XPF-ERCC1 acts in Unhooking DNA interstrand crosslinks in cooperation with FANCD2 and FANCP/SLX4

Abstract

DNA interstrand crosslinks (ICLs), highly toxic lesions that covalently link the Watson and Crick strands of the double helix, are repaired by a complex, replication-coupled pathway in higher eukaryotes. The earliest DNA processing event in ICL repair is the incision of parental DNA on either side of the ICL ("unhooking"), which allows lesion bypass. Incisions depend critically on the Fanconi anemia pathway, whose activation involves ubiquitylation of the FANCD2 protein. Using Xenopus egg extracts, which support replication-coupled ICL repair, we show that the 3' flap endonuclease XPF-ERCC1 cooperates with SLX4/FANCP to carry out the unhooking incisions. Efficient recruitment of XPF-ERCC1 and SLX4 to the ICL depends on FANCD2 and its ubiquitylation. These data help define the molecular mechanism by which the Fanconi anemia pathway promotes a key event in replication-coupled ICL repair.

Figure 1. XPF depletion, but not MUS81 or FAN1 depletion, abrogates ICL repair

(A) Schematic representation of ICL repair in Xenopus egg extract (Raschle et al., 2008). (B) Mock- and MUS81-depleted nucleoplasmic egg extract (NPE) and high-speed supernatant (HSS) were analyzed by Western blot using α-MUS81 antibody. A dilution series of undepleted extract was loaded on the same blot to determine the degree of depletion. A relative volume of 100 corresponds to 0.2 μl NPE or HSS. Line within blot indicates position where irrelevant lanes were removed. (C) As in (B) but using α-FAN1 antibody. To reduce the level of FAN1 further, HSS was diluted in the replication reaction (see supplemental experimental procedures). (D) As in (B) but using α-XPF antibody. (E) pICL was replicated in Xenopus egg extract that was mock-depleted or MUS81-depleted. Replication intermediates were isolated and digested with HincII, or HincII and Sap1, and separated on agarose gel. Repair efficiency was calculated and plotted. (F) As in (E) but for mock- versus FAN1-depleted extract. (G) As in (E) but for mock- versus XPF-depleted extract. Of note: Due to differences in depletion conditions total repair efficiency may vary between experiments. *, background band. #, SapI fragments from contaminating un-crosslinked plasmid present in varying degree in different pICL preparations. See also Figure S1.

Figure 2. ICL repair defect after XPF-ERCC1 depletion is rescued by addition of XPF-ERCC1 in combination with SLX4

(A) Recombinant xlXPF-hsERCC1 complex was isolated via affinity purification using a FLAG-tag on xlXPF and stained with Coomassie blue. (B) Undepleted, mock-depleted, XPF-depleted (ΔXE), and XPF-depleted NPE complemented with the XPF-ERCC1 (ΔXE+XE), were analyzed by Western blot using α-XPF antibody (upper panel). These extracts, with similarly treated HSS, were used to replicate pICL. Repair efficiency was calculated and plotted (lower panel). (C) Undepleted, mock-depleted and ERCC1-depleted NPE were analyzed by Western blot using α-ERCC1, α-XPF, α-FANCD2 and α-SLX4 antibodies. (D) Xenopus laevis SLX4 was purified via a FLAG-tag and stained with Coomassie blue. (E) Mock-depleted, ERCC1-depleted (ΔXE), and ERCC1-depleted NPE complemented with XPF-ERCC1 (ΔXE+XE) or XPF-ERCC1 and SLX4 (ΔXE+XE+S), were analyzed by Western blot using α-XPF and α-SLX4 antibodies (upper panel). These extracts, with similarly treated HSS, were used to replicate pICL. Repair efficiency was calculated and plotted (lower panel). (F) Undepleted, ERCC1-depleted (ΔXE), and ERCC1-depleted NPE complemented with XPF^wt-ERCC1 and SLX4 (ΔXE+XE+S) or XPF^D668A-ERCC1 and SLX4 (ΔXE+XE^mt+S), were analyzed by Western blot using α-XPF or α-SLX4 antibodies (upper panel). These extracts, with similarly treated HSS, were used to replicate pICL. Repair efficiency was calculated and plotted (lower panel). *, background band. #, SapI fragments from contaminating un-crosslinked plasmid present in varying degree in different pICL preparations. See also Figures S2 and S3.

Figure 3. XPF-ERCC1 depletion inhibits unhooking incisions during ICL repair

(A) Schematic representation of the incision assay. ³²P-α-deoxycytosine triphosphate labeled parental strands are in red. Products before and after ICL incisions are indicated. (B) Mock-depleted and XPF-depleted NPE were analyzed by Western blot using α-XPF antibody (upper panel). Pre-labeled pICL and pQuant were replicated in these extracts with similarly treated HSS. Replication products were isolated, digested by HincII, separated on a denaturing agarose gel, and visualized via autoradiography (lower panel). (C) X-structures (left) and linear products (right) visualized in (B) were quantified and plotted. (D) Mock-depleted, XPF-depleted (ΔXE), and FANCD2-depleted (ΔFD2) NPE were analyzed by Western blot using α-XPF and α-FANCD2 antibodies (upper panel). Pre-labeled pICL and pQuant were replicated in these extracts with similarly treated HSS. Replication products were analyzed as in (B), and X-structures and linear products were quantified and plotted (lower panels). *, background band. §, Linear fragments from contaminating un-crosslinked plasmid present in pICL preparation. See also Figure S4.

Figure 4. Unhooking incisions during ICL repair require both SLX4 and nuclease active XPF-ERCC1

(A) Mock-depleted, ERCC1-depleted (ΔXE), and ERCC1-depleted NPE complemented with XPF-ERCC1 (ΔXE+XE) or XPF-ERCC1 and SLX4 (ΔXE+XE+S) were analyzed by Western blot using α-XPF and α-SLX4 antibodies (left panel). Pre-labeled pICL and pQuant were replicated in these extracts with similarly treated HSS. Replication products were analyzed as in Figure 3B and X-structures and linear products were quantified and plotted (right panels). (B) Undepleted, ERCC1-depleted (ΔXE), or ERCC1-depleted NPE complemented with XPF^wt-ERCC1 and SLX4 (ΔXE+XE^wt+S) or XPF^D668A-ERCC1 and SLX4 (ΔXE+XE^mut+S) were analyzed by Western blot using α-XPF and α-SLX4 antibodies (left panel). Pre-labeled pICL and pQuant were replicated in these extracts with similarly treated HSS. Replication products were analyzed as in Figure 3B, and X-structures and linear products were quantified and plotted (right panels). §, Linear fragments from contaminating un-crosslinked plasmid present in pICL preparation. See also Figure S3.

Figure 5. XPF and SLX4 are recruited to ICLs shortly after FANCD2

(A) Schematic representation showing the primer locations on pICL: ICL (25–132 bp from ICL), MID (663–775 bp from ICL), FAR (2523–2622 bp from ICL), and on pQuant. (B) pICL was replicated in the presence of ³²P-α-deoxycytidine triphosphate. Repair intermediates were isolated, digested with AflIII, separated on a sequencing gel and visualized via autoradiography. The leading strand first stalls 20 to 40 nucleotides from the ICL (−20 to −40 products), after approach to the ICL it again stalls 1 nucleotide from the ICL (−1 products). Full lesion bypass results the ‘extension product’. Timing of incisions is indicated below the gel (‘incisions’)(Knipscheer et al., 2009). (C) Samples from parallel pICL replication reaction as in (B) were analyzed by MCM7 ChIP using ICL and pQuant primers. Initial MCM7 signal on pQuant is a result of MCM2-7 loading in HSS, during replication the signal decreases as the MCM2-7 helicase is displaced(Fu et al., 2011). (D) Samples from parallel pICL replication reaction as in (B) were analyzed by FANCD2 ChIP using ICL, MID, FAR and pQuant primers. (E+F) Samples from parallel pICL replication reaction as in (B) were analyzed by XPF (E) and SLX4 (F) ChIP using ICL and pQuant primers. (G) Combined graph for MCM7, FANCD2, XPF and SLX4 ChIP using ICL primers. ChIP data are plotted as the percentage of peak value with the highest value set to 100%. See also Figures S5 and S6.

Figure 6. Efficient recruitment of XPF and SLX4 to ICLs depends on FANCD2

(A) Mock-depleted, FANCD2-depleted (ΔFD2), and FANCD2-depleted NPE complemented with FANCI-FANCD2 (ΔFD2+ID) were analyzed by Western blot using α-FANCD2, α-XPF and α-SLX4 antibodies. (B) Extracts as described in (A), and similarly treated HSS, were used to replicate pICL. Repair intermediates were isolated, digested with AflIII, separated on a sequencing gel and visualized via autoradiography. Based on this lesion bypass assay we infer that the majority of the incisions take place between 60 and 150 minutes in this experiment. (C–E) Samples from parallel pICL replication reaction as in (B) were analyzed by FANCD2 (C), XPF (D) and SLX4(E) ChIP using ICL primers. ChIP data are plotted as the percentage of peak value with the highest value set to 100%. *, background band. See also Figure S7.

Figure 7. Model for XPF-ERCC1-dependent incision(s) mediated by SLX4 and ubiquitylated FANCI-FANCD2

XPF-ERCC1 is recruited to the ICL by SLX4 and most likely makes the 3′ flap incision. The 5′ flap incision could also be made by XPF-ERCC1, or by a currently unknown endonuclease (yellow). FANCD2 and its ubiquitylation are required for the efficient recruitment of XPF-ERCC1 and SLX4 to the ICL. This could be through a direct interaction, through interaction mediated by an additional protein, or by creating an appropriate DNA structure. After loading at the ICL FANCD2 spreads throughout the surrounding DNA (faded FANCI-FANCD2 complexes).