively narrow and focused on His442, whereas DinB2 is similar to DnaE2 in its wider mutagenic spectrum. DinB2 is highly prone to backward slippage in runs of A and T in vitro Our previous study revealed that mycobacterial DinB1 mediates –1 and +1 frameshift mutations in runs of homo-oligonucleotides in vivo through a slippage activity of the polymerase (Dupuy et al., 2022). To determine whether DinB2 might have similar properties, we measured the ability of DinB2 to perform slippage in vitro. We reacted purified recombinant DinB2 with a 5' 32P-labeled primer-template DNA in the presence of dTTP or dATP. The DNA substrate consisted of a 13 bp duplex and a 5'-template tail in which the length of the homo-oligonucleotide contains 4, 6, or 8A or T followed by 3C or 3G (Figure 3A and B). Figure 3 with 1 supplement see all Download asset Open asset DinB2 efficiently promotes –1 and +1 frameshifts in short and long runs of A and T. (A and B) Reaction mixtures containing 10 mM Tris-HCl, pH 7.5, 5 mM MnCl2,1 pmol 5' 32P-labeled primer-template DNAs with indicated runs in the template strand (depicted below, and included as indicated above the lanes), 125 µM dTTP and ddGTP as specified, and 10 pmol DinB2 were incubated at 37°C for 15 min. The reaction products were analyzed by urea-PAGE and visualized by autoradiography. DinB2 was omitted from reactions in lanes –. (C–H) kanR frequencies in the indicated strains carrying the indicated mutation reporters in presence of inducer (50 nM ATc). Results shown are means (± SEM) of data obtained from biological replicates symbolized by gray dots. Stars above the bars mark a statistical difference with the reference strain (empty) (*, p<0.05; *, p<0.001). Relative (pie chart) and absolute (bar chart) frequencies of nucleotide changes detected in kan of kanR cells represented with colors: pink=–1 or +1 frameshift in the homo-oligonucleotide run, blue=-2 frameshift in the run, green=+1 frameshift localized outside of the run and gray=no detected mutation. The number of sequenced kanR colonies is given in the center of each pie chart. Empty=empty vector, tet = Atc-inducible promoter, DinB2=M. smegmatis DinB2, DinB3=M. smegmatis DinB3, ST=streptavidin tag, D107A=catalytically inactive M. smegmatis DinB2. Figure 3—source data 1 Original autoradiograms (Figure 3A and B). https://cdn.elifesciences.org/articles/83094/elife-83094-fig3-data1-v1.zip Download elife-83094-fig3-data1-v1.zip In presence of the A4 template and dTTP only, the reaction generated predominant +5 and minority +6 products (Figure 3A) that could be the consequence of either: the fill-in of the A4 run followed by the misincorporation of one or two Ts opposite C; or a backward slippage reaction catalyzed by DinB2 incorporating one or two extra Ts in the 4A template tract. DinB2’s slippage activity was more evident during synthesis on the A6 primer-template, generating a ladder of products elongated by 8–17 nucleotides. Progression to the A8 template generated a longer array of slippage products extended by 11–23 nucleotides. Similar results were obtained with template runs of 6 or 8 Ts (Figure 3B). The finding that DinB2 is capable of iterative slippage synthesis on a homo-oligomeric tract when the only dNTP available is that templated by the homo-oligomer does not reflect the situation in vivo where DinB2 will have access to the next correctly templated dNTP. To query whether provision of the next templated nucleotide in vitro suppresses slippage, we included a dideoxy NTP (ddNTP): either ddGTP templated by the run of three C nucleotides following the A4, A6, and A8 tracts or ddCTP templated by the run of three G nucleotides flanking the T4, T6, and T8 tracts. ddNTPs are employed to force termination upon incorporation of the first templated nucleotide following the homo-oligomeric tract. Inclusion of a templated ddGTP with dTTP generated a +5 on the A4 as for addition of nucleotides and a (Figure DinB2 reaction with the template in the presence of ddCTP and a predominant +5 4 of templated incorporation and a addition (Figure 3B). In both the +6 products with dTTP or only were by of a templated The for slippage was by the effects of ddGTP on DinB2 activity on the A6 and A8 where of products longer or nucleotides results of templated were evident (Figure to 10 extra were detected on the A8 template when the next templated nucleotide was Similar results to the T8 ddCTP but did not the slippage ladder with (Figure 3B). to extra were observed on the T8 template in the presence of These experiments that DinB2 is prone to backward slippage in A and T runs, a that is exacerbated in longer tracts. The distribution of the slippage ladder is consistent with either of two multiple slippage in which the backward on the template by a and one or several of backward of the on the template by more one nucleotide the length of the template homo-oligomeric followed by fill-in to the of the homo-oligomeric tract. is that reaction of DinB2 with the T4, T6, and T8 in the presence of and ddCTP also generated a that was the predominant of a of slippage on the T runs prior to incorporation. DinB2 promotes –1 and +1 frameshifts in short and long runs of A and T in vivo The slippage activity of DinB2 in homo-oligonucleotide runs detected in vitro that the polymerase may incorporate mutations in vivo. To test this we used a in our previous study (Dupuy et al., a chromosomal kan conferring the resistance to by an homo-oligonucleotide run of the in the run can a and the mutation frequency can be on To determine the effect of sequence as well as run we used reporters carrying different 4A and 8A the we detected colonies in the control strain, carrying the empty vector (Figure A of the sequenced kanR had a –1 mutation in the run of The overexpression of DinB2 increased the kanR frequency more compared to the control strain, but DinB3 had no kanR obtained after DinB2 overexpression had –1 localized in the kan run, revealing that the polymerase promotes this of mutation. The of the mutation frequency was by DinB2 active site mutation (Figure indicating that DinB2 –1 through its polymerase activity. The –1 mutation frequency observed with the in the control strain was with the vs with of the sequenced kanR containing a –1 in the run (Figure revealing that the of the run the mutation frequency in mycobacteria. with this of overexpression of DinB2 enhanced the –1 mutation frequency in the run by A similar was observed with runs of 4A and (Figure supplement and B). and to detect +1 the strain carrying the empty vector had and 24 with and of sequenced kanR +1 mutations in the run (Figure and Figure supplement Whereas DinB3 overexpression had no effect on the kanR DinB2 overexpression increased the +1 frequency by in both runs (Figure and Figure supplement 1C) and this effect was on its polymerase activity (Figure of the run vs enhanced the +1 detected in the run of A by and these were still induced by DinB2 overexpression (Figure supplement 1D). In contrast, the kanR frequency in the strain carrying the empty vector and the was similar to the frequency observed with the but in contrast to the the sequenced kanR had no kan mutations (Figure because the encoded three amino acid the function of the protein or expression of the kanR gene. Despite this we observed an of the kanR frequency after DinB2 overexpression which was due to an of mutations in the run, revealing that DinB2 can also stimulate these results that the of homo-oligonucleotide runs the frequency of mutations incorporated in A and T runs of the mycobacterial chromosome and that DinB2 can and mutagenesis in these low DinB2 slippage activity is enhanced on C and G homo-oligonucleotide We to DinB2 with a of containing runs of G or C in vitro. Similar to our findings with 4A and runs, DinB2 synthesis the run generated +5 products (Figure However, the polymerase was more prone to on the generating a of extended by of addition (Figure on was by of which the ladder with a +5 (Figure the of the of the +5 the +5 a of slippage (Figure A +6 in the presence of is of +1 slippage on the run in the presence of the next templated Figure 4 with 2 supplements see all Download asset Open asset DinB2 slippage activity is enhanced on C and G homo-oligonucleotide (A and B) Reaction mixtures containing 10 mM Tris-HCl, pH 7.5, 5 mM MnCl2,1 pmol 5' 32P-labeled primer-template DNAs with indicated runs in the template strand (depicted below, and included as indicated above the lanes), 125 µM and or and as specified, and 10 pmol DinB2 were incubated at 37°C for 15 min. DinB2 was omitted from reactions in lanes –. The reaction products were analyzed by urea-PAGE and visualized by autoradiography. The of the strand and 5' 32P-labeled and analyzed in are indicated on the kanR frequencies in the indicated strains carrying indicated mutation reporters in presence of inducer (50 nM [ATc]). Results shown are means (± SEM) of data obtained from biological replicates symbolized by gray dots. Stars above the means mark a statistical difference with the reference strain (empty) (*, p<0.001). Relative (pie chart) and absolute (bar chart) frequencies of nucleotide changes detected in kan of kanR cells represented with colors: pink=–1 or +1 frameshift in the homo-oligonucleotide run, or +1 localized outside of the run, localized outside of the run, substitution mutation and gray=no detected mutation. The number of sequenced kanR colonies is given in the center of each pie chart. Empty=empty vector, promoter, DinB2=M. smegmatis DinB2, DinB3=M. smegmatis DinB3. Figure data 1 Original autoradiograms (Figure 4A and B). Download the template or to 6 or 8 nucleotides enhanced the slippage activity of DinB2, generating to Although the very slippage products were or by of or DinB2 to long products by iterative slippage that tracts of to or (Figure 4A and B). The finding that DinB2 is more during DNA synthesis a template G run on A runs or T runs of length the that and of the on the template homo-oligonucleotide run is