MDC1 PST-repeat region promotes histone H2AX-independent chromatin association and DNA damage tolerance

Abstract

Histone H2AX and MDC1 are key DNA repair and DNA-damage signalling proteins. When DNA double-strand breaks (DSBs) occur, H2AX is phosphorylated and then recruits MDC1, which in turn serves as a docking platform to promote the localization of other factors, including 53BP1, to DSB sites. Here, by using CRISPR-Cas9 engineered human cell lines, we identify a hitherto unknown, H2AX-independent, function of MDC1 mediated by its PST-repeat region. We show that the PST-repeat region directly interacts with chromatin via the nucleosome acidic patch and mediates DNA damage-independent association of MDC1 with chromatin. We find that this region is largely functionally dispensable when the canonical γH2AX-MDC1 pathway is operative but becomes critical for 53BP1 recruitment to DNA-damage sites and cell survival following DSB induction when H2AX is not available. Consequently, our results suggest a role for MDC1 in activating the DDR in areas of the genome lacking or depleted of H2AX.

Fig. 1

MDC1 loss causes greater IR sensitivity than H2AX loss. a Diagram depicting the main events in the signal transduction pathway that leads to 53BP1 accumulation on chromatin at DNA-damage sites: (1) DSB induction and MRN/ATM recruitment/activation. (2) ATM phosphorylates H2AX and this is recognised by MDC1, which brings in more MRN and ATM. (3) MDC1-mediated accumulation of ATM results in amplification of the γH2AX signal and, consequently, further recruitment of MDC1/MRN/ATM. (4) ATM also phosphorylates the TQXF cluster of MDC1. (5) Phosphorylated TQXF motifs are bound by RNF8, which ubiquitylates another protein(s). (6) This ubuiqitylated protein(s) serves as a docking site(s) for RNF168. (7) RNF168 ubiquitylates H2A/H2AX and this, together with constitutive histone H4-K20-methylation, creates a platform that recruits 53BP1 and the Shieldin complex. b Clonogenic survival assays towards ionising radiation (IR) in specified RPE-1 genetic backgrounds. The tendency for MDC1^−/− H2AX^−/− double knockout cells to be slightly more IR sensitive than MDC1^−/− single knockout cells might be explained by 53BP1 binding γH2AX in a MDC1-independent fashion^,, and/or by replication stress caused by the lack of H2AX ; n = 6/genotype (except for MDC1 KO n = 4); error bars s.e.m. Additional supporting data, including validation, genotyping and cell cycle profiling of knockouts, are presented in Supplementary Fig. 1

Fig. 2

53BP1 localisation to DNA-damage sites in H2AX^−/− cells depends on MDC1. a Representative immunofluorescence images of 53BP1 NB formation after 24 h of 0.4 μM aphidicolin (APH) treatment, and of 53BP1 IRIF 1 h after IR (3 Gy) exposure in wild-type RPE-1 and knockout cell lines. b Quantification of 53BP1-NBs and 53BP1 IRIF in cells treated as in a. Cyclin A staining was used to differentiate G1 from S/G2 cells; n = 4/genotype (except for non-treated H2AX KO and IR MDC1 KO n = 3); error bars s.e.m.; ****p < 0.0001; two-tailed Student’s t test. c Representative immunofluorescence images showing MDC1 localisation at NBs or IRIF after APH or IR treatments (as in a) in the RPE-1 H2AX^+/+ and H2AX^−/− cell lines. Cyclin A staining was used to distinguish between G1 and S/G2 cells. d Quantifications showing MDC1 localisation at NBs in the RPE-1 H2AX^+/+ and H2AX^−/− cell lines after 24 h treatment with 0.4 μM APH. Cells were depleted for 53BP1, RNF8, SHLD1 or SHLD2 by siRNA for 48 h before the APH treatment; n = 3/genotype; error bars s.e.m.; **p < 0.01; two-tailed Student’s t test. Supporting data, including quantifications of number of NBs and IRIF per cell and validation of 53BP1, RNF8, SHLD1 and SHLD2 depletion are shown in Supplementary Fig. 2. Scale bars, 10 μm

Fig. 3

The PST region of MDC1 mediates key DDR functions in cells lacking H2AX. a Schematic representation of MDC1 architecture and the different mutant versions used in this work. b Quantification of 53BP1-NB formation after 24 h of 0.4 μM APH treatment in RPE-1 MDC1^−/– H2AX^+/+ and MDC1^−/− H2AX^−/− cells complemented with indicated mutant versions of GFP-tagged MDC1; n = 3/genotype; error bars s.e.m.; ****p < 0.0001; two-tailed Student’s t test. c Quantification of MDC1-NB formation after APH treatment (as in b) of RPE-1 MDC1^−/− H2AX^+/+ and MDC1^−/− H2AX^−/− cells complemented either with the WT or ΔPST versions of GFP-MDC1; n = 3/genotype; error bars s.e.m.; ***p < 0.001; two-tailed Student’s t test. d Quantification of 53BP1 IRIF 1 h after IR (3 Gy) treatment of RPE-1 MDC1^−/− H2AX^+/+ and MDC1^−/− H2AX^−/− cells complemented with WT or ΔPST versions of GFP-MDC1; n = 4/genotype; error bars s.e.m.; ****p < 0.0001, *p < 0.05; two-tailed Student’s t test. e Clonogenic survivals after IR treatments of RPE-1 MDC1^−/− cells complemented with WT or ΔPST versions of GFP-MDC1; n = 5/genotype; error bars s.e.m. f Same as in (e) but for RPE-1 MDC1^−/− H2AX^−/− cells; n = 3/genotype; error bars s.e.m. Supporting information, as representative images and quantifications of number of 53BP1 foci per cell are presented in Supplementary Fig. 3

Fig. 4

MDC1 PST region binds nucleosomes and promotes DNA damage-independent association of MDC1 with chromatin. a Chromatin fractionation of U2OS MDC1^−/− and MDC1^−/− H2AX^−/− cells transfected with plasmids expressing a GFP-PST construct or GFP-only. b GFP pulldowns from extracts of U2OS MDC1^−/− H2AX^−/− cells expressing GFP-PST or GFP-only were analysed by western blotting using antibodies against the four core histones and GFP. c Representative image of chromatin fractionation of U2OS MDC1^−/− H2AX^−/− cells transfected with plasmids expressing GFP-tagged full-length or ΔPST versions of MDC1 (top panel) and quantification of the relative MDC1 abundance in each fraction (lower panel); n = 4; error bars s.e.m. As previously reported, we found that GFP-MDC1ΔPST mutant protein was expressed in much higher amounts than wild-type GFP-MDC1, implying that the PST region promotes MDC1 turnover. To normalise protein levels, we harvested cells at 48 h after transfection for full-length MDC1 and at 8 h after transfection in the case of ΔPST-MDC1 (Supplementary Fig. 4f). d Western blot to detect the 4 core histones in samples derived from biochemical binding assays between GFP-PST or GFP-only (both purified from transfected HEK293 cells) and core histones purified from calf thymus or native mono-nucleosomes purified from HeLa cells. Additional data can be found in Supplementary Fig. 4, which includes GFP-PST protein localisation after DNA damage, validation of knockouts and PST-DNA binding assays

Fig. 5

Binding of MDC1 PST region to chromatin is mediated by the nucleosome acidic patch and requires several PST repeats. a Depiction of the nucleosome and its acidic patch. Top: electrostatic potential view of the nucleosome surface, coloured according to coulombic surface charge. Bottom: view of the nucleosome surface and localisation of the mutated acidic patch residues. b Biochemical binding assay between GFP-PST or GFP-only and mono-nucleosomes reconstituted using recombinant histones. WT, all core histones are wild-type; AP, acidic patch mutant with the H2A E61/91/92 A and H2B E105A point mutations. c Biochemical binding assay between different GFP constructs, containing 0, 1, 2 or 5 repeats or the full-length (FL) PST fragment, and native mono-nucleosomes purified from HeLa cells. d Chromatin fractionation of U2OS MDC1^−/− H2AX^−/− cells transfected with plasmids expressing the indicated GFP-PST constructs or GFP-only