SPRTN patient variants cause global-genome DNA-protein crosslink repair defects

Abstract

DNA-protein crosslinks (DPCs) are pervasive DNA lesions that are induced by reactive metabolites and various chemotherapeutic agents. Here, we develop a technique for the Purification of x-linked Proteins (PxP), which allows identification and tracking of diverse DPCs in mammalian cells. Using PxP, we investigate DPC repair in cells genetically-engineered to express variants of the SPRTN protease that cause premature ageing and early-onset liver cancer in Ruijs-Aalfs syndrome patients. We find an unexpected role for SPRTN in global-genome DPC repair, that does not rely on replication-coupled detection of the lesion. Mechanistically, we demonstrate that replication-independent DPC cleavage by SPRTN requires SUMO-targeted ubiquitylation of the protein adduct and occurs in addition to proteasomal DPC degradation. Defective ubiquitin binding of SPRTN patient variants compromises global-genome DPC repair and causes synthetic lethality in combination with a reduction in proteasomal DPC repair capacity.

Fig. 1. A strategy for the purification of crosslinked proteins.

a Schematic depiction of the Purification of x-linked Proteins (PxP) assay. Cells are harvested and embedded in low-melt agarose plugs. Plugs are transferred to denaturing lysis buffer. Upon completion of lysis, DNA is optionally digested using a nuclease. Next, plugs are transferred to an SDS-PAGE gel and subjected to electro-elution. For DPC detection, plugs are melted following electro-elution, digested with nuclease and analysed using SDS-PAGE followed by western blotting or silver-staining. Alternatively, plugs are fixed and subjected to in-plug tryptic digestion for quantitative proteomics. b Camptothecin (CPT)-induced TOP1-DPC formation assessed by PxP. HeLa T-REx Flp-In cells were treated for 30 min with the indicated doses of CPT prior to isolation of DPCs using PxP and analysis by western blotting. c Untreated or formaldehyde (FA)-treated (2 mM, 1 h) HeLa cells were processed as depicted in (a) and analysed by SDS-PAGE and silver staining. Asterisk indicates Benzonase nuclease used to digest all samples prior to running the final SDS-PAGE. d Mass spectrometry analysis of PxP samples comparing untreated and FA-treated (2 mM, 1 h) HeLa cells. Six plugs per condition were subjected to in-plug tryptic digestion followed by label-free quantitative mass spectrometry. Volcano plot depicting fold change (FC, log2) between conditions plotted against FDR-adjusted P-value (-log10). See also Supplementary Data 1, Supplementary Data 2. e Heatmap showing normalized intensities of six replicates of statistically significant FA-induced DPCs (FDR-adjusted P < 0.01, FC > 2) identified in (d) ranked by average intensity upon FA-treatment. See also Supplementary Data 1, Supplementary Data 2. f PxP analysis of FA-induced histone crosslinks. Cells were treated for 1 h with 2 mM FA and subjected to PxP extraction including a nuclease digestion as indicated and analysed by western blotting. The experiment was repeated twice and similar results were obtained. g PxP analysis of histone H3 crosslinks induced by increasing concentrations of FA. Cells were treated for 1 h with the indicated doses of FA and subjected to PxP analysis including a nuclease digestion as indicated and analysed by western blotting. The experiment was repeated three times and similar results were obtained. Source data are provided as a Source Data file.

Fig. 2. Global-genome repair of 5-azadC-induced DNMT1-DPCs monitored by PxP.

a Schematic depiction of the experimental workflow used to monitor the repair of 5-azadC-induced DNMT1-DPCs. Cells were synchronized via a double thymidine block and released into early/mid S-phase for 3 h prior to induction of DNMT1-DPCs by a 30-min pulse of 5-azadC. Samples were collected either immediately after 5-azadC exposure or following a chase in drug-free media. Proteasome inhibitor (MG132, 5 µM), p97 inhibitor (p97i CB-5083, 5 µM) and SUMOylation inhibitor (SUMO-E1i ML-792, 5 µM) were added 1 h prior to induction of DPCs and kept during the chase with 5-azadC-free medium. Ubiquitylation inhibitor (Ub-E1i TAK-243, 1 µM) was added together with 5-azadC. b 5-azadC-induced DNMT1-DPC formation assessed by PxP. HeLa T-REx Flp-In cells were treated as depicted in (a) with the indicated doses of 5-azadC for 30 min prior to immediate isolation of DPCs using PxP and western blotting analysis. c–f 5-azadC-induced DNMT1-DPC formation and repair upon proteasome inhibition (c), inhibition of SUMOylation (d), inhibition of ubiquitylation (e), or knock-out of RNF4 (f) assessed by PxP. HeLa T-REx Flp-In cells were treated as depicted in (a) prior to extraction of DPCs using PxP, and analysis of samples by western blotting using the indicated antibodies. g HeLa WT and RNF4 knock-out (KO) cells were treated and analysed as depicted including an optional treatment with SUMOylation inhibitor (SUMO-E1i ML-792, 5 µM), prior to extraction of DPCs using PxP and analysis of samples by western blotting using the indicated antibodies. Experiments in (c–g) were repeated three times and similar results were obtained. Source data are provided as a Source Data file.

Fig. 3. The metalloprotease SPRTN cleaves DNMT1-DPCs during global-genome repair.

a HeLa T-REx Flp-In cells transfected with the indicated siRNAs were treated as depicted in Fig. 2a. DNMT1-DPCs were isolated using PxP and analysed by western blotting using the indicated antibodies. b HeLa T-REx Flp-In cells stably expressing siRNA-resistant SPRTN variants (wildtype (WT) or catalytically inactive E112Q (EQ)) were transfected with the indicated siRNAs and treated as in Fig. 2a. DPCs were isolated by PxP and analysed by western blotting using the indicated antibodies. c 5-azadC-induced DNMT1-DPC repair upon inhibition of DNA synthesis assessed by PxP. HeLa T-REx Flp-In cells were treated as depicted including an optional addition of aphidicolin (3 µM) during the chase (left). DNMT1-DPCs were isolated using PxP and analysed by western blotting using the indicated antibodies (right). d–f HeLa WT or RNF4 KO cells were treated with formaldehyde (FA, 250 µM) (d), camptothecin (CPT, 500 nM) (e) or etoposide (ETO, 50 µM) (f), including a 2-h pre-treatment with aphidicolin, as indicated, before whole cell lysates were analysed by western blotting using the indicated antibodies. Source data are provided as a Source Data file.

Fig. 4. SPRTN patient variants affect global-genome DNA-protein crosslink repair.

a Domain structure of SPRTN wildtype and the C-terminally truncated SPRTN-ΔC variant observed in Ruijs-Aalfs syndrome patients indicating SPRTN’s two DNA binding domains (zinc-binding domain, ZBD, and basic region, BR) and interaction motifs/domains for binding to p97 (SHP), PCNA (PIP) and ubiquitin (UBZ) (left). Western blot analysis of HeLa T-REx Flp-In cells genetically-engineered to express patient-like SPRTN-ΔC variants (right). b, c 5-azadC-induced DNMT1-DPC repair assessed by PxP in WT and SPRTN-ΔC cells. HeLa T-REx Flp-In cells were treated as depicted in Fig. 2g including a pre-treatment with proteasome inhibitor MG132 (b) or inhibition of p97 (p97i) (c), as indicated. Upon isolation of DPCs by PxP, samples were analysed using western blotting using the indicated antibodies. d HeLa T-REx Flp-In SPRTN-ΔC cells were complemented with YFP, YFP-SPRTN-Strep variants as indicated (wildtype (WT), E112Q (EQ), R185A (ZBD*)) and treated as shown in Fig. 2a, before DPCs were isolated using PxP and analysed by western blotting using the indicated antibodies. e U2OS T-REx Flp-In WT and SPRTN-ΔC cells were transfected with the indicated siRNAs. Cell confluency was monitored over 5-days using IncuCyte live cell imaging, (top, values represent the mean ± SD of 3 technical replicates), before cells were stained with crystal violet (bottom). Source data are provided as a Source Data file.

Fig. 5. Compromised ubiquitin binding is the main defect of SPRTN patient variants.

a Schematic depiction of YFP-SPRTN-Strep variants used to complement HeLa T-REx Flp-In SPRTN-ΔC cells (top). The localisation of each variant (wildtype (WT), E112Q (EQ), ΔC, and ΔC including an orthogonal N-terminal nuclear localisation signal (NLS-ΔC)) was determined using immunofluorescence staining with anti-GFP antibodies recognizing YFP. Please note that different exposure times are shown for different variants due to different expression levels, see western blot analysis in (b). b HeLa T-REx Flp-In SPRTN-ΔC cells complemented with the indicated SPRTN variants were treated as depicted in Fig. 2a, DPCs were isolated using PxP, and samples were analysed by western blotting using the indicated antibodies. c HeLa T-REx Flp-In SPRTN-ΔC cells complemented with indicated SPRTN variants (full length or an internally-truncated version lacking amino acids 241–400 (Δ241–400) either WT or in combination with an amino acid replacement within the UBZ domain (D473A, UBZ*), were treated as depicted in Fig. 2a. DPCs were isolated using PxP and analysed by western blotting using the indicated antibodies. d Two clonal HeLa T-REx Flp-In SPRTN-ΔUBZ cell lines (see Supplementary Fig. 7a) were treated as depicted in Fig. 2a, before isolation of DPCs and analysis by western blotting using the indicated antibodies. e U2OS T-REx Flp-In WT and SPRTN-ΔUBZ cells were transfected with the indicated siRNAs. Cell confluency was monitored over 5-days using IncuCyte live cell imaging, (left, values represent the mean ± SD of 3 technical replicates), before cells were stained with crystal violet (right). Experiments in (a) and (c) were performed three times and experiments in (b) and (d) twice with similar results. Source data are provided as a Source Data file.

Fig. 6. Model of global-genome DNMT1-DPCs repair.

Repair of 5-azadC-induced DNMT1-DPCs is initiated by SUMOylation, followed by subsequent ubiquitylation by the SUMO-targeted ubiquitin ligase RNF4. Modified DNMT1-DPCs are either targeted by p97- and proteasome-dependent degradation or cleaved by SPRTN. The DPC fragment generated by SPRTN cleavage may be subjected to further degradation by p97 and the proteasome.