Ubiquitin and TFIIH-stimulated DDB2 dissociation drives DNA damage handover in nucleotide excision repair

Abstract

DNA damage sensors DDB2 and XPC initiate global genome nucleotide excision repair (NER) to protect DNA from mutagenesis caused by helix-distorting lesions. XPC recognizes helical distortions by binding to unpaired ssDNA opposite DNA lesions. DDB2 binds to UV-induced lesions directly and facilitates efficient recognition by XPC. We show that not only lesion-binding but also timely DDB2 dissociation is required for DNA damage handover to XPC and swift progression of the multistep repair reaction. DNA-binding-induced DDB2 ubiquitylation and ensuing degradation regulate its homeostasis to prevent excessive lesion (re)binding. Additionally, damage handover from DDB2 to XPC coincides with the arrival of the TFIIH complex, which further promotes DDB2 dissociation and formation of a stable XPC-TFIIH damage verification complex. Our results reveal a reciprocal coordination between DNA damage recognition and verification within NER and illustrate that timely repair factor dissociation is vital for correct spatiotemporal control of a multistep repair process.

Fig. 1. DDB2 and XPC are differently regulated by downstream factors.

a Fluorescence Recovery After Photobleaching (FRAP) analysis of DDB2 mobility in mock or UV-C irradiated (10 J/m²) VH10 cells stably expressing GFP-DDB2 and transfected with control (CTRL) or GTF2H1 siRNAs. GFP-DDB2 fluorescence recovery was measured in a strip across the nucleus after bleaching, normalized to bleach depth, and the average pre-bleach intensities (1.0). b Percentage of GFP-DDB2 immobile fraction in VH10 fibroblasts treated with control (CTRL), GTF2H1 or XPG siRNAs, determined from FRAP analyses as depicted in (a). Percentage immobile fraction represents the ratio between the average recovered fluorescence intensity of UV- and mock-treated cells, over the last 10 s of the measurements, as explained in the methods. c FRAP analysis of XPC mobility in mock or UV-C irradiated (10 J/m²) XP4PA cells stably expressing XPC-GFP and transfected with control (CTRL) or GTF2H1 siRNAs. XPC-GFP-fluorescence recovery was measured and normalized as described in (a). d Percentage of XPC-GFP immobile fraction in XP4PA cells treated with control (CTRL), GTF2H1 or XPG siRNAs, determined by FRAP analysis as depicted in (c) and described in (b). Graphs and FRAP curves depict the mean & S.E.M. of >30 cells from three independent experiments. **P < 0.01, ***P < 0.001, relative to siCTRL control 10 J/m², analyzed by unpaired, two-tailed t-test (adjusted for multiple comparisons, see “Methods”). Source data are provided as a Source Data file.

Fig. 2. Persistent damage detection in absence of lesion excision.

a Representative immunofluorescence (IF) images of endogenous DDB2 accumulation at local UV-C damage (LUD) in U2OS wild-type (WT) and U2OS XPF knockout (XPF KO) cells. Cells were fixed 10 min, 40 min, 2 h and 8 h after LUD (marked by CPD staining) induced with UV-C irradiation (60 J/m²) through a microporous membrane (8 µm). b Quantification of DDB2 accumulation at LUD, as depicted in (a). DDB2 accumulation was normalized to the nuclear background and U2OS WT 10 min after UV-C, which was set to 1.0. c Total DDB2 protein levels determined by measuring total nuclear fluorescent signal intensities in nuclei such as depicted in a and normalized to U2OS WT 10 min, which was set to 1.0. d Representative IF images of endogenous XPC accumulation at LUD in U2OS WT and XPF KO cells, as described in (a). e Quantification of XPC accumulation at LUD, as depicted in (d) and described in (b). Mean and S.E.M. of, respectively, n = 348, 313, 383, 334, 355, 334, 316, 247 cells for DDB2 and n = 305, 276, 413, 272, 339, 383, 266, 339 cells for XPC from five independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, analyzed by one-way ANOVA (see “Methods”). Scale bars in (a, d): 5 µm. Source data are provided as a Source Data file.

Fig. 3. TFIIH promotes DDB2 dissociation and stable XPC binding.

a IF images of endogenous DDB2 LUD accumulation in U2OS WT and XPF KO cells treated with control (CTRL) or GTF2H1 siRNAs, 40 min and 8 h after UV-C irradiation (60 J/m²) Scale bar: 5 µm. b, c Quantification of (b) DDB2, or (c) XPC accumulation at LUD in U2OS WT and XPF KO cells treated with CTRL or GTF2H1 siRNAs, 10 min, 40 min, 2 h and 8 h after damage, as described in (a), normalized to the nuclear background and U2OS siCTRL 10 min. Mean and S.E.M. of, respectively, n = 322, 321, 317, 374, 335, 364, 291, 299, 370, 314, 318, 342, 307, 315, 283, 287 cells in (b) or 217, 225, 210, 234, 220, 193, 165, 173, 218, 238, 218, 213, 217, 202, 194, 156 cells in (c) from two independent experiments. IF images of XPC are shown in Supplementary Fig. 2e. d Immunoblot showing DDB2 ubiquitylation in GFP-DDB2 VH10 cells, before or 15 min after UV-C irradiation (30 J/m²) and treated with CTRL or GTF2H1 siRNAs. Total cell lysates (Input) were analyzed with DDB2, GTF2H1 and Tubulin antibodies. GFP-DDB2 immunoprecipitation (IP) fractions were analyzed using anti-ubiquitin (FK2) and DDB2 antibodies. Control IP is shown in Supplementary Fig. 2f. e Quantification of ubiquitin levels shown in d, normalized to DDB2 levels and non-irradiated siCTRL samples. Mean and S.E.M. of three independent experiments. f Immunoblot showing UV-induced DDB2 proteolysis in total cell lysates of U2OS cells treated with CTRL or GTF2H1 siRNAs in the absence and presence of MG132 proteasome inhibitor, at the indicated time points after UV irradiation (30 J/m²) and analyzed by DDB2 and tubulin antibody. g Quantification of DDB2 proteolysis as depicted in (f), normalized to tubulin and non-irradiated samples. Mean and S.E.M. of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. non-significant, analyzed by one-way ANOVA in (b), (c) and by unpaired, two-tailed t-test (adjusted for multiple comparisons) in (e), (g) (see “Methods”). Source data are provided as a Source Data file.

Fig. 4. DDB2 retention impairs stable XPC and TFIIH damage binding.

a Representative IF images of endogenous DDB2 accumulation at LUD in U2OS WT cells in the absence or presence of VCP inhibitor (VCPi). 1 h before LUD induction, 10 µM VCPi was added and 40 min and 8 h after local UV irradiation (60 J/m²) through a microporous membrane (8 µm), cells were fixed and analyzed by IF. Scale bars: 5 µm. b Quantification of endogenous DDB2 accumulation at LUD, normalized to the nuclear background and mock-treated U2OS WT 10 min after UV-C, which was set to 1.0. U2OS cells mock- or VCPi-treated were fixed 10 min, 40 min, 2 h and 8 h after LUD induction. Mean and S.E.M. of, respectively, n = 104,96, 150, 91, 124, 145, 120, 68 cells from two independent experiments. c FRAP analysis of GFP-DDB2 mobility in VH10 cells before and immediately after UV irradiation (10 J/m²), in the presence or absence of VCPi (10 μM) added 1 h before irradiation. GFP-DDB2 fluorescence recovery was measured in a strip across the nucleus after bleaching and normalized to the average pre-bleach intensity (1.0). Curves represent the average of >30 cells per condition from three independent experiments. d, e Recruitment of endogenous (d) XPC and (e) XPB to LUD in U2OS WT cells in the absence or presence of VCP inhibitor (VCPi), as described in (a). Scale bars: 5 µm. f, g Quantification of endogenous accumulation of (f) XPC and (g) XPB at LUD as described in (b). Mean and S.E.M. of, respectively, n = 206, 291, 290, 348, 234, 226, 146, 72 cells for XPC and n = 145, 93, 140, 119, 144, 161, 139, 48 cells for XPB from three and two independent experiments, respectively. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. non-significant, analyzed by one-way ANOVA (see “Methods”). Source data are provided as a Source Data file.

Fig. 5. VCP-mediated DDB2 extraction facilitates the stable binding of XPC and TFIIH.

a Representative IF images and b quantification of endogenous DDB2 accumulation at LUD in U2OS WT and DDB2 KO cells, at the indicated time points after UV irradiation (60 J/m²). Data were normalized to nuclear background and U2OS WT 10 min. Mean and S.E.M. of, respectively, n = 200, 132, 120, 102 cells from two independent experiments. c Representative IF and d quantification of endogenous XPC accumulation at LUD in U2OS WT and DDB2 KO cells. Mean and S.E.M. of, respectively, n = 106, 93, 117, 103, 97, 101, 148, 150 cells from two independent experiments. e Representative IF and f quantification of endogenous XPB accumulation at LUD in U2OS WT and DDB2 KO cells. Mean and S.E.M. of, respectively, n = 140, 143, 117, 113, 188, 145, 142, 132 cells from two independent experiments. g FRAP analysis of endogenously GFP-tagged XPB mobility before and 1 h after UV irradiation (10 J/m²), in the presence and absence of DDB2 and/or VCP activity. MRC-5 cells with GFP knock-in at the ERCC3/XPB locus were transfected with control (CTRL) or DDB2 siRNAs and incubated with mock or VCPi (10 µM). GFP-XPB fluorescence recovery was measured in a strip across the nucleus for 30 s after bleaching and normalized to the average pre-bleach intensity (1.0). h Percentage of endogenous XPB immobile fraction in MRC-5 cells treated with CTRL or DDB2 siRNAs and/or VCPi, determined from FRAP analysis as depicted in (g). Mean and S.E.M. of >30 cells per condition from three independent experiments. i Percentage of endogenous XPC immobile fraction in HCT116 cells with (WT) or without DDB2 (DDB2 KO), mock or VCPi treated, determined from FRAP analysis as depicted in Supplementary Fig. 3g. Mean and S.E.M. of >25 cells per condition from two independent experiments.*P < 0.05, **P < 0.01, ***P < 0.001, n.s. non-significant, analyzed by one-way ANOVA in (b), (d) and (f) and by unpaired, two-tailed t-test (adjusted for multiple comparison) in (h) and (i) (see “Methods”). Scale bars: 5 µm. Source data are provided as a Source Data file.

Fig. 6. DDB2 ubiquitylation facilitates its damage extraction to promote damage handover to XPC.

a Overview of DDB2 wild-type (WT, 427 amino acids) and deletion mutants lacking first 40 amino acids (white stripes; ΔNT) or carrying additionally five lysine to arginine substitutions (red dots; ΔNT/BP5KR). b Immunoblot of UV-induced DDB2 proteolysis in U2OS DDB2 KO cells expressing GFP-tagged WT, ΔNT or ΔNT/BP5KR DDB2, analyzed in total cell lysates with DDB2 and CSN5 antibodies. c Immunoblot of DDB2 immunoprecipitation showing binding partners and UV-induced ubiquitylation in U2OS DDB2 KO cells expressing WT, ΔNT or ΔNT/BP5KR DDB2, before and 15 min after UV-C irradiation (30 J/m²), analyzed using FK2, GFP, DDB1, CUL4A, and CSN5 antibodies. d Quantification of ubiquitylated XPC in whole-cell lysates of U2OS DDB2 KO cells expressing WT, ΔNT or ΔNT/BP5KR GFP-DDB2, analyzed by immunoblot in Supplementary Fig. 4b, c and normalized to Tubulin and mock-treated WT DDB2. Mean and S.E.M. of four independent experiments. e Scheme of inverse FRAP (iFRAP) method. Accumulation of a fluorescent protein to local UV-C-laser-induced damage was measured until reaching a steady-state level, after which the GFP-fluorescence outside the UV-damaged and control area was bleached. The loss of fluorescence in the control and UV-damaged areas was measured. f iFRAP of WT (gray), ΔNT (green) and ΔNT/BP5KR (orange) GFP-DDB2 dissociation from local UV-damage in U2OS DDB2 KO cells. Fluorescence loss, reflecting DDB2 dissociation, was measured over time, normalized to the background and to fluorescence levels before bleaching. Mean and S.E.M. of >30 cells per condition from three independent experiments. g IF images and h quantification of endogenous XPC (cyan) accumulation at LUD (CPD, red) in U2OS DDB2 KO cells expressing WT, ΔNT or ΔNT/BP5KR GFP-DDB2 (green), 40 min after UV irradiation (60 J/m²). Data were normalized to the nuclear background and WT. Mean and S.E.M. of, respectively, n = 163, 207, 225 cells from three independent experiments. *P < 0.05, ***P < 0.001, n.s., non-significant, analyzed by unpaired, two-tailed t-test (adjusted for multiple comparisons) in (d), by ROC curve analysis in (f) and by one-way ANOVA in (h) (see “Methods”). Scale bars: 5 µm. Source data are provided as a Source Data file.

Fig. 7. Reciprocal coordination of DNA damage detection and handover in GG-NER.

DDB2 binds directly to UV-photoproducts, thereby stimulating XPC recruitment to CPDs and 6-4PPs. The CRL4 E3 ubiquitin ligase is activated upon DDB2 binding and ubiquitylates DDB2 and XPC. TFIIH is recruited via an interaction between its subunit XPB with XPC (interaction depicted with dotted lines). Upon TFIIH binding, its trimeric CDK7-activating kinase (CAK) sub-complex is released and allows XPA binding, which further stimulates TFIIH’s XPD helicase that unwinds the DNA in the 5′–3′ direction while scanning for helicase blocking lesions. This configuration facilitates further interaction between TFIIH and XPC by allowing GTF2H1 to interact with XPC. Recruitment of TFIIH and ensuing damage verification promote the stable association of XPC with the undamaged strand and simultaneously facilitate the displacement of DDB2, which is also promoted by ubiquitylation-mediated extraction by VCP (1). The subsequent degradation of DDB2 (2) regulates its availability to rebind to lesions, possibly to avoid competition with the emerging NER pre-incision complex. The formation of a stable ternary XPC-TFIIH-XPA damage verification complex on the lesion and the unpaired DNA surrounding the lesion (created by this complex) provide substrate for the structure-specific endonucleases XPF-ERRC1 and XPG (the latter coinciding with XPC dissociation), which completes the formation of the pre-incision complex.