🧬 Aging Wiki

xpf

Properties1
aliasesERCC4, XPF, DNA repair endonuclease XPF, ERCC11

15 min read

XPF (ERCC4)

XPF (encoded by ERCC4) is the catalytic subunit of the XPF-ERCC1 structure-specific endonuclease, the enzyme that makes the 5’ incision during nucleotide excision repair (NER) and the crosslink-unhooking cut during interstrand crosslink (ICL) repair. Hypomorphic and null mutations in XPF/ERCC4 cause a spectrum of human disorders — from mild xeroderma pigmentosum (XP-F) to the severe XFE progeroid syndrome — making this protein one of the clearest genetic demonstrations that accumulating DNA damage drives accelerated aging. The catalytic GDxxS nuclease motif resides in XPF; its non-catalytic partner ercc1 provides scaffolding and DNA-damage recognition.

Identity

  • UniProt: Q92889 (ERCC4_HUMAN)
  • NCBI Gene: 2072
  • HGNC symbol: ERCC4 (alias XPF, ERCC11)
  • Ensembl: ENSG00000175595
  • Mouse ortholog: Ercc4 / Xpf (one-to-one ortholog; all progeroid mouse models use Ercc1 or Ercc4 alleles)
  • GenAge: entry 261 (listed for humans, mice, and C. elegans; evidence quality: disease/progeroid model)
  • Length: 916 amino acids (canonical isoform per UniProt Q92889; Sijbers 1996 original cDNA clone encoded 905 aa per EMBL U64315 — likely reflects updated reference sequence annotation)
  • Complex partner: ERCC1 (UniProt P07992) — obligate heterodimer; XPF is unstable without ERCC1 in cells

Domain architecture

XPF is a mosaic protein retaining structural elements of ancestral ERCC4-family helicases, though the helicase ATPase activity is defunct in the vertebrate protein 1:

DomainResidues (approx.)Function
Helicase-like (N-terminal)1–457Catalytically inactive vestige; contains two leucine-zipper regions (~233–254, ~270–298); contributes to protein folding and likely to substrate interaction
Nuclease / ERCC4 domain658–813 (region); 683–763 (ERCC4 domain core per UniProt)Active site carrying the GDxxS metal-binding motif; Mg²⁺ cofactor; executes the 5’ phosphodiester cleavage; catalytic residues Asp-687/Glu-689 per Enzlin & Schärer 2002
HhH2 dimerization~837–905Helix-hairpin-helix fold; interlocks with ERCC1’s HhH2 domain to form the heterodimer interface

Active-site residues: The ERCC4 nuclease domain employs a GDxxS catalytic motif coordinating Mg²⁺ for phosphodiester hydrolysis; the catalytic residues were mapped to Asp-687 and Glu-689 by Enzlin and Schärer 2002 (EMBO J 21:2045–2053) using site-directed mutagenesis of the conserved nuclease motif. UniProt Q92889 (verified 2026-05-07) does not annotate explicit “Active site” typed features at these positions in its feature table — the domain annotation lists the ERCC4 domain as residues 683–763, which is consistent. The catalytic mechanism involves a one-metal-ion nucleophilic attack on the phosphate 5’ of the scissile bond 1. PTMs confirmed by UniProt Q92889: N6-acetyllysine at Lys-289 (within the N-terminal helicase-like region) and Lys-911 (within or near the HhH2 domain); phosphoserine at Ser-521 and Ser-764. The regulatory significance of these PTMs in vivo is not established. needs-replication

Structural specificity: Like its archaeal ERCC4 homologs, human XPF-ERCC1 is a structure-specific endonuclease — it incises at single-strand/double-strand junctions (ssDNA-dsDNA boundaries) rather than recognizing a specific sequence. This allows one nuclease to operate across all NER substrates and ICL intermediates.

NER mechanism — XPF’s role

nucleotide-excision-repair removes helix-distorting bulky adducts (UV photoproducts, platinum-DNA adducts, bulky chemical adducts). The pathway culminates in a dual incision that excises a ~27–29 nt damage-containing oligonucleotide:

  1. 5’ incision by XPF-ERCC1 — ~20–22 nt 5’ of the lesion on the damaged strand; XPF’s catalytic ERCC4 domain performs the cut; ERCC1 positions the complex at the bubble via interactions with xpa and rpa 2
  2. 3’ incision by XPG — ~6–8 nt 3’ of the lesion; XPG is a separate ERCC2/XPD-family nuclease unrelated to XPF

Sub-pathway distinction: In GG-NER (global-genome NER, initiated by XPC-RAD23B damage recognition), XPF-ERCC1 acts before XPG cleavage. In TC-NER (transcription-coupled NER, triggered by stalled RNA Pol II at lesions), the order may differ, though both incisions are required for gap formation. ERCC1 is recruited to the open bubble via xpa binding the ERCC1 central domain 2.

After dual incision:

  • Gap-filling DNA synthesis by Pol delta or epsilon + pcna + RFC
  • Ligation by LIG1 (replicating cells) or LIG3 (non-replicating)
  • The excised 27–29 nt fragment is degraded

ICL repair — XPF as the unhooking nuclease

Interstrand crosslinks (ICLs) covalently link both strands of DNA, blocking replication forks and transcription. XPF-ERCC1 is the primary ICL “unhooking” nuclease in vertebrates 3. The canonical vertebrate ICL repair mechanism:

  1. Replication fork collision with the ICL triggers arrest and recruitment of the Fanconi Anemia (FA) core complex → FANCD2/FANCI monoubiquitination
  2. XPF-ERCC1 (scaffolded by SLX4/FANCP) makes dual incisions flanking the crosslink on one strand, “unhooking” it and creating a DSB intermediate 3
  3. The unhooked ICL-oligonucleotide remnant on the template is bypassed by TLS polymerases (Pol zeta, Rev1)
  4. The DSB is repaired by homologous recombination

Loss of XPF causes Fanconi anemia complementation group Q (FANCQ) — the FA classification reflects XPF-ERCC1’s essential ICL-repair function. Severe XPF mutations (e.g., R689S at the catalytic site) in the FA-Q patient context produce marrow failure, congenital abnormalities, and cancer predisposition distinct from XP-F or XFE.

Single-strand annealing (SSA)

After DSB-flanking end resection creates long 3’ ssDNA tails, RAD52-mediated strand annealing can pair flanking direct repeats. XPF-ERCC1 then removes the resulting 3’ nonhomologous flaps to allow ligation. SSA is a mutagenic pathway (it deletes the intervening sequence); XPF-ERCC1’s role here is to enable completion of an error-prone pathway rather than error-free repair. The aging relevance of SSA flap-trimming by XPF-ERCC1 is not established. no-mechanism

Role in aging

XFE progeroid syndrome — the human genetic argument

The XFE progeroid syndrome (OMIM #278760 for XP-F; XFE features are driven by severe ERCC4 mutations) is the most phenotypically extreme form of xeroderma pigmentosum group F 4. The Niedernhofer 2006 Nature paper described a patient with compound heterozygous severe XPF mutations exhibiting:

  • Profound crosslink sensitivity
  • Dramatic multi-organ progeroid symptoms (neurodegeneration, cachexia, growth failure, dysmorphic features)
  • Transcriptomic profile of the progeroid mouse liver overlapping significantly with the transcriptome of aged wild-type mice — establishing that XPF/ERCC1 deficiency recapitulates normal aging at the gene-expression level 4

The same paper showed that Ercc1-deficient mouse liver transcriptomes (the murine model) closely matched aged wild-type liver expression patterns, providing the molecular-correlation evidence that this progeroid model is mechanistically relevant to normal aging rather than an artifact of UV hypersensitivity.

For the full Ercc1-/- mouse phenotype (lifespan, organ-level progeroid features, quantitative somatotrophic axis suppression), see ercc1 § Role in aging, which holds the primary quantitative claims from Niedernhofer 2006. The progeroid phenotype is caused by loss of the ERCC1-XPF heterodimer — both subunits’ pages cross-reference the same study data to avoid duplication.

Somatotrophic axis suppression — genotoxic stress as an aging signal

The most conceptually significant finding from Niedernhofer 2006 was that genotoxic stress suppresses the GH/IGF-1 axis — phenocopying the hormonal milieu of caloric restriction and long-lived mutant mice. XPF/ERCC1-deficient mice showed significantly lower circulating IGF-1 and insulin, and reduced blood glucose 4. This was interpreted as an adaptive survival response: cells detect unresolved DNA damage and shift from growth to maintenance programs by downregulating anabolic somatotrophic signaling.

Garinis et al. 2009 extended this to a general principle: persistent transcription-blocking DNA lesions (which accumulate when NER is impaired) reduce IGF-1 receptor and GH receptor expression, generating cellular resistance to growth signals and oxidative stress across proliferating and differentiated cell types 5. This “growth attenuation” response is proposed to promote longevity at the organism level by redirecting resources from growth to somatic maintenance.

DimensionStatusNotes
Pathway conserved in humans?yesNER, ICL repair, and IGF-1 signaling are highly conserved
Phenotype conserved in humans?partialXFE and XP-F provide human genetic support; IGF-1 suppression in normal human aging via DNA damage load not directly demonstrated
Replicated in humans?noGenetic models only; human XFE too rare for quantitative population analysis

needs-human-replication — The specific claim that endogenous DNA damage (repaired by XPF-ERCC1) suppresses the somatotrophic axis in normal human aging has not been directly demonstrated.

Dietary restriction rescues XPF-ERCC1-deficient mice — Vermeij 2016

Vermeij et al. 2016 demonstrated that dietary restriction (DR) — starting at 10% food reduction from week 7 and reaching 30% from week 9 onward — extended median remaining lifespan by ~250% in male and ~200% in female Ercc1^Δ/− compound heterozygous mice (natural median lifespan 10–13 weeks to 35–39 weeks; p<0.0001) 6. The result was replicated in a second animal facility at ~180% median extension (p<0.0001). DR also rescued Xpg-/- (Cockayne-syndrome model) mice, suggesting the rescue is pathway-level, not allele-specific. Crucially, DR reduced γH2AX foci (DNA damage accumulation) in Purkinje neurons and retained ~50% more neocortical neurons versus ad libitum controls — consistent with the interpretation that DR attenuates rather than merely compensates for the DNA damage load.

For full quantitative detail on the DR rescue, see ercc1 § Vermeij 2016 — dietary restriction rescues the Ercc1^Δ/− mouse.

A subsequent rapamycin study (Birkisdottir et al. 2021) found that rapamycin failed to extend lifespan or reduce transcription stress in DNA repair-deficient progeroid mice, despite lowering mTOR signaling — indicating that dietary restriction’s benefit operates through mechanisms beyond mTOR inhibition alone, and that XPF-ERCC1 deficiency is not simply equivalent to mTOR pathway insufficiency 7.

Deng et al. 2017 measured ERCC4/XPF mRNA in PBMCs from 147 donors and found a steep age-dependent decline: r = −0.844; p<0.001 — steeper than the parallel ERCC1 decline (r = −0.578) 8. If XPF expression declines with age in other tissues (not yet established), this would reduce the heterodimer pool and NER capacity, creating a positive feedback loop: less XPF-ERCC1 → more unrepaired bulky lesions → greater genomic instability → more activation of senescence/apoptosis programs.

needs-human-replication — Tissue-level XPF expression decline with age has not been directly measured; only PBMC data is available 8. no-mechanism — The driver of age-dependent XPF mRNA decline is unknown.

Disease spectrum

XPF/ERCC4 mutations cause a phenotypic spectrum reflecting residual heterodimer activity 4 1:

SyndromeResidual NERKey features
Xeroderma pigmentosum F (XP-F)~10–25%UV hypersensitivity; skin cancers; mild to moderate neurological involvement; no progeroid aging
XP-F/Cockayne Syndrome overlap~5–10%XP-F features + Cockayne-like neurodegeneration, cachexia, photosensitivity
XFE progeroid syndrome<5%Profound progeroid features; ICL hypersensitivity; death in infancy to early childhood
Fanconi anemia (FANCQ)ICL-specificBone marrow failure; congenital abnormalities; cancer predisposition; caused by XPF mutations specifically affecting ICL-unhooking function

The correlation between residual activity and phenotypic severity is consistent with the model that endogenous DNA damage (mainly ICLs and transcription-blocking adducts accumulating throughout life) is the pathogenic substrate, not UV-induced lesions per se.

Pharmacology and druggability

Druggability tier: 3 (aging-context): XPF-ERCC1 is predicted druggable via the ERCC1-binding HhH2 surface (a protein-protein interaction, or PPI) or via allosteric targeting of the nuclease domain, but no clinical-stage drug exists for any indication.

Cancer chemosensitization: XPF-ERCC1 repairs platinum-DNA adducts; high expression predicts resistance to cisplatin and oxaliplatin in multiple tumor types. This has motivated small-molecule XPF-ERCC1 PPI inhibitor programs (targeting the ERCC1-XPF interface to prevent heterodimerization and sensitize tumors to platinum chemotherapy). Several nanomolar compounds have been described in academic laboratories, but none have advanced to clinical trials as of 2026-05-07. needs-canonical-id — specific compound page for XPF-ERCC1 inhibitors not yet seeded.

Aging-directed therapeutic hypothesis: Restoring XPF expression or heterodimer stability in aged tissues could slow accumulation of endogenous DNA damage and delay the somatotrophic response. This remains speculative; no pre-clinical aging data supports it. A gene therapy delivering wild-type ERCC4 to aged rodents is conceptually the next experiment. no-mechanism

Pathway membership

Key interactors

  • ercc1 — obligate non-catalytic partner; XPF is unstable and non-functional without ERCC1; the heterodimer is the functional unit for all XPF repair activities
  • xpa — recruits ERCC1-XPF to the NER bubble via direct ERCC1 interaction; indispensable for 5’ positioning 2
  • rpa — coats the undamaged strand at the bubble; contributes to lesion discrimination and heterodimer orientation
  • xpg — performs the 3’ incision; coordinates with XPF-ERCC1 for dual excision; XPG binding may stimulate XPF incision in TC-NER
  • slx4 (FANCP) — scaffold for XPF-ERCC1 in ICL repair; SLX4 mutations cause Fanconi anemia complementation group P; SLX4IP stabilizes the SLX4-XPF-ERCC1 interaction following DNA damage 3

Limitations and gaps

  • #gap/needs-human-replication — Progeroid and DR-rescue data are entirely mouse (Ercc1/Xpf alleles); human XFE cases are too rare for quantitative extrapolation.
  • #gap/needs-replication — The Vermeij 2016 DR rescue (200–250% median lifespan extension in primary cohort; ~180% in second-facility replication within the same study) has not been independently replicated by a separate laboratory.
  • #gap/no-mechanism — The mechanism driving age-dependent decline of XPF mRNA in human PBMCs is unknown.
  • #gap/needs-human-replication — Tissue-level XPF-ERCC1 expression and NER capacity decline with age in non-PBMC human tissues are not established.
  • #gap/needs-replication — Active-site residue significance (Asp-687, Glu-689 mapped by Enzlin & Schärer 2002 mutagenesis; not annotated as active-site residues in UniProt Q92889 feature table as of 2026-05-07) and PTM in-vivo significance (Lys-289 and Lys-911 acetylation, Ser-521 and Ser-764 phosphorylation) have not been validated in aging-context studies.
  • #gap/no-mechanism — SSA flap-trimming by XPF-ERCC1: aging relevance of this error-prone repair activity in aged tissues is not established.
  • GenAge HAGRID 261 confirmed (ERCC4, Homo sapiens) via direct genomics.senescence.info query on 2026-05-07 verification pass. Open Targets druggability tier 3 not confirmed via API call; ENSG00000175595 lookup pending. needs-canonical-id (Open Targets only)

Footnotes

Footnotes

  1. doi:10.1016/s0092-8674(00)80155-5 · Sijbers AM, de Laat WL, Ariza RR et al. (Hoeijmakers JH, Wood RD labs) · Cell 1996 · in-vitro (biochemical reconstitution + cDNA cloning) · n=n/a · model: human cell lines complementation + purified ERCC4/XPF protein; cDNA encodes 905 aa (EMBL U64315; note current UniProt Q92889 canonical isoform is 916 aa — minor discrepancy likely reflects updated annotation); XPF purified as obligate ERCC1-XPF heterodimer (~42 kDa ERCC1 + 115 kDa XPF); complex is structure-specific endonuclease cleaving 2–4 nt 5’ of ssDNA-dsDNA junction; established XPF as catalytic 5’ incision subunit of NER; gene mapped to chromosome 16p13.1–13.2 · PDF locally available ↩ ↩2 ↩3

  2. doi:10.1038/sj.emboj.7601894 · Tsodikov OV, Ivanov D, Orelli B, Staresincic L, Shoshani I, Oberman R, Schärer OD, Wagner G, Ellenberger T · EMBO J 2007 · in-vitro (NMR + X-ray crystallography) · n=n/a · model: XPA peptide (residues 67–80; KIIDTGGGFILEEE) in complex with ERCC1 central domain (residues 92–214); KD ~0.78 µM; XPA inserts Gly72/73/74 into V-shaped groove of ERCC1; XPA peptide inhibits NER in HeLa cell-free extracts; establishes XPA–ERCC1 interaction as essential for ERCC1-XPF recruitment to NER bubble · PDF locally available — NOTE: originally miskeyed as “park2019” and misattributed to “Park CJ, Choi BS” in auto-extraction; corrected to authors Tsodikov OV et al. on 2026-05-07 verification pass. ↩ ↩2 ↩3

  3. doi:10.1093/nar/gkz769 · Zhang H, Chen Z, Ye Y et al. (Chen J lab, MD Anderson) · Nucleic Acids Research 2019 · in-vitro + cellular · model: SLX4IP depletion (CRISPR KO + shRNA) in HEK293A/T cells sensitizes to MMC and camptothecin; SLX4IP binds SLX4 and XPF-ERCC1 simultaneously; disrupting one interaction disrupts both; SLX4IP stabilizes the SLX4-XPF-ERCC1 ternary complex especially after DNA damage; identifies SLX4IP as a regulatory scaffold maintaining efficient XPF-ERCC1 ICL unhooking · PDF locally available ↩ ↩2 ↩3

  4. doi:10.1038/nature05456 · Niedernhofer LJ, Garinis GA et al. · Nature 2006 · in-vivo + case-report · n=27 (Ercc1-/- lifespan cohort) + 1 human XFE patient · model: Ercc1-/- mice in F1 hybrid background (C57BL/6 cross); severe ERCC4/XPF mutation (R153P) causing XFE progeroid syndrome; Ercc1-/- mouse liver transcriptome overlaps aged WT (Spearman r=0.32, P≤0.0001 vs old mice; r=−0.03 vs young); somatotrophic axis suppression: serum IGF-1 significantly lower P<0.001, blood glucose P<0.001, serum insulin P<0.03 · PDF locally available; KEY aging paper ↩ ↩2 ↩3 ↩4

  5. doi:10.1038/ncb1866 · Garinis GA, Uittenboogaard LM, Stachelscheid H et al. (Hoeijmakers JH, Schumacher B labs) · Nature Cell Biology 2009 · in-vitro + in-vivo · model: primary mouse dermal fibroblasts from Csb^m/m/Xpa-/- (severe progeroid), Csb^m/m (mild), Xpa-/- and wild-type mice, plus primary chondrocytes and neurons; UV doses 0.6–4 J/m²; IGF-1R and GHR mRNA suppressed within 6 h of UV in a dose-dependent manner in all cell types including post-mitotic neurons and quiescent fibroblasts; suppression exacerbated and prolonged in progeroid Ercc1-/- cells; attenuation is transcription-coupled (Xpc-/- GG-NER-only cells recover normally); resistance to oxidative stress induced; establishes cell-autonomous attenuation of IGF-1/GHR as primary response to persistent transcription-blocking lesions · PDF locally available ↩

  6. doi:10.1038/nature19329 · Vermeij WP et al. · Nature 2016 · in-vivo (dietary restriction intervention) · model: Ercc1^Δ/− compound heterozygote (F1 C57BL6J/FVB hybrid background); DR protocol: 10% restriction from week 7, reaching 30% from week 9 onward; primary cohort: males median lifespan 10→35 wks (250% extension; p<0.0001), females 13→39 wks (200% extension; p<0.0001); second-facility replication: ~180% median lifespan extension (p<0.0001); γH2AX foci reduced in DR animals; ~50% more neurons retained in neocortex vs ad libitum controls; Xpg-/- mice also rescued by same DR protocol · PDF locally available ↩

  7. doi:10.1111/acel.13302 · Birkisdottir MB, Jaarsma D, Brandt RMC et al. (Hoeijmakers JH, Vermeij WP, Dollé MET labs) · Aging Cell 2021 · in-vivo · model: Ercc1^Δ/− female mice (F1 C57BL6J/FVB background); rapamycin at 14 ppm (standard ITP dose) from 4 weeks of age failed to extend median or maximal lifespan (log-rank p=0.71 vs control); dose-response tested at 4.7, 14, and 42 ppm — all failed; genetic mTOR modulation via Rheb^+/− and Tsc1^+/− also ineffective; rapamycin did reduce S6 phosphorylation confirming mTOR target engagement; 30% DR in parallel cohort extended lifespan ~200% (p<0.0001); rapamycin did not reduce transcription stress or P53+ neurons; DR and rapamycin differ mechanistically — DR benefit not primarily mTOR-dependent · PDF locally available ↩

  8. cite from ercc1 § Deng 2017 — doi:10.1016/j.jflm.2017.05.005 · Deng XD et al. · J Forensic Leg Med 2017 · observational · n=147 donors · XPF mRNA r = −0.844 with age (steeper than ERCC1 r = −0.578) — citation inherited from ercc1 page; DOI-title match not independently verified against DOI lookup on this verification pass (PDF not in archive); needs-replication — verify DOI-title match on next lint pass ↩ ↩2

Graph View

Backlinks