Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 52 (3), 274-303

Translesion DNA Polymerases in Eukaryotes: What Makes Them Tick?


Translesion DNA Polymerases in Eukaryotes: What Makes Them Tick?

Alexandra Vaisman et al. Crit Rev Biochem Mol Biol.


Life as we know it, simply would not exist without DNA replication. All living organisms utilize a complex machinery to duplicate their genomes and the central role in this machinery belongs to replicative DNA polymerases, enzymes that are specifically designed to copy DNA. "Hassle-free" DNA duplication exists only in an ideal world, while in real life, it is constantly threatened by a myriad of diverse challenges. Among the most pressing obstacles that replicative polymerases often cannot overcome by themselves are lesions that distort the structure of DNA. Despite elaborate systems that cells utilize to cleanse their genomes of damaged DNA, repair is often incomplete. The persistence of DNA lesions obstructing the cellular replicases can have deleterious consequences. One of the mechanisms allowing cells to complete replication is "Translesion DNA Synthesis (TLS)". TLS is intrinsically error-prone, but apparently, the potential downside of increased mutagenesis is a healthier outcome for the cell than incomplete replication. Although most of the currently identified eukaryotic DNA polymerases have been implicated in TLS, the best characterized are those belonging to the "Y-family" of DNA polymerases (pols η, ι, κ and Rev1), which are thought to play major roles in the TLS of persisting DNA lesions in coordination with the B-family polymerase, pol ζ. In this review, we summarize the unique features of these DNA polymerases by mainly focusing on their biochemical and structural characteristics, as well as potential protein-protein interactions with other critical factors affecting TLS regulation.

Keywords: DNA polymerase families; PCNA; Translesion synthesis; mutagenesis; protein interaction network; replicative bypass; ubiquitin.

Conflict of interest statement

Disclosure statement

The authors declare no conflict of interest


Figure 1
Figure 1. DNA polymerases: 60 years of discoveries
1956 – The first enzyme capable of copying DNA was discovered in E.coli extracts and was assumed at that time to be the only bacterial DNA polymerase (Kornberg et al., 1956). Later, when a second E. coli DNA polymerase was purified, this enzyme which plays an important role in prokaryotic DNA replication and repair was named pol I. The polA gene was sequenced in 1982 (Joyce et al., 1982) (accession P00582a). 1957 – The first eukaryotic DNA polymerase was identified (Bollum and Potter, 1957). When the uniform nomenclature was adopted in 1975, this enzyme was appropriately designated as pol α. Originally assumed to bear the sole responsibility for DNA synthesis in mammalian cells, this polymerase instead plays a key role in the initiation of chromosomal replication. The POLA gene was sequenced in 1988 (Wong et al., 1988) (accession CAA29920). 1968 – An enzyme with DNA polymerase activity was isolated from the rat liver mitochondria (Kalf and Ch’ih, 1968; Meyer and Simpson, 1968). This enzyme, known since 1977 as pol γ (Bolden et al., 1977) is a major polymerase dealing with all transactions involving mitochondrial DNA in mammalian cells. The POLG gene was sequenced in 1995 (Ropp and Copeland, 1995) (accession CAA88012). 1970 – A second DNA polymerase was discovered in E.coli by several groups in the USA and Germany (Knippers, 1970; Kornberg and Gefter, 1970; Moses and Richardson, 1970). According to the chronological order of discovery, it was named pol II. The sequencing of the polB gene was accomplished in 1990 (Chen et al., 1990) (accession P21189). 1971 – During the course of purification of E.coli pol II, a third prokaryotic DNA polymerase was detected (Kornberg and Gefter, 1971). DNA polymerase III is now known to be the main prokaryotic replicative polymerase. The dnaE gene encoding the catalytic α-subunit of pol III was sequenced in 1987 (Tomasiewicz and McHenry, 1987) (accession P10443). - A second eukaryotic nuclear DNA polymerase later named pol β, was identified in mammalian cells and tissues practically simultaneously by several laboratories in the USA and England (Baril et al., 1971; Berger et al., 1971; Chang and Bollum, 1971; Haines et al., 1971; Weissbach et al., 1971). Very soon, it became apparent that this polymerase does not play a direct role in DNA replication. Instead, extensive research conducted by various groups revealed a major role for DNA pol β in base-excision repair. The POLB gene was sequenced in 1986 (Zmudzka et al., 1986) (accession P06766). 1974 – A formal nomenclature designating each mammalian DNA polymerase with a Greek symbol was proposed in 1974 and accepted by attendees of the international conference on eukaryotic DNA polymerases in 1975 (Weissbach et al., 1975). According to the established nomenclature, the first two mammalian DNA polymerases were designated pols α and β. It is worth noting that the third DNA polymerase which was given the name pol γ was and a forth DNA polymerase found in mitochondria (designated DNA polymerase-mt) were later shown to be identical and the name pol γ has been retained for this mitochondrial polymerase (Bolden et al., 1977). 1976 – The first nuclear polymerase containing an associated 3′→5′ exonuclease activity was purified and called pol δ (Byrnes et al., 1976). Later, this polymerase was shown to be an essential component of the eukaryotic replication machinery. The sequencing of the POLD1 gene was accomplished in 1989 (Boulet et al., 1989) (accession P15436). 1987 – It was proposed that DNA polymerases should be classified into discrete families based on their evolutionary relatedness. The first two evolutionary groups of DNA polymerases were designated as polymerase families A- and B- according to the amino acid homology to E coli pols I and II, respectively (Jung et al., 1987). In 1991 two additional groups typified by the catalytic subunit of E coli pols III and eukaryotic pol β were designated as families C- and X-, respectively (Ito and Braithwaite, 1991). In 1999 family D- was proposed to group polymerases involved in the DNA replication machinery of the Euryarchaeota (Cann and Ishino, 1999). In 2001, proteins originally defined as belonging to the UmuC/DinB/Rev1/Rad30 superfamily and involved in mutagenesis and TLS DNA synthesis were designated as Y-family polymerases (Ohmori et al., 2001). 1989 – The fourth nuclear DNA polymerase in mammalian cells, pol ε, was first reported as a PCNA-independent form of pol δ (Focher et al., 1989). Subsequently, this enzyme was recognized as a distinct DNA polymerase and accordingly it was named pol ε. Similar to pol δ, pol ε is equipped with 3′→5′ exonuclease proofreading activity and is essential for replication of the eukaryotic genome. The POLE1 gene was sequenced in 1990 (Morrison et al., 1990) (accession P21951). 1996Saccharomyces cerevisiae DNA pol ζ was characterized as a complex of Rev3 and Rev7 proteins (Nelson et al., 1996b). These studies confirmed the hypothesis that the Rev3 gene long known to be involved in damage-induced and spontaneous mutagenesis, encodes the first DNA polymerase specializing in TLS (Morrison et al., 1989) (accession P14284). This prediction was made based on the homology of Rev3 to other genes encoding B-family DNA polymerases. - In the same year, the same group discovered dCMP transferase activity for the S. cerevisiae REV1 protein (Nelson et al., 1996a) that was known at the time to be required for the damage-induced mutagenesis and having ~25% identity with the E.coli UmuC protein (Larimer et al., 1989) (accession P12689). It was proposed that the CMP transferase function was important for mutagenic TLS involving pol ζ. However, Rev1 was not recognized as belonging to the broad superfamily of DNA-dependent DNA polymerases until 1999, when deoxynucleotidyl transferase activity was detected in several enzymes homologous to REV1. 1998 – The first evidence suggesting that the E. coli UmuD’2C complex consisting of the umuDC gene products (Perry et al., 1985; Kitagawa et al., 1985) (accession P04152), is a DNA polymerase was demonstrated (Tang et al., 1998). At that time, it was shown that in vitro the UmuD’2C complex could copy an abasic site-containing DNA template without the assistance of any other polymerase, although the possibility of contamination with trace amounts of other DNA polymerase were not entirely ruled out. A year later, additional biochemical studies unequivocally confirmed that the UmuD’2C complex is a bona fide DNA polymerase designated as E.coli pol V (Tang et al., 1999). 1999 –The S.cerevisiae RAD30 gene which was previously identified by sequence homology to prokaryotic UmuC and DinB in 1996 (Kulaeva et al., 1996; McDonald et al., 1997) was shown to encode DNA polymerase η (Johnson et al., 1999b). Shortly thereafter, human polymerase η was characterized (Masutani et al., 1999b). Polη became one of the founding members of the new Y-family of DNA polymerases (Ohmori et al., 2001). Nonsense, or frameshift mutations in the gene (RAD30A, POLH, XPV) encoding pol η are responsible for the Xeroderm Pigmentosum Variant syndrome in humans (Johnson et al., 1999a; Masutani et al., 1999b). - The same week that the DNA polymerase activity of the UmuD’2C encoded pol V was confirmed, a manuscript describing the TLS activity of E.coli DinB was published (Wagner et al., 1999). This polymerase became known as E.coli DNA pol IV. The dinB gene was originally identified as dinP in 1995 (Ohmori et al., 1995) (accession BAA07593) and despite being shown to be allelic with dinB in 1999 the name of dinP, rather than the correct name of dinB, is still often used in Genbank data files describing related proteins. 2000 –The TLS activity of two eukaryotic Y-family polymerases ι (Tissier et al., 2000b) and κ (Ohashi et al., 2000a) (products of the POLI (RAD30B) and POLK (DINB1) genes, respectively) was demonstrated a year after they were cloned (Gerlach et al., 1999; McDonald et al., 1999) [accession numbers AAD50381 (pol ι) and AAF02541 pol κ)]. - Three X-family polymerases implicated in participating in different types of DNA transactions (such as BER, non-homologous end joining repair, V(D)J recombination, TLS, and sister chromatid cohesion) were discovered. This includes pol λ (Garcia-Diaz et al., 2000) (the sequence was first submitted in 1998 (Blanco, 1998) [accession number CAB65241]), pol μ (Dominguez et al., 2000) (accession CAB65075), and pol σ (Wang et al., 2000) (first sequenced in 1995 (Sadoff et al., 1995) accession P53632). It should be noted that the DNA polymerase activity of pol σ has been contested by Haracska et al., (Haracska et al., 2005b), who suggest that the protein is actually a poly-A RNA polymerase, rather than a bona fide DNA polymerase. 2003 – Two mammalian A-family DNA polymerases that are homologous to the DNA cross-link sensitivity protein Mus308 and implicated in different defense pathways against DNA damage were identified and characterized; pol θ (Seki et al., 2003) and pol ν (Marini et al., 2003). It worth mentioning that pol θ, the only polymerase known to contain a helicase domain, was first identified in 1997 in the genomes of a variety of eukaryotic organisms based on sequence homology to E. coli DNA pol I (Harris et al., 1996; Sharief et al., 1999; Sonnhammer and Wootton, 1997) (accession numbers AAB67306 and AAC33565). The polymerase was originally named pol η (Burtis and Harris, 1997), but was later renamed pol θ (Burgers et al., 2001). Pol ν was sequenced in 2003 (Marini et al., 2003) (accession NP_861524). 2013 – An ability to replicate both damaged and undamaged DNA templates was detected in PrimPol, an enzyme belonging to the archaeal-eukaryotic primase superfamily (Bianchi et al., 2013; García-Gómez et al., 2013). The gene encoding PrimPol was first sequenced in 2005 (Iyer et al., 2005) (accession NP_689896). The ability to catalyze TLS is only one of the broad enzymatic activities of the PrimPol enzymes that have been implicated in a large variety of cellular functions. a GeneBank Sequence identifiers are listed (Clark et al., 2016).
Figure 2
Figure 2. Structural organization of catalytic domains of TLS polymerases
A. Comparison of yeast (S. cerevisiae) and human (H. sapiens) REV1 and REV3 gene structures. The domain arrangement in yeast and human Rev proteins are very similar (the percent identity for the related regions are indicated within the gray areas between the human and yeast proteins). The human enzymes are larger than the yeast enzymes due to long inserts in the respective genes (one in Rev3 and two in Rev1 [I1 and I2], see the text for details). The conserved sequence motifs in Rev3 characteristic for the B-family DNA polymerases are indicated by roman numerals (I-VI). B. Arrangement of the functional domains in human Y-family polymerases. All polymerases (panels A & B) are aligned relative to the N-terminus of the catalytic core and all diagrams are drawn to the scale. The name of each polymerase and its length (number of amino acids) are indicated on the right-hand side of the figure and the gene designation is indicated on the left side of each protein. The catalytic domains of the polymerases are color-coded as follows: red – palm; green – thumb; light blue – fingers; purple – little finger; aquamarine –catalytic core of pol ζ with polymerase (light pink) and inactive 3′-5′ exonuclease (dark pink) domains. C. Interaction map of the proteins important for TLS. In all panels (A-C) the domains involved in protein-protein interactions are color-coded: NTD – N-terminal domain; MLS – mitochondrial localization signal; BRCT – breast cancer-associated protein-1 carboxyl-terminal domain; N-Clasp – structural feature of pol κ; NLS – nuclear localization signal in pols ζ, η, ι, and κ (pol ι contains non-classical NLS motif; in pol ζ the signal is shown for the yeast protein, but is not shown for the human enzyme since it is located within the omitted ~1550 amino acid region); UBM – ubiquitin binding motif; UBZ – ubiquitin binding zinc-finger motif; CTD – C-terminal domain of Rev3 which in addition to the N-terminal zinc finger [ZF] motif contains C-terminal iron–sulfur [FS] cluster, a binding site for other polymerase subunits; dRP – dRP lyase domain (the ~40 kDa region in pol ι to which the dRP lyase activity has been mapped (Prasad et al., 2003) is indicated below the schematic representation of the primary protein structure); PIP – PCNA-interacting protein motif; RIR – Rev1-interacting region motif; PIR–protein interaction regions (two PIR sites [PIR1 & 2]) in human Rev3 and yeast Rev1, as well as one PIR site in yeast Rev3 and human Rev1 are involved in the interaction with Rev7; in addition the PIR site in the CTD of human Rev1 is required for the binding to pols η, κ, and ι, while in yeast the PIR site in the catalytic domain of Rev1 [PIR1] is involved in the interaction with pol η, Rev3 and Rev7. The PIR in the CTD of human Rev1 [PIR2] has two interfaces: the C-terminal part (C) is required for the interaction with Rev7, while the N-terminal part (N) is involved in an interaction with PolD3 and pols η, ι and κ. The PIP and RIR motifs have similar consensus sequences, but have traditionally been viewed as separate entities. In several recent studies, it has been proposed that these motifs be considered as parts of a single “PIP-like” motif capable of binding multiple target proteins (Boehm and Washington, 2016). The position of the star (*) above the pol η sequence corresponds to the F1 motif necessary for the interaction with the POLD2 subunit of pol δ (Baldeck et al., 2015).
Figure 3
Figure 3. Basic mechanism of TLS in eukaryotic cells
TLS is a multi-step process involving (A) stalling of the replicative DNA polymerase (here shown for pol δ consisting of PolD1, PolD2, PolD3, and PolD4 subunits) at damaged sites (red); (B) PCNA ubiquitination, recruitment of a TLS polymerase to the primer terminus (here shown for pol η) and nucleotide (mis)insertion opposite the lesion; (C) extension (catalyzed by pol ζ) of the resulting primer termini by several nucleotides; (D) de-ubiquitination of PCNA and release of a TLS polymerase after bypass completion allowing the normal replication restart. TLS is regulated through an elaborate network of protein-protein interactions and posttranslational modifications (shown for different polymerases in A-E; see text for the details). TLS polymerases are recruited to the replication factories via interactions with PCNA (E), which due to the trimeric structure and various binding sites can accommodate multiple partners (A-E). The initial interactions through the PIP boxes (1) are too unstable, which makes them hard to be detected in replication foci. DNA damage reinforces the PCNA/TLS polymerases interactions due to mono-ubiquitination of PCNA (2) that is catalyzed by the RAD6/RAD18 complex (3). In normal cells, Rad18 forms a dimer consisting of a mono-ubiquitinated and non-ubiquitinated proteins (4). Non-modified Rad18 (5) that is required for PCNA ubiquitination is released from the dimer via an interaction with Rev1 (6) and is stabilized by an interaction with polη (7). Damage-induced phosphorylation of pol η and Rad18 facilitates their interaction and localization to replication foci (8). PCNA ubiquitination is further promoted by the binding of the polymerase molecules via their UBD domains (2). Pol η has a higher affinity for ubiquitinated PCNA and this interaction is sufficiently stable for pol η to form detectable foci, while the accumulation of pol ι requires its interaction with pol η (9 in the panel A). Mono-ubiquitination of either polη or polι enhances the interaction between these two polymerases (9). Cells use another protein binding platform, Rev1, which serves as a “polymerase bridge” for the assembly of two TLS polymerases [shown in panel B for pol ζ (through Rev7 as shown here, and through the PolD3 subunit) and pol κ (10)] and might facilitate the switch from the “inserter” (here shown for pol η) to the “extender” (pol ζ). Interactions between Rev3 (through Rev7 and PolD3) and Rev1 are possibly facilitated by Rev1 phosphorylation (11). Such interactions stabilize binding of pol ζ (through the PIP motif of PolD3) to ubiquitinated PCNA and thereby enhances pol ζ-dependent TLS.

Similar articles

See all similar articles

Cited by 42 articles

See all "Cited by" articles

Publication types