IntroductionGenomic DNA is copied by DNA polymerase enzymes. Cells possess myriad polymerases that coordinate to ensure timely and accurate duplication of the genetic material. In eukaryotes, bulk replication progresses as a replication fork structure, where Pol ε synthesises the continuous leading strand while Pol α and Pol δ coordinate Okazaki fragment synthesis on the lagging strand (reviewed in ref. 1). Other polymerases play more specialised roles, offering flexibility to tolerate complications that may arise during replication, such as DNA damage, and are therefore crucial mediators of genomic stability. The most well-studied of these is Pol η, encoded by the POLH gene, which is mutated in Xeroderma Pigmentosum Variant (XPV) patients2,3. Pol η was primarily identified as the enzyme bypassing UV-induced damage, particularly cyclobutane pyrimidine dimers (CPD), with high accuracy via translesion synthesis (TLS)4,5,6. Later studies showed that Pol η’s enlarged active site also enables accurate TLS over cisplatin-induced lesions7,8,9. Pol η is also the key DNA polymerase responsible for A/T mutations during somatic hypermutation in B cells, acting downstream of cytosine deaminase–induced mismatch repair10,11. Interestingly, Pol η mRNA is comparably expressed across multiple tissues12, not solely in sun-exposed skin or immune cells (Supplementary Fig. 1a), suggesting functions in replication beyond TLS. Previous observations support this, with Pol η travelling with the replication fork13 and forming foci in unperturbed cells14. Indeed, additional roles of Pol η have been identified, for instance, Pol η is required to maintain common fragile site (CFS) stability, even in the absence of exogenous stress15. These structure-forming repeat sequences are inefficiently replicated by Pol δ, but well tolerated by Pol η, which is recruited to these sites16,17. The use of Pol η at these regions of non-B DNA reduces fork pausing18. These roles highlight the versatility of Pol η during DNA synthesis, a property that may also help cells tolerate replication stress associated with oncogenic transformation (reviewed in ref. 19).Given the diverse functions of Pol η and other error-prone polymerases, elucidating their roles in genome replication and their coordination with replicative polymerases remains a key focus. To examine polymerase coordination during human replication, we recently established the polymerase usage sequencing (Pu-seq) assay in HCT116 cells20. Analysis of Pol ε and Pol α usage in this system revealed unexpected deviations from canonical patterns in specific regions, influenced by transcribed genes and large-scale chromosomal domains20. These observations hint at the involvement of other polymerases, which were not included in the analyses. To date, global analyses of error-prone polymerase usage in eukaryotes are limited to budding yeast Pol η, which was determined to contribute to lagging-strand synthesis21. However, the yeast genome is ~300-fold smaller than that of humans and lacks distinct chromosomal domains. Applying Pu-seq in human cells, therefore, provides an opportunity to determine roles of polymerases that may be overlooked in the yeast genome. Originally developed in fission yeast22,23, Pu-seq maps polymerase activity by detecting ribonucleotides (rNMPs or rNTPs) incorporated into nascent DNA by mutant polymerases with reduced nucleotide selectivity. To prevent removal of these rNMPs, human Pu-seq assays typically employ the auxin-inducible degron 2 (AID2) degron system to degrade RNaseH2A, thereby inactivating the ribonucleotide excision repair (RER) pathway during the experiment24. In principle, Pu-seq can be applied to any polymerase, provided a mutant can be generated that incorporates rNMPs at a higher rate than the wild-type (WT) protein. Therefore, we set out to establish a Pol η Pu-seq system to map Pol η synthesis events and investigate its role during human DNA replication.ResultsThe Pol η-F18A mutation increases ribonucleotide incorporationTo map polymerase usage with Pu-seq, a polymerase mutant with an increased propensity to incorporate rNMPs is required. The specificity of polymerases for deoxy-ribonucleotides (dNTPs) over the more abundant rNTPs is primarily governed by a steric gate residue (tyrosine or phenylalanine in B and Y family polymerases), which sterically excludes the occupancy of nucleotides with a 2′-hydroxyl moiety25. In budding yeast Pol η, strict rNTP discrimination is predominantly governed by F35, and substituting this residue for alanine (Pol η-F35A) dramatically increases rNTP incorporation in vitro26, which has been used to map Pol η usage in budding yeast21. Although human Pol η is less selective than its yeast counterpart27, the steric gate residue (F18) demonstrates high sequence and structural conservation with yeast (Fig. 1a, Supplementary Fig. 1b). Structural models indicate that steric hindrance between this residue and an incoming rNTP distances the α-phosphate from the nascent strand terminus, rendering phosphodiester bond formation inefficient27. We therefore selected the Pol η-F18A mutation as a candidate for use in Pu-seq assays to investigate how Pol η is utilised in human cells.Fig. 1: Human Pol η-F18A mutant incorporates ribonucleotides during DNA synthesis.Full size imagea Structural comparison of Pol η active sites from Saccharomyces cerevisiae: S.c (PDB: 3MFH) and Homo sapiens: H.s. (PDB: 3MR2). Conserved steric gate phenylalanine residues (red) sterically hinder the 2′-hydroxyl group of incoming nucleotides (yellow sticks). Divalent metal ions (red spheres) and the primed DNA template (pink) are displayed. b Polymerase activity of Pol η-F18A over a CPD lesion. A γ-32P-labelled 13-mer primer annealed to a 30-mer template containing a cis-syn thymidine dimer was extended by recombinant His-tagged Pol η or Pol η-F18A in the presence of dNTPs and MgCl2. c Relative efficiency of rNTP versus dNTP incorporation by Pol η and Pol η-F18A. Relative efficiency was calculated as [kcat/Km rNTP]/[kcat/Km dNTP] for each base, as determined by single-nucleotide incorporation assays. d Sensitivity of Pu-seq cells to 7 days of 5-Ph-IAA treatment. Survival was assessed by crystal violet staining quantification and normalised to untreated controls. Data represent the mean ± standard deviation from n = 3 independent experiments. e Representative western blots confirming RNaseH2A-mAC degradation after 48 h of 5-Ph-IAA treatment (top) and RNaseH2A-dTAG degradation after 48-h dTAGV-1 treatment (bottom). f DNA fragmentation following NaOH hydrolysis of DNA harvested from cells after 48 h of 5-Ph-IAA treatment. g Strand-specific polymerase usage of Pol ε, Pol α and Pol η across a representative region of Chromosome 3. Red: Watson strand. Blue: Crick strand. Data were smoothed with a moving average (m = 30; see Methods). h Heatmap of Pearson correlations between genome-wide strand-specific profiles of Pol ε, Pol α and Pol η usage. Red: positive. Blue: negative. Source data are provided as a Source Data file.To evaluate the effect of the F18A mutation, recombinant Pol η was expressed and purified from Escherichia coli for in vitro characterisation (Supplementary Fig. 2a). Human Pol η-F18A retained polymerase activity with slightly reduced processivity compared to the wild-type enzyme, while performing TLS across a CPD lesion with similar efficiency (Fig. 1b). To assess rNTP incorporation, the steady-state kinetic parameters for single-nucleotide incorporation were determined (Table 1, Supplementary Fig. 2b). Calculation of catalytic efficiency (kcat/Km) revealed a striking difference in nucleotide differentiation between the WT and mutant enzymes. The WT enzyme exhibited much higher efficiency in incorporating dNTPs than rNTPs. In contrast, the Pol η-F18A mutant exhibited a much less distinct preference, with altered kinetic parameters for both dNTPs and rNTPs. Specifically, the mutant enzyme exhibited an increase in Km for dNTPs and a decrease in Km for rNTPs, accompanied by an elevation in kcat for rNTPs, dramatically enhancing the relative efficiency of rNTP incorporation compared to the WT enzyme (Fig. 1c). These results demonstrate that the Pol η-F18A mutation alters nucleotide incorporation dynamics, which could serve as a detectable signature of Pol η activity in human cells.Table 1 Steady-state kinetic parameters for one-base incorporation by WT and F18A Pol ηFull size tableEstablishing Pol η-F18A cell linesHaving established that Pol η-F18A incorporates rNTPs more efficiently than the WT protein, we introduced this mutation into both alleles of the endogenous POLH gene in RNASEH2A-degron HCT116 cells using CRISPR-Cas9 (Supplementary Fig. 3a). The mutation did not affect Pol η protein levels (Supplementary Fig. 3b) or cell proliferation rate (Supplementary Fig. 3c). Pol η-F18A cells also exhibited a cell cycle profile similar to control (Pol+) cells, even in the presence of 5-Ph-IAA, an auxin compound that induces degradation of the RNaseH2A subunit via the AID2 system (Supplementary Fig. 3c, d).To assess the functionality of the polymerase mutation in cells, we examined cell fitness under conditions promoting chronic persistence of rNMPs in the genome. We previously demonstrated that short-term ( ~ 48 h) depletion of RNaseH2A had no significant impact on cell growth, even in cells harbouring rNMP-incorporating replicative polymerases (Pol ε-M630F and Pol α-Y865F). In contrast, long-term (7 days) 5-Ph-IAA exposure decreased survival of RNASEH2A-degron cells by ~20% (Fig. 1d). The effect was more severe in Polα-Y865F or Polε-M630F cells, which lost viability, consistent with a high rNMP incorporation rate from the mutated replicative polymerases. Pol η-F18A cells also exhibited reduced viability, though to a lesser extent, which was not observed in Pol η knockout (KO) cells (Supplementary Fig. 3b), reflecting moderate rNTP incorporation by this polymerase.Given the role of Pol η in DNA damage tolerance, we next assessed cell survival after UV exposure. Despite the TLS proficiency of Pol η-F18A (Fig. 1b), Pol η-F18A cells displayed high UV sensitivity, exceeding that of Pol η-KO cells (Supplementary Fig. 3e). These findings mirror results in budding yeast26, suggesting rNMP incorporation opposite or near UV lesions is deleterious for cell survival.Detection of genome-wide Pol η usage via Pu-seqFor Pu-seq assays, cells were harvested 48 h after degrading RNASEH2A to levels barely detectable by western blot (Fig. 1e) and genomic DNA was extracted. Genomic DNA was alkaline-treated to hydrolyse rNMPs, and single-stranded DNA was separated by electrophoresis. Unlike replicative polymerase mutants, which produce a marked increase in small DNA fragments20, the electrophoresis patterns for Pol η-F18A and control (Pol+) samples showed no obvious differences (Fig. 1f). This indicates low levels of rNMP incorporation, consistent with the moderate viability defect observed during RNaseH2A depletion (Fig. 1d). For Pu-seq, alkaline-induced small DNA fragments ( G mutations (n = 333,228) relative to replication timing (RT)73 in gastric adenocarcinoma patient samples (n = 31). c Average R.A.S. at most efficient 25% of IZs (see definition in Fig. 2c: WL represents the number of A > G changes on the Watson strand to the left of IZs, while CL is the number of A > G changes on the Crick strand) of A-to-G mutations in individual patients, scaled and grouped by cancer type. Sample numbers for each cancer type are shown below the horizontal axis. Boxes: 25th–75th percentiles, centre lines: medians, whiskers: minima and maxima within 1.5× the interquartile range. d Fraction of A > G mutations in late-replicating regions (bins 12-16) relative to all regions in individual patients, scaled to z values and grouped by cancer type. Sample numbers for each cancer type are shown below the horizontal axis. Boxplots are shown in the same manner as in (c). e Correlation between gene expression (TPM) and the fraction of A > G mutations occurring in late-replicating domains. Points represent mean values for each cancer type. Grey shaded area: 95% confidence interval for the fitted linear regression line. p value: a two-sided Spearman’s rank correlation test without adjustment for multiple comparisons.Given this variation in late-replication bias, we next asked whether these patterns were associated with the expression of genes related to Pol η function. To investigate this, we correlated mRNA expression levels with the mutational distributions. POLH expression itself did not correlate with the A-to-G mutation enrichment in late-replicating regions across cancer types when taking the mean of each cohort (Fig. 6e), consistent with previous work showing that overexpression of Pol η alone does not increase spontaneous mutation rates40. By contrast, the expression of PCNA and both RAD18 and UBE2A (the E3 and E2 ligases responsible for PCNA-K164 ubiquitylation) correlated positively with enrichment of A-to-G mutations in late-replicating regions. This association is consistent with PCNA ubiquitylation-dependent error-prone synthesis at these regions. To assess the specificity of this association, we extended the analysis to other substitutions. RAD18 and UBE2A expression correlated with the late-replication enrichment of substitution classes arising from A/T bases, but not those from C/G bases (Supplementary Fig. 10). Thus, the association was not specific to A-to-G mutations, although its restriction to substitutions at A/T bases is broadly compatible with the mutational preferences of Pol η observed in vitro39. However, it is important to note that these results do not rule out the involvement of other Y-family error-prone polymerases, which are also recruited via PCNA ubiquitylation.Expression of Y-family polymerases Pol κ and Pol ι, which also interact with ubiquitylated PCNA, negatively correlated with A-to-G mutation enrichment in late-replicating regions (Fig. 6e). Because our findings suggest competition between the error-prone polymerases for access to PCNA-bound nascent termini (Fig. 2d, f, Supplementary Fig. 6b, c), increased expression of one polymerase could reduce the contribution of another. Since error-prone polymerases can be regulated in a transcription-dependent manner41,42 and Pol κ has been implicated in early S phase replication43, competition among polymerases may influence the timing and genomic distribution of mutation occurrence. Together, these associations identify mutation distributions consistent with PCNA-regulated error-prone polymerase activity during late replication, suggesting that the dynamics of polymerase usage may contribute to mutational patterns across S phase.DiscussionHere we establish a method to directly monitor Pol η synthesis in human cells and identify two features of genome replication that correlate with Pol η function under unperturbed conditions: (i) lagging-strand replication and (ii) replication timing. While the former has been observed for the budding yeast Pol η21 (Rad30) and the E. coli homolog (Pol V) under SOS response44, this work provides evidence that the lagging-strand bias is conserved in mammalian cells. These findings suggest that the association of Pol η with the lagging strand is either an evolutionarily conserved property or an inherent consequence of discontinuous lagging strand synthesis. In contrast, the bias towards late-replicating regions appears to be specific to large and complex genomes, since such domains are absent in compact genomes, underscoring the importance of studying replication in mammalian systems to capture the full complexity of eukaryotic replication dynamics.Stochastic Pol η usage at replication forksThe eukaryotic replisome exhibits substantial heterogeneity due to the stochastic nature of multiprotein assemblies, providing opportunities for concentration-dependent polymerase exchange45. To evaluate potential roles for Pol η in lagging-strand synthesis, we consider how intrinsic mechanisms of leading and lagging-strand replication influence polymerase usage. As a core component of the replisome, Pol ε is directly coupled to the CMG helicase. In contrast, Pol δ maintains only a weak association with DNA and can readily dissociate from PCNA46,47,48. Because PCNA dissociation from DNA is relatively slow46, Pol δ release can leave a PCNA-bound primer-template end, which represents a potential substrate for error-prone polymerases (Fig. 7). Pol η travels with the replication fork13 and interacts with PCNA through three redundant PIP-box motifs32, providing a basis for competition to access nascent DNA termini. Pol η may also access lagging strand intermediates at an earlier stage, as its ability to extend RNA–DNA junctions in vitro suggests that it could extend RNA primers following inefficient primase-Pol α handoff49 (Fig. 7). Previous studies have shown that Okazaki fragment intermediates can trigger PCNA ubiquitylation to promote their synthesis and maturation50,51,52. Consistent with this model, we found that PCNA ubiquitylation enhances the lagging-strand bias of Pol η. Because PCNA-bound Pol η can itself promote PCNA ubiquitylation, which further increases its local accumulation32, this interaction may represent a local feedback mechanism that increases Pol η recruitment. Thus, elevated polymerase turnover facilitated by PCNA on the lagging strand may increase the likelihood that error-prone polymerases, including Pol η, stochastically engage and extend exposed DNA ends. We also note that Pol η can function in post-replicative gap filling, thereby enabling post-replicative repair to complete lagging strand replication (Fig. 7).Fig. 7: Roles of Pol η during lagging strand DNA synthesis.Full size imageAt the replication fork, the MCM helicase complex (grey) unwinds DNA. Pol ε (orange) synthesises the leading strand continuously while tethered to the MCM. On the lagging strand, Pol α-primase (pink) synthesises short RNA-DNA primers (light blue), which Pol δ (blue) usually extends. Pol η (purple) may also extend from RNA or DNA primers. Pol δ, Pol η and Pol κ (green) all associate with PCNA (turquoise), suggesting they compete for binding and access to the lagging strand termini. Formation of ssDNA gaps between Okazaki fragments triggers PCNA ubiquitylation: Ub (dark blue), enhancing Pol η recruitment for gap filling.However, it remains unclear whether Pol η actively serves as a designated backup for Pol δ or whether its usage simply reflects stochastic polymerase exchange on PCNA. Notably, unperturbed Pol η-KO HCT116 cells showed no overt defects, potentially because other error-prone polymerases, including Pol κ, Pol ι or Pol ζ, can redundantly support Okazaki fragment synthesis. In support of this model, Pol η usage on the lagging strand increased in Pol κ-deficient cells. This redundancy may obscure the contribution of Pol η to unperturbed replication. Indeed, when lagging strand synthesis was strongly perturbed, loss of Pol η increased ssDNA-associated replication stress, indicating that Pol η contributes to limiting ssDNA accumulation under these conditions (Fig. 2e). Previous studies have similarly shown that Pol η and Pol κ become more important during cell cycle dysregulation and replication stress, including that induced by oncogene activation, which elevates fork stalling and PCNA ubiquitylation53,54. Thus, error-prone polymerases appear to provide a flexible option whose contribution to replication increases under stress. As a future direction, we are extending Pu-seq to various error-prone polymerases to define their contributions to replication in human cells.Pol η usage under unperturbed conditions is unlikely to reflect lesion bypassStochastic dynamics at replication forks provide a simple explanation for Pol η engagement during unperturbed replication, but an alternative is that the observed lagging-strand bias reflects asymmetric tolerance of endogenous lesions. Analyses of large-scale mutagenesis in human genomes revealed elevated instances of error-prone synthesis on the lagging strand and similar mutagenic spectra on the lagging and non-transcribed strands, suggesting a shared error-prone bypass pathway in both contexts55. In our study, Pol η activity exhibited a clear transcriptional strand bias only after UV irradiation, with no bias observed under unperturbed conditions (Fig. 4d). This pattern implies that, under normal growth conditions, Pol η is not preferentially used at DNA lesions recognised by TC-NER. Instead, its activity predominantly reflects synthesis on undamaged templates. We therefore favour the interpretation that the lagging strand bias of Pol η usage is largely independent of DNA damage tolerance and does not imply biased usage during DNA damage bypass. Consistent with this interpretation, damage-induced mutagenesis during mouse liver tumorigenesis produces TLS-dependent clustered mutations with no discernible leading and lagging strand asymmetry56.Despite observing transcriptional strand asymmetry after UV exposure, we detected no substantial change in the replicative strand bias of Pol η synthesis relative to untreated cells. We interpret this as evidence that, under mild UV exposure (5 J/m², a dose that permits continued cell cycle progression), TLS over UV-induced lesions constitutes a minor contribution versus the predominant replication-associated Pol η signal. Taken together, our data support a model in which Pol η usage is largely governed by the stochastic architecture of replication forks.PCNA ubiquitylation-dependent polymerase usage correlates with mutagenesis during late replicationPol η usage was biased towards late-replicating regions of the genome enriched with the repressive histone modification H3K9me3, characteristic of constitutive heterochromatin. These regions lack defined replication initiation zones and do not exhibit defined replication fork directionality20,30. In our previous Pu-seq analyses with replicative polymerases Pol ε and Pol α, these regions showed no reciprocal strand asymmetry20. While this could represent random initiation of DNA replication, it could also reflect distinct polymerase assignment relative to other genomic domains.In addition to accessibility, late-replicating heterochromatin differs from early-replicating euchromatin in several respects. For instance, replication forks progress more rapidly through these domains57. Elevated MCM unwinding rates at higher fork speeds may increase the spatial separation between lagging-strand priming and synthesis, raising the likelihood that synthesis becomes uncoupled from the replisome. Since PCNA ubiquitylation has been implicated in post-replicative gap filling58,59, such uncoupling could leave ssDNA gaps that recruit Pol η in a PCNA ubiquitylation–dependent manner (Fig. 7). Supporting this model, recent single-molecule imaging in unperturbed cells revealed chromatin-associated Pol η spatially distinct from sites of active DNA synthesis, with increased foci observed in cells with high DNA content (i.e. late S or G2 phase)60. Moreover, error-prone Pol ζ (REV3L) has been implicated in facilitating replication through heterochromatin61, suggesting cooperative roles for these polymerases in mitigating under-replication (ssDNA gaps, etc.) in compact, late-replicating chromatin domains. Although the precise molecular mechanisms remain to be defined, accumulating evidence supports the notion that replication dynamics vary over the course of S phase in human cells.Late-replicating regions are associated with high mutation loads in various biological systems (e.g. comparisons of human and chimpanzee genomes62, and analyses of lymphoblastoid cell lines63), yet the underlying mechanisms remain incompletely understood. Consistent with PCNA ubiquitylation-dependent Pol η usage in late-replicating regions, we observed positive correlations between the enrichment of A/T-derived substitutions in these domains in cancer genomes and the expression of RAD18 and UBE2A, key components of the PCNA ubiquitylation pathway. Our analysis of cancer mutations is consistent with RAD18-dependent PCNA ubiquitylation promoting the usage of error-prone polymerases in these regions, which could contribute to the elevated mutation load in late-replicating domains (Supplementary Fig. 10). Supporting a potential mechanistic link, RAD18 overexpression enhanced the lagging-strand bias of Pol η usage, demonstrating that RAD18 levels can influence Pol η usage patterns. However, it did not further increase the enrichment of Pol η in late-replicating regions. Because RAD18 was overexpressed independently of cell cycle stage in our experiments, elevated RAD18 may promote Pol η usage more broadly throughout S phase rather than specifically during late replication. Notably, variation in RAD18 expression is known to correlate with tumour aggressiveness and poor clinical outcomes (reviewed in ref. 19).In addition to monoubiquitylating PCNA, RAD18 promotes recruitment of Pol η to replication forks through a physical interaction and by stimulating its SUMOylation independently of RAD18 ubiquitin-ligase activity13,64. We therefore cannot conclusively determine whether altered Pol η usage following RAD18 overexpression results from increased PCNA ubiquitylation or a combination of these mechanisms. Given the critical role of PCNA ubiquitylation in post-replication repair58, this pathway may represent an unavoidable trade-off of this last resort mechanism for ensuring timely completion of genome replication, at the expense of mutagenesis. While our mutational associations are broadly compatible with the biochemical preferences of Pol η, we must acknowledge that other polymerases or processes may generate similar patterns. Establishing the specific contribution of Pol η will require controlled mutation-accumulation experiments to determine whether Pol η loss alters the mutational spectrum or burden of late-replicating regions. Such experiments represent an important direction for future research.Comparison with other methodologiesDetecting polymerase usage with Pu-seq offers a distinct advantage, as it captures nucleotide incorporation directly in vivo, generating high-resolution, strand-specific maps of genome-wide activity. Here, we demonstrate that Pu-seq is sufficiently sensitive to detect the activity of low-abundance polymerases such as Pol η, a sensitivity that is difficult to achieve with conventional methods. For example, our attempt to profile Pol η enrichment in late-replicating regions by ChIP-seq of endogenously tagged 3×FLAG-Pol η failed to detect enrichment over the untagged control. This likely reflects a combination of Pol η’s low abundance, its transient or weak DNA association and its lack of strong site-specific localisation (unpublished data). These limitations highlight the advantage of directly measuring DNA synthesis events for detecting low-abundance polymerase activity in vivo. A deeper understanding of the roles of these polymerases will elucidate the mechanisms safeguarding genome stability and may help explain how the mutational landscape is shaped during human genome replication.MethodsProtein purificationN-terminally hexahistidine-tagged WT Pol η (his-Pol η) and its F18A mutant were purified as follows. E. coli BL21(DE3) harbouring pET-h6-POLH65 and its F18A mutant derivative were grown in 2 L and 4 L of LB medium, respectively, containing 250 μg/mL ampicillin at 15 °C to an OD600 of 0.6, and then induced with 0.2 mM isopropyl β-D−1-thiogalactopyranoside for 6.5 h. Cell pellets were resuspended in 2 mL/g lysis buffer (50 mM HEPES-NaOH (pH 7.5), 0.1 mM EDTA, 10 mM β-mercaptoethanol and 500 mM NaCl), frozen in liquid nitrogen and subsequently thawed in ice water. Cells were lysed by addition of the lysis buffer supplemented with 100 mM spermidine (pH 7.5), 4 mg/mL lysozyme and 1 mM phenylmethylsulfonyl fluoride at one-sixth volume of the initial lysis buffer, and then by incubation on ice for 30 mins, in a 37 °C water bath for approximately 1 min with continuous mixing, and then on ice for a further 30 mins. Lysates were clarified by centrifugation at 53,000 × g for 30 mins at 0 °C and further clarified by a second centrifugation at 53,000 × g for 1 h at 0 °C, and then 1/19 volume of 1 M imidazole (pH 7.5) was added to the clarified lysate. Proteins were purified at 4 °C by nickel-affinity chromatography using a 1-mL HiTrap Chelating column (Cytiva) pre-charged with 100 mM NiSO4 and equilibrated with buffer A (50 mM HEPES-NaOH (pH 7.5), 10 mM β-mercaptoethanol and 10% glycerol) containing 500 mM NaCl and 50 mM imidazole (pH 7.5). The column was washed with the equilibration buffer, and proteins were eluted using buffer A containing 500 mM NaCl and 100 mM imidazole (pH 7.5). Protein-containing fractions were pooled, diluted with buffer A to 300 mM NaCl for WT Pol η or 150 mM NaCl for the F18A mutant, and applied to a 1-mL HiTrap Q FF column (Cytiva) equilibrated with buffer A containing 300 mM NaCl for WT Pol η or 150 mM NaCl for the F18A mutant. The flow-through fraction was applied to a 1-mL HiTrap Heparin HP column (Cytiva) equilibrated with buffer A containing 300 mM NaCl for WT Pol η or 150 mM NaCl for the F18A mutant. The column was washed with the equilibration buffer, and the proteins were eluted with a linear gradient of 300-800 mM NaCl in buffer A for WT Pol η or 150–650 mM NaCl in buffer A for the F18A mutant. Protein-containing fractions were pooled and applied onto a Superdex 200 10/300 GL column (GE Healthcare Life Science) equilibrated with buffer A containing 300 mM NaCl. The peak fraction was frozen in liquid nitrogen and stored at −80 °C.Primer extension assaysNucleotide sequences of oligonucleotide templates and the primer are provided in Supplementary Table 1. The CPD-containing template66 was purchased from Operon Biotech. The primer was labelled using polynucleotide kinase (New England Biolabs) and [γ-32P] ATP (Revvity Health Sciences, Inc. BLU502Z) and subsequently annealed to the templates. The standard reaction mixture (10 μL) contained 20 mM HEPES-NaOH (pH 7.5), 50 mM NaCl, 0.2 mg/mL bovine serum albumin (BSA), 5 mM DTT, 5 mM MgCl2, 100 μM of each dNTP, 1 pmol primer/template and the indicated amounts of his-Pol η. Reactions were incubated at 30 °C for 5 mins and terminated by adding 10 μL of stop solution (30 mM EDTA, 94% formamide, 0.05% bromophenol blue, 0.05% xylene cyanol). Reaction products were resolved on 8% polyacrylamide gels containing 7 M urea and visualised and analysed using a Typhoon FLA 7000 (GE Healthcare). Kinetic assays were performed at 30 °C for 5 mins in 10 μL standard reaction mixture in the presence of one dNTP or rNTP at the indicated concentrations. Reaction products were resolved on 20% polyacrylamide gels containing 7 M urea. Km and kcat were determined from plots of initial velocity versus NTP concentration, obtained from three or four independent experiments using a hyperbolic curve-fitting program (OriginLab Corporation).Cell cultureCell lines derived from HCT116 cells (ATCC #CCL-247) or hTERT RPE-1 (ATCC #CRL-4000) were cultured in McCoy’s 5 A (modified) Medium (Gibco) or DMEM/F-12 (Gibco), respectively, supplemented with 10% fetal bovine serum (FBS, Gibco), 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin. Cells were maintained at 37 °C in a humidified atmosphere containing 5% CO2. 5-Ph-IAA was prepared as a 250 μM stock solution in dimethyl sulfoxide (DMSO) and stored at −20 °C. dTAGV−1 (R&D Systems) was prepared as a 500 μM stock in DMSO and stored at −20 °C. Drugs were added directly to the culture medium to achieve final working concentrations of 0.4 μM (5-Ph-IAA) and 0.5 μM (dTAGV−1).Generation of endogenous gene mutationsCRISPR-associated (Cas) proteins (Alt-RTM A.s. Cas12a (Cpf1) Ultra or Alt-RTM S.p. HiFi Cas9 Nuclease V3) were purchased from IDT (Integrated DNA Technologies, USA). For modifications with CRISPR-Cas12a, ribonucleoprotein (RNP) complexes were assembled by combining equal volumes of 63 μM Cas12a protein and 75 μM Alt-RTM A.s. Cas12a crRNA (IDT; Custom production; Supplementary Table 1). For modifications with CRISPR-Cas9, guide RNA (gRNA) molecules were assembled by mixing 50 μM Alt-RTM CRISPR-Cas9 crRNA (IDT; Custom production; Supplementary Table 1) with 50 μM Alt-R™ CRISPR-Cas9 tracrRNA (IDT) in Nuclease-Free Duplex Buffer (IDT), heating at 96 °C for 5 mins, then cooling to RT. RNP complexes were assembled by combining 125 pmol Cas9 enzyme with 150 pmol gRNA. For electroporation, 1 × 105 cells were mixed with 1.5 μL RNP and either 1.5 μL single-stranded Alt-R™ HDR Donor Oligo (IDT; Custom production; Supplementary Table 1) or 2 μL of 500 nM double-stranded Alt-R™ HDR Donor Block (IDT; Custom production; Supplementary Table 1) to 15 μL in R buffer. Electroporation was performed using the Neon Transfection System (Invitrogen) under the following conditions: HCT116: one pulse of 1,530 V for 20 ms; RPE1: two pulses of 1,350 V for 20 ms. Cells were plated in media containing 1 μM Alt-R™ HDR Enhancer V2 (IDT) for 24 h, then medium was replaced and cells were incubated for a further 48 h at 37 °C to allow recovery. Subsequently, individual clones were isolated, expanded and screened for the genetic modifications.Generation of gene knockout mutationsFrameshift mutations were introduced into genes using the CRISPR-Cas9 system. Specifically, Cas9 and a gRNA (Supplementary Table 1) targeting POLH (exon 3, amino acid 83/713), POLK (exon 5, amino acid 164/870), or REV1 (exon 5, amino acid 232/1250) were expressed from the pSpCas9 vector67. The plasmid was transfected into cells using FuGENE HD (Promega) according to the manufacturer’s instructions. After 24 h, puromycin (Gibco) was added to a final concentration of 0.25 μg/mL to select for transfected cells. Upon complete selection, individual clones were isolated, expanded, and screened for the presence of knockout mutations. Gene knockouts were verified by western blotting.Generation of RAD18 overexpression cellsTo establish cells expressing the tet repressor (tetR), pcDNA6/TR was linearised with FspI (New England Biolabs), and 500 ng was introduced into Pu-seq-competent cells by electroporation using the Neon Transfection System (Invitrogen; one pulse at 1,530 V for 20 ms). After 72 h, cells were seeded for clonal isolation in Blasticidin S (InvivoGen) to a final concentration of 10 μg/mL. Following expansion, resistant clones were screened for tetR expression.The RAD18 CDS was synthesised as a gBlocks HiFi Gene Fragment (IDT; custom production) and cloned into pcDNA4/TO vector using In-Fusion cloning. The resulting plasmid was linearised with ScaI-HF (New England Biolabs) and 200 ng was transfected into cells using FuGENE HD (Promega) according to the manufacturer’s instructions. After 48 h, Zeocin (InvivoGen) was added to a final concentration of 100 μg/mL to select for transfected cells. Upon complete selection, individual clones were isolated, expanded, and screened for doxycycline-inducible RAD18 expression.Cell growth assayCells were seeded in a 6-well plate and allowed to adhere for 48 h. Following attachment, cells were treated with 0.4 μM 5-Ph-IAA, as required. At the indicated timepoints, cells were harvested and counted using a Countess 3 Automated Cell Counter (Invitrogen). To assess relative cell growth, cell counts at each time point were normalised to the cell number at 0 h.Cell survival assayCells (2 × 103 per well) were seeded in a 6-well plate and allowed to adhere for 48 h. Following attachment, treated cells were either exposed to the indicated dose of UV radiation using a crosslinker (UVP CX-2000) or continuously grown in 0.4 μM 5-Ph-IAA. Colonies were allowed to form for approximately 10 (UV) or 7 (5-Ph-IAA) days, with culture medium and drugs replaced as needed. To assess cell survival, colonies were stained with a 0.5% crystal violet solution (0.5% crystal violet in 20% methanol) for 20 mins. Excess stain was removed by washing with Milli-Q H2O, and plates were dried overnight. For quantification, the crystal violet stain was dissolved in 33% acetic acid, and absorbance at 590 nm was measured using a plate reader (Thermo Scientific). Relative cell survival was calculated by comparing the treated conditions to the untreated control.Cell synchronisation with nocodazoleCells (1 × 105 per well) were seeded in a 6-well plate and allowed to adhere for 48 h. Synchronisation in mitosis was achieved by treatment with 50 ng/mL nocodazole for 15 h. To release cells from arrest, cultures were washed twice with fresh medium. Samples were harvested at the indicated timepoints and processed for analysis by western blotting and flow cytometry.Chromatin extractionCells were harvested and washed once with ice-cold PBS. The pellet was resuspended by gentle pipetting in 50 µL of 0.2% Triton X-100 in phosphate-buffered saline (PBS) supplemented with protease inhibitor and incubated on ice for 5 mins with occasional gentle mixing. Lysates were centrifuged at 17,800 × g at 4 °C for 10 mins and the supernatant was collected as the soluble fraction. The pellet was resuspended in 200 µL of 0.2% Triton X-100 in PBS with protease inhibitor, underlayered with 200 µL of 30% (w/v) sucrose and centrifuged at 17,800 × g at 4 °C for 10 mins. The supernatant was carefully aspirated and the sucrose separation step was repeated. The final pellet was resuspended in 50 µL of 0.2% Triton X-100 in PBS with protease inhibitor and treated with benzonase (Millipore 70746) on ice for 15 mins to yield the chromatin fraction. Samples were mixed with 4× Bolt LDS Sample Buffer and 10× Bolt Reducing Agent and heated for 10 mins at 70 °C.Western blottingCells were lysed in RIPA buffer (Sigma-Aldrich) supplemented with protease inhibitors and sonicated (30 sec on, 30 sec off for 15 cycles) using a Bioruptor II (BMBio). Following sonication, lysates were centrifuged at 10,000 × g at 4 °C for 20 mins. The supernatant was collected and mixed with 4× BoltTM LDS sample buffer (Invitrogen) and 10× BoltTM Sample Reducing Agent (Invitrogen). Samples were incubated at 70 °C for 10 mins and proteins were separated by SDS-PAGE on BoltTM 8% or 4–12% Bis-Tris Plus gels (Invitrogen). Following electrophoresis, proteins were transferred to Immun-Blot PVDF membranes (Bio-Rad) using a Power Blotter XL (Invitrogen). Membranes were blocked with 5% milk in PBS containing 0.1% Tween-20 (PBST). Membranes were incubated with primary antibodies at 4 °C overnight in 0.5% milk in PBST. Secondary antibody incubations were performed at room temperature for 1 h in 0.5% milk in PBST. Detection was conducted using Western Lightning® Plus-ECL (Perkin Elmer), and images were acquired with a ChemiDocTM Touch MP Imaging system (Bio-Rad).The following antibodies were used: anti-RNASEH2A (dilution 1:2000, Bethyl Laboratories® A304-149A), anti-POLH (dilution 1:2000, Bethyl Laboratories A301-231A), anti-POLK (dilution 1:2000, Bethyl Laboratories A301-977A), anti-PCNA (dilution 1:1000, Sigma-Aldrich MAB424), anti-Ubiquityl-PCNA (Lys164) (dilution 1:2000, Cell Signalling Technology D5C7P), anti-Rad18 (dilution 1:2000, Cell Signalling Technology D2B8), anti-REV1 NTD (dilution 1:2000, Santa Cruz Biotechnology sc-393022), anti-REV1 CTD68 (dilution 1:2000), anti-FLAG (dilution 1:2000, Sigma-Aldrich F1804), anti-α-tubulin (dilution 1:5000, Sigma-Aldrich T5168), anti-Histone H3 (dilution 1:2000, Abcam ab1791), anti-NAK/TBK1 (dilution 1:2000, Abcam ab40676), anti-mouse IgG HRP conjugate (dilution 1:10,000, Agilent P026002-2), anti-rabbit IgG HRP conjugate (dilution 1:10,000, Agilent P044801-2) and anti-guinea pig IgG HRP conjugate (dilution 1:5,000, Millipore AP108P).Flow cytometryCells were harvested, fixed in 70% ethanol, and stored at −20 °C for 1-7 days. Prior to analysis by flow cytometry, cells were washed with PBS containing 1% BSA and resuspended in 1% BSA/PBS containing 40 μg/mL propidium iodide (PI) and 50 μg/mL RNase A. The suspension was incubated for 30 mins at 37 °C, then on ice for two mins. Samples were analysed on a BD AccuriTM C6 Plus Flow Cytometer (BD Biosciences) and approximately 1 × 105 cells were analysed using FlowJoTM software (BD Biosciences).ImmunofluorescenceCells (3 × 105 per well) were seeded onto sterilised glass coverslips in a 6-well plate and allowed to adhere for 48 h. Cells were then treated with either 1 µM Pol α inhibitor ST1926 (POLAi; MedChemExpress HY-14808) or an equivalent volume of DMSO for 50 mins. Cells were washed with Tris-buffered saline (TBS), fixed with 2% paraformaldehyde in TBS for 10 mins at room temperature, then permeabilised with 0.2% Triton X-100 in TBS for 10 mins. After washing with TBS, cells were blocked with 3% BSA in TBS for 10 mins and incubated with anti-phospho-RPA32 (Ser33) antibody (dilution 1:1,000; Bethyl Laboratories A300-246A) diluted in 3% BSA/TBS for 2 h at room temperature. Coverslips were washed with TBS and incubated for 1 h at room temperature in the dark with Alexa-Fluor 568 goat anti-rabbit IgG (H + L) antibody (dilution 1:1,000; Invitrogen A11036) and DAPI (1 μg/mL; Sigma-Aldrich D9564), diluted in 3% BSA/TBS. After further washing with TBS, coverslips were mounted onto glass slides using ProLong Gold antifade mounting medium (Invitrogen P36934).Slides were imaged using a BZ-X800L microscope (KEYENCE) with a 100× objective, acquiring images from at least ten fields of view per condition. Nuclei were automatically segmented using CellProfiler69, with segmentation parameters validated manually to achieve an identification error rate below 0.5%. For each nucleus, normalised pRPA intensity was calculated by dividing the mean nuclear pRPA fluorescence intensity by the mean nuclear DAPI intensity. pRPA-positive nuclei were defined as those with a normalised pRPA intensity exceeding the 99th percentile of the corresponding DMSO-treated control for each cell line.ChIP-seqSemi-confluent cells in 10 cm dishes were crosslinked in 1.42% formaldehyde (final concentration) in PBS for 15 mins at room temperature with gentle agitation. Crosslinking was quenched with 127 mM glycine for 5 mins. Cells were washed with PBS, lysed in buffer (10 mM Tris-HCl pH 8, 10 mM NaCl, 0.5% IGEPAL® CA-630) and pelleted (2000 × g, 5 mins, 4 °C). Pellets were resuspended in SDS lysis buffer (50 mM Tris-HCl pH 8, 1% SDS, 10 mM EDTA), followed by addition of ice-cold RIPA buffer (50 mM HEPES, 1 mM EDTA, 140 mM NaCl, 1% Triton X-100, 0.1% sodium deoxycholate) supplemented with protease inhibitors (Roche Complete and 1 mM AEBSF).Chromatin was sheared using a Bioruptor (30 sec on, 30 sec off for 60 cycles), yielding ~330 μL clarified lysate per sample after centrifugation (12,000 × g, 10 mins, 4 °C). A 30 μL aliquot was reserved as input. Immunoprecipitation was carried out using 2 μg of anti-Histone H3K9me3 antibody (Abcam ab8898) per 300 μL chromatin, incubated for 3 h at 4 °C with rotation, followed by addition of 20 μL pre-washed Protein G Dynabeads and further incubation for 1 h. Beads were sequentially washed with RIPA buffer, wash buffer 1 (50 mM HEPES, 1 mM EDTA, 500 mM NaCl, 1% Triton X-100, 0.1% Sodium deoxycholate), wash buffer 2 (10 mM Tris-HCl (pH 8), 250 mM LiCl, 0.5% IGEPAL® CA-630, 0.5% Sodium deoxycholate, 1 mM EDTA), and TE, each with rotation at 4 °C. Elution was performed in elution buffer (50 mM Tris-HCl (pH 8), 10 mM EDTA, 1% SDS) at 65 °C for 15 mins, followed by overnight reverse crosslinking at 65 °C.Input and immunoprecipitated DNA samples were treated with RNase A and Proteinase K, extracted with phenol/chloroform, and precipitated with ethanol in the presence of sodium acetate and glycogen. DNA pellets were washed with 80% ethanol, air-dried, and resuspended in 10 mM Tris-HCl, pH 7.5. DNA concentrations were determined with PicoGreen, and libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit for Illumina according to the manufacturer’s instructions. Libraries were sequenced as 150-bp paired-end (PE) reads on an Illumina NovaSeq X platform (Macrogen, Tokyo, Japan). One biological replicate was analysed.Pu-seq assaysStandard Pu-seq assays were conducted as described previously20. Briefly, 1.25 × 104 cells/ml were seeded in 15 cm petri dishes (Greiner). After 48 h, RNaseH2A was degraded by adding either 0.4 μM 5-Ph-IAA (BioAcademia, Japan #30-003) or 0.5 μM dTAGV−1 (Tocris Bioscience 6914). Control cells were left untreated. For Pu-seq assays examining RAD18 overexpression, 0.5 μg/mL doxycycline was added 16 h before the addition of 5-Ph-IAA. Following an additional 48 h, cells were harvested by centrifugation. For Pu-seq assays with UV exposure, 4 × 104 cells/mL were seeded. After 48 h, 0.4 μM 5-Ph-IAA was added. After an additional 4 h, cells were washed with PBS and exposed to 5 J/m2 of UV. Medium containing 0.4 μM 5-Ph-IAA was replaced and cells were grown for 24 h before harvesting by centrifugation.Genomic DNA was extracted using the Blood & Cell Culture DNA Midi Kit 100/G Genomic-tips (Qiagen). Small-scale alkaline treatment experiments were conducted by mixing 3 μg genomic DNA with 0.3 M NaOH and incubating the mixture for 2 h at 55 °C. Reactions were neutralised by adding 0.3 M Tris-HCl (pH 7.5) and the distribution of ssDNA was analysed using TapeStation RNA ScreenTape (Agilent). For library preparation, 25 μg of genomic DNA was subjected to alkaline treatment (0.3 M NaOH for 2 h) and subsequently separated on a 1.5% agarose gel for 100 mins at 100 V. The gel was incubated with 5 μg/mL acridine orange (Invitrogen) for 2 h at room temperature with gentle shaking to stain DNA. Destaining was performed overnight in water. DNA fragments ranging from 300-2,000 bp were excised from the gel and purified using the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel). Library preparation followed the protocol described previously22,23. The libraries were sequenced as 150-bp paired-end (PE) reads on an Illumina NovaSeq X platform (Macrogen, Tokyo, Japan) to generate ~200 million reads per sample. Two biological replicates were acquired for each condition, except for Pol η-KO and Pu-seq in RPE1 cells, which were each conducted once.Analysis of polymerase usagePaired-end Illumina reads from Pu-seq were processed as previously described20. Briefly, reads were aligned to the GRCh38 genome with Bowtie2 (version 2.3.5), with reads aligning to multiple locations removed. The 5′ ends of the R1 reads, corresponding to NTP locations, were counted for each 1 kb bin of the genome to provide bin counts for both the Watson and Crick strands. These counts were normalised to the total number of reads for each strand. Pol η-F18A data were then expressed as a ratio of reads in each bin compared to a control sample by dividing each bin count by the corresponding count for wild-type Pol η. The resulting data were smoothed using a moving average of the 2 m + 1 bins, where m is the moving average parameter when specified for individual analyses.Strand bias scores were calculated as (PuW-PuC)/(PuW+PuC), where PuW and PuC are the normalised ribonucleotide counts on the Watson and Crick strands, respectively (Fig. 2a). Replication asymmetry scores (R.A.S.) at initiation zones were calculated within a 200 kb window on either side of the initiation index peak as ((ΣPuW,R + ΣPuC,L)-(ΣPuW,L + ΣPuC,R))/(ΣPuW,R + ΣPuC,L + ΣPuW,L + ΣPuC,R), where ΣPuW,L denotes the sum of Watson strand scores on the left flank and ΣPuW,R the right flank, etc (Fig. 2c).Analysis of ChIP-seq dataPaired-end Illumina reads were aligned to the human reference genome (GRCh38 primary assembly) with Bowtie2 (v2.3.5). After converting the mapped data from SAM to BAM format with Samtools70 (https://www.htslib.org/), these data were subjected to analysis using Model-based Analysis of ChIP-Seq (MACS)71. For each sample, duplicate fragments were removed using MACS3 (filterdup, paired-end mode) and genome-wide coverage was estimated with MACS3 pileup. To match library sizes, the input coverage track was scaled to the immunoprecipitated (IP) track by multiplying the input pileup by the ratio of post-deduplication fragment counts in IP versus input (IP/Input) using MACS3 bdgopt. Coverage was then aggregated into non-overlapping 100 bp bins and the IP-to-input ratio was computed per bin using custom R scripts. Resulting tracks were exported for downstream visualisation.Analysis of cancer mutationsCancer genome mutation data and transcriptome datasets were acquired from TCGA. For mutation data, VCF files from TCGA, each containing single-nucleotide variant data for individual patients across multiple cancer types (801 files, released in May 2016), were processed. These VCF files were converted to the GRCh38 reference genome using GATK LiftoverVcf (https://gatk.broadinstitute.org)72. Only variants included in the TCGA “whitelist” (i.e., those passing quality control filters) were retained. Data corresponding to specific mutation motifs were extracted, accounting for strand orientation (e.g., A > G substitutions, which appear in the VCF file as T > C when on the reverse strand). The extracted data were then used for downstream analyses. R.A.S. scores were calculated as described above, using binned mutation counts in place of ribonucleotide counts. Late replication timing ratio was calculated for each mutation type as the number of mutations occurring within replication timing bins 12-16 divided by the total number of mutations for that type.Statistics and reproducibilityRepresentative images are from experiments that were independently repeated at least three times with similar results, unless otherwise stated.Data availabilityThe Pu-seq data generated in this study, including raw sequencing reads and processed data, have been deposited in the NCBI GEO under accession code GSE308763. Processed genome-wide Pol η usage profiles are available for interactive visualisation using JBrowse through the website of our laboratory (https://www.jfcr.or.jp/laboratory/department/genome_dynamics/www/resource_eng.html). ChIP-seq data for histone modification in HCT116 have been deposited under accession GSE308762. High-resolution RT data in HCT116 are available from GSE137764 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE137764]73. All other data that support the findings of this study are available in figshare under https://doi.org/10.6084/m9.figshare.3312842074. Source data are provided with this paper.Code availabilityThe genome-wide profiles of polymerase usage and associated analysed data were created using R. The code is available in GitHub (https://github.com/yasukasu/Human_Pu-seq)75.ReferencesBainbridge, L. J. & Daigaku, Y. Bulk synthesis and beyond: The roles of eukaryotic replicative DNA polymerases. DNA Repair (Amst.) 141, 103740 (2024).Article CAS PubMed Google Scholar Masutani, C. et al. The XPV (xeroderma pigmentosum variant) gene encodes human DNA polymerase eta. Nat. 399, 700–704 (1999).Article ADS CAS Google Scholar Johnson, R. E., Prakash, S. & Prakash, L. Efficient bypass of a thymine-thymine dimer by yeast DNA polymerase, Polη. Sci. 283, 1001–1004 (1999).Article ADS CAS Google Scholar Masutani, C., Kusumoto, R., Iwai, S. & Hanaoka, F. Mechanisms of accurate translesion synthesis by human DNA polymerase η. EMBO J. 19, 3100–3109 (2000).Article ADS CAS PubMed PubMed Central Google Scholar Biertumpfel, C. et al. Structure and mechanism of human DNA polymerase eta. Nature 465, 1044–1048 (2010).Article ADS PubMed PubMed Central Google Scholar McCulloch, S. D. et al. Preferential cis-syn thymine dimer bypass by DNA polymerase η occurs with biased fidelity. Nature 428, 97–100 (2004).Article ADS CAS PubMed Google Scholar Ummat, A. et al. Structural basis for cisplatin DNA damage tolerance by human polymerase η during cancer chemotherapy. Nat. Struct. Mol. Biol. 19, 628–632 (2012).Article CAS PubMed PubMed Central Google Scholar Zhao, Y. et al. Structural basis of human DNA polymerase η-mediated chemoresistance to cisplatin. Proc. Natl. Acad. Sci. USA 109, 7269–7274 (2012).Article ADS CAS PubMed PubMed Central Google Scholar Bassett, E. et al. The role of DNA polymerase η in translesion synthesis past platinum-DNA adducts in human fibroblasts. Cancer Res. 64, 6469–6475 (2004).Article CAS PubMed Google Scholar Zeng, X. et al. DNA polymerase η is an A-T mutator in somatic hypermutation of immunoglobulin variable genes. Nat. Immunol. 2, 537–541 (2001).Article CAS PubMed Google Scholar Lerner, L. K. et al. Large deletions in immunoglobulin genes are associated with a sustained absence of DNA Polymerase η. Sci. Rep. 10, 1311 (2020).Article ADS CAS PubMed PubMed Central Google Scholar Fagerberg, L. et al. Analysis of the human tissue-specific expression by genome-wide integration of transcriptomics and antibody-based proteomics. Mol. Cell Proteom. 13, 397–406 (2014).Article CAS Google Scholar Despras, E. et al. Rad18-dependent SUMOylation of human specialized DNA polymerase eta is required to prevent under-replicated DNA. Nat. Commun. 7, 13326 (2016).Article ADS CAS PubMed PubMed Central Google Scholar Kannouche, P. et al. Domain structure, localization, and function of DNA polymerase η, defective in xeroderma pigmentosum variant cells. Genes Dev. 15, 158–172 (2001).Article CAS PubMed PubMed Central Google Scholar Rey, L. et al. Human DNA polymerase η is required for common fragile site stability during unperturbed DNA replication. Mol. Cell Biol. 29, 3344–3354 (2009).Article CAS PubMed PubMed Central Google Scholar Bergoglio, V. et al. DNA synthesis by Pol η promotes fragile site stability by preventing under-replicated DNA in mitosis. J. Cell Biol. 201, 395–408 (2013).Article CAS PubMed PubMed Central Google Scholar Barnes, R. P., Hile, S. E., Lee, M. Y. & Eckert, K. A. DNA polymerases eta and kappa exchange with the polymerase delta holoenzyme to complete common fragile site synthesis. DNA Repair (Amst.) 57, 1–11 (2017).Article CAS PubMed PubMed Central Google Scholar Twayana, S. et al. Translesion polymerase eta both facilitates DNA replication and promotes increased human genetic variation at common fragile sites. Proc. Natl. Acad. Sci. USA 118, e2106477118 (2021).Bainbridge, L. J. & Daigaku, Y. Adaptive use of error-prone DNA polymerases provides flexibility in genome replication during tumorigenesis. Cancer Sci. 115, 2125–2137 (2024).Article CAS PubMed PubMed Central Google Scholar Koyanagi, E. et al. Global landscape of replicative DNA polymerase usage in the human genome. Nat. Commun. 13, 7221 (2022).Article ADS CAS PubMed PubMed Central Google Scholar Kreisel, K. et al. DNA polymerase η contributes to genome-wide lagging strand synthesis. Nucleic Acids Res. 47, 2425–2435 (2019).Article CAS PubMed PubMed Central Google Scholar Daigaku, Y. et al. A global profile of replicative polymerase usage. Nat. Struct. Mol. Biol. 22, 192–198 (2015).Article CAS PubMed PubMed Central Google Scholar Keszthelyi, A., Daigaku, Y., Ptasinska, K., Miyabe, I. & Carr, A. M. Mapping ribonucleotides in genomic DNA and exploring replication dynamics by polymerase usage sequencing (Pu-seq). Nat. Protoc. 10, 1786–1801 (2015).Article CAS PubMed Google Scholar Yesbolatova, A. et al. The auxin-inducible degron 2 technology provides sharp degradation control in yeast, mammalian cells, and mice. Nat. Commun. 11, 5701 (2020).Article ADS CAS PubMed PubMed Central Google Scholar Vaisman, A. & Woodgate, R. Ribonucleotide discrimination by translesion synthesis DNA polymerases. Crit. Rev. Biochem. Mol. Biol. 53, 382–402 (2018).Article CAS PubMed PubMed Central Google Scholar Donigan, K. A. et al. Unlocking the steric gate of DNA polymerase eta leads to increased genomic instability in Saccharomyces cerevisiae. DNA Repair (Amst.) 35, 1–12 (2015).Article CAS PubMed PubMed Central Google Scholar Su, Y., Egli, M. & Guengerich, F. P. Mechanism of Ribonucleotide Incorporation by Human DNA Polymerase η. J. Biol. Chem. 291, 3747–3756 (2016).Article CAS PubMed PubMed Central Google Scholar Nabet, B. et al. Rapid and direct control of target protein levels with VHL-recruiting dTAG molecules. Nat. Commun. 11, 4687 (2020).Article ADS CAS PubMed PubMed Central Google Scholar Zach, R. & Carr, A. M. Increased expression of Polδ does not alter the canonical replication program in vivo. Wellcome Open Res. 6, 44 (2021).Article PubMed PubMed Central Google Scholar Petryk, N. et al. Replication landscape of the human genome. Nat. Commun. 7, 10208 (2016).Article ADS CAS PubMed PubMed Central Google Scholar Ercilla, A. et al. Physiological Tolerance to ssDNA Enables Strand Uncoupling during DNA Replication. Cell Rep. 30, 2416–2429.e7 (2020).Article CAS PubMed Google Scholar Masuda, Y. et al. Different types of interaction between PCNA and PIP boxes contribute to distinct cellular functions of Y-family DNA polymerases. Nucleic Acids Res. 43, 7898–7910 (2015).Article CAS PubMed PubMed Central Google Scholar Daigaku, Y. Roadworks of DNA Damage Bypass during and after Replication. Genes Environ. 34, 77–88 (2012).Article CAS Google Scholar Pohler, J. R., Otterlei, M. & Warbrick, E. An in vivo analysis of the localisation and interactions of human p66 DNA polymerase delta subunit. BMC Mol. Biol. 6, 17 (2005).Article PubMed PubMed Central Google Scholar Bertoletti, F. et al. Phosphorylation regulates human polη stability and damage bypass throughout the cell cycle. Nucleic Acids Res. 45, 9441–9454 (2017).Article CAS PubMed PubMed Central Google Scholar Mao, P., Smerdon, M. J., Roberts, S. A. & Wyrick, J. J. Chromosomal landscape of UV damage formation and repair at single-nucleotide resolution. Proc. Natl. Acad. Sci. USA 113, 9057–9062 (2016).Article ADS CAS PubMed PubMed Central Google Scholar Mao, P. et al. ETS transcription factors induce a unique UV damage signature that drives recurrent mutagenesis in melanoma. Nat. Commun. 9, 2626 (2018).Article ADS PubMed PubMed Central Google Scholar Lopez Martinez, D. & Svejstrup, J. Q. Mechanisms of RNA Polymerase II Termination at the 3’-End of Genes. J. Mol. Biol. 437, 168735 (2025).Article CAS PubMed Google Scholar Matsuda, T., Bebenek, K., Masutani, C., Hanaoka, F. & Kunkel, T. A. Low fidelity DNA synthesis by human DNA polymerase-eta. Nature 404, 1011–1013 (2000).Article ADS CAS PubMed Google Scholar King, N. M. et al. Overproduction of DNA polymerase eta does not raise the spontaneous mutation rate in diploid human fibroblasts. DNA Repair (Amst.) 4, 714–724 (2005).Article CAS PubMed Google Scholar Wang, Y. et al. Elevated expression of DNA polymerase kappa in human lung cancer is associated with p53 inactivation: Negative regulation of POLK promoter activity by p53. Int. J. Oncol. 25, 161–165 (2004).PubMed Google Scholar Yuan, F. et al. Overexpressed DNA polymerase iota regulated by JNK/c-Jun contributes to hypermutagenesis in bladder cancer. PLoS One 8, e69317 (2013).Article ADS CAS PubMed PubMed Central Google Scholar Tang, F. et al. DNA polymerase κ participates in early S-phase DNA replication in human cells. Proc. Natl. Acad. Sci. USA 121, e2405473121 (2024).Article CAS PubMed PubMed Central Google Scholar Faraz, M., Woodgate, R. & Clausen, A. R. Tracking Escherichia coli DNA polymerase V to the entire genome during the SOS response. DNA Repair (Amst.) 101, 103075 (2021).Article CAS PubMed PubMed Central Google Scholar Lewis, J. S. et al. Tunability of DNA Polymerase Stability during Eukaryotic DNA Replication. Mol. Cell 77, 17–25.e15 (2020).Article CAS PubMed Google Scholar Hedglin, M., Pandey, B. & Benkovic, S. J. Stability of the human polymerase δ holoenzyme and its implications in lagging strand DNA synthesis. Proc. Natl. Acad. Sci. USA 113, E1777–E1786 (2016).Article ADS CAS PubMed PubMed Central Google Scholar Masuda, Y. et al. Dynamics of human replication factors in the elongation phase of DNA replication. Nucleic Acids Res. 35, 6904–6916 (2007).Article CAS PubMed PubMed Central Google Scholar Masuda, Y., Piao, J. & Kamiya, K. DNA replication-coupled PCNA mono-ubiquitination and polymerase switching in a human in vitro system. J. Mol. Biol. 396, 487–500 (2010).Article CAS PubMed Google Scholar Su, Y., Egli, M. & Guengerich, F. P. Human DNA polymerase η accommodates RNA for strand extension. J. Biol. Chem. 292, 18044–18051 (2017).Article CAS PubMed PubMed Central Google Scholar Becker, J. R. et al. Genetic Interactions Implicating Postreplicative Repair in Okazaki Fragment Processing. PLoS Genet. 11, e1005659 (2015).Article PubMed PubMed Central Google Scholar Daigaku, Y. et al. PCNA ubiquitylation ensures timely completion of unperturbed DNA replication in fission yeast. PLoS Genet. 13, e1006789 (2017).Article PubMed PubMed Central Google Scholar Thakar, T. et al. Ubiquitinated-PCNA protects replication forks from DNA2-mediated degradation by regulating Okazaki fragment maturation and chromatin assembly. Nat. Commun. 11, 2147 (2020).Article ADS CAS PubMed PubMed Central Google Scholar Kurashima, K. et al. Polη, a Y-family translesion synthesis polymerase, promotes cellular tolerance of Myc-induced replication stress. J. Cell Sci. 131, jcs212183 (2018).Yang, Y. et al. DNA repair factor RAD18 and DNA polymerase Polκ confer tolerance of oncogenic DNA replication stress. J. Cell Biol. 216, 3097–3115 (2017).Article CAS PubMed PubMed Central Google Scholar Seplyarskiy, V. B. et al. Error-prone bypass of DNA lesions during lagging-strand replication is a common source of germline and cancer mutations. Nat. Genet. 51, 36–41 (2019).Article CAS PubMed Google Scholar Anderson, C. J. et al. Strand-resolved mutagenicity of DNA damage and repair. Nature 630, 744–751 (2024).Article ADS CAS PubMed PubMed Central Google Scholar van den Berg, J. et al. Quantifying DNA replication speeds in single cells by scEdU-seq. Nat. Methods 21, 1175–1184 (2024).Article PubMed PubMed Central Google Scholar Daigaku, Y., Davies, A. A. & Ulrich, H. D. Ubiquitin-dependent DNA damage bypass is separable from genome replication. Nature 465, 951–955 (2010).Article ADS CAS PubMed PubMed Central Google Scholar Edmunds, C. E., Simpson, L. J. & Sale, J. E. PCNA ubiquitination and REV1 define temporally distinct mechanisms for controlling translesion synthesis in the avian cell line DT40. Mol. Cell 30, 519–529 (2008).Article CAS PubMed Google Scholar Egger, T., Aze, A. & Maiorano, D. Detection of endogenous translesion DNA synthesis in single mammalian cells. Cell Rep. Methods 3, 100501 (2023).Article CAS PubMed PubMed Central Google Scholar Ben Yamin, B. et al. DNA polymerase zeta contributes to heterochromatin replication to prevent genome instability. EMBO J. 40, e104543 (2021).Article PubMed PubMed Central Google Scholar Stamatoyannopoulos, J. A. et al. Human mutation rate associated with DNA replication timing. Nat. Genet. 41, 393–395 (2009).Article CAS PubMed PubMed Central Google Scholar Koren, A. et al. Differential relationship of DNA replication timing to different forms of human mutation and variation. Am. J. Hum. Genet. 91, 1033–1040 (2012).Article CAS PubMed PubMed Central Google Scholar Watanabe, K. et al. Rad18 guides polη to replication stalling sites through physical interaction and PCNA monoubiquitination. EMBO J. 23, 3886–3896 (2004).Article ADS CAS PubMed PubMed Central Google Scholar Masuda, Y. & Kamiya, K. Role of single-stranded DNA in targeting REV1 to primer termini. J. Biol. Chem. 281, 24314–24321 (2006).Article CAS PubMed Google Scholar Taylor, J. S., Brockie, I. R. & O’Day, C. L. A building block for the sequence-specific introduction of cis-syn thymine dimers into oligonucleotides. Solid-phase synthesis of TpT[c,s]pTpT. J. Am. Chem. Soc. 109, 6735–6742 (1987).Article ADS CAS Google Scholar Ran, F. A. et al. Genome engineering using the CRISPR-Cas9 system. Nat. Protoc. 8, 2281–2308 (2013).Article CAS PubMed PubMed Central Google Scholar Akagi, J. et al. Interaction with DNA polymerase η is required for nuclear accumulation of REV1 and suppression of spontaneous mutations in human cells. DNA Repair (Amst.) 8, 585–599 (2009).Article CAS PubMed Google Scholar Stirling, D. R. et al. CellProfiler 4: improvements in speed, utility and usability. BMC Bioinforma. 22, 433 (2021).Article Google Scholar Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).Article PubMed PubMed Central Google Scholar Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).Article PubMed PubMed Central Google Scholar McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20, 1297–1303 (2010).Article CAS PubMed PubMed Central Google Scholar Zhao, P. A., Sasaki, T. & Gilbert, D. M. High-resolution Repli-Seq defines the temporal choreography of initiation, elongation and termination of replication in mammalian cells. Genome Biol. 21, 76 (2020).Article CAS PubMed PubMed Central Google Scholar Bainbridge, L.J. et al. Datasets for Strand- and replication timing-dependent functions of DNA polymerase η in human DNA replication. figshare https://doi.org/10.6084/m9.figshare.33128420 (2026).Bainbridge, L.J. et al. Codes for Strand- and replication timing-dependent functions of DNA polymerase η in human DNA replication. Zenodo https://doi.org/10.5281/zenodo.22188940 (2026).Download referencesAcknowledgementsThe authors thank Dr John Diffley (The Francis Crick Institute, London, UK) for providing a pcDNA4-derived plasmid, Mr Takahiro Yamaguchi for supporting the development of tetR-expressing cell lines, Dr Ryosuke Nozawa (Japanese Foundation for Cancer Research, Tokyo, Japan) for providing technical assistance and reagents for immunostaining and microscopy and Dr Rie Kanao for information on the REV1 (CTD) antibody. 5-Ph-IAA was kindly provided by Masato Kanemaki (National Institute of Genetics, Mishima, Japan). Experiments with radioisotopes were performed in the Radioisotope Research Center of Nagoya University. The results shown here are in part based upon data generated by the TCGA Research Network: https://www.cancer.gov/tcga.FundingThis work was supported by: JST PRESTO grant and FOREST program (JPMJPR18K7 and JPMJFR204X) to Y.D.; JSPS KAKENHI grants (JP21K19203, JP23H02463, and JP26K01989) to Y.D.; (JP24K18094) to L.J.B.; (JP23K11418) to Y.M.; (JP26K09283) to C.M.; AMED ASPIRE grant (JP25jf01260150001, 26jf0126015h0002) to Y.D.; Naito Foundation, Takeda Science Foundation, the Vehicle Racing Commemorative Foundation, and Yamada Science Foundation to Y.D. T.M., M.T. and K.K. declare no relevant funding.Author informationAuthors and AffiliationsCancer Genome Dynamics Project, Cancer Institute, Japanese Foundation for Cancer Research, Tokyo, 135-8550, JapanLewis J. Bainbridge, Tamiko Minamisawa, Mami Takahashi, Kanae Kaneko & Yasukazu DaigakuDepartment of Genome Dynamics, Research Institute of Environmental Medicine, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8601, JapanYuji Masuda & Chikahide MasutaniDepartment of Molecular Pharmaco-Biology, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya, 466-8550, JapanYuji Masuda & Chikahide MasutaniResearch Institute for Radiation Biology and Medicine, Hiroshima University, Hiroshima, 734-8553, JapanYuji MasudaAuthorsLewis J. BainbridgeView author publicationsSearch author on:PubMed Google ScholarYuji MasudaView author publicationsSearch author on:PubMed Google ScholarTamiko MinamisawaView author publicationsSearch author on:PubMed Google ScholarMami TakahashiView author publicationsSearch author on:PubMed Google ScholarKanae KanekoView author publicationsSearch author on:PubMed Google ScholarChikahide MasutaniView author publicationsSearch author on:PubMed Google ScholarYasukazu DaigakuView author publicationsSearch author on:PubMed Google ScholarContributionsY.D. conceived the study. L.J.B. and Y.D. designed the experimental approaches. Y.D. and L.J.B. designed informatics approaches. L.J.B., Y.M., T.M., M.T., K.K., and C.M. performed experiments and analysed the data. L.J.B. and Y.D. performed computational analysis. L.J.B. and Y.D. wrote the manuscript. L.J.B., Y.M., and C.M. edited the manuscript. All authors reviewed and accepted the manuscript before submission.Corresponding authorCorrespondence to Yasukazu Daigaku.Ethics declarationsCompeting interestsThe authors declare no competing interests.Peer reviewPeer review informationNature Communications thanks Kouji Hirota, Cyrus Vaziri and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationSource dataRights and permissionsOpen Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.Reprints and permissionsAbout this article