Different types of interaction between PCNA and PIP boxes contribute to distinct cellular functions of Y-family DNA polymerases

Translesion DNA synthesis (TLS) by the Y-family DNA polymerases Polη, Polι and Polκ, mediated via interaction with proliferating cell nuclear antigen (PCNA), is a crucial pathway that protects human cells against DNA damage. We report that Polη has three PCNA-interacting protein (PIP) boxes (PIP1, 2, 3) that contribute differentially to two distinct functions, stimulation of DNA synthesis and promotion of PCNA ubiquitination. The latter function is strongly associated with formation of nuclear Polη foci, which co-localize with PCNA. We also show that Polκ has two functionally distinct PIP boxes, like Polη, whereas Polι has a single PIP box involved in stimulation of DNA synthesis. All three polymerases were additionally stimulated by mono-ubiquitinated PCNA in vitro. The three PIP boxes and a ubiquitin-binding zinc-finger of Polη exert redundant and additive effects in vivo via distinct molecular mechanisms. These findings provide an integrated picture of the orchestration of TLS polymerases.


INTRODUCTION
Translesion DNA synthesis (TLS), a DNA damage tolerance mechanism, is a crucial biological function that protects cells from various genotoxic agents. Particularly in humans, DNA polymerase (Pol), a Y-family DNA polymerases (1), plays an important role in preventing cell death and mutagenesis after ultraviolet (UV) light irradiation, and malfunction of Pol causes the inherited genetic disorder, xeroderma pigmentosum variant (XP-V) (2)(3)(4).
Interactions between proliferating cell nuclear antigen (PCNA) and the three Y-family human DNA polymerases (Pol, Pol and Pol) are critically involved in regulation of TLS. Pol and Pol are known to contain two PCNAinteracting protein (PIP) boxes (PIP1 and PIP2) in their central and C-terminal regions, respectively, whereas Pol is known to contain only one functional PIP box (PIP1, in the central region) (5)(6)(7)(8)(9)(10)(11)(12). In DNA-damaged cells, PCNA is mono-ubiquitinated at residue K164 by the RAD6-RAD18 complex (13)(14)(15), and poly-ubiquitinated by additional factors including UBC13, MMS2, and RAD5/HLTF or SH-PRH (13,(16)(17)(18)(19). Each of the three Y-family DNA polymerases described above has one or two copies of the ubiquitin-binding domain (UBD), called UBZ (ubiquitinbinding zinc-finger) in Pol and Pol and UBM (ubiquitinbinding motif) in Pol (6). These findings support the notion that mono-ubiquitination of PCNA plays a key role in switching from replicative DNA polymerase stalled at a site of DNA damage to a DNA polymerase (such as Pol, or ) capable of carrying out TLS (14)(15)(20)(21). However, this idea is still controversial (5), and more recent publications report that Pol and Pol are able to carry out TLS independently of PCNA ubiquitination in some circumstances (22)(23)(24).
The intracellular functions of the various motifs of Pol are monitored in two ways: formation of nuclear foci containing Pol co-localized with PCNA and complementation of UV sensitivity of XP-V cells. Mutations in PIP2 strongly impair the localization of Pol, indicating that PIP2 plays a crucial role in the accumulation of this protein in replication foci (23,(25)(26)(27). However, ubz mutants also failed to accumulate in replication foci (6,26), and accumulation is barely detectable in a human cell line expressing the PCNA K164R mutant instead of endogenous PCNA (28) and in a PCNA K164R knock-in murine cell line (23), demonstrating that the PIP2-PCNA interaction itself is not sufficient for foci formation. Recently, Durando and colleagues reported an additional function of PIP2 of Pol, namely, that Pol promotes mono-ubiquitination of PCNA in a PIP2-dependent manner (29). It remains unclear how the two PIP2-mediated activities, co-localization of Pol with PCNA and promotion of PCNA mono-ubiquitination, are linked at the molecular level.
The pip2 and ubz mutants of Pol exhibit pronounced defects in foci formation, but retain the capacity to complement UV sensitivity of XP-V cells; however, the levels of complementation activity vary amongst studies by different groups (5)(6)(25)(26)30), including one report that showed no role for PIP2 in survival after UV irradiation (27). Even a pip2 ubz double mutant (25,26) and a mutant in which PIP2 and UBZ were deleted (30) still exhibited significant levels of complementation activity. By contrast, the pip1 pip2 double mutant and a mutant lacking the entire PIP1, UBZ and PIP2 domains were severely defective in complementation (5,30).
The biochemical activities of the motifs of human Pol have been studied in vitro. In primer extension assays, both PIP1 and PIP2 contributed to some extent to the stimulation of DNA synthesis in the presence of PCNA (5,8,30). Using a reconstitution system to investigate switching between Pol␦ and Pol at DNA lesions, we demonstrated that the weak enhancement of the recruitment of Pol to the 3 -end of primer DNA by the UBZ domain depended on mono-ubiquitination of PCNA (21). These biochemical properties of mutant proteins defective for PIP1, PIP2 or UBZ are well correlated with their abilities to complement UV sensitivity of XP-V cells, but not with their abilities to promote accumulation into the replication foci.
Here, we describe the molecular functions of Pol's PIP1, PIP2 and UBZ domains, together with the newly found PIP3, in UV tolerance, and present findings that resolve the controversies raised in previous reports. Our key finding is that Pol has two functionally distinct types of PIP box, one that stimulates DNA synthesis, and another that promotes PCNA mono-ubiquitination and accumulation into replication foci. Both PIP functions, together with the UBZ domain, are redundantly required for survival after UV irradiation. Additionally, we show that Pol has two PIP boxes with different functions, whereas Pol has only one PIP box involved in stimulation of DNA synthesis. Taking the results of previous reports together with the in vivo and in vitro data obtained in this study, we propose a model for the cellular functions of the various motifs of Pol// in orchestrating TLS.

Proteins
Expression plasmids were constructed as follows. Human POLH, POLI (encoding 740 amino-acid residues) (11) and POLK were cloned into pET20b(+) (Novagen) to obtain untagged proteins, and into pET15b (Novagen) to obtain N-terminally histidine-tagged proteins. Plasmids for expression of Pol, Pol ubz and His-Pol in E. coli were described previously (21,31). A truncated gene encoding Pol C (32) was cloned into pET21a(+) (Novagen). A gene encoding UBCH5c S22R was cloned into pET15b. Mutations were created by PCR, and nucleotide sequences were verified after cloning.
Pol C was purified in the same way as Pol (21), except that HiTrap Phenyl HP was used instead of an Econopack methyl column (Bio-Rad).
Pol and Pol pip1 were purified as follows. BL21 (DE3) harbouring each of the expression plasmids was grown in 5 l of LB supplemented with ampicillin (250 g/ml) at 15 • C. Pol was induced with 0.2 mM IPTG for 5 h, and then purified by sequential chromatography on HiTrap Capto MMC, Ni 2+ -charged HiTrap chelating HP, HiTrap SP HP, HiTrap Q HP and Superdex 200 columns. Note that Pol itself (without the His-tag) has weak affinity for the Ni 2+ -charged Hi-Trap chelating column. For Pol pip1 , the gel-filtration chromatography step was omitted. The peak fraction containing Pol was frozen in liquid nitrogen and stored at −80 • C.
A histidine-tagged Pol (His-Pol) was purified as follows. BL21 (DE3) harbouring the expression plasmid was grown in 3 l of LB supplemented with ampicillin (250 g/ml) at 15 • C. Pol was induced with 0.2 mM IPTG for 8 h, and then purified by sequential chromatography on Ni 2+ -charged Hi-Trap chelating HP, HiTrap SP HP, HiTrap Q HP and Superdex 200 columns. The peak fraction containing His-Pol was frozen in liquid nitrogen and stored at −80 • C.
His-UBCH5c S22R was purified as follows. BL21 (DE3) harbouring the expression plasmid was grown in 2 l of LB supplemented with ampicillin (250 g/ml) at 15 • C. His-UBCH5c S22R was induced with 0.2 mM IPTG for 16 h, and then purified by sequential chromatography on Ni 2+charged HiTrap chelating HP and HiTrap SP HP columns. The peak fraction containing His-UBCH5c S22R was frozen in liquid nitrogen and stored at −80 • C.

PCNA pull-down assays
Four microlitres of MagneticHis TM Ni Particles (Promega V8560) were re-suspended in 10 l of a binding buffer containing 20 mM HEPES-NaOH (pH 7.5), 50 mM NaCl, 10 mM imidazole, 0.2 mg/ml BSA and 1 mM DTT, and then incubated at 4 • C for 5 min with 10 pmol of each of the polymerases. After washing the beads twice with 50 l of the binding buffer, 2.5 pmol of PCNA or mUb-PCNA was introduced and incubated at 4 • C for 5 min in 25 l of binding buffer. After the beads were washed twice with 50 l of binding buffer, proteins that bound to the beads were analysed by western blotting with anti-PCNA antibody as described above.

Cell lines and cultures
SV-40 immortalized XP-V fibroblasts (XP2SASV3) and a SV-40 immortalized normal human fibroblasts (WI38VA13) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 0.584 g/L L-glutamine, 0.07 g/L penicillin and 0.15 g/L streptomycin. To obtain stably expressing cells, either wild-type or mutant Pol expression constructs were transfected into XP-V cells using the Neon R transfection system (Invitrogen), followed by 0.2 mg/ml G418 selection. For construction of expression plasmids in human cells, the indicated genes were cloned into pIRESneo2 (Clontech) to create N-terminally FLAGtagged proteins or pAcGFP1-Hyg-C1 (Clontech) to create GFP fusion proteins, as described previously (28).

Preparation of cellular fractions and western blotting
XP2SASV3 cells were transfected with expression constructs encoding either wild-type or mutant FLAG-Pol using the Neon R transfection system (Invitrogen) and incubated for 24 h. Three hours after 15 J/m 2 UVC irradiation, cells were harvested and lysed in 1% SDS in PBS to obtain whole cell lysates (WCL). In the case of fractionation, cells were suspended in lysis buffer [20 mM Tris-HCl (pH 7.5), 150 mM KCl, 25% glycerol, 0.5% NP-40, 1.5 mM MgCl 2 , 1× Complete Protease Inhibitor Cocktail (Roche), 1× Phosphatase Inhibitor Cocktail Set II (Calbiochem)], and a portion was withdrawn as WCL. Next, soluble materials (soluble fractions) were separated by centrifugation. The precipitants were resuspended in micrococcal nuclease buffer [20 mM Tris-HCl (pH 7.5), 100 mM KCl, 300 mM sucrose, 0.1% Triton X-100, 2 mM MgCl 2 , 1 mM CaCl 2 , 1× EDTA-free Complete protease inhibitor cocktail (Roche)] and incubated with 2.5 U of micrococcal nuclease (Roche) at room temperature for 10 min. After centrifugation, soluble materials were collected as chromatin fractions, and precipitates (insoluble fractions) were resuspended in 1% SDS in PBS and solubilized by sonication.

Analysis of co-localization of Pol with PCNA
XP2SASV3 cells were transfected with expression constructs encoding either wild-type or mutant FLAG-Pol using the Neon R transfection system (Invitrogen). Fortyeight hours after transfection, cells were irradiated with 15 J/m 2 UVC and incubated for 3 h. Triton-soluble materials were removed by incubation with extraction buffer (0.5% Triton X-100, PBS, 0.4 g/ml antipain, 0.4 g/ml aprotinin, 0.2 g/ml leupeptin, 0.16 g/ml pepstatin, 0.1 mM EGTA and 0.5 mM phenylmethylsulfonyl fluoride), and then the cells were fixed with 3.5% formaldehyde in PBS and permeabilized with 0.1% Triton X-100 and 3.5% formaldehyde in PBS. After sequential treatments with 70% EtOH, 100% EtOH, and acetone on ice, cells were incubated with anti-POLH (Santa Cruz Biotechnology, sc-5592) and anti-PCNA (Santa Cruz Biotechnology, sc-56) antibodies. Alexa Fluor 488-and 594-conjugated secondary antibodies (Invitrogen) were used to visualize the immune-conjugated proteins. Nuclei were visualized by staining with 2 g/ml Hoechst 33342. Images were collected using an LSM710 confocal microscope (Zeiss).

Promotion of PCNA mono-ubiquitination by Pol is dependent on PIP2 and PIP3, but independent of PIP1
The three human Y-family DNA polymerases (Pol, and ) share a basic architecture. Each protein contains a catalytic domain in the N-terminal half and various motifs/domains involved in interactions with other proteins in the C-terminal half (see Figures 1A and 4A, D). Human Pol contains multiple PIP boxes and a single copy of UBZ, which are believed to be involved in the interaction with mUb-PCNA in DNA-damaged cells. Recently, Durando and co-workers reported that depletion of endogenous Pol decreases the levels of damage-induced mUb-PCNA, and ectopically expressed Pol promotes mono-ubiquitination of PCNA in cells in a manner that depends on PIP2 at the C-terminus (29). However, those authors did not examine the contribution of another PIP box, PIP1, which is located in an internal region (5) ( Figure 1A). To determine whether PIP1 plays any role in the promotion of PCNA mono-ubiquitination, we introduced the pip1 or pip2 mutation into FLAG-tagged Pol ( Figure 1A), expressed the mutant proteins in XP-V cells, and analysed the levels of mUb-PCNA in cells by western blotting. As shown in Figure 1B, the result indicated that the pip1 mutant and the wild type promoted mono-ubiquitination of PCNA to similar extents. Importantly, PCNA ubiquitination was observed in the presence or absence of UV irradiation, although it was more extensive when the cells were UV-irradiated. By contrast, the pip2 mutant lost most of the ability to promote mono-ubiquitination. Because the pip1 pip2 double mutant still exhibited weak activity, similar to that of the single pip2 mutant, the residual activity could be attributable to additional PIP box(es). To identify another PIP box in Pol, we employed yeast two-hybrid assays and found one additional PIP box (hereafter, referred to as PIP3), which overlaps with a REV1-interacting region (RIR) (11,40) ( Figure  1A; Supplementary Figure S1). Although short peptides carrying the PIP2 sequence interact with PCNA strongly enough for detailed physicochemical and structural analyses (10), the PCNA-binding activity of short peptides carrying the PIP1 sequence has never been detected, even using very sensitive yeast two-hybrid assay (11). Similarly, the PCNA-binding activity of PIP3 is not detected using short fragments; however, the activity of PIP3 appears stronger than that of PIP1, because the pip3 mutation caused a much more drastic reduction in the positive signal in the yeast two-hybrid assay than the pip1 mutation (Supplementary Figure S1). Subsequently, we made a series of pip3 mutants and expressed them in XP-V cells. As expected, the residual activities of the pip2 single and pip1 pip2 double mutants were diminished further by the additional introduction of the pip3 mutation. The levels of mUb-PCNA in pip3 pip2 double and pip1 pip3 pip2 triple mutants were similar to that in the vector control. On the other hand, the levels of mUb-PCNA in pip3 single and pip1 pip3 double mutants exhibited marginal differences from those in the wild type only in UV-unirradiated samples. Because all of these mutants were similarly detected in the chromatin fraction, the defects in the mutants were not attributed to alteration in sub-cellular localization ( Figure 1B, bottom panel). These results indicate that PIP2 and PIP3 play a major and minor role, respectively, whereas PIP1 has no or little role, in promoting mono-ubiquitination of PCNA in cells, suggesting that two functional types of PIP boxes play distinct roles in the regulation of Pol.
Additionally, we showed that the ubz mutants severely reduced the ability to promote PCNA mono-ubiquitination (Supplementary Figure S2). However, these mutants had a significantly reduced ability to accumulate in chromatin (26) (Supplementary Figure S2), indicating that the mutations have additional effects, as postulated previously (5). Because the defect in promoting PCNA mono-ubiquitination could be also attributed to additional effects, such as alteration of sub-cellular localization, it remains unclear whether UBZ has the potential to promote ubiquitination of PCNA in vivo.

Reconstitution of Pol-dependent mono-ubiquitination of PCNA in vitro
To study the molecular mechanisms underlying the promotion of PCNA mono-ubiquitination, we sought to develop in vitro experimental conditions for recapitulating the in vivo situation using purified enzymes (Supplementary Figure S3A) (21,34). Because Pol interacts with both RAD18 and PCNA (8,15,41), we first checked the possibility that such protein-protein interactions could themselves promote PCNA ubiquitination. However, we observed no ubiquitination of PCNA under such conditions (Figure 2A, lane 1). When poly(dA)-oligo(dT) was added into the reaction mixture, PCNA ubiquitination occurred and it was dependent on Pol (lanes 2 and 3) as well as on E1, RAD6-(His-RAD18) 2 and ubiquitin (lanes 4-6). No modification was observed with the PCNA K164R mutant, confirming that ubiquitination of PCNA takes place at Lys164 (lane 7). Note that PCNA is spontaneously loaded from the ends of poly(dA)-oligo(dT) without RFC in these experiments (21). Additionally, we found that the specific interaction between RAD18 and Pol was not required for promotion of PCNA mono-ubiquitination in the in vitro reactions, because a mutant of RAD18 lacking the C-terminal region required for its interaction with Pol (15,34) could promote PCNA ubiquitination as efficiently as full-length RAD18 (Supplementary Figure S4A). These results suggested that a certain mode of Pol-PCNA interaction on DNA is required for the promotion of PCNA ubiquitination, inspiring us to further investigate the mechanism.  The results demonstrated that in vitro promotion of ubiquitination is dependent on PIP3 and PIP2, but independent of PIP1 ( Figure 2B, C), in good agreement with the in vivo observations ( Figure 1B). Together, these results support the interpretation that the in vivo accumulation of mUb-PCNA is a consequence of directly promoting de novo ubiquitination by Pol.
Nucleic Acids Research, 2015, Vol. 43, No. 16 7903 In addition, we demonstrated that the ubz mutant promoted mono-ubiquitination as efficiently as the wild type (Supplementary Figure S4B), indicating that the UBZ function is dispensable for promotion of ubiquitination in our in vitro system.

Different roles of the three PIP boxes of Pol in stimulation of DNA synthesis
Next, to study how PCNA stimulates the polymerase activity of Pol, we employed the primer extension assay using M13 mp18 ssDNA as a template in the presence of RPA and RFC. As shown in Figure 3A, we clearly detected stimulation of Pol polymerase activity by PCNA. When the pip mutants were examined, the stimulation was slightly reduced in each of the single pip mutants relative to the wild type ( Figure 3B), indicating that all of the PIP boxes contribute to stimulation to some extent. The activities of the pip1 pip2 and pip3 pip2 were further reduced. Surprisingly, the activity of the pip1 pip3 double mutant was inhibited by addition of PCNA, despite the fact that it still contains a PIP2 domain ( Figure 3B). Similar inhibition was also observed with the pip1 pip3 pip2 triple mutant ( Figure 3B). Based on these results, we conclude that PCNA binding to PIP1 or PIP3, both located in the adjacent region of the Pol catalytic domain, is critical for the stimulation of DNA polymerase activity, and that even if PCNA binds to PIP2 at the C-terminus, PCNA does not stimulate Pol polymerase activity in vitro unless PIP1 or PIP3 is present. These results suggest that the function of PIP2 in stimulation of DNA synthesis is largely PIP1-and PIP3-dependent.

Functional roles of PIP boxes of Pol and Pol
The findings described above regarding the PIP boxes of Pol prompted us to investigate the PIP boxes of two other Y-family DNA polymerases, Pol and Pol. Pol has a PIP box at the C-terminus, which is required for formation of nuclear foci in cells with DNA damage (42), and has the potential to promote mono-ubiquitination of PCNA in vivo when Pol is ectopically expressed (29). More recently, another PIP box was found adjacent to the catalytic domain ( Figure 4A; Supplementary Figure S5); therefore, the internal PIP box was named PIP1, and the C-terminal one was renamed PIP2, following the example of Pol (Figure 4A) (11). Unlike the internal PIP1 and PIP3 boxes in Pol, the PCNA-binding activity of the internal PIP1 box in Pol was detected in short fragments by the yeast twohybrid assay, implying that the activity is equivalent to that of the C-terminal PIP2 in Pol (Supplementary Figure S5). When the wild-type Pol protein was introduced into the in vitro PCNA ubiquitination reaction, a large amount of ubiquitinated PCNA was observed (the leftmost panel in Figure 4B), as in the case of Pol. To study the roles of the respective PIP boxes, we examined the pip1, pip2 and double mutants ( Figure 4A; Supplementary Figure S3B). As shown in Figure 4B, the pip1 and pip2 mutants exhibited reduced levels of PCNA ubiquitination: the pip2 mutant retained a relatively higher level, whereas the double mutant lost the activity. These results indicate that each of the two PIP boxes functions independently and have similar affinity for PCNA, and that both are required to promote the maximum level of ubiquitination ( Figure 4B). Next, we examined effects of PCNA on DNA polymerase activity of the wild-type and mutant Pol proteins. As shown in Figure 4C, the pip2 mutant was stimulated as efficiently as the wild type by PCNA, but pip1 and the double mutants failed to be stimulated by PCNA, indicating that the internal PIP1 box is responsible for stimulation by PCNA but the C-terminal PIP2 is not. Therefore, we conclude that the multiple PIP boxes of Pol serve different functions: PIP1 stimulates DNA synthesis by PCNA, and PIP1 and PIP2 promote PCNA ubiquitination.
Next, we examined Pol ( Figure 4D; Supplementary Figure S3C). In contrast to Pol and Pol, Pol did not promote PCNA ubiquitination in vitro ( Figure 4E). On the other hand, PCNA stimulated DNA synthesis of Pol in vitro in a PIP1-dependent manner ( Figure 4F), in line with previous reports that Pol has only one functional PIP box (for stimulation of DNA synthesis) immediately adjacent to the catalytic domain (7,10,12).
Subsequently, we examined the levels of mUb-PCNA in cells with ectopic expression of Pol or Pol, with or without UV irradiation ( Figure 5). When GFP-Pol was expressed in Pol-deficient (XP-V) or proficient cells, promotion of PCNA mono-ubiquitination was observed in both types of cells ( Figure 5A, B). The promoting effect of ectopic expression of Pol is weaker than that of Pol reported previously (29,43). The difference in the extent to which PCNA monoubiquitination was promoted by these enzymes could be attributed to differences in the expression system and/or cell types used in these experiments. Indeed, a difference is evident between the two types of cells used in our study (Figure 5A, B), indicating that it may not be appropriate to compare and draw conclusions from differences in mUb-PCNA levels between different cell lines. By contrast to Pol and Pol, ectopic expression of FLAG-Pol in both cells did not increase the levels of mUb-PCNA ( Figure 5C, D). These results are consistent with the in vitro properties described above ( Figure 4B, E), implying that Pol, but not Pol, can promote de novo mono-ubiquitination of PCNA in vivo.

Interactions of Pol, and with mUb-PCNA
PCNA is mono-ubiquitinated in DNA-damaged cells. Consequently, Y-family DNA polymerases with UBD(s), as well as PIP box(es), could interact with modified PCNA in preference to unmodified PCNA. To test this in vitro, we compared the stimulatory effects of mUb-PCNA and unmodified PCNA on DNA synthesis in vitro, using a primer extension assay (Figure 6; Supplementary Figure S3D). As shown in Figure 6A, mUb-PCNA stimulated DNA synthesis by Pol more effectively than unmodified PCNA under our assay conditions. As expected, no additional stimulation was observed with the ubz mutant of Pol ( Figure 6B). DNA synthesis by either the pip1 pip2 or pip3 pip2 double mutants was stimulated by mUb-PCNA, although it was only marginally stimulated by unmodified PCNA (Figure 6C, D). The activity of the pip1 pip3 double mutant was slightly higher in the presence of mUb-PCNA than in the absence of PCNA ( Figure 6E). The pip1 pip3 pip2 triple mutant exhibited a negative effect by the addition of mUb-   PCNA ( Figure 6F). These results suggested that UBZ function requires at least one PIP.
Next, we examined the stimulation of DNA synthesis by Pol and Pol by mUb-PCNA. Additional stimulation, albeit marginal, by mUb-PCNA was reproducibly observed for Pol ( Figure 6G) as well as its pip1 mutant ( Figure 6H). Similarly, mUb-PCNA stimulated Pol and its pip mutants to a slightly greater extent than unmodified PCNA (Figure 6I, J). The relatively lower contributions of ubiquitin moieties to Pol and Pol activity than to Pol activity could be attributed to the weaker affinity of Pol and Pol for mUb-PCNA than Pol (Supplementary Figure S6).

Cellular functions of the motifs of Pol
After UV irradiation, Pol forms nuclear foci that colocalize with PCNA (6,15,(25)(26)(27)39,44). To determine the roles of each of Pol's three PIP boxes in foci formation, wild type and pip mutants bearing a FLAG-tag, all of which were used in the experiments shown in Figure 1, were transiently expressed in XP-V cells. After UV irradiation, localization of Pol and PCNA was visualized using anti-Pol and anti-PCNA antibodies, respectively. As shown in Figure 7A, Pol foci co-localized with PCNA were observed in all of the samples except for the pip2 mutant. Because all of the proteins, including the pip2 mutant, could be detected with similar efficiency by western blotting ( Figure 1B) and immunostaining (Supplementary Figure S7), we conclude that PIP2, but not PIP1 or PIP3, plays a crucial role in foci formation along with PCNA.
Next, we examined the abilities of the Pol mutants to complement the UV sensitivity of XP-V cells ( Figure 7B). Although the pip and ubz mutants of Pol have been analysed previously using such assays, the levels of complementation were inconsistent among studies (5)(6)(25)(26)(27)30). Complementation of the UV sensitivity of XP-V cells differs among clones stably expressing a particular mutant Pol, but is not correlated with the expression levels of Pol (26). To avoid such complexities due to differences among clones, we used for our survival assays a mixture of the cells that were transfected with pIRESneo2 carrying wild-type or mutant Pol and selected by G418. As shown in Figure 7B, the results indicated that the single and double pip mutants could complement the UV sensitivity of XP-V cells as efficiently as the wild type. By contrast, the pip1 pip3 pip2 triple mutant exhibited clearly reduced complementation activity ( Figure 7B).
In contrast to such subtle phenotypes of individual single and double pip mutants, the ubz single mutant exhibited a severe defect (Supplementary Figure S8) (6,26). However, because the ubz mutant accumulated poorly in the chromatin fraction (see Supplementary Figure S2) (26), we hypothesized that this defect could be attributed to secondary effects due to the ubz mutation. To investigate this possibility, we made use of a deletion mutant carrying the 1-511 region of Pol (Pol C) (2-3,32) ( Figure 1A). Because Pol C lacks the nuclear localization signal (NLS) as well as PIP2 and UBZ, we introduced an artificial NLS at the C-terminus. Because of the lack of PIP2, Pol C was expected to have a lower ability to promote PCNA ubiquitination and fail to form foci in co-localization with PCNA. Those properties were confirmed in vivo and in vitro (Supplementary Figure S9). Nevertheless, Pol C retained the ability to accumulate in the chromatin fraction (Supplementary Figure S9D), in contrast to the ubz mutants of fulllength Pol (Supplementary Figure S2), supporting the idea that some ubz mutations provoke secondary effects (5,30). More importantly, Pol C could complement the UV sensitivity of XP-V cells much better than the ubz mutant (Supplementary Figure S8). As shown in Figure 7B, the pip1 and pip3 derivatives of Pol C exhibited reduced complementation activity, and the pip1 pip3 derivative of Pol C exhibited a severer defect than the pip1 pip3 pip2 triple mutant of full-length Pol. The defects were not attributable to alterations in the sub-cellular localization (Supplementary Figure S9D). Together, these results suggest that PIP1, PIP3, PIP2 and UBZ exert additive and redundant effects that protect cells from the lethal effects of UV irradiation.

DISCUSSION
Cellular functions of the respective motifs of Pol are routinely monitored in two ways: co-localization with PCNA and complementation of UV sensitivity of XP-V cells. In this study, we demonstrated that these two phenotypes are mediated by different PIP boxes and distinct modes of interaction with PCNA. We also showed that the diverse functions of PIP boxes are conserved in Pol, but not in Pol.

Functions of PIP boxes in ubiquitination of PCNA and foci formation
In this study, we found that Pol promotes monoubiquitination of PCNA in a manner dependent on PIP2 and to a lesser extent on PIP3, but independent of PIP1 in vivo ( Figure 1B). These findings were perfectly correlated with the in vitro observations regarding promotion of PCNA mono-ubiquitination by purified proteins (Figure 2B, C). Together, the data strongly suggest that the intracellular accumulation of mUb-PCNA with ectopically expressed Pol is a consequence of direct promotion of de novo ubiquitination. Importantly, we found that DNA is an absolute requirement for Pol-dependent PCNA ubiquitination reactions in vitro (Figure 2A). We suggest that Nucleic Acids Research, 2015, Vol. 43, No. 16 7907  the mode of interaction between PIP3 or PIP2 and PCNA on DNA for the promotion of PCNA ubiquitination could act in such a way as an appropriate substrate for RAD6-(RAD18) 2 catalysis, which was independent of the interaction between RAD18 and Pol in vitro. The partial involvement of the interaction in the promotion of PCNA ubiquitination in vivo (29) could be attributed to an additional function, such as recruitment of Pol to damage sites (15).
The contribution of individual Pol PIP boxes to colocalization with PCNA was correlated with the effect on promotion of PCNA ubiquitination. These effects could be largely attributed to PIP2 (Figures 1B, 2B, C and 7A). Pol also promotes mono-ubiquitination of PCNA (Fig-ures 4B and 5A, B) (29) and co-localizes with PCNA (42). Pol failed to promote mono-ubiquitination of PCNA (Figures 4E and 5C, D) and failed to co-localize with PCNA by itself (45,46). These results suggest that a large part of the function of the PIP2 box in foci formation is the promotion of PCNA mono-ubiquitination, which is a prerequisite for co-localization of pol with PCNA in nuclear foci. The following observation supports this idea: first, accumulation is dependent on RAD18 and its catalytic activity (15,29). Second, accumulation of Pol is barely detectable in a human cell line in which the PCNA K164R mutant is expressed instead of endogenous PCNA (28), or in a PCNA K164R knock-in murine cell line (23). Therefore, we suggest that one of the functions of PIP2 in foci formation is to promote mono-ubiquitination of PCNA. The resultant mUb-PCNA could stabilize Pol via interaction with UBZ, because mUb-PCNA has a higher affinity for Pol than unmodified PCNA (Supplementary Figure S6). Thus, even though UBZ is dispensable for the promotion of PCNA ubiquitination (at least in vitro), UBZ may still contribute to stable foci formation. This idea does not exclude another role for PIP2 in foci formation via direct interaction with PCNA. Indeed, among the three PIPs of Pol, only the PIP2 peptide has been shown to interact directly with PCNA in yeast two-hybrid and structural analyses (10,11). Furthermore, enhancement of the PIP2-PCNA interaction of Pol by manipulations of its PIP2 sequence (10) improves foci formation (29,47). These results suggest that stable interactions with mUb-PCNA via both PIP and UBZ are required for detectable focus formation.
Interestingly, PIP3 and RIR1 share the same FF residues for binding to PCNA and REV1, respectively (40). Therefore, it is unlikely that PIP3 could act as a PIP box and RIR at the same time. However, Pol has two RIRs, RIR1 and RIR2 (40). So far, we have been unable to detect any defect in the pol-REV1 interaction in the rir1 single mutant in cells (39) or in yeast-two hybrid assays (40) (Supplemental Figure S1). We believe that the PIP3-PCNA interaction or pip3 mutation does not interfere with REV1-related function(s) because of the presence of RIR2.

Functions of PIP and UBD in stimulating DNA synthesis
In this study, we demonstrated that all three PIP boxes of Pol play roles in stimulating DNA synthesis, although the mechanisms are different. We showed that PIP-less Pol has an intrinsic defect in accessing the PCNA-loaded 3 -end; the presence of PIP1 or PIP3, but not PIP2, compensates for this inhibition ( Figure 3B). It is likely that both PIP1 and PIP3 stimulate DNA synthesis via direct interaction with PCNA, and that PIP2 facilitates the PIP1-PCNA and PIP3-PCNA interactions. Because PIP1 and PIP3 are proximal to the catalytic domain of Pol, the PCNA-Pol interaction via PIP1 or PIP3 may enable the active site of Pol to effectively bind the 3 -OH end of the primer-terminus, whereas the interaction via PIP2, located distant from the C-terminus, may be less efficient in this respect. This mechanism seems conserved in Pol and Pol, since each PIP1 box located adjacent to the catalytic domain can stimulate DNA synthesis. More recently, during the preparation of this manuscript, another group reported that PIP1 in Pol is responsible for PCNA stimulation of in vitro DNA synthesis (48).
We demonstrated for the first time that mUb-PCNA stimulates DNA synthesis by human Pol, and (Figure 6A-H). We showed that the UBZ-function of Pol is largely PIP1-and PIP3-dependent and slightly PIP2dependent ( Figure 6A-F), suggesting that adequate stimulation by mUb-PCNA requires interaction with at least one PIP box. Alternatively, an increased local concentration of Pol around the primer end, achieved through PIP-PCNA interactions, indirectly promotes the interaction between UBZ and the ubiquitin moiety. Collectively, these results suggest that interactions between mUb-PCNA and Pol stimulate DNA synthesis via diverse mechanisms.

Functions of PIP and UBZ in Pol in enhancing survival after UV irradiation
All three PIP boxes and UBZ of Pol serve additive and redundant functions in enhancing survival after UV irradiation ( Figure 7B). The survival rate in vivo ( Figure 7B) correlated well with the level of stimulation of DNA synthesis by these mutants by PCNA or mUb-PCNA in vitro ( Figure 6A-H). However, there were two exceptions. One was the ubz point mutant that exhibited severe defects (Supplementary Figure S8), which might be attributed to a secondary effect related to accumulation on chromatin (Supplementary Figure S2). Indeed, the deletion mutation of UBZ and PIP2 in Pol C restored the sub-cellular localization (Supplementary Figure S9D) and increased the ability to complement the UV sensitivity of XP-V cells relative to the ubz point mutant (Supplementary Figure S8). This partial contribution of UBZ is consistent with reports of PCNA ubiquitination-independent TLS in PCNA K164R knock-in murine cells (22,23). The other exception was the pip1 pip3 double mutant. The defect in complementation of UV sensitivity of XP-V cells by the pip1 pip3 mutant was marginal ( Figure 7B), although its DNA synthesis was poorly stimulated by mUb-PCNA in vitro ( Figure 6E). Because the mutant is proficient in the promotion of PCNA ubiquitination and foci formation ( Figures 1B, 2B and 7A), the elevated local concentration of the mutant could compensate for the defect in stimulation, and this could explain the minor defect of the pip1 pip3 mutant in vivo. Overall, the data suggest that the level of stimulation of DNA synthesis by mUb-PCNA could directly affect the efficiency of TLS in vivo.

A model for the cellular functions of PIP and UBD of Pol // in the orchestration of TLS
Taking the results of previous reports together with the in vivo and in vitro data obtained in this study, we propose a model for cellular functions of the various motifs of Pol// ( Figure 7C). We suggest that interactions between Pol/ and PCNA constitute a network that regulates promotion of mono-ubiquitination and accumulation of Pol//. PCNA is concentrated on DNA in close proximity to replication forks (21,35,49). In the model, Pol/ are recruited to these locations via interactions with PIP boxes together with RAD18 (15,41,43). The initial PIP-PCNA interactions are too transient and unstable to be detected as foci, but the interaction turns PCNA into an appropriate substrate for RAD6-(RAD18) 2 catalysis, which promotes ubiquitination of PCNA. The resultant mUb-PCNA recruits additional Pol// molecules via interaction with both PIPs and UBDs. Because Pol has a much higher affinity for mUb-PCNA than Pol and Pol (Supplementary Figure S6), it is likely that Pol is predominantly recruited. In the case of Pol, the interactions are sufficiently stable for detectable foci formation, but in the case of Pol they are relatively weak. By contrast, Pol binds mUb-PCNA too weakly to form stable foci by itself. Increased local concentrations of Pol/ further promote ubiquitination of PCNA. Consequently, Pol/ and ubiquitinated PCNA robustly accumulate around the stalled primer ends until accumulation of Pol/ is saturated. Any mutation that disrupts the regulatory network should attenuate the response in vivo via activities of de-ubiquitination enzymes. This model is compatible with the dynamic mobile properties of Pol/ in cells (50).
Accumulation of Pol in replication foci plays a minor role in survival, as reflected by the observation that accumulation-defective PIP2 mutant could perfectly complement UV sensitivity of XP-V cells ( Figure 7B) (27). In addition, Pol C with a C-terminal deletion encompassing UBZ and PIP2 also exhibited considerable complementation activity ( Figure 7B). These results suggest that Pol can perform TLS of UV lesions without accumulating in foci. The minor contribution of foci formation to cellular function is also true in other members of the Y family. Although accumulation of Pol in replication foci is dependent on Pol (45,46), Pol appears to be functional in TLS of UV lesions in Pol-knockout mice (51). Deficiency in the Poldependent co-localization of REV1 with UV lesions does not affect survival, but does modulate mutagenesis (39). In this scenario, one of physiological functions of the accumulated mUb-PCNA could be to establish an order of recruitment for TLS polymerases, via dynamic interactions between mUb-PCNA and their PIPs and UBDs, depending on their respective affinities around the stalled primer ends (50). This possibility is compatible with the observation that Pol can be induced by specific agents that produce DNA damage cognate for Pol (52). Increasing the ratio of Pol forces it to predominantly access specific DNA damage. We believe that this model provides an integrated picture of the cellular functions of various motifs of the Y-family DNA polymerases.

SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.