The Functions of Serine 687 Phosphorylation of Human DNA Polymerase η in UV Damage Tolerance

Abstract

DNA polymerase η (polη) is a Y-family translesion synthesis polymerase that plays a key role in the cellular tolerance toward UV irradiation-induced DNA damage. Here, we identified, for the first time, the phosphorylation of serine 687 (Ser⁶⁸⁷), which is located in the highly conserved nuclear localization signal (NLS) region of human polη and is mediated by cyclin-dependent kinase 2 (CDK2). We also showed that this phosphorylation is stimulated in human cells upon UV light exposure and results in diminished interaction of polη with proliferating cell nuclear antigen (PCNA). Furthermore, we demonstrated that the phosphorylation of Ser⁶⁸⁷ in polη confers cellular protection from UV irradiation and increases the efficiency in replication across a site-specifically incorporated cyclobutane pyrimidine dimer in human cells. Based on these results, we proposed a mechanistic model where Ser⁶⁸⁷ phosphorylation functions in the reverse polymerase switching step of translesion synthesis: The phosphorylation brings negative charges to the NLS of polη, which facilitates its departure from PCNA, thereby resetting the replication fork for highly accurate and processive DNA replication. Thus, our study, together with previous findings, supported that the posttranslational modifications of NLS of polη played a dual role in polymerase switching, where Lys⁶⁸² deubiquitination promotes the recruitment of polη to PCNA immediately prior to lesion bypass and Ser⁶⁸⁷ phosphorylation stimulates its departure from the replication fork immediately after lesion bypass.

Living cells are constantly exposed to various DNA-damaging agents, such as UV irradiation, chemical carcinogens, and endogenous reactive oxygen species (1). The resulting DNA damage can block the progression of replication forks and/or induce mutations, which result in cell transformation, senescence, or apoptosis (2). To overcome replication blockages, cells are equipped with multiple DNA damage surveillance and repair systems, including translesion synthesis (TLS) (3). A polymerase switching model has been proposed for TLS, where the replicative DNA polymerase δ or ε is switched out for specialized TLS polymerases, a process thought to be mediated by the monoubiquitination of proliferating cell nuclear antigen (PCNA) (4). Replicative DNA polymerases are highly accurate in DNA synthesis, whereas DNA synthesis mediated by TLS polymerases is often error-prone (5, 6). Thus, timely removal of TLS polymerases from the replication fork and restoration of replicative polymerases after TLS are essential for avoiding undesirable mutagenesis during genome replication (7).

DNA polymerase η is a Y-family polymerase that plays a critical role in the TLS across UV-induced cyclobutane pyrimidine dimers (CPDs) (8–10). Deficiency in the gene encoding polη¹ results in xeroderma pigmentosum variant syndrome in humans (11, 12). Xeroderma pigmentosum variant patients manifest high sensitivity to sunlight irradiation, and they are prone to developing skin cancer (11, 12). During TLS, PCNA becomes monoubiquitinated, which recruits polη to the replication fork through polη's ubiquitin-binding zinc finger (UBZ) and PCNA-interacting peptide (PIP) domains (13–16). A recent study has extended the PCNA-interacting region of polη to amino acid residues 682–713, and PCNA-interacting region of polη can be monoubiquitinated, which masks its PCNA-interacting region and inhibits its interaction with PCNA (17). After UV irradiation, the monoubiquitination of polη is down-regulated to expose the PCNA-interacting surface, thereby stimulating polη's interaction with PCNA and facilitating TLS (17). In addition, previous studies showed that polη can be phosphorylated at serine 587, threonine 617, and serine 601 (Ser⁶⁰¹), which are required for normal survival and efficient recovery from UV damage (18, 19). It is therefore important to identify other possible posttranslational modification(s) of human polη for further understanding the regulatory mechanisms of this important TLS polymerase.

MATERIALS AND METHODS

Cell Culture

The HEK293T human embryonic kidney epithelial cells (ATCC) and XP30RO human skin fibroblasts (20, 21) were cultured at 37 °C in Dulbecco's modified Eagle medium (DMEM) containing 10% fetal bovine serum (Invitrogen) and 100 IU/ml penicillin (ATCC) in an incubator containing 5% CO₂.

Construction of Vectors

Three tandem repeats of the FLAG epitope tag (DYKDDDDK) were inserted into BamHI and EcoRI sites of pRK7 vector to produce the pRK7–3×FLAG vector. Human POLH gene was amplified from HEK293T cells by reverse transcription-PCR to introduce a 5′ XbaI site and a 3′ BamHI site and subcloned into pRK7–3×FLAG vector. The pEGFPC3-polη vector was a gift from Prof. Alan R. Lehmann (22). GST-polη_C vector was generated by cloning the C-terminal portion of polη encoding the last 112 amino acids into pGEX-4T-1 vector that were pretreated with BamHI and XhoI. The polη-S687A, polη-S687D, PIP (L704A, FF708–709AA), and PIP/UBZ (D652A, L704A, FF708–709AA) mutants were amplified by site-directed mutagenesis using primers containing the indicated mutations. The primers are listed in Supplemental Table S1.

Preparation of FLAG-Tagged Polη Protein and Mass Spectrometric Analysis

HEK293T cells were grown to 70–80% confluency in a six-well plate and transfected with 1.5 μg FLAG-tagged polη plasmid using Lipofectamine 2000 (Invitrogen). After a 48-h incubation, the protein was isolated and purified using anti-FLAG M2 beads (Sigma). The purified protein was digested with trypsin (Roche) at an enzyme/substrate ratio of 1:50 and subjected to LC-MS/MS analysis.

LC-MS/MS experiments were performed as previously described (23). Briefly, the peptides were separated on an EASY-nLC II and analyzed on an LTQ Orbitrap Velos mass spectrometer equipped with a nanoelectrospray ionization source (Thermo). The trapping column (150 μm × 50 mm) and separation column (75 μm × 120 mm) were packed with ReproSil-Pur C18-AQ resin (3 μm in particle size, Dr. Maisch HPLC GmbH, Germany). The peptide samples were first loaded onto the trapping column in CH₃CN/H₂O (2:98, v/v) at a flow rate of 4.0 μl/min and resolved on the separation column with a 120-min linear gradient of 2–40% acetonitrile in 0.1% formic acid and at a flow rate of 300 nl/min. The LTQ-Orbitrap Velos mass spectrometer was operated in the positive-ion mode, and the spray voltage was 1.8 kV. The full-scan mass spectra (m/z 300–2000) were acquired with a resolution of 60,000 at m/z 400 after accumulation to a target value of 500,000 in the linear ion trap. MS/MS data were obtained in a data-dependent scan mode where one full MS scan was followed with 20 MS/MS scans. The LC-MS/MS data were employed for the identification of polη and its posttranslational modifications, which was conducted using Maxquant, version 1.2.0.18 against International Protein Index (IPI) database, version 3.68 with 87,061 entries to which contaminants and reverse sequences were added. The maximum number of miscleavages for trypsin was two per peptide. Cysteine carbamidomethylation was set as fixed modifications. Methionine oxidation, lysine ubiquitination, and serine/threonine phosphorylation were set as variable modifications. The tolerances in mass accuracy for MS and MS/MS were 25 ppm and 0.6 Da, respectively. The required false positive discovery rate was set to 1% at both the peptide and protein levels, with the minimal required peptide length being set at six amino acids. For quantitative analysis, MS/MS experiments were conducted in selected-ion monitoring mode, and the peak areas found in the selected-ion chromatograms for monitoring the formation of a specific fragment ion from the unmodified and the corresponding modified peptide were employed for estimating the levels of modifications. In particular, the y₅ ion (m/z 593.3) and the [MH-Pho]²⁺ ion (m/z 606.3) were used to assess the relative levels of unmodified (₆₈₇SPLACTNK₆₉₄) and phosphorylated (₆₈₄NPKS_phoPLACTNK₆₉₄) forms of Ser⁶⁸⁷-containing peptides, respectively.

UV Irradiation

The wild-type or mutant polη plasmids were transfected into HEK293T cells as described above. After a 48-h incubation, the medium was replaced with PBS, and the cells were treated with 30 J/m² UV light (254 nm) and recovered for 6 h in DMEM medium in a CO₂ incubator at 37 °C, unless otherwise noted.

Preparation of Recombinant Proteins

PCNA-His₆ (24), and the GST-polη_C proteins were obtained by inducing transformed Rosetta (DE3) pLysS Escherichia coli cells with 1 mm isopropyl 1-thio-β-D-galactopyranoside when the culture reached OD₆₀₀ ≈ 0.6 and culturing at 37 °C for 4 h. Subsequently, the PCNA-His₆ and GST-polη_C proteins were purified by using Talon affinity resin (Clontech) and glutathione agarose (Pierce), respectively, following the manufacturers' recommended procedures.

In Vitro Protein Kinase Assay

The assay was performed with purified CDK2/Cyclin A2 complex (Sigma) according to the manufacturer's instructions. Briefly, 10 μg recombinant GST-polη_c was incubated with 250 μm ATP and a serial dilution of CDK2/Cyclin A2 in kinase assay buffer, which contained 5 mm MOPS (pH 7.2), 2.5 mm glycerol 2-phosphate, 1 mm EGTA, 0.4 mm EDTA, 5 mm MgCl₂ and 0.05 mm DTT. The reaction mixture (25 μl) was incubated at 30 °C for 15 min and boiled in 1× SDS loading buffer for 5 min to terminate the reaction. The resulting protein mixture was reduced, alkylated, and digested with trypsin, and the peptide mixture was subjected to LC-MS/MS analysis, as described above.

Assay for Monitoring the In Vitro Interaction between Polη and PCNA

The assay for monitoring the physical interaction between polη and PCNA was performed as previously described (13). Briefly, wild-type or mutant GST-polη_C protein (4 μg) was mixed with PCNA-His₆ (4 μg) in a binding buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 0.01% Nonidet P-40, 10% glycerol) and incubated at 4 °C for 30 min followed by 10 min at 25 °C. Subsequently, 10 μl of glutathione agarose (Pierce) was added to bind GST-polη_C. The samples were further incubated with rocking at 4 °C for 1 h. The beads were then washed five times using the binding buffer, and the bound proteins were eluted using the binding buffer containing 10 mm reduced glutathione and analyzed using Western blotting. The antibody that specifically recognized human PCNA (Cell Signaling Technology) was used at 1:2,000,000 dilution.

RNA Interference and Western blotting

RNA interference was carried out as described previously (23). Briefly, HEK293T cells were seeded in six-well plates at 40–60% confluence level and transfected with 100 pmol siRNAs using Lipofectamine 2000 (Invitrogen). After a 48-h incubation, 1.5 μg polη expression plasmids were cotransfected into the cells together with another aliquot of siRNA. The knockdown efficiency of CDK2 was evaluated at 48 h after transfection by Western blotting as described previously (25). The antibodies that specifically recognized human CDK2 (Cell Signaling Technology) and β-actin (Abcam) were used at 1:20,000 and 1:5,000 dilutions, respectively.

Flow Cytometry-Based Assay for Monitoring the Levels of CPD Lesion

The cellular levels of CPD lesion were monitored by flow cytometry according to previously described procedures (26, 27). Briefly, HEK293T cells were irradiated with 10 J/m² UV light and cultured in DMEM containing 100 ng/ml nocodazole. At various time intervals following UV irradiation, the cells were harvested and incubated first with mouse anti-CPD antibody (Kamiya Biomedical Co., 1:1000 dilution) and then with Alexa Fluor 647 goat anti-mouse secondary antibody (Invitrogen, 1:200 dilution). The cells were then analyzed by flow cytometry (BD Biosciences FACSAria I).

Fluorescence Microscopy

For assessing the nuclear localization of EGFP-polη proteins, HEK293T cells transiently expressing EGFP-tagged wild-type polη, polη-S687A, or polη-S687D were exposed with 30 J/m² UVC or mock-treated. At 6 h after UV treatment, the cells were harvested, fixed with 4% paraformaldehyde or ice-cold methanol, mounted in Vectashield Mounting Medium (Vector Laboratories) containing DAPI, and imaged using a Leica TCS SP2/UV confocal microscope (Leica Microsystems).

Selection of Stable Transfectants

XP30RO cells were transfected with EGFP-polη, EGFP-polη-S687A, or EGFP-polη-S687D plasmid as described above. After a 24-h incubation, the cells were cultured in DMEM containing 600 μg/ml G418 (Sigma). The cells were selected for 2 weeks, and the stable transfectants were isolated. The stable transfectants expressing wild-type or mutant polη were harvested for flow cytometry analysis (BD Biosciences FACSAria I), and the proteins were isolated for Western blotting analysis as described above. The antibodies that specifically recognized human polη (Bethyl Laboratories) and β-actin (Abcam) were used at 1:2,000 and 1:10,000 dilutions, respectively.

Clonogenic Survival Assay

XP30RO cells stably expressing wild-type or mutant EGFP-polη were plated in six-well plates in triplicate. On the next day, the cells were exposed to various doses of UV light, after which the cells were incubated in DMEM containing 0.4 mm caffeine to inhibit DNA repair. Cell colonies grown for 10–14 days were then fixed with 6% (v/v) glutaraldehyde and stained with 0.5% (w/v) crystal violet. Colonies containing at least 50 cells were subsequently counted.

Cell-Based Translesion Synthesis Assay

The construction of CPD-bearing double-stranded shuttle vector and its lesion-free control and competitor vectors was described recently (28). The strand-specific PCR-based competitive replication and adduct bypass assay was performed following previously published procedures (28, 29). Briefly, the CPD-bearing and the corresponding nonlesion control plasmids were premixed individually with the competitor genome for the cell-based TLS assay, with the molar ratios of competitor vector to control or lesion-bearing genome being 1:1 and 1:30, respectively. The XP30RO cells stably expressing wild-type or mutant EGFP-polη were transfected with ∼300 ng mixed plasmids by using Lipofectamine 2000 (Invitrogen). At 24 h after transfection, the progenies of the plasmid were isolated by using a modified alkali lysis method (30). The residual unreplicated plasmid was further processed by combined digestion with DpnI and exonuclease III, as described previously (31–33). The progeny genomes arising from in vivo replication were PCR amplified with primers spanning the lesion site; the resulting PCR products were restriction-digested with NcoI and SfaNI and subsequently analyzed by PAGE and LC-MS/MS, as described elsewhere (28, 29).

RESULTS

Serine 687 (Ser⁶⁸⁷) is Phosphorylated in Human Polη

We employed a combined mass spectrometric and cell biological approach to identify and functionally characterize novel posttranslational modification(s) of polη in human cells. To this end, we expressed FLAG-tagged human polη in HEK293T cells and purified the FLAG-tagged protein using anti-FLAG M2 beads. The purified protein was then digested with trypsin and the resulting peptides subjected to LC-MS/MS analysis. Database search of the LC-MS/MS results led to the identification of a number of peptides derived from polη with a sequence coverage of 93%. Our LC-MS/MS analysis also gave rise to the identification of the phosphorylation of serine residues 601 and 687 as well as the ubiquitination of lysine 682 (Lys⁶⁸², Fig. 1A, Supplemental Fig. S1, and Supplemental Table S2). In this vein, Lys⁶⁸² ubiquitination and Ser⁶⁰¹ phosphorylation were identified previously and found to be involved in UV-induced DNA damage response (17, 18). In addition, Ser⁶⁸⁷, similar as Lys⁶⁸², resides in the highly conserved nuclear localization signal (NLS) region of human polη (Fig. 1B).

Fig. 1.

Polη is phosphorylated at Ser⁶⁸⁷ in human cells. (A) The MS/MS of the [M+2H]²⁺ ion of the peptide ₆₈₄NPKSPLACTNK₆₉₄ from polη with Ser⁶⁸⁷ being phosphorylated. Pho, phosphoric acid; Car, carbamidomethylation. (B) Partial amino acid sequence of the C terminus of polη from human, mouse, and platypus, with the Ser⁶⁸⁷ site being highlighted in bold. (C) The relative abundance of unmodified (₆₈₇SPLACTNK₆₉₄) and phosphorylated peptide (₆₈₄NPKS_phoPLACTNK₆₉₄) in selected-ion monitoring mode showing that CDK2/Cyclin A2 can phosphorylate polη at Ser⁶⁸⁷. The relative phosphorylation levels (phosphorylated/unmodified peptide) are 0.0% for polη only, 1.8% for CDK2/Cyclin A2 (50 ng), and 3.2% for CDK2/Cyclin A2 (100 ng). MA, peak area found in the selected-ion chromatogram for monitoring the formation of [M-Pho]²⁺ ion for the phosphorylated peptide and y₅ ion from the corresponding unmodified peptide (see Materials and Methods). (D) Relative levels of Ser⁶⁸⁷ phosphorylation of polη in HEK293T cells treated with control or CDK2 siRNA. Inset shows Western blotting results of CDK2. (E) Ser⁶⁸⁷ phosphorylation displays a dose-dependent increase at 6 h following UV irradiation. Error bar represents the S.E. (n = 3). **, p < 0.01; ***, p < 0.001. The p values were calculated using unpaired two-tailed Student's t test.

Fig. 1A depicts the MS/MS for the [M+2H]²⁺ ion of the peptide ₆₈₄NPKS_phoPLACTNK₆₉₄ with Ser⁶⁸⁷ being phosphorylated. The Ser⁶⁸⁷ phosphorylation was unambiguously identified from fragment ions found in the MS/MS, with the consideration of the mass shift introduced by phosphorylation, i.e. 80.0 Da. The existence of the b₄-Pho, y₈-Pho, y₉-Pho, and y₁₀-Pho ions and the absence of the b₃-Pho and y₇-Pho ions (“Pho” designates a phosphoric acid) in the MS/MS supported the phosphorylation of Ser⁶⁸⁷ (Fig. 1A).

Trypsin cleaves the amide bonds on the C-terminal sides of lysine and arginine. Thus, the phosphorylated peptide ₆₈₄NPKS_phoPLACTNK₆₉₄ contains a miscleavage site. We failed to observe the phosphorylated peptide without miscleavage, though the corresponding unmodified peptide, ₆₈₇SPLACTNK₆₉₄, could be readily detected (MS/MS shown in Supplemental Fig. S2). Previous studies showed that phosphorylation can produce a greater degree of miscleavage by trypsin near the phosphorylation site compared with the unphosphorylated peptide (34). Thus, the miscleavage of the amide bond on the C-terminal side of Lys⁶⁸⁶ provided another line of evidence to support the phosphorylation of Ser⁶⁸⁷ in polη.

Ser⁶⁸⁷ Phosphorylation Involves Cyclin-Dependent Kinase 2 (CDK2)

CDK2 plays a major role in the S phase of cell cycle when DNA replication and repair occur (35, 36). To explore whether CDK2 is involved in Ser⁶⁸⁷ phosphorylation in polη, we first assessed the ability of CDK2 to phosphorylate directly Ser⁶⁸⁷ in polη with an in vitro kinase assay (Fig. 1C, Supplemental Fig. S3). Our results demonstrated that the recombinant CDK2/Cyclin A2 complex could indeed directly phosphorylate Ser⁶⁸⁷ in recombinant human polη (Fig. 1C, Supplemental Fig. S3).

We next asked whether CDK2 is involved in this phosphorylation in cells. Toward this end, we knocked down the expression of CDK2 in HEK293T cells by using siRNA and transfected the cells with the plasmid for expressing FLAG-tagged polη. Upon efficient CDK2 knockdown (Fig. 1D), we observed a significant reduction in the level of Ser⁶⁸⁷ phosphorylation in polη relative to polη isolated from cells treated with nontargeting control siRNA (Fig. 1D). In this vein, it is worth noting that CDK2 depletion was previously shown not to alter cell cycle progression (37, 38). Thus, the reduced phosphorylation of Ser⁶⁸⁷ in polη upon CDK2 depletion is unlikely attributed to cell cycle perturbation.

Ser⁶⁸⁷ Phosphorylation Is Stimulated by UV Irradiation

Considering that polη's major functional role resides in its capability in bypassing UV-induced CPD lesions, we asked whether the level of Ser⁶⁸⁷ phosphorylation is modulated by UV irradiation. To this end, we exposed HEK293T cells expressing FLAG-tagged polη with UV irradiation, immunoprecipitated polη by using anti-FLAG M2 beads, digested it with trypsin, and subjected the resulting peptide mixture to LC-MS/MS analysis in the selected-ion monitoring mode. We then estimated the relative level of phosphorylation based on the relative abundances of fragment ions formed from the unmodified and phosphorylated peptides (see Materials and Methods). The results showed that the phosphorylation level of Ser⁶⁸⁷ increased with the dose of the UV irradiation (Fig. 1E). Along this line, our results also revealed an elevated phosphorylation of Ser⁶⁰¹ and a diminished ubiquitination of Lys⁶⁸² after UV irradiation, which are in keeping with previous findings (17, 18), thereby validating our method in quantifying the posttranslational modifications of human polη (Supplemental Figs. S1C and S1D).

Ser⁶⁸⁷ Phosphorylation Does Not Alter the Nuclear Localization of Human Polη

As stated above, Ser⁶⁸⁷ is situated in the highly conserved NLS region of polη. In the viewpoint that phosphorylation at or near the NLS may regulate the protein's nuclear import (39), we next investigated whether Ser⁶⁸⁷ phosphorylation affects polη's subcellular localization. Our fluorescence microscopy result showed that, similar as its wild-type counterpart, both the phosphorylation-deficient (i.e. EGFP-polη-S687A) and phosphomimetic (i.e. EGFP-polη-S687D) mutants were exclusively nuclear in both untreated and UV-treated HEK293T cells (Fig. 2 and Fig. S4). Thus, we conclude that Ser⁶⁸⁷ phosphorylation does not alter the nuclear import of human polη, suggesting that Ser⁶⁸⁷ phosphorylation likely occurs after polη's nuclear import.

Fig. 2.

Ser⁶⁸⁷ Phosphorylation does not alter the nuclear localization of human polη. HEK293T cells transiently expressing wild-type EGFP-polη, EGFP-polη-S687A and EGFP-polη-S687D were treated using 30 J/m² UVC, incubated for 6 h at 37 °C, and analyzed by confocal microscopy. Scale bars represent 20 μm.

Increase in Ser⁶⁸⁷ Phosphorylation Is Correlated with a Decrease in the Cellular Levels of CPD Lesion and Impedes Polη's Interaction with PCNA

We examined the levels of Ser⁶⁸⁷ phosphorylation of polη in HEK293T cells at various time intervals following UV irradiation. It turned out that Ser⁶⁸⁷ phosphorylation exhibited a progressive increase after UV irradiation (Fig. 3A and Supplemental Fig. S5), which is accompanied with a concomitant reduction in CPD level, as revealed by a flow cytometry-based method (26) (Fig. 3B and Supplemental Fig. S6).

Fig. 3.

The level of Ser⁶⁸⁷ phosphorylation is negatively correlated with the level of unrepaired CPD, and the relationship between Ser⁶⁸⁷ phosphorylation in polη and its interaction with PCNA. (A) The level of Ser⁶⁸⁷ phosphorylation in polη without UV treatment or at 3, 6, 12, and 24 h following UV irradiation (30 J/m²). (B) Flow cytometry results showing the level of CPD at various time intervals following UV irradiation (10 J/m²). (C) The level of Ser⁶⁸⁷ phosphorylation in wild-type polη and the PIP/UBZ domain mutant of polη after UV irradiation (30 J/m²). (D) GST pull-down of PCNA by wild-type and various forms of polη with mutations in the C terminus. Bottom panels show the quantification data. Error bar represents the S.E. (n = 3). *, p < 0.05; ***, p < 0.001. The p values were calculated using unpaired two-tailed Student's t test.

Previous studies demonstrated that interaction with PCNA plays an indispensable role in the recruitment of polη to UV-induced CPD site to carry out TLS, and the dissociation of polη from PCNA after TLS is essential for maintaining high-fidelity DNA replication (4–7). We next asked whether Ser⁶⁸⁷ phosphorylation in human polη is modulated by its interaction with PCNA. To this end, we generated a polη variant with mutations in both the PIP and UBZ domains, which is defective in PCNA interaction (16, 20), and measured the level of Ser⁶⁸⁷ phosphorylation in the mutant protein following UV irradiation. Our result showed that the mutation of PIP and UBZ domains of polη did not affect appreciably the level of Ser⁶⁸⁷ phosphorylation (Fig. 3C), suggesting that interaction with PCNA is not essential for polη's phosphorylation at Ser⁶⁸⁷.

The Role of Ser⁶⁸⁷ Phosphorylation of Polη in Its Interaction with PCNA

Considering that Ser⁶⁸⁷ resides in the PCNA-interacting region of polη that is essential for polη-PCNA interaction during UV damage response (17), we investigated the potential role of Ser⁶⁸⁷ phosphorylation in the interaction between these two proteins. To this end, we performed GST pull-down experiments using GST fusion proteins of the C-terminal portion of polη (amino acids 602–713, GST-polη_C) and recombinant PCNA. It turned out that, similar to what was reported previously (17), wild-type GST-polη_C could effectively capture PCNA, whereas mutation of PIP box in polη compromised its binding to PCNA (Fig. 3D). Interestingly, mutation of Ser⁶⁸⁷ to an alanine did not influence the binding; however, mutation of Ser⁶⁸⁷ to an aspartic acid diminished significantly the polη-PCNA interaction (Fig. 3D). It is of note that wild-type GST-polη_C purified from E. coli cells was not phosphorylated at Ser⁶⁸⁷ (Fig. 1C); thus, recombinant wild-type polη and the S687A mutant should carry the same charges in the NLS, whereas the S687D mutant introduced a permanent negative charge to the NLS. Our observation of the compromised interaction between the phosphomimetic mutant of polη and PCNA is consistent with the previous finding that the positively charged lysines in the NLS are necessary for the binding of polη to PCNA (17).

Ser⁶⁸⁷ Phosphorylation Confers Cellular Tolerance toward UV DNA Damage and Promotes TLS across CPD Lesion

We next examined the role of Ser⁶⁸⁷ phosphorylation in UV damage tolerance and in TLS across UV-induced CPD lesions. To this end, we first generated a set of stable cell lines by transfecting wild-type EGFP-polη and the corresponding S687A and S687D mutant plasmids into polη-deficient XP30RO fibroblasts and confirmed that the expression levels of the wild-type and mutant polη were similar in the transfected cells (Supplemental Fig. S7). We next assessed, by using clonogenic survival assay, the sensitivities of these cells toward UV light exposure. Our results showed that the cellular sensitivity toward UV light emanating from deficiency in polη could be restored to a similar extent by wild-type polη and polη-S687D whereas polη-S687A only partially rescued the cells from polη deficiency (Fig. 4A). This result suggests that the Ser⁶⁸⁷ phosphorylation of polη conferred cellular resistance toward the cytotoxic effects of UV irradiation. In this regard, the moderate survival defect introduced by deficient Ser⁶⁸⁷ phosphorylation is in keeping with previous observations for the corresponding deficiencies in Ser⁶⁰¹ phosphorylation or Lys⁶⁸² deubiquitination (17, 18), suggesting that the functions of polη in cellular tolerance toward UV light exposure are regulated through multiple mechanisms, including Ser⁶⁰¹/Ser⁶⁸⁷ phosphorylation, Lys⁶⁸² ubiquitination and perhaps some other as yet unidentified mechanism(s).

Fig. 4.

The role of Ser⁶⁸⁷ phosphorylation in UV damage tolerance. (A) UV survival curves of XP30RO-derived cells stably expressing EGFP-polη, EGFP-polη-S687A or EGFP-polη-S687D. (B) The percentage of relative bypass efficiency of the CPD lesion in the indicated cells. The relative bypass efficiency of the CPD lesion in the cells expressing wild-type polη protein was designated as 100%. Error bar represents the S.E. (n = 3). *, p < 0.05; ***, p < 0.001. The p values were calculated using unpaired two-tailed Student's t test.

We also employed a cell-based replication assay to directly quantify how the efficiency of TLS across a site-specifically inserted CPD lesion in human cells is affected by Ser⁶⁸⁷ phosphorylation (28, 29). Our results revealed that the extent of TLS across a CPD lesion was markedly higher in cells stably expressing wild-type polη than in the parental XP30RO cells (Fig. 4B and Supplemental Fig. S8). In agreement with our cell survival data, the cells stably expressing polη-S687A mutant showed a moderate and significant decrease in TLS efficiency across a CPD lesion relative to the cells expressing the wild-type polη or its phosphomimetic mutant (S687D Fig. 4B and Supplemental Fig. S8). These results support that the phosphorylation of polη at Ser⁶⁸⁷ promotes DNA replication past UV-induced CPD lesion in human cells, perhaps through facilitating polη's departure from PCNA after TLS across the UV-induced CPD lesion (vide infra).

DISCUSSION

Genetic deficiency in polη causes xeroderma pigmentosum variant syndrome in humans, which is manifested by sunlight sensitivity and elevated susceptibility to develop sunlight-induced skin cancer (11, 12). The C terminus of polη plays a critical role in protein–protein interactions and in intracellular localization of polη in response to UV damage. There are three important regulatory motifs in the C-terminal region of polη, including the NLS, UBZ, and PIP domains (13–17). In this vein, PCNA is known to be monoubiquitinated in response to UV light exposure, and this ubiquitination stimulates its interaction with polη through the UBZ domain of the latter protein (16). This, together with the binding between PCNA and polη through the latter's PIP and NLS domains, enables the recruitment of polη to stalled replication forks to perform TLS (10, 14, 22).

In this study, we found that the phosphorylation of Ser⁶⁸⁷, which resides in the NLS of polη, was stimulated upon UV irradiation and this phosphorylation could diminish polη's interaction with PCNA (Fig. 1 and Fig. 3). This observation is in line with the previous finding that the positively charged amino acids in the NLS of polη are required for the interaction between polη and PCNA (17). We also observed a time-dependent elevation in the level of Ser⁶⁸⁷ phosphorylation following UV light exposure, and this increase is accompanied with a concomitant drop of the level of CPD lesion in cells (Fig. 3). On the basis of these findings, in conjunction with the importance of Ser⁶⁸⁷ phosphorylation in the cellular tolerance toward UV damage and in promoting TLS across a site-specifically incorporated CPD lesion (Fig. 4), we propose a tentative mechanistic model for the role of this phosphorylation in the reverse polymerase switching step of TLS: Ser⁶⁸⁷ phosphorylation introduces negative charges to NLS of polη. The increased negative charge to the NLS diminishes polη's binding affinity with PCNA and promotes its departure from the replication fork, thereby resetting the replication fork for highly accurate and processive replication of the downstream undamaged template. This is reminiscent of deubiquitination of Lys⁶⁸² in the NLS of polη during the initial stage of TLS, where the removal of the ubiquitination mark unshields the UBZ domain of polη and renders it available for binding with monoubiquitinated PCNA (17, 28). Therefore, our study, along with the previous studies (17, 28), supports that the posttranslational modifications of the NLS of polη functions in both its recruitment to, and departure from replication fork, which involve Lys⁶⁸² deubiquitination and Ser⁶⁸⁷ phosphorylation, respectively.

Taken together, we identified Ser⁶⁸⁷ as a novel phosphorylation site of human polη by mass spectrometric analysis. We found that the phosphorylation of Ser⁶⁸⁷ in polη can be stimulated by UV irradiation, and this phosphorylation may function in the reverse polymerase switching step of TLS, which promotes cellular tolerance against UV irradiation. These findings reveal novel knowledge into the functional regulation of polη in UV damage tolerance at the posttranslational level.