Complex Formation of Yeast Rev1 and Rev7 Proteins: a Novel Role for the Polymerase-Associated Domain

Abstract

The Rev1 protein of Saccharomyces cerevisiae functions in translesion synthesis (TLS) together with DNA polymerase (Pol) ζ, which is comprised of the Rev3 catalytic and the Rev7 accessory subunits. Rev1, a member of the Y family of Pols, differs from other members in its high degree of specificity for incorporating a C opposite template G as well as opposite an abasic site. Although Rev1 is indispensable for Polζ-dependent TLS, its DNA synthetic activity is not required for many of the Polζ-dependent lesion bypass events. This observation has suggested a structural role for Rev1 in this process. Here we show that in yeast, Rev1 forms a stable complex with Rev7, and the two proteins copurify. Importantly, the polymerase-associated domain (PAD) of Rev1 mediates its binding to Rev7. These observations reveal a novel role for the PAD region of Rev1 in protein-protein interactions, and they raise the possibility of a similar involvement of the PAD of other Y family Pols in protein-protein interactions. We discuss the possible roles of Rev1 versus the Rev1-Rev7 complex in TLS.

Rad6, a ubiquitin-conjugating enzyme, exists in vivo in a tight complex with Rad18, which is a DNA binding protein (1, 2) and also acts as an E3 in the ubiquitin conjugation process (11). Rad6-Rad18-mediated ubiquitin conjugation promotes replication through DNA lesions via at least three different pathways: DNA polymerase (Pol) η-dependent translesion synthesis (TLS), Polζ dependent TLS, and a Rad5-Mms2-Ubc13-dependent pathway, the mechanism of which is not known (41).

Polη promotes efficient and relatively error-free synthesis through UV-induced cyclobutane pyrimidine dimers which form at TT, TC, and CC sites (12, 16, 45, 46); consequently, inactivation of Polη in Saccharomyces cerevisiae and human cells confers enhanced UV mutagenesis (13, 30, 39, 44, 47, 48) and in humans causes the variant form of xeroderma pigmentosum (12, 29). Polζ, which is comprised of the Rev3 catalytic subunit and the Rev7 accessory subunit (33), is indispensable for UV mutagenesis in yeast (19, 21, 23) as well as human cells (6, 18, 24), and genetic studies in yeast have indicated its requirement for mutagenesis resulting from TLS occurring through abasic sites (14) and also through bases damaged upon treatment with certain chemical agents (35). Polζ promotes lesion bypass primarily via its role as an extender, wherein following the insertion of a nucleotide opposite the DNA lesion by an inserter polymerase, Polζ performs the extension of the nascent primer terminus (9, 15, 17, 37). For its role in TLS, however, Polζ requires the Rev1 protein, which like the Rev3 and Rev7 proteins is indispensable for UV mutagenesis (19, 20, 22, 23, 36) and for mutagenesis resulting from TLS occurring through abasic sites (14) and through other damaged bases (3).

Rev1, although a member of the Y family of DNA polymerases, differs from the other members in its specificity for predominantly inserting a C opposite template G and also opposite the other template nucleotides as well as an abasic site. Thus, Rev1 is a highly specialized polymerase, being specific for the incorporation of a C not only opposite the G template but also opposite the other template bases or even when the template base is missing.

Although Rev1 is indispensable for most Polζ-dependent TLS (20), its DNA synthetic activity is not required for many of these lesion bypass events. Thus, even though Rev1 is necessary for most base substitution mutations induced by UV light, inactivation of its polymerase activity has no effect on UV mutagenesis, and moreover, C insertion occurs rarely opposite UV lesions (4, 5). Likewise, Rev1 is required for mutagenesis resulting from TLS through abasic sites (14), but its DNA synthetic activity makes little contribution to its bypass, particularly when the lesion is in chromosomal DNA (9). Also, Rev1 is required for TLS through N-2-acetylaminofluorene bound to the C8 of a G, but its DNA synthesis activity has no role in its bypass (3). Mutagenic TLS occurring through UV lesions or abasic sites and predominantly error-free TLS through the N-2-acetylaminofluorene-adducted Gs, then, are dependent upon the Rev1 protein but not upon its DNA synthetic activity. The lack of requirement of Rev1 DNA synthetic activity in many of the Polζ-dependent TLS events has suggested a role for Rev1 in coordinating the assembly of Polζ at the replication fork (9).

Two hybrid analyses have provided evidence for the interaction of mouse and human Rev1 with Pols η, ι, and κ, and all these Pols bind the same ∼100 C-terminal amino acid residues of Rev1 (7, 34, 40). Human Rev1 has also been shown to interact with Rev7, and this interaction involves the same ∼100 C-terminal residues of Rev1 as those needed for the binding of Pols η, ι, and κ (26, 31). These observations have suggested that the binding of Rev7 excludes the binding of Pol η, ι, or κ to mammalian Rev1. Although the structure and the specificity for C incorporation have been conserved in the Rev1 protein from yeast to humans, the C terminus of mammalian Rev1, where Rev7 or Pol η, ι, or κ binds, shows little similarity to the C terminus of yeast Rev1 protein (27, 28). This has raised the possibility that yeast Rev1 is not involved in similar protein-protein interactions, and the yeast and mammalian Rev1 proteins may therefore differ in their mode of action.

To better understand the roles that Rev1 plays in the lesion bypass processes, we have begun a systematic analysis of the physical and functional interactions of yeast Rev1 with the other proteins involved in TLS, including Polζ. Here we show that the yeast Rev1 and Rev7 proteins form a stable complex which is resistant to high salt concentrations. Importantly, the region of Rev1 which mediates its interaction with Rev7 corresponds to the polymerase-associated domain (PAD) and the linker region of the Y family polymerases. In all the Y family polymerases, whose crystal structures have been determined, the PAD is joined to the thumb by a flexible linker (25, 32, 38, 42, 43, 49). The PAD is essential for the DNA synthetic activity of these polymerases, because it greatly enhances the DNA binding surface area of the polymerase. Our finding that the Rev1 PAD effects binding to Rev7 identifies a novel role for this region. We consider the implications of the PAD's involvement in protein-protein interactions in modulating the activity and lesion bypass ability of Rev1 as well as of other Y family polymerases. Also, we discuss the possible roles of Rev1 versus the Rev1-Rev7 complex in TLS.

MATERIALS AND METHODS

Yeast strains, plasmids, and DNA substrates.

S. cerevisiae strain BJ5464 and its isogenic derivatives rev7Δ and rev1Δ were used for protein expression. For in vitro binding assays, the open reading frames of full-length or truncated REV1 and of full-length REV7 were cloned in frame with glutathione-S-transferase (GST) in plasmid pBJ842 to produce GST fusion proteins. Oligonucleotides were synthesized by Midland Certified Reagent Co. (Midland, TX). DNA substrates were generated by annealing a 52-nucleotide-long oligonucleotide template (5′-TTC GTA TAA TGC CTA CAC TXG AGT ACC GGA GCA TCG TCG TGA CTG GGA AAA C-3′), which contained a C or a G at the position indicated by an X, to the 32-nucleotide 5′-³²P-labeled oligonucleotide primer N4456 (5′-GTT TTC CCA GTC ACG ACG ATG CTC CGG TAC TC-3′).

Purification of proteins.

S. cerevisiae Rev1 protein as a fusion with GST was expressed in the yeast strain BJ5464 and purified on a glutathione-Sepharose 4B column followed by MiniS (Amersham Biosciences) chromatography as described previously (9). For in vitro binding studies, wild-type and mutant GST-Rev1 were purified from yeast strain BJ5464, which lacks the genomic copy of REV7, on glutathione-Sepharose beads by using a protocol described previously (10). GST-Rev7 protein was purified from the yeast strain BJ5464 with a deletion of the genomic REV1 gene by using a similar approach. To obtain untagged proteins, GST fused proteins bound to glutathione-Sepharose beads were treated overnight at 4°C with PreScission protease (Amersham Pharmacia) to cleave the GST tag from the Rev1 or Rev7 protein. These proteins contain a 7-amino-acid leader peptide attached to the N terminus.

In vitro interaction of Rev1 and Rev7 proteins.

GST pull-down assays were carried out using a protocol described previously (10). Briefly, full-length or truncated GST-Rev1proteins (3 μg) were incubated with Rev7 (3 μg) in 20 μl buffer I (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM dithiothreitol, 0.01% NP-40, 10% glycerol) at 4°C for 30 min and then for 10 min at 25°C. To such a mixture, 20 μl glutathione-Sepharose beads was added and further incubated for 1 h with constant rocking at 4°C. The beads were spun down, and the unbound protein was collected. Further, the beads were thoroughly washed three times with 10 volumes of buffer I. Finally, the bound proteins were eluted with 20 μl of sodium dodecyl sulfate (SDS) loading buffer. Various fractions were resolved on a 12% SDS-polyacrylamide gel; this was followed by Coomassie blue R-250 staining. A similar approach was taken to check the interaction of GST-Rev7 with Rev1 proteins.

Yeast two-hybrid analysis.

The HF7c yeast cell line was transformed with the GAL4 BD-Rev7 and GAL4 AD-Rev1 fusion constructs. Transformants harboring both the GAL4 BD-Rev7 and GAL4 AD-Rev1 fusion constructs were grown on synthetic complete medium lacking leucine and tryptophan. β-Galactosidase activity was examined to determine the interaction between Rev1 and Rev7 as described in the Clontech yeast protocols handbook (chapter 6). Experiments were performed at least three times with triplicate samples.

Deoxynucleotide incorporation assays.

The standard primer extension reaction mixture (10 μl) contained 40 mM Tris-HCl (pH 7.5), 8 mM MgCl₂, 1 mM dithiothreitol, 100 μg/ml bovine serum albumin, 10% glycerol, 20 nM 5′-³²P-labeled oligonucleotide primer annealed to an oligonucleotide template, 100 μM of a single or each of all four deoxynucleoside triphosphates (dNTPs), and 5 nM Rev1 or Rev1-Rev7 complex. Assay mixtures were assembled on ice and incubated at 30°C for 10 min, and reactions were stopped by the addition of loading buffer (40 μl) containing EDTA (20 mM), 95% formamide, 0.3% bromphenol blue, 0.3% cyanol blue. The reaction products were resolved on 10% polyacrylamide gels containing 8 M urea.

Analysis of steady-state kinetics.

Steady-state kinetic analyses for deoxynucleotide incorporation opposite template G was performed as described previously (8). Rev1 or Rev1-Rev7 complex (2 nM) was incubated with 20 nM DNA substrate in the presence of increasing concentrations of dCTP for 10 min. Gel band intensities of the substrate and product were quantitated by use of a PhosphorImager, and the observed rate of deoxynucleotide incorporation was plotted as a function of dCTP concentration. The data were fitted by nonlinear regression using SigmaPlot 8.0 to the Michaelis-Menten equation describing a hyperbola, v = (V_max × [dCTP]/(K_m + [dCTP]). Apparent K_m and V_max steady-state parameters were obtained from the fit and were used to calculate the efficiency of deoxynucleotide incorporation (V_max/K_m).

RESULTS

Yeast Rev1 and Rev7 proteins form a stable complex in vivo and in vitro.

During purification of Rev1 from yeast cells, we observed an ∼30 kDa protein which copurified with GST-Rev1 and which had mobility on a denaturing polyacrylamide gel similar to that of the purified Rev7 protein (Fig. 1A, panel i, compare lanes 1 and 2), and Western blot analysis using anti-Rev7 antibodies confirmed the copurified protein to be Rev7 protein (Fig. 1A, panel ii). Since we used highly stringent purification conditions, in which the GST-Rev1 protein was bound to glutathione-Sepharose affinity beads followed by washing with a buffer containing 1 M NaCl before elution with glutathione, and since Rev1-containing fractions were then subjected to S-Sepharose ion-exchange chromatography, the presence of Rev7 in the Rev1 preparation strongly indicated that Rev1 forms a stable complex with Rev7 in vivo. Next, we examined whether the Rev1-Rev7 protein complex could be formed in vitro from purified proteins. For this purpose, we added purified Rev7 protein to the GST-Rev1-containing protein extract obtained from a rev7Δ yeast strain, and this was followed by incubation with glutathione-Sepharose affinity beads. The Rev1 protein was then released from the beads by treatment with PreScission protease. As shown in Fig. 1B (lane 3), Rev1 was eluted in a complex with Rev7, thus confirming a strong interaction of Rev1 with Rev7.

FIG. 1.

Purification of Rev1-Rev7 complex. (A) Rev1 copurifies with the Rev7 protein. N-terminal GST-Rev1 fusion protein was overexpressed in a wild-type yeast strain and was purified first on glutathione-Sepharose beads and then on an S-Sepharose column. (i) The GST-Rev1-containing protein sample was analyzed on a 10% denaturing polyacrylamide gel and stained with Coomassie blue. The positions of Rev7 present in the purified Rev1 sample (lane 1) and of purified Rev7 alone (lane 2) are indicated. The positions of molecular weight standards are indicated. (ii) Western blot analysis with anti-Rev7 antibodies to purified GST Rev1-Rev7 (lane 1) and to purified Rev7 (lane 2). (B) In vitro reconstitution of the Rev1-Rev7 complex. Purified Rev7 (9 μg) was added to the GST-Rev1-containing yeast extract (100 μg). After incubation, the reaction mixture was bound to glutathione-Sepharose beads; this was followed by washings and elution of the bound proteins with PreScission protease (2 units), which cleaves the GST tag. Fractions enriched for the Rev1-Rev7 complex were pooled, concentrated, and analyzed on an SDS-12% polyacrylamide gel and developed by Coomassie blue staining. Lane 2, Rev1 (200 ng); lane 3, Rev1-Rev7 complex (200 ng); lane 4, Rev7 (200 ng). *, unidentified but possibly heat shock proteins.

Interaction of Rev1 with Rev7 by two-hybrid analysis.

We used the yeast two-hybrid system to examine the interaction of Rev1 and Rev7 proteins in vivo. In one of the plasmids, the GAL4 activation domain (AD) was fused with the REV1 open reading frame, and in the other plasmid, the GAL4 binding domain (BD) was fused with the REV7 open reading frame. The HF7c yeast reporter strain harboring the GAL4 BD-REV7 plasmid was transformed with the GAL4 AD-REV1 plasmid. The expression of GAL4 AD-Rev1 fusion protein was confirmed by immunoblotting using anti-Rev1 antibodies (data not shown). The interaction of Rev1 with Rev7 in these transformants was analyzed by a β-galactosidase liquid assay. Compared to the low level of β-galactosidase activity detected when the GAL4 AD and GAL4 BD-Rev7 proteins were expressed together, a ninefold higher level of β-galactosidase activity was present when the GAL4 AD-Rev1 protein and GAL4 BD-Rev7 proteins were expressed together, indicating an interaction of Rev1 and Rev7 in yeast cells (Table 1).

TABLE 1.

Interaction of Rev1 with Rev7 in the yeast two-hybrid system

DNA binding domain fusion	Activation domain fusion	Mean β-galactosidase activity ± SD	Fold activation
GAL4 BD-Rev7	GAL4 AD	1.37 ± 0.1	1
GAL4 BD-Rev7	GAL4 AD-Rev1 (wt)a	12.07 ± 0.5	8.8
GAL4 BD-Rev7	GAL4 AD-Rev1-6 (567-767)	6.72 ± 0.2	4.9

^awt, wild type.

Mapping the Rev1 region mediating interaction with Rev7.

To map the region of Rev1 which mediates its binding to Rev7, the wild-type Rev1 protein and an array of Rev1 proteins where different portions of the protein had been deleted were purified from a rev7Δ yeast strain. Rev1 proteins with the amino-terminal GST fusion or without it were incubated with the Rev7 protein or the GST-Rev7 protein, respectively, and pull-down assays were performed on glutathione-Sepharose affinity beads as described previously (10); the results are summarized in Fig. 2A. In such an assay, the GST fusion protein will bind tightly to the beads, and the interacting protein will be pulled down only if it forms a complex with the protein bound to the beads. In accordance with the data shown in Fig. 1, both the GST-Rev1 and GST-Rev7 proteins were able to pull down most of the Rev7 and Rev1 proteins, respectively, from solution (Fig. 2B, panel i). In the control experiments, neither the Rev1 or Rev7 protein showed any evidence of interaction with the GST protein alone (Fig. 2B, panel vi). Interestingly, deletion of the C-terminal 239 amino acids, as in the Rev1-2 protein, or simultaneous deletion of this C-terminal region and the N-terminal BRCT domain, as in the Rev1-3 protein, had no effect on Rev1 binding to Rev7 (Fig. 2A). Also, further deletion of the N terminus beyond the BRCT domain, as in the Rev1-4 protein, which retains residues 329 to 746 (Fig. 2A), did not affect Rev1 interaction with Rev7 (Fig. 2B, panel ii). Further shortening of the C terminus to 640 amino acids, as in the Rev1-7 protein, which has residues 1 to 640 (Fig. 2A), however, abolished binding of Rev7 protein (Fig. 2B, panel iv), suggesting the involvement of the Rev1 region between residues 640 and 746 in the binding of Rev7. Importantly, the Rev1-6 protein, which has only the amino acids from 567 to 767 (Fig. 2A), is sufficient for interaction with the Rev7 protein (Fig. 2B, panel iii). In this case, however, we observed only the binding of Rev1-6 to GST-Rev7 (Fig. 2B, panel iii, lanes 21 to 24) but not the binding of Rev7 to GST-Rev1-6 protein (Fig. 2B, panel iii, lanes 17 to 20). Very likely, the lack of Rev7 binding to GST Rev1-6 is due to steric interference resulting from the proximity of GST to the Rev1 binding region. We confirmed the interaction of Rev1-6 with Rev7 by two-hybrid analysis, where we observed a fivefold increase in β-galactosidase activity in yeast cells expressing the GAL4 AD-Rev1-6 and GAL4 BD-Rev7 proteins (Table 1). Also, and as expected, the Rev1-8 protein, which carries the C-terminal 200 residues of Rev1 (Fig. 2A), does not bind to Rev7 (Fig. 2B, panel v). From these results, we conclude that the region of Rev1 from residues 567 to 767, which encompasses the linker and PAD regions of the protein, is both necessary and sufficient for interactions with Rev7. Also, the observation that while the Rev1-2 or Rev1-4 protein binds Rev7, but the Rev1-7 protein does not (Fig. 2), provides clear evidence for the requirement of the Rev1 PAD region for Rev7 binding.

FIG. 2.

Mapping the Rev1 region needed for interaction with Rev7. (A) Schematic representation of wild-type Rev1 and various Rev1 mutant proteins. The yeast Rev1 protein is comprised of 985 amino acids, and it shares conserved motifs I to V with the other Y family DNA polymerases. Although not as conserved as these five motifs, all Y family polymerases have another domain, the PAD, which follows after motif V. Motifs I and II constitute the fingers, motifs III and IV form the palm, and motif V forms the thumb. The PAD is joined to the thumb via a flexible linker. In addition to motifs I to V and PAD, which are present in all Y family polymerases, Rev1 also contains an amino-terminal BRCT domain and a carboxyl-terminal domain (CTD). For the Rev1 proteins with different portions deleted, the amino acids that remain in the protein are shown, and the results of direct physical interaction with Rev7 are summarized. (B) The linker and PAD regions of Rev1 are necessary and sufficient for interaction with Rev7. Yeast Rev7 or Rev1 protein was mixed with GST Rev1 or GST Rev7 proteins, respectively. About 3 μg of each protein was used for this study. After incubation, samples were bound to glutathione-Sepharose beads; this was followed by washings and elution of the bound proteins by SDS-sample buffer. Aliquots of each sample, taken before addition to the beads (L), from the flowthrough fraction (F), from the last washing fraction (W), and from the eluted proteins (E), were analyzed on an SDS-12% polyacrylamide gel stained with Coomassie blue.

Enzymatic properties of Rev1 and Rev1-Rev7 complex.

Although Rev1 is most proficient at inserting a C opposite template G, it also inserts a C opposite templates A, T, and C. First, we examined whether interaction with Rev7 modifies the ability of Rev1 to incorporate a C opposite template C. DNA substrates containing a C template nucleotide next to the primer-template junction were incubated with Rev1 in the presence of just one dNTP (Fig. 3A). Similar to the Rev1 protein, the Rev1-Rev7 complex predominantly incorporated a C residue across from the template C. Thus, interaction with Rev7 has no effect on the nucleotide incorporation specificity of Rev1. Next, by analysis of steady-state kinetics, we determined whether interaction with Rev7 affects the catalytic efficiency of Rev1 (Fig. 3B). The kinetics of insertion of C opposite template G by the Rev1 and the Rev1-Rev7 complex was determined as a function of dCTP concentration under steady-state conditions. From the kinetics of deoxynucleotide incorporation, the steady-state apparent K_m and V_max values for C incorporation were obtained from the curve fitted to the Michaelis-Menten equation. As indicated by the V_max/K_m values, the Rev1 protein and the Rev1-Rev7 complex incorporated C opposite template G with about the same efficiency (Fig. 3B). Thus, interaction with Rev7 has no significant effect on the enzymatic properties of Rev1.

FIG. 3.

Nucleotide incorporation by Rev1 and Rev1-Rev7. (A) Specificity of deoxynucleotide incorporation by Rev1 and Rev1-Rev7 enzymes. The nucleotide sequence adjacent to the primer-template junction is shown for the DNA substrate. Rev1 (5 nM) or Rev1-Rev7 complex (5 nM) were incubated with the DNA substrate (20 nM) in the absence (−) or presence of a single nucleotide (G, A, T, or C) or all four deoxynucleotides (N) (100 μM each) for 10 min at 30°C. The reaction products were resolved on 10% polyacrylamide gels containing 8 M urea; this was followed by autoradiography. (B) Steady-state kinetic analysis of dCTP incorporation opposite an undamaged G by Rev1 and Rev1-Rev7. Rev1 (2 nM) and Rev1-Rev7 complex (2 nM) were incubated with the primer template DNA (20 nM) and with the indicated concentrations of dCTP for 10 min at 30°C. The rate of incorporation as a function of dCTP concentration was measured, and the data were fitted to the Michaelis-Menten equation. The V_max and K_m parameters were obtained from the fit, and the efficiency (V_max/K_m) of dCTP incorporation was calculated.

DISCUSSION

Here we show that the yeast Rev1 and Rev7 proteins form a stable complex and that the region of Rev1 which encompasses the linker and the PAD mediates its interaction with Rev7. The observation that the Rev1 and Rev7 proteins form a stable complex in yeast as they do in human cells points to the conservation of Rev1 function among eukaryotes. However, the involvement of different regions of Rev1, the PAD in yeast versus the C terminus in humans, in the binding of Rev7, is surprising, given the overall conservation of Rev1 in eukaryotes. Perhaps, the different methodologies used for the demarcation of interaction domains in yeast and human Rev1 proteins account for this discrepancy. In contrast to our studies with the yeast Rev1-Rev7 complex, where direct physical interactions have been examined using purified proteins, for the human Rev1-Rev7 study, only the yeast two-hybrid analyses were performed (31), in which case the possibility of other proteins affecting the interaction results cannot be excluded. Another possibility which cannot be discounted from the published results is that the human Rev1 protein uses two sites for Rev7 binding, the region of the PAD and the C terminus (26, 31).

DNA polymerases of the Y family contain five conserved motifs, I to V (Fig. 2A). These motifs collectively comprise the palm domain, which harbors the invariant acidic residues necessary for catalysis, and the fingers and thumb domains. All these regions affect dNTP incorporation activity and are essential for DNA synthetic activity. In addition, a region of ∼120 amino acids which lies just C terminal to motif V forms the PAD which is joined to the thumb by a flexible linker (36). Although the PAD is not directly involved in any of the reactions for dNTP incorporation, it greatly increases the DNA binding capacity of Y family Pols and is therefore indispensable for their DNA synthetic ability (36). Furthermore, because of its connection with the thumb via the flexible linker and its proximity to the fingers domain, the PAD could modulate the dNTP incorporation efficiency/specificity opposite undamaged and damaged DNA templates.

The crystal structures of Y family polymerases have indicated that despite a lack of sequence conservation, the PAD shares the same structural features of β sheets buttressed by two long α-helicases (36). Because the β sheets directly interact with the DNA backbone, it is unlikely that they are involved in binding to Rev7. Instead, the long α-helices, in addition to maintaining the structural integrity of the PAD, could function in mediating specific protein-protein interactions.

The involvement of PAD in direct protein-protein interactions, as shown here for yeast Rev1, reveals another potentially important role for this region in Y family polymerases. Although the binding of Rev7 to Rev1 PAD has no significant effect on Rev1 activity, there remains the possibility that the binding of the Rev1 PAD to other protein factors affects its activity and damage bypass ability.

The involvement of the Rev1 PAD in protein-protein interactions raises the interesting possibility that in other Y family polymerases also, this region mediates functionally important protein-protein interactions. Because of the proximity of the PAD to the fingers domain, its connection with the thumb via the flexible linker, its nonspecific binding to the major groove of DNA, and its additional role as a landing pad for the other proteins that we discuss here, the PAD region could potentially affect polymerase function in many ways.

What might be the biological significance of Rev1/Rev7 interaction? One possibility is that Rev1 mediates its function in Polζ-dependent TLS through direct physical interactions with Rev7 in Polζ, and the Rev1 interaction with Rev7 that we report here is merely reflective of that association. This notion, however, is not supported by our recent observations, which indicate that although Rev1 physically interacts with Polζ, this interaction occurs through the Rev3 subunit and not through the Rev7 subunit of Polζ (N. Acharya, R. E. Johnson, S. Prakash, and L. Prakash, unpublished observations). Since Rev7 is an integral structural and functional component of Polζ, this observation is more in line with the view that Rev1-Rev7 and Rev3-Rev7 complexes do not physically interact and that they represent separate functional entities.

The cellular abundance of Rev7 is much greater than that of either Rev1 or Rev3, and Rev7 forms heterodimeric complexes with Rev1 and Rev3, respectively. However, the role that Rev7 performs in these two complexes appears to be quite different, suggestive of a dual role. In Polζ, Rev7 is needed for Rev3 to express its DNA polymerase activity, whereas in the Rev1-Rev7 complex, Rev7 produces no significant effect upon Rev1 DNA synthetic activity. This raises the possibility that in Rev1-Rev7, the role of Rev7 is to modulate protein-protein interactions at the replication fork.

How might Rev1 function in TLS compared to Rev1-Rev7? Since Rev1 directly interacts with Polζ, Rev1 alone may be instrumental in the targeting of Polζ to the major replicative polymerase, Polδ, stalled at a lesion site. Although the manner by which the Rev1-Rev7 complex may contribute to TLS remains unclear, it is unlikely to be involved in the targeting of Polζ to the replication fork stalled at a lesion site. Instead, we presume that Rev1-Rev7 functions in TLS independently of Polζ and in association with a polymerase other than Polδ, such as, for example, Polɛ.