Histone H2A variants alpha1-extension helix directs RNF168-mediated ubiquitination

Abstract

Histone ubiquitination plays an important role in the DNA damage response (DDR) pathway. RNF168 catalyzes H2A and H2AX ubiquitination on lysine 13/15 (K13/K15) upon DNA damage and promotes the accrual of downstream repair factors at damaged chromatin. Here, we report that RNF168 ubiquitinates the non-canonical H2A variants H2AZ and macroH2A1/2 at the divergent N-terminal tail lysine residue. In addition to their evolutionarily conserved nucleosome acidic patch, we identify the positively charged alpha1-extension helix as essential for RNF168-mediated ubiquitination of H2A variants. Moreover, mutation of the RNF168 UMI (UIM- and MIU-related UBD) hydrophilic acidic residues abolishes RNF168-mediated ubiquitination as well as 53BP1 and BRCA1 ionizing radiation-induced foci formation. Our results reveal a juxtaposed bipartite electrostatic interaction utilized by the nucleosome to direct RNF168 orientation towards the target lysine residues in proximity to the H2A alpha1-extension helix, which plays an important role in the DDR pathway.

Fig. 1. H2AZ and macroH2A1/2 are RNF168 substrates.

a Schematic illustration of the human H2A family. Consensus sequences were analyzed in comparison to H2A. For macroH2A1 and macroH2A2, the macro domain is excluded from the sequence homology analyses. Green lines represent lysine residues in the histone tail regions. b RNF168 ubiquitinates H2A variants. Cells were co-transfected with Myc-RNF168 and SFB-H2A variants in HEK293T cells, then harvested in SDS-PAGE sample buffer followed by western blot analysis with indicated antibodies. H2AX was used as a positive control. Repeated three times independently with similar results. c Coomassie staining of macroH2A-containing octamer. d Analysis of in vitro reconstitution of macroH2A-containing nucleosome core particles (NCPs). The 207 bp 601 DNA fragment was analyzed alone or after NCP reconstitution. 100 bp DNA ladder indicates size. e in vitro ubiquitination assay of macroH2A-conatining nucleosome was incubated with E1, E2, RNF168 (1-113), ATP and ubiquitin in 1× ubiquitination buffer at 30 ^oC overnight. The reactions were stopped by adding 2× SDS sample buffer. Samples were analyzed by western blot and macroH2A antibody. Repeated three times independently with similar results. Source data are provided as Source Data file.

Fig. 2. RNF168 ubiquitinates H2A variants at a specific lysine residue.

a Schematic diagram of the lysine residue distribution of H2AZ. b H2AZ C-terminus is the major Ub-acceptor. SFB-H2AZ wildtype and mutants were transfected in HEK293T cells for 24 h, harvested with SDS-PAGE sample buffer, and followed by western blot analysis using Flag antibody and tubulin as loading control. Repeated two times independently with similar results. c RNF168 specifically ubiquitinates H2AZ at K15. Co-transfection of Myc-RNF168 and SFB-H2AZ wildtype and mutants with lysine to arginine mutation in HEK293T cells as indicated. Repeated three times independently with similar results. d–f RNF168 ubiquitinates H2A variants at a specific lysine residue. SFB-H2AZ, SFB-macroH2A1, and SFB-macroH2A2 wildtype and mutants were co-transfected with Myc-RNF168 in HEK293T cells followed by western blot analysis using antibodies as indicated. Repeated at least three times independently with similar results. g–i RNF168 is required for site-specific ubiquitination of H2A variants. HEK293T cells with stable expression of CMV-SFB-H2AZ 9K to R-R15K, CMV-SFB-MacroH2A1 8K to R-R11K, EF1α -SFB-MacroH2A2 8K to R-R11K were transfected with siRNAs targeting RNF168. Cells were harvested and pulled down using streptavidin agarose. Pull-down samples were analyzed by western blot with indicated antibodies. Repeated four times for g-h and two times for i with similar results. Source data are provided as Source Data file.

Fig. 3. Mutation of the acidic patches of H2A variants impairs ubiquitination.

a H2A variants have an evolutionary conserved acidic patch. The graphic represents sequence alignment of H2A variants. Acidic residues required to form an intact acidic patch were highlighted in red. b Structural illustration of the H2A-containing, H2AZ-containing and MacroH2A-containing nucleosome. H4 is in light gray, H3 is in dark gray, H2B is in teal, and H2A variants are in the following colors: H2A–light blue; H2AZ–pink; MacroH2A–yellow. The evolutionary and structurally conserved acidic patch is in red and the RNF168-targeted lysine residues are in green. c–e The acidic patch of H2A variants is required for RNF168-mediated site-specific ubiquitination. HEK293T cells were transfected with Myc-RNF168 and SFB-H2A variants and their acidic patch mutants (H2AZ-D93A; MacroH2A1-D87A; MacroH2A2-D87A) as indicated. Cells were harvested 24 h after transfection and ubiquitinations were analyzed by western blot with indicated antibodies. Repeated three times independently with similar results. Source data are provided as Source Data file.

Fig. 4. RNF168-mediated H2A ubiquitination requires the histone tail alpha1-extension helix.

a Sequence alignment of the N-terminal tail and the alpha-1 extension helices of H2A variants. Residues highlighted in blue represent the evolutionary conserved alpha1-extension helix and those in green represent the RNF168-targeted lysine residues. b Analysis of the superimposition of H2A-containing, H2AZ-containing, and macroH2A-containing nucleosome, RNF168-targeted lysine residues were labeled green. c–e Individual H2A-containing (PDB:6FQ5), H2AZ-containing (PDB: 5Z30), and macroH2A-containing (2F8N) nucleosomes at their N-terminal tails. H2B is colored in teal, H2A is in light blue, H2AZ is in pink, and MacroH2A is in yellow. The RNF168-targeted lysines sidechains were labeled green and presented as a stick. f Schematic illustration of the mutants used are presented as in g–k. g–h Conserved residues on alpha1-extension helix are required for RNF168-mediated H2AX ubiquitination. Myc-RNF168 and SFB-H2AX alpha-1 extension helix mutants as indicated were co-transfected in HEK293T cells followed by western blot analysis. Repeated three times independently with similar results. i–k Alpha1-extension helix is required for RNF168-dependent ubiquitination of H2A variants. SFB-H2A variants and their alpha1-extension helix mutants were co-transfected with Myc-RNF168 in HEK293T cells. Cells were harvested 24 h after transfection and analyzed by western blot. Repeated three times with similar results. Source data are provided as Source Data file.

Fig. 5. Ubiquitination requirement for RNF168 selectivity on H2A variants.

a Superimposition of H2A-containing, H2AZ-containing, and macroH2A-containing nucleosome illustrates the detailed residues required for RNF168-mediated site-specific ubiquitination. Zoom on the acidic patch (red), alpha1-extension helix region (blue) and RNF168-targeted lysine residues (green). b Proximity restriction of RNF168-mediated site-specific ubiquitination on H2A. Sequence alignment of human H2AX and yeast HTA (yHTA) at the N-terminal tail. The N-terminal sequence of the yHTA K13m mutant (ΔASQ) used as in c. The alpha1-extension helix residues are highlighted in blue. The proximal lysine residues are highlighted in green. c SFB-yHTA and mutant were co-transfected with Myc-RNF168 in HEK293T cells for 24 h and analyzed by western blot. Repeated three times independently with similar results. d Electrostatic potential (red-negative, blue-positive) analyses of H2A variant-containing nucleosomes. Zoomed illustration of the N-terminus and C-terminus of H2A variants. Histone tail lysines were labeled. H2AZ C-terminal lysines were absent in the illustration and G119 residue was marked. Source data are provided as Source Data file.

Fig. 6. UDM1 is required for RNF168-mediated 53BP1 recruitment.

a Schematic illustration of RNF8 and RNF168 structural domain organization, RNF168 fragments, and chimeric proteins used as in b and c. FHA–forkhead-association domain; coiled-coil–coiled-coil domain; RING–ubiquitin E3 ligase RING domain; Rs-arginine anchor; LRM–LR motif; UMI–UIM-and MIU-related ubiquitin binding domain; MIU–motif interacting with ubiquitin. b–c Transient transfection of GFP-tagged RNF168 fragments and chimeric proteins in U2OS RNF168 KO cells. Cells were irradiated with 10 Gy and allowed to recover for 1 h followed by immunofluorescence analysis using 53BP1 antibody as indicated. Cells were counterstained with DAPI. Repeated at least two times independently with similar results. Source data are provided as Source Data file.

Fig. 7. UMI acidic residues are required for RNF168-mediated site-specific ubiquitination and the DDR.

a Sequence alignment analysis of RNF168 UMI domains across species. The evolutionary conserved E143/E144 are highlighted in red. b Deletion of residues 143-144 attenuates 53BP1 IRIF formation. U2OS RNF168 KO cells were transiently transfected with GFP-RNF168 or RNF168 Δ143-144. Cells were irradiated with 10 Gy, allowed to recover for 1 h, followed by immunofluorescence analysis with indicated antibodies. Repeated three times with similar results. c Deletion of 143-144 abolishes RNF168-mediated ubiquitination. Co-transfection of GFP-RNF168 or mutant with Flag-H2AX K13/K15 only vector in HEK293T cells followed by western blot analysis. Repeated three times with similar results. d RNF168 Δ143-144 is recruited to laser damage. U2OS RNF168 KO cells expressing GFP-RNF168 and RNF168 Δ143-144 were damaged using a 405 nm laser source and analyzed 15 min later by confocal microscopy. Dotted line indicates the laser path. Repeated three times independently with similar results. e–f U2OS cells were transfected with GFP-RNF168 wildtype and mutants. Proximity ligation assay (PLA) was performed using specific antibodies against GFP and γH2AX. Mean intensity of PLA signals in GFP-positive cells were measured and analyzed using ImageJ software. Data presented as mean ± SD. One-way ANOVA, Dunn’s multiple comparison test was used in statistical analysis, wildtype vs. R57D P = 0.1455, wildtype vs. Δ143-144 P = 0.1589, R57D vs. Δ143-144, P > 0.9999, ****P < 0.0001 with 95% confidence interval. n = 75 (wildtype), 48 (R57D), 74 (Δ143-144), 53 (R57D/Δ143-144) over 3 independent experiments. Source data are provided as Source Data file.