Translesion synthesis of acetylaminofluorene-dG adducts by DNA polymerase zeta is stimulated by yeast Rev1 protein

Translesion synthesis is an important mechanism in response to unrepaired DNA lesions during replication. The DNA polymerase ζ (Polζ) mutagenesis pathway is a major error-prone translesion synthesis mechanism requiring Polζ and Rev1. In addition to its dCMP transferase, a non-catalytic function of Rev1 is suspected in cellular response to certain types of DNA lesions. However, it is not well understood about the non-catalytic function of Rev1 in translesion synthesis. We have analyzed the role of Rev1 in translesion synthesis of an acetylaminofluorene (AAF)-dG DNA adduct. Purified yeast Rev1 was essentially unresponsive to a template AAF-dG DNA adduct, in contrast to its efficient C insertion opposite a template 1,N⁶-ethenoadenine adduct. Purified yeast Polζ was very inefficient in the bypass of the AAF-dG adduct. Combining Rev1 and Polζ, however, led to a synergistic effect on translesion synthesis. Rev1 protein enhanced Polζ-catalyzed nucleotide insertion opposite the AAF-dG adduct and strongly stimulated Polζ-catalyzed extension from opposite the lesion. Rev1 also stimulated the deficient synthesis by Polζ at the very end of undamaged DNA templates. Deleting the C-terminal 205 aa of Rev1 did not affect its dCMP transferase activity, but abolished its stimulatory activity on Polζ-catalyzed extension from opposite the AAF-dG adduct. These results suggest that translesion synthesis of AAF-dG adducts by Polζ is stimulated by Rev1 protein in yeast. Consistent with the in vitro results, both Polζ and Rev1 were found to be equally important for error-prone translesion synthesis across from AAF-dG DNA adducts in yeast cells.     

Complex formation of yeast Rev1 with DNA polymerase eta

In Saccharomyces cerevisiae, Rev1 functions in translesion DNA synthesis (TLS) together with polymerase ζ (Polζ), comprised of the Rev3 catalytic and Rev7 accessory subunits. Rev1 plays an indispensable structural role in promoting Polζ function, and deletion of the Rev1-C terminal region that is involved in physical interactions with Rev3 inactivates Polζ function in TLS. In humans, however, Rev1 has been shown to physically interact with the Y-family polymerases Polη, Polι, and Polκ, and the Rev1 C terminus mediates these interactions. Since all the available genetic and biochemical evidence in yeast support the requirement of Rev1 as a structural element for Polζ and not for Polη, these observations have raised the possibility that in its structural role, Rev1 has diverged between yeast and humans. Here we show that although in yeast a stable Rev1-Polη complex can be formed, this complex formation involves the polymerase-associated domain of Rev1 and not the Rev1 C terminus as in humans. We also found that the DNA synthesis activity of Rev1 is enhanced in this complex. We discuss the implications of these and other observations for the possible divergence of Rev1's structural role between yeast and humans.     

Cryo-EM reveals conformational flexibility in apo DNA polymerase ζ

The translesion synthesis (TLS) DNA polymerases Rev1 and Polζ function together in DNA lesion bypass during DNA replication, acting as nucleotide inserter and extender polymerases, respectively. While the structural characterization of the Saccharomyces cerevisiae Polζ in its DNA-bound state has illuminated how this enzyme synthesizes DNA, a mechanistic understanding of TLS also requires probing conformational changes associated with DNA- and Rev1 binding. Here, we used single-particle cryo-electron microscopy to determine the structure of the apo Polζ holoenzyme. We show that compared with its DNA-bound state, apo Polζ displays enhanced flexibility that correlates with concerted motions associated with expansion of the Polζ DNA-binding channel upon DNA binding. We also identified a lysine residue that obstructs the DNA-binding channel in apo Polζ, suggesting a gating mechanism. The Polζ subunit Rev7 is a hub protein that directly binds Rev1 and is a component of several other protein complexes such as the shieldin DNA double-strand break repair complex. We analyzed the molecular interactions of budding yeast Rev7 in the context of Polζ and those of human Rev7 in the context of shieldin using a crystal structure of Rev7 bound to a fragment of the shieldin-3 protein. Overall, our study provides new insights into Polζ mechanism of action and the manner in which Rev7 recognizes partner proteins.     

Role of DNA polymerase eta in the bypass of abasic sites in yeast cells

Abasic (AP) sites are major DNA lesions and are highly mutagenic. AP site-induced mutagenesis largely depends on translesion synthesis. We have examined the role of DNA polymerase η (Polη) in translesion synthesis of AP sites by replicating a plasmid containing a site-specific AP site in yeast cells. In wild-type cells, AP site bypass resulted in preferred C insertion (62%) over A insertion (21%), as well as −1 deletion (3%), and complex event (14%) containing multiple mutations. In cells lacking Polη (rad30), Rev1, Polζ (rev3), and both Polη and Polζ, translesion synthesis was reduced to 30%, 30%, 15% and 3% of the wild-type level, respectively. C insertion opposite the AP site was reduced in rad30 mutant cells and was abolished in cells lacking Rev1 or Polζ, but significant A insertion was still detected in these mutant cells. While purified yeast Polα effectively inserted an A opposite the AP site in vitro, purified yeast Polδ was much less effective in A insertion opposite the lesion due to its 3′→5′ proofreading exonuclease activity. Purified yeast Polη performed extension synthesis from the primer 3′ A opposite the lesion. These results show that Polη is involved in translesion synthesis of AP sites in yeast cells, and suggest that an important role of Polη is to catalyze extension following A insertion opposite the lesion. Consistent with these conclusions, rad30 mutant cells were sensitive to methyl methanesulfonate (MMS), and rev1 rad30 or rev3 rad30 double mutant cells were synergistically more sensitive to MMS than the respective single mutant strains.     

A Missense Mutation in Rev7 Disrupts Formation of Polζ, Impairing Mouse Development and Repair of Genotoxic Agent-induced DNA Lesions

Repro22 is a mutant mouse produced via N-ethyl-N-nitrosourea-induced mutagenesis that shows sterility with germ cell depletion caused by defective proliferation of primordial germ cells, decreased body weight, and partial lethality during embryonic development. Using a positional cloning strategy, we identified a missense mutation in Rev7/Mad2l2 (Rev7^C70R) and confirmed that the mutation is the cause of the defects in repro22 mice through transgenic rescue with normal Rev7. Rev7/Mad2l2 encodes a subunit of DNA polymerase ζ (Polζ), 1 of 10 translesion DNA synthesis polymerases known in mammals. The mutant REV7 did not interact with REV3, the catalytic subunit of Polζ. Rev7^C70R/C70R cells showed decreased proliferation, increased apoptosis, and arrest in S phase with extensive γH2AX foci in nuclei that indicated accumulation of DNA damage after treatment with the genotoxic agent mitomycin C. The Rev7^C70R mutation does not affect the mitotic spindle assembly checkpoint. These results demonstrated that Rev7 is essential in resolving the replication stalls caused by DNA damage during S phase. We concluded that Rev7 is required for primordial germ cell proliferation and embryonic viability and development through the translesion DNA synthesis activity of Polζ preserving DNA integrity during cell proliferation, which is required in highly proliferating embryonic cells.     

In vivo evidence for translesion synthesis by the replicative DNA polymerase δ

The intolerance of DNA polymerase δ (Polδ) to incorrect base pairing contributes to its extremely high accuracy during replication, but is believed to inhibit translesion synthesis (TLS). However, chicken DT40 cells lacking the POLD3 subunit of Polδ are deficient in TLS. Previous genetic and biochemical analysis showed that POLD3 may promote lesion bypass by Polδ itself independently of the translesion polymerase Polζ of which POLD3 is also a subunit. To test this hypothesis, we have inactivated Polδ proofreading in pold3 cells. This significantly restored TLS in pold3 mutants, enhancing dA incorporation opposite abasic sites. Purified proofreading-deficient human Polδ holoenzyme performs TLS of abasic sites in vitro much more efficiently than the wild type enzyme, with over 90% of TLS events resulting in dA incorporation. Furthermore, proofreading deficiency enhances the capability of Polδ to continue DNA synthesis over UV lesions both in vivo and in vitro. These data support Polδ contributing to TLS in vivo and suggest that the mutagenesis resulting from loss of Polδ proofreading activity may in part be explained by enhanced lesion bypass.     

Poleta, Polzeta and Rev1 together are required for G to T transversion mutations induced by the (+)- and (-)-trans-anti-BPDE-N2-dG DNA adducts in yeast cells

Benzo[a]pyrene is an important environmental mutagen and carcinogen. Its metabolism in cells yields the mutagenic, key ultimate carcinogen 7R,8S,9S,10R-anti-benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide, (+)-anti-BPDE, which reacts via its 10-position with N²-dG in DNA to form the adduct (+)-trans-anti-BPDE-N²-dG. To gain molecular insights into BPDE-induced mutagenesis, we examined in vivo translesion synthesis and mutagenesis in yeast cells of a site-specific 10S (+)-trans-anti-BPDE-N²-dG adduct and the stereoisomeric 10R (−)-trans-anti-BPDE-N²-dG adduct. In wild-type cells, bypass products consisted of 76% C, 14% A and 7% G insertions opposite (+)-trans-anti-BPDE-N²-dG; and 89% C, 4% A and 4% G insertions opposite (−)-trans-anti-BPDE-N²-dG. Translesion synthesis was reduced by ~26–37% in rad30 mutant cells lacking Polη, but more deficient in rev1 and almost totally deficient in rev3 (lacking Polζ) mutants. C insertion opposite the lesion was reduced by ~24–33% in rad30 mutant cells, further reduced in rev1 mutant, and mostly disappeared in the rev3 mutant strain. The insertion of A was largely abolished in cells lacking either Polη, Polζ or Rev1. The insertion of G was not detected in either rev1 or rev3 mutant cells. The rad30 rev3 double mutant exhibited a similar phenotype as the single rev3 mutant with respect to translesion synthesis and mutagenesis. These results show that while the Polζ pathway is generally required for translesion synthesis and mutagenesis of the (+)- and (−)-trans-anti-BPDE-N²-dG DNA adducts, Polη, Polζ and Rev1 together are required for G→T transversion mutations, a major type of mutagenesis induced by these lesions. Based on biochemical and genetic results, we present mechanistic models of translesion synthesis of these two DNA adducts, involving both the one-polymerase one-step and two-polymerase two-step models.     

Human DNA polymerase kappa synthesizes DNA with extraordinarily low fidelity

Escherichia coli DNA polymerase IV encoded by the dinB gene is involved in untargeted mutagenesis. Its human homologue is DNA polymerase κ (Polκ) encoded by the DINB1 gene. Our recent studies have indicated that human Polκ is capable of both error-free and error-prone translesion DNA synthesis in vitro. However, it is not known whether human Polκ also plays a role in untargeted mutagenesis. To examine this possibility, we have measured the fidelity of human Polκ during DNA synthesis from undamaged templates. Using kinetic measurements of nucleotide incorporations and a fidelity assay with gapped M13mp2 DNA, we show that human Polκ synthesizes DNA with extraordinarily low fidelity. At the lacZα target gene, human Polκ made on average one error for every 200 nucleotides synthesized, with a predominant T→G transversion mutation at a rate of 1/147. The overall error rate of human Polκ is 1.7-fold lower than human Polη, but 33-fold higher than human Polβ, a DNA polymerase with very low fidelity. Thus, human Polκ is one of the most inaccurate DNA polymerases known. These results support a role for human Polκ in untargeted mutagenesis surrounding a DNA lesion and in DNA regions without damage.     

Translesion synthesis by yeast DNA polymerase ζ from templates containing lesions of ultraviolet radiation and acetylaminofluorene

In the yeast Saccharomyces cerevisiae, DNA polymerase ζ (Polζ) is required in a major lesion bypass pathway. To help understand the role of Polζ in lesion bypass, we have performed in vitro biochemical analyses of this polymerase in response to several DNA lesions. Purified yeast Polζ performed limited translesion synthesis opposite a template TT (6-4) photoproduct, incorporating A or T with similar efficiencies (and less frequently G) opposite the 3′ T, and predominantly A opposite the 5′ T. Purified yeast Polζ predominantly incorporated a G opposite an acetylaminofluorene (AAF)-adducted guanine. The lesion, however, significantly inhibited subsequent extension. Furthermore, yeast Polζ catalyzed extension DNA synthesis from primers annealed opposite the AAF-guanine and the 3′ T of the TT (6-4) photoproduct with varying efficiencies. Extension synthesis was more efficient when A or C was opposite the AAF-guanine, and when G was opposite the 3′ T of the TT (6-4) photoproduct. In contrast, the 3′ T of a cis–syn TT dimer completely blocked purified yeast Polζ, whereas the 5′ T was readily bypassed. These results support the following dual-function model of Polζ. First, Polζ catalyzes nucleotide incorporation opposite AAF-guanine and TT (6-4) photoproduct with a limited efficiency. Secondly, more efficient bypass of these lesions may require nucleotide incorporation by other DNA polymerases followed by extension DNA synthesis by Polζ.     

Roles of PCNA ubiquitination and TLS polymerases κ and η in the bypass of methyl methanesulfonate-induced DNA damage

Translesion synthesis (TLS) provides a highly conserved mechanism that enables DNA synthesis on a damaged template. TLS is performed by specialized DNA polymerases of which polymerase (Pol) κ is important for the cellular response to DNA damage induced by benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE), ultraviolet (UV) light and the alkylating agent methyl methanesulfonate (MMS). As TLS polymerases are intrinsically error-prone, tight regulation of their activity is required. One level of control is provided by ubiquitination of the homotrimeric DNA clamp PCNA at lysine residue 164 (PCNA-Ub). We here show that Polκ can function independently of PCNA modification and that Polη can function as a backup during TLS of MMS-induced lesions. Compared to cell lines deficient for PCNA modification (Pcna^K164R) or Polκ, double mutant cell lines display hypersensitivity to MMS but not to BPDE or UV-C. Double mutant cells also displayed delayed post-replicative TLS, accumulate higher levels of replication stress and delayed S-phase progression. Furthermore, we show that Polη and Polκ are redundant in the DNA damage bypass of MMS-induced DNA damage. Taken together, we provide evidence for PCNA-Ub-independent activation of Polκ and establish Polη as an important backup polymerase in the absence of Polκ in response to MMS-induced DNA damage.     

Roles of Saccharomyces cerevisiae DNA polymerases Poleta and Polzeta in response to irradiation by simulated sunlight

Sunlight causes lesions in DNA that if unrepaired and inaccurately replicated by DNA polymerases yield mutations that result in skin cancer in humans. Two enzymes involved in translesion synthesis (TLS) of UV-induced photolesions are DNA polymerase η (Polη) and polymerase ζ (Polζ), encoded by the RAD30A and REV3 genes, respectively. Previous studies have investigated the TLS roles of these polymerases in human and yeast cells irradiated with monochromatic, short wavelength UVC radiation (254 nm). However, less is known about cellular responses to solar radiation, which is of higher and mixed wavelengths (310–1100 nm) and produces a different spectrum of DNA lesions, including Dewar photoproducts and oxidative lesions. Here we report on the comparative cytotoxic and mutagenic effects of simulated sunlight (SSL) and UVC radiation on yeast wild-type, rad30Δ, rev3Δ and rev3Δ rad30Δ strains. The results with SSL support several previous interpretations on the roles of these two polymerases in TLS of photodimers and (6–4) photoproducts derived from studies with UVC. They further suggest that Polη participates in the non-mutagenic bypass of SSL-dependent cytosine-containing Dewar photoproducts and 8-oxoguanine, while Polζ is mainly responsible for the mutagenic bypass of all types of Dewar photoproducts. They also suggest that in the absence of Polζ, Polη contributes to UVC- and SSL-induced mutagenesis, possibly by the bypass of photodimers containing deaminated cytosine.     

DNA polymerase λ promotes error-free replication through Watson–Crick impairing N1-methyl-deoxyadenosine adduct in conjunction with DNA polymerase ζ

In a previous study, we showed that replication through the N1-methyl-deoxyadenosine (1-MeA) adduct in human cells is mediated via three different Polι/Polθ, Polη, and Polζ-dependent pathways. Based on biochemical studies with these Pols, in the Polι/Polθ pathway, we inferred a role for Polι in the insertion of a nucleotide (nt) opposite 1-MeA and of Polθ in extension of synthesis from the inserted nt; in the Polη pathway, we inferred that this Pol alone would replicate through 1-MeA; in the Polζ pathway, however, the Pol required for inserting an nt opposite 1-MeA had remained unidentified. In this study, we provide biochemical and genetic evidence for a role for Polλ in inserting the correct nt T opposite 1-MeA, from which Polζ would extend synthesis. The high proficiency of purified Polλ for inserting a T opposite 1-MeA implicates a role for Polλ—which normally uses W-C base pairing for DNA synthesis—in accommodating 1-MeA in a syn confirmation and forming a Hoogsteen base pair with T. The potential of Polλ to replicate through DNA lesions by Hoogsteen base pairing adds another novel aspect to Polλ’s role in translesion synthesis in addition to its role as a scaffolding component of Polζ. We discuss how the action mechanisms of Polλ and Polζ could be restrained to inserting a T opposite 1-MeA and extending synthesis thereafter, respectively.     

The effect of sequence context on spontaneous Polzeta-dependent mutagenesis in Saccharomyces cerevisiae

The Polζ translesion synthesis (TLS) DNA polymerase is responsible for over 50% of spontaneous mutagenesis and virtually all damage-induced mutagenesis in yeast. We previously demonstrated that reversion of the lys2ΔA746 −1 frameshift allele detects a novel type of +1 frameshift that is accompanied by one or more base substitutions and depends completely on the activity of Polζ. These ‘complex’ frameshifts accumulate at two discrete hotspots (HS1 and HS2) in the absence of nucleotide excision repair, and accumulate at a third location (HS3) in the additional absence of the translesion polymerase Polη. The current study investigates the sequence requirements for accumulation of Polζ-dependent complex frameshifts at these hotspots. We observed that transposing 13 bp of identity from HS1 or HS3 to a new location within LYS2 was sufficient to recapitulate these hotspots. In addition, altering the sequence immediately upstream of HS2 had no effect on the activity of the hotspot. These data support a model in which misincorporation opposite a lesion precedes and facilitates the selected slippage event. Finally, analysis of nonsense mutation revertants indicates that Polζ can simultaneously introduce multiple base substitutions in the absence of an accompanying frameshift event.     

The ubiquitin-binding domain of DNA polymerase η directly binds to DNA clamp PCNA and regulates translesion DNA synthesis

DNA polymerase eta (Polη) is a unique translesion DNA synthesis (TLS) enzyme required for the error-free bypass of ultraviolet ray (UV)-induced cyclobutane pyrimidine dimers in DNA. Therefore, its deficiency confers cellular sensitivity to UV radiation and an increased rate of UV-induced mutagenesis. Polη possesses a ubiquitin-binding zinc finger (ubz) domain and a PCNA-interacting-protein (pip) motif in the carboxy-terminal region. The role of the Polη pip motif in PCNA interaction required for DNA polymerase recruitment to the stalled replication fork has been demonstrated in earlier studies; however, the function of the ubz domain remains divisive. As per the current notion, the ubz domain of Polη binds to the ubiquitin moiety of the ubiquitinated PCNA, but such interaction is found to be nonessential for Polη''s function. In this study, through amino acid sequence alignments, we identify three classes of Polη among different species based on the presence or absence of pip motif or ubz domain and using comprehensive mutational analyses, we show that the ubz domain of Polη, which intrinsically lacks the pip motif directly binds to the interdomain connecting loop (IDCL) of PCNA and regulates Polη''s TLS activity. We further propose two distinct modes of PCNA interaction mediated either by pip motif or ubz domain in various Polη homologs. When the pip motif or ubz domain of a given Polη binds to the IDCL of PCNA, such interaction becomes essential, whereas the binding of ubz domain to PCNA through ubiquitin is dispensable for Polη''s function.     

Redundancy of mammalian Y family DNA polymerases in cellular responses to genomic DNA lesions induced by ultraviolet light

Short-wave ultraviolet light induces both mildly helix-distorting cyclobutane pyrimidine dimers (CPDs) and severely distorting (6–4) pyrimidine pyrimidone photoproducts ((6–4)PPs). The only DNA polymerase (Pol) that is known to replicate efficiently across CPDs is Polη, a member of the Y family of translesion synthesis (TLS) DNA polymerases. Phenotypes of Polη deficiency are transient, suggesting redundancy with other DNA damage tolerance pathways. Here we performed a comprehensive analysis of the temporal requirements of Y-family Pols ι and κ as backups for Polη in (i) bypassing genomic CPD and (6–4)PP lesions in vivo, (ii) suppressing DNA damage signaling, (iii) maintaining cell cycle progression and (iv) promoting cell survival, by using mouse embryonic fibroblast lines with single and combined disruptions in these Pols. The contribution of Polι is restricted to TLS at a subset of the photolesions. Polκ plays a dominant role in rescuing stalled replication forks in Polη-deficient mouse embryonic fibroblasts, both at CPDs and (6–4)PPs. This dampens DNA damage signaling and cell cycle arrest, and results in increased survival. The role of relatively error-prone Pols ι and κ as backups for Polη contributes to the understanding of the mutator phenotype of xeroderma pigmentosum variant, a syndrome caused by Polη defects.     

DNA-damage tolerance mediated by PCNA?Ub fusions in human cells is dependent on Rev1 but not Polη

In response to replication-blocking lesions, proliferating cell nuclear antigen (PCNA) can be sequentially ubiquitinated at the K164 residue, leading to two modes of DNA-damage tolerance, namely, translesion DNA synthesis (TLS) and error-free lesion bypass. Although the majority of reported data support a model whereby monoubiquitinated PCNA enhances its affinity for TLS polymerases and hence recruits them to the damage sites, this model has also been challenged by several observations. In this study, we expressed the PCNA-164R and ubiquitin (UB) fusion genes in an inducible manner in an attempt to mimic PCNA monoubiquitination in cultured human cells. It was found that expression of both N- and C-terminal PCNA•Ub fusions conferred significant tolerance to ultraviolet (UV)-induced DNA damage. Surprisingly, depletion of Polη, a TLS polymerase dedicated to bypassing UV-induced pyrimidine dimers, did not alter tolerance conferred by PCNA•Ub. In contrast, depletion of Rev1, another TLS polymerase serving as a scaffold for the assembly of the TLS complex, completely abolished PCNA•Ub-mediated damage tolerance. Similar genetic interactions were confirmed when UV-induced monoubiquitination of endogenous PCNA is abolished by RAD18 deletion. Hence, PCNA•Ub fusions bypass the requirement for PCNA monoubiquitination, and UV damage tolerance conferred by these fusions is dependent on Rev1 but independent of Polη.     

DNA Polymerase ζ-Dependent Lesion Bypass in Saccharomyces cerevisiae Is Accompanied by Error-Prone Copying of Long Stretches of Adjacent DNA

Translesion synthesis (TLS) helps cells to accomplish chromosomal replication in the presence of unrepaired DNA lesions. In eukaryotes, the bypass of most lesions involves a nucleotide insertion opposite the lesion by either a replicative or a specialized DNA polymerase, followed by extension of the resulting distorted primer terminus by DNA polymerase ζ (Polζ). The subsequent events leading to disengagement of the error-prone Polζ from the primer terminus and its replacement with an accurate replicative DNA polymerase remain largely unknown. As a first step toward understanding these events, we aimed to determine the length of DNA stretches synthesized in an error-prone manner during the Polζ-dependent lesion bypass. We developed new in vivo assays to identify the products of mutagenic TLS through a plasmid-borne tetrahydrofuran lesion and a UV-induced chromosomal lesion. We then surveyed the region downstream of the lesion site (in respect to the direction of TLS) for the presence of mutations indicative of an error-prone polymerase activity. The bypass of both lesions was associated with an approximately 300,000-fold increase in the mutation rate in the adjacent DNA segment, in comparison to the mutation rate during normal replication. The hypermutated tract extended 200 bp from the lesion in the plasmid-based assay and as far as 1 kb from the lesion in the chromosome-based assay. The mutation rate in this region was similar to the rate of errors produced by purified Polζ during copying of undamaged DNA in vitro. Further, no mutations downstream of the lesion were observed in rare TLS products recovered from Polζ-deficient cells. This led us to conclude that error-prone Polζ synthesis continues for several hundred nucleotides after the lesion bypass is completed. These results provide insight into the late steps of TLS and show that error-prone TLS tracts span a substantially larger region than previously appreciated.     

A four-subunit DNA polymerase ζ complex containing Pol δ accessory subunits is essential for PCNA-mediated mutagenesis

DNA polymerase ζ (Pol ζ) plays a key role in DNA translesion synthesis (TLS) and mutagenesis in eukaryotes. Previously, a two-subunit Rev3–Rev7 complex had been identified as the minimal assembly required for catalytic activity in vitro. Herein, we show that Saccharomyces cerevisiae Pol ζ binds to the Pol31 and Pol32 subunits of Pol δ, forming a four-subunit Pol ζ₄ complex (Rev3–Rev7–Pol31–Pol32). A [4Fe-4S] cluster in Rev3 is essential for the formation of Pol ζ₄ and damage-induced mutagenesis. Pol32 is indispensible for complex formation, providing an explanation for the long-standing observation that pol32Δ strains are defective for mutagenesis. The Pol31 and Pol32 subunits are also required for proliferating cell nuclear antigen (PCNA)-dependent TLS by Pol ζ as Pol ζ₂ lacks functional interactions with PCNA. Mutation of the C-terminal PCNA-interaction motif in Pol32 attenuates PCNA-dependent TLS in vitro and mutagenesis in vivo. Furthermore, a mutant form of PCNA, encoded by the mutagenesis-defective pol30-113 mutant, fails to stimulate Pol ζ₄ activity, providing an explanation for the observed mutagenesis phenotype. A stable Pol ζ₄ complex can be identified in all phases of the cell cycle suggesting that this complex is not regulated at the level of protein interactions between Rev3-Rev7 and Pol31-Pol32.