Chromatin retention of DNA damage sensors DDB2 and XPC through loss of p97 segregase causes genotoxicity

Abstract

DNA damage recognition subunits such as DDB2 and XPC protect the human skin from ultraviolet (UV) light-induced genome instability and cancer, as demonstrated by the devastating inherited syndrome xeroderma pigmentosum. Here we show that the beneficial DNA repair response triggered by these two genome caretakers critically depends on a dynamic spatiotemporal regulation of their homeostasis. The prolonged retention of DDB2 and XPC in chromatin, because of a failure to readily remove both recognition subunits by the ubiquitin-dependent p97/VCP/Cdc48 segregase complex, leads to impaired DNA excision repair of UV lesions. Surprisingly, the ensuing chromosomal aberrations in p97-deficient cells are alleviated by a concomitant downregulation of DDB2 or XPC. Also, genome instability resulting from an excess of DDB2 persisting in UV-irradiated cells is prevented by concurrent p97 overexpression. Our findings demonstrate that DNA damage sensors and repair initiators acquire unexpected genotoxic properties if not controlled by timely extraction from chromatin.

Fig. 1. Ubiquitin-dependent p97 recruitment to UV lesions

(a) Co-localisation of p97 with CPDs in human cells. Mild expression of myc-tagged p97 (wild-type or p97 EQ) was induced with doxycycline (+Dox). Cells were probed 15 min after UV irradiation through micropore filters, whereby endogenous p97 was detected with anti-p97 antibodies and p97-myc with anti-myc antibodies. Scale bar, 10 μm. (b) UV-induced p97 re-localisation shown as quantitative ratio of fluorescence at CPDs against surrounding nuclear areas (200 nuclei from two independent experiments); error bars, s.e.m, ***P<0.001 relative to wild-type (the unpaired two-tailed t-test was used for all P-value determinations). (c) Reduced recruitment of myc-p97 EQ to CPDs after treatment with MG132 or siRNA targeting the indicated cullins; siNC, non-coding control. (d) Quantification of p97 EQ at CPDs over three experiments (>300 nuclei), ***P<0.001 relative to siNC control (unpaired two-tailed t-test). (e) Increased K48-linked ubiquitin after treatment with siRNA targeting p97, CUL4A/B or DDB2. (f, g) Quantification (>300 nuclei from three experiments) of K48-linked ubiquitin in nuclei overall and at lesion sites, respectively.

Fig. 2. Recruitment of p97 to UV lesions requires DDB2

(a) Co-localisation of p97 EQ with CPDs in human cells treated with siRNA targeting the indicated proteins. Samples were probed 15 min after UV irradiation through micropore filters; scale bar, 10 μm. (b) Quantification of p97 EQ accumulation at CPDs as illustrated in Fig. 2a. Error bars, s.e.m. (>300 nuclei from three experiments); *P<0.05, ***P<0.001 relative to siNC control (unpaired two-tailed t-test). (c) UV-stimulated interaction of p97 with a DDB2-containing complex. HEK293 cells were transfected to produce DDB2-FLAG and doxycycline-treated (Dox) for mild p97 EQ expression. DDB2-FLAG was pulled-down from cell lysates using beads coated with anti-FLAG antibodies and eluates were analysed by immunoblotting against K48-ubiquitin, FLAG epitope and p97. Myc-tagged p97 migrates slower than the endogenous counterpart. (d) Interaction of DDB2 with the p97 complex. HEK293 cells were doxycycline-treated (Dox) for mild p97 EQ expression and p97 EQ was pulled-down from cell lysates using beads coated with Strep-Tactin. The resulting protein mixture was analysed by immunoblotting against p97 and DDB2.

Fig. 3. The p97 segregase complex extracts DDB2 and XPC from chromatin

(a) Chromatin retention of DDB2 and XPC visualised 15 and 180 min after induction of UV damage spots in human cells treated with siRNA targeting p97 or non-coding (NC) control. Scale bar, 10 m. (b, c) Quantification of DDB2 and XPC, respectively, in chromatin over three experiments (>300 nuclei). Error bars, s.e.m.; ***P<0.001 relative to siNC control (unpaired two-tailed t-test). (d) XPC hyperubiquitination and DDB2 degradation revealed by whole-cell immunoblot analysis after treatment with siRNA against p97 or non-coding control. (e) Levels of soluble (non-chromatin) and chromatin-bound DDB2 assessed in the respective cellular fractions.

Fig. 4. Processing of DDB2 by the p97 segregase triggered by adapters

(a) Co-localisation of p97 EQ with CPDs in human cells treated with siRNA against the indicated adapter proteins. The samples were probed 15 min after UV irradiation through micropore filters; scale bar, 10 μm. (b) Quantification of p97 EQ accumulation at CPDs as illustrated in Fig. 4a. Error bars, s.e.m. (>300 nuclei from three experiments); **P<0.01, relative to siNC control (unpaired two-tailed t-test). (c) DDB2 degradation revealed by whole-cell immunoblot analysis after treatment with two different siRNA sequences against DVC1; siNC, non-coding control. GAPDH, glyceraldehyde 3-phosphate dehydrogenase as loading standard. (d) Inhibition of UV-triggered DDB2 degradation in cells treated with siRNA targeting p97 or the indicated adapters.

Fig. 5. An excess of damage sensors attenuates repair and causes genomic instability

(a) Excision of 6-4PPs in human cells treated with siRNA targeting p97 or XPC, used as comparator inflicting a severe repair defect; siNC, non-coding RNA control. Error bars, s.e.m. (n=3, each experiment with four replicates), **P<0.01 (unpaired two-tailed t-test). (b) Excision of CPDs upon treatment with siRNA targeting p97 or DDB2 (n=3, each experiment with four replicates), *P<0.05 (unpaired two-tailed t-test). In the case of CPDs, depletion of DDB2 is sufficient to compromise excision efficiency. (c) Metaphase spreads showing a progressively increasing genomic instability. Representative chromosomes are magnified in the inserts. (d) Chromosomal aberrations in cells treated with siRNA targeting the indicated proteins (n=3, 60 cells in a blinded analysis). **P<0.01, *P<0.05 relative to siNC and cells with double depletion (unpaired two-tailed t-test). (e) Chromosomal aberrations in cells overexpressing the indicated proteins (n=3, 60 cells in a blinded analysis). *P<0.05 relative to cells expressing FLAG only and the double over-expressing cells (unpaired two-tailed t-test).