Recruitment and positioning determine the specific role of the XPF-ERCC1 endonuclease in interstrand crosslink repair

Abstract

XPF-ERCC1 is a structure-specific endonuclease pivotal for several DNA repair pathways and, when mutated, can cause multiple diseases. Although the disease-specific mutations are thought to affect different DNA repair pathways, the molecular basis for this is unknown. Here we examine the function of XPF-ERCC1 in DNA interstrand crosslink (ICL) repair. We used Xenopus egg extracts to measure both ICL and nucleotide excision repair, and we identified mutations that are specifically defective in ICL repair. One of these separation-of-function mutations resides in the helicase-like domain of XPF and disrupts binding to SLX4 and recruitment to the ICL A small deletion in the same domain supports recruitment of XPF to the ICL, but inhibited the unhooking incisions most likely by disrupting a second, transient interaction with SLX4. Finally, mutation of residues in the nuclease domain did not affect localization of XPF-ERCC1 to the ICL but did prevent incisions on the ICL substrate. Our data support a model in which the ICL repair-specific function of XPF-ERCC1 is dependent on recruitment, positioning and substrate recognition.

Keywords: Fanconi anemia; XPF‐ERCC1; Xenopus egg extract; interstrand crosslink repair; nucleotide excision repair.

Figure 1. Characterization of mutant XPF‐ERCC1 complexes

1. Schematic representation of the domain organization of the XPF protein. Domain boundaries of human and Xenopus laevis XPF are indicated. Relevant mutations of the human protein, and the Xenopus laevis equivalents, are indicated on top and bottom, respectively.

2. Superdex 200 gel filtration column elution profile of wild‐type XPF‐ERCC1 and indicated mutant complexes. Aggregates eluted in the void volume of the column (˜45 ml) while the active XPF‐ERCC1 heterodimer eluted at ˜65 ml. The peak eluting at ˜105 ml contains the FLAG peptide used to elute the protein from the FLAG affinity resin. The heterodimer peak was isolated, and proteins were separated on SDS–PAGE and stained with Coomassie blue (inset).

3. As in (B) but for different mutant complexes that showed more aggregation.

4. Wild‐type and indicated mutant XPF‐ERCC1 complexes were incubated with a 5′‐FAM‐labeled stem‐loop DNA substrate (10 nM) at room temperature for 30 min. Reaction products were separated on a 12% urea–PAGE gel and visualized using a fluorescence imaging system. Red arrow indicates position of incision by XPF‐ERCC1.

5. Wild‐type and mutant XPF‐ERCC1 complexes at various concentrations were incubated with a 5′‐FAM‐labeled 3′ flap DNA substrate (10 nM) and fluorescent anisotropy was measured. Graphs were fitted to calculate dissociation constants (K _ds) as described in the Materials and Methods section. The error bars represent s.d. from three measurements. Experimental replicates are shown in Fig EV2.

Source data are available online for this figure.

Figure EV1. Schematic representation of replication‐dependent ICL repair in Xenopus egg extract

Arrow heads represent 3' ends of leading strands.

Figure EV2. Characterization of mutant XPF‐ERCC1 complexes

1. Superdex 200 gel filtration column elution profile of XPF^L219R‐ERCC1. The heterodimeric fraction depicted in Fig 1A was collected, concentrated, and rerun on the same column. The protein did not aggregate and eluted as a heterodimer at ˜65 ml.

2. Replicate of Fig 1D. Wild‐type and indicated mutant XPF‐ERCC1 complexes were incubated with a 5′‐FAM‐labeled stem‐loop DNA substrate (10 nM) at room temperature for 30 min. Reaction products were separated on a 12% urea–PAGE gel and visualized using a fluorescence imaging system. Red arrow indicates position of incision by XPF‐ERCC1.

3. As in (B) but using a 5′‐FAM‐labeled 3′ flap DNA substrate.

4. As in (B) but using higher concentrations of the XE^R670S mutant.

5. Replicate of Fig 1E. Wild‐type and mutant XPF‐ERCC1 complexes at various concentrations were incubated with a 5′‐FAM‐labeled 3′ flap DNA substrate (10 nM) and fluorescent anisotropy was measured. Graphs were fitted to calculate dissociation constants (K _ds) as described in the Materials and Methods section. The error bars represent s.d. from three measurements.

Source data are available online for this figure.

Figure 2. Effect of mutations in XPF‐ERCC1 on ICL repair in Xenopus egg extract

1. A

   Schematic representation of repair of a plasmid containing a site‐specific cisplatin ICL (pICL) in Xenopus egg extract. The SapI site that is blocked by the ICL becomes available on one of the replicated molecules after full repair via HR using the sister molecule (Fig EV1). The sister molecule is repaired by lesion bypass, but retains the unhooked ICL that is not removed efficiently in Xenopus egg extract (Räschle et al, 2008).

2. B

   XPF‐ERCC1‐depleted (ΔXE) and XPF‐ERCC1‐depleted extracts complemented with wild‐type (XE^WT) or indicated mutant XPF‐ERCC1 (XE^MUT) were analyzed by Western blot using α‐XPF antibodies (left panel). Line within blot indicates position where irrelevant lanes were removed. These extracts were used to replicate pICL. Replication intermediates were isolated and digested with HincII, or HincII and SapI, and separated on agarose gel. Repair efficiency, represented by SapI regeneration, was calculated as described (Räschle et al, 2008) and plotted (right panel).

3. C, D

   As in (B) but analyzing different XPF‐ERCC1 mutant complexes. Experimental replicates are shown in Fig EV3. Note: repair levels can differ per batch of individually prepared extract or per depletion experiment and can only be compared within an experiment. *, background band.

Data information: (B–D) #, SapI fragments from contaminating uncrosslinked plasmid present in varying degrees in pICL preparations.Source data are available online for this figure.

Figure EV3. Effect of mutations in XPF‐ERCC1 on ICL repair in Xenopus egg extract

1. Mock‐depleted, XPF‐ERCC1‐depleted (ΔXE), and XPF‐ERCC1‐depleted NPE complemented with SLX4 (ΔXE+S) or XPF‐ERCC1 and SLX4 (ΔXE+SXE) were analyzed by Western blot using α‐XPF or α‐SLX4 antibodies. A dilution series of undepleted NPE was loaded on the same blot to determine the degree of depletion. A relative volume of 100 corresponds to 0.2 μl of NPE.

2. Replicates of Fig 2B. XPF‐ERCC1‐depleted (ΔXE) and XPF‐ERCC1‐depleted extracts complemented with wild‐type (XE^WT) or indicated mutant XPF‐ERCC1 (XE^MUT) were analyzed by Western blot using α‐XPF antibodies (left panel). These extracts were used to replicate pICL. Replication intermediates were isolated and digested with HincII, or HincII and SapI, and separated on an agarose gel. Repair efficiency was calculated and plotted (right panels).

3. As in (B) but analyzing different mutant complexes. Note: repair levels can differ per batch of individually prepared extract or per depletion experiment and can only be compared within an experiment.

Data information: (B, C) #, SapI fragments from contaminating uncrosslinked plasmid present in varying degrees in pICL preparations. (A, C) *, background band. Source data are available online for this figure.

Figure 3. XPF‐ERCC1 mutant complexes are active in nucleotide excision repair (NER)

1. Schematic representation of unscheduled DNA synthesis (UDS) during NER on a UV‐treated template in high‐speed supernatant (HSS) egg extract.

2. Mock‐depleted, XPF‐ERCC1‐depleted (ΔXE), and XPF‐ERCC1‐depleted extracts complemented with wild‐type (XE^WT) or mutant XPF‐ERCC1 (XE^MUT) were incubated with untreated or UV‐treated plasmids for 2 h at room temperature in the presence of ³²P‐α‐dCTP. Reaction products were isolated, linearized with HincII, and separated on a 0.8% agarose gel. The DNA was visualized by autoradiography to show incorporation of ³²P‐α‐dCTP during UDS (upper panel) and stained with SYBR gold for total DNA (lower panel).

3. The incorporation of ³²P‐α‐dCTP was quantified, the background signal from non‐damaged plasmid was subtracted, and the signal for the mock depletion condition was set to 100% to normalize the data. Error bars represent s.e.m. of three independent experiments. **P = 0.003, ***P = 0.0004, paired t‐test comparing all conditions to the mock. All non‐marked conditions did not show a statistical difference from the mock condition.

Source data are available online for this figure.

Figure EV4. XPF‐ERCC1 mutant complexes are active in NER

1. Mock‐depleted and XPF‐ERCC1‐depleted (ΔXE) high‐speed supernatant (HSS) egg extracts used in Fig 3B were analyzed by Western blot using α‐XPF antibodies. A dilution series of undepleted NPE was loaded on the same blot to determine the degree of depletion. A relative volume of 100 corresponds to 0.2 μl of NPE. *, background band.

2. Mock‐depleted, XPF‐ERCC1‐depleted (ΔXE), and XPF‐ERCC1‐depleted HSS complemented with wild‐type (XE^WT) or mutant XPF‐ERCC1 (XE^MUT) were analyzed by Western blot using α‐XPF antibodies. *, background band.

3. Untreated or UV‐treated (10 J/m² _, left panel, 350 J/m² right panel) plasmid DNA was incubated in HSS for 2 h. Samples were taken at time 0 and 2 h, and DNA was extracted and analyzed by an enzyme‐linked immunosorbent assay (ELISA) for the presence of CPDs. The highest value within one experiment was set to 100%. Error bars represent s.e.m. of three independent experiments. **P = 0.0061, paired t‐test. ns, not significant.

4. Mock‐depleted, PCNA‐depleted (ΔPCNA), and PCNA‐depleted HSS complemented with recombinant His‐xlPCNA (ΔPCNA + PCNA) were analyzed by Western blot using α‐PCNA antibodies (Kochaniak et al, 2009). A dilution series of undepleted NPE was loaded on the same blot to determine the degree of depletion. A relative volume of 100 corresponds to 0.2 μl of NPE (top panel). These extracts were incubated with untreated or UV‐treated (350 J/m²) plasmids for 0 or 2 h at room temperature in the presence of ³²P‐α‐dCTP. Reaction products were isolated, linearized with HincII, and separated on a 0.8% agarose gel. The DNA was visualized by autoradiography to show incorporation of ³²P‐α‐dCTP during UDS (bottom left panel). The signal was quantified, the background signal from non‐damaged plasmid was subtracted, and the signal for the mock depletion condition was set to 100% to normalize the data. Error bars represent s.e.m. of three independent experiments. **P = 0.0059, paired t‐test compared to the mock condition. ns, not significant.

5. As in (D) but using mock‐depleted or XPA‐depleted HSS. **P = 0.0014.

Source data are available online for this figure.

Figure 4. XPF‐ERCC1 separation‐of‐function mutants are all defective in ICL unhooking

1. A

   Schematic representation of the assay used to directly measure unhooking incisions. ³²P‐labeled parental stands are indicated in red. Products before and after ICL unhooking during repair are indicated. HincII digestion of repair intermediates yields X‐structures, arms, and linears under denaturing conditions.

2. B

   XPF‐ERCC1‐depleted (ΔXE) or XPF‐ERCC1‐depleted egg extract complemented with wild‐type (XE^WT) or mutant XPF‐ERCC1 (XE^MUT) were incubated with pre‐labeled pICL. Repair products were isolated at indicated times, linearized with HincII, separated on a denaturing agarose gel, and visualized by autoradiography. The X‐structures and linear products were quantified and plotted.

3. C, D

   As in (B) but using different XPF‐ERCC1 mutant complexes. Experimental replicates are shown in Appendix Fig S1.

Figure 5. Recruitment of XPF‐ERCC1 mutants to the ICL during repair

1. A

   Schematic representation showing the primer locations on pICL and pQuant.

2. B

   pICL was replicated in XPF‐ERCC1‐depleted (ΔXE) or XPF‐ERCC1‐depleted egg extract supplemented with wild‐type (XE^WT) or mutant XPF‐ERCC1 (XE^MUT; see Appendix Fig S2). Samples were taken at various times and immunoprecipitated with α‐XPF antibodies. Co‐precipitated DNA was isolated and analyzed by quantitative PCR using the primers depicted in (A). The qPCR data were plotted as the percentage of peak value with the highest value within one experiment set to 100%.

3. C–F

   As in (B) but using the indicated XPF‐ERCC1 mutant complexes. Experimental replicates are shown in Appendix Fig S2.

Figure 6. XPF leucine 219 is part of the major interaction site between XPF and SLX4

1. pICL was replicated in XPF‐ERCC1‐depleted (ΔXE) extract or in XPF‐ERCC1‐depleted extract supplemented with wild‐type XPF‐ERCC1 only (+XE^WT), wild‐type XPF‐ERCC1 and SLX4 (+SXE^WT), or XPF^L219R‐ERCC1 and SLX4 (+SXE^L219R; see Fig EV5A). Samples were taken at the indicated times and immunoprecipitated with α‐XPF (left panel) or α‐SLX4 antibodies (right panel). Co‐precipitated DNA was isolated and analyzed by quantitative PCR using ICL or pQuant primers. The qPCR data were plotted as the percentage of peak value with the highest value set to 100%.

2. Wild‐type and mutant FLAG‐XPF‐ERCC1 were co‐expressed with His‐SLX4 in Sf9 insect cells. Cells were lysed and XPF was immunoprecipitated via the FLAG‐tag. Samples were analyzed by Western blot using α‐FLAG and α‐His antibodies. In, input; FT, flow‐through fraction; B, fraction bound to beads.

3. Schematic representation of xlSLX4 proteins, with the MLR and BTB domains indicated. Experimental replicates are shown in Fig EV5.

4. Purified wild‐type FLAG‐SLX4 and FLAG‐SLX4^∆MLR were added to Xenopus egg extract. SLX4 was immunoprecipitated via the FLAG‐tag. Samples were analyzed by Western blot using α‐FLAG and α‐XPF antibodies. Line within blot indicates position where irrelevant lanes were removed. *, background band.

Source data are available online for this figure.

Figure EV5. XPF leucine 219 is part of the major interaction site between XPF and SLX4

1. XPF‐ERCC1‐depleted (ΔXE) and XPF‐ERCC1‐depleted NPE supplemented with XPF‐ERCC1 (+XE^WT), XPF‐ERCC1 and SLX4 (+SXE^WT), or XPF^L219R‐ERCC1 and SLX4 (+SXE^L219R) were analyzed by Western blot using α‐XPF and α‐SLX4 antibodies. Extracts were used for Fig 6A.

2. As in (A).

3. Replicate of Fig 6A. The extracts from (B), with similarly treated HSS, were used to replicate pICL. Samples were taken at the indicated times and analyzed by XPF (left) and SLX4 (right) ChIP using pICL and pQuant primers. The qPCR data were plotted as the percentage of peak value with the highest value set to 100%.

4. Replicate of Fig 6C. Wild‐type and mutant FLAG‐XPF‐ERCC1 were co‐expressed with His‐SLX4 in Sf9 insect cells. Cells were lysed and XPF was immunoprecipitated via the FLAG‐tag. Samples were analyzed by Western blot using α‐FLAG and α‐His antibodies. In, input; FT, flow‐through fraction; B, fraction bound to beads.

5. Size exclusion chromatography of recombinant XPF‐ERCC1 and BTB domain of SLX4. Superdex 200 gel filtration column elution profile of FLAG‐XPF‐ERCC1, His‐tagged BTB domain, and both proteins combined (top panel). The XPF‐ERCC1 heterodimer eluted at ˜12 ml, while His‐BTB eluted around ˜16 ml. Collected fractions during elution were analyzed by Western blot using α‐XPF and α‐His antibodies (bottom panel). The BTB domain protein does not shift to a higher elution volume when incubated with XPF‐ERCC1 indicating the affinity is not high enough to show binding between the two proteins.

Source data are available online for this figure.

Figure 7. Model for ICL repair‐specific features of XPF

Leucine 219 in the helicase‐like domain of XPF is essential for the interaction of XPF with the MLR domain of SLX4. This interaction mediates the recruitment of XPF to an ICL. Residues 312–315 transiently interact with the BTB domain of SLX4 and are required for the incisions of an ICL by XPF. Arginine 670 and serine 767 in the nuclease domain of XPF are crucial for the recognition of the ICL substrate.