nucleotide-excision-repair

Nucleotide excision repair (NER)

Nucleotide excision repair is a versatile, multi-step DNA repair pathway that removes bulky lesions distorting the DNA double helix — principally UV-induced cyclobutane pyrimidine dimers (CPDs) and 6-4 photoproducts (6-4PPs), as well as chemical adducts from environmental carcinogens (e.g., benzo[a]pyrene diol epoxide / BPDE, aflatoxin B1, AAF), and intrastrand crosslinks from platinum-based drugs (cisplatin). NER is the primary repair route for helix-distorting damage and operates through two sub-pathways — Global Genome NER (GG-NER), which surveys the entire genome, and Transcription-Coupled NER (TC-NER), which prioritizes the transcribed strand of active genes. Inherited loss-of-function mutations in NER genes underlie three classic human syndromes — Xeroderma pigmentosum (XP), Cockayne syndrome (CS), and Trichothiodystrophy (TTD) — and NER capacity declines with normal aging, linking repair insufficiency causally to the genomic instability hallmark ¹.

Naming note: [[nucleotide-excision-repair]] is this pathway page. The sub-pathways (GG-NER: Reactome R-HSA-5696399; TC-NER: Reactome R-HSA-6781827) are covered in dedicated sub-sections here rather than as separate pages, following CLAUDE.md “extend vs new-page” guidance.

Two sub-pathways: GG-NER and TC-NER

NER splits at the damage-recognition step into two functionally distinct sub-pathways:

Feature	GG-NER	TC-NER
Scope	Entire genome (transcribed + non-transcribed, coding + non-coding)	Transcribed strand of RNA Pol II-active genes only
Damage sensor	xpc–RAD23B heterodimer (+ UV-DDB / xpe–DDB1 for CPDs)	Stalled RNA Polymerase II + csb (ERCC6) + csa (ERCC8)
Speed relative to GG-NER	Slower	Faster — active genes repaired in hours vs days for silent regions
Syndrome when deficient	XP (cancer-prone but not progeroid)	CS, some XP variants, TTD (progeroid)
Reactome ID	R-HSA-5696399	R-HSA-6781827

The divergence in clinical outcomes — cancer predisposition (GG-NER loss) vs premature aging (TC-NER loss) — is a key observation: unrepaired transcription-blocking lesions produce persistent transcription stress, stochastic gene expression failure, and cell death / senescence in post-mitotic tissues, rather than the mutagenic outcome that drives cancer ¹.

NER machinery and reaction steps

NER requires >30 polypeptides. The core reaction is shared between GG-NER and TC-NER after the recognition step diverges.

Step 1 — Damage recognition (sub-pathway-specific)

GG-NER: xpc–RAD23B detects helical distortion by probing for single-stranded character on the undamaged strand opposite the lesion. For CPDs (weakly distorting), UV-DDB (the xpe / DDB2–DDB1 complex) is required as an accessory damage sensor that hands off to XPC. 6-4PPs are more distorting and recognized by XPC directly ¹.

TC-NER: RNA Pol II stalling at a lesion recruits csb (ERCC6) first; csb stabilizes the stalled complex and recruits the CUL4–RBX1 E3 ligase adaptor csa (ERCC8; the ERCC8–DDB1–CUL4A–RBX1 complex). The CSA complex ubiquitylates CSB among other substrates, enabling downstream factor recruitment. UVSSA–USP7 protects CSB from proteasomal degradation at this stage ¹.

Step 2 — Verification and TFIIH recruitment

Both sub-pathways converge on recruitment of TFIIH, a 10-subunit complex that includes two ATP-dependent DNA helicases:

• xpb (ERCC3) — 3’→5’ helicase; anchors TFIIH at the lesion and is required for NER but its helicase activity is dispensable for the incision step (acts structurally)
• xpd (ERCC2) — 5’→3’ helicase; unwinds ~20-25 nt of DNA around the lesion; scans the damaged strand to verify the lesion and functions as a gatekeeper

xpa binds with high affinity to the unwound bubble and coordinates damage verification in concert with the single-stranded DNA-binding complex rpa (RPA1–RPA2–RPA3 heterotrimer). RPA coats the undamaged strand, positioning the complex for dual incision ¹.

Step 3 — Dual incision

Two structure-specific endonucleases excise a ~25-30 nucleotide oligonucleotide containing the lesion:

• xpf–ercc1 (ERCC4–ERCC1 heterodimer): cleaves 5’ of the lesion (~15-24 nt 5’ of the damage)
• xpg (ERCC5): cleaves 3’ of the lesion (~2-8 nt 3’ of the damage)

XPG must bind before XPF–ERCC1 cuts; XPG cleavage follows. Incision order and spacing determine the excised fragment size ¹.

Steps 4-5 — Gap-filling and ligation

The ~25-30 nt single-stranded gap is filled by:

• DNA polymerases: Pol δ or Pol ε (replication-coupled in S phase); Pol κ (in non-dividing cells, G1/G0)
• Accessory factors: pcna (sliding clamp), RFC (clamp loader), RPA (stabilizes template)
• Ligation: LIG1 (S-phase, replication-coupled) or LIG3–XRCC1 (non-dividing cells)

(Reactome sub-pathway IDs: GG-NER gap-filling R-HSA-5696397; TC-NER gap-filling R-HSA-6782210.)

Key protein table

Protein	Gene	Role	Syndrome if mutated
XPA	XPA	Damage verification, scaffold	XP-A (severe, with neurodegeneration)
XPB	ERCC3	TFIIH 3’→5’ helicase; anchoring	XP-B, XP/CS (combined), TTD
XPC	XPC	GG-NER damage sensor	XP-C (cancer-prone, no neurodegen.)
XPD	ERCC2	TFIIH 5’→3’ helicase; verification	XP-D, XP/CS, TTD, COFS
XPE / DDB2	DDB2	GG-NER accessory CPD sensor	XP-E (mild)
XPF	ERCC4	5’ endonuclease (with ERCC1)	XP-F, XFE progeroid syndrome
XPG	ERCC5	3’ endonuclease	XP-G, XP/CS
ERCC1	ERCC1	XPF partner; 5’ incision	COFS; no standalone XP; XFE
CSA	ERCC8	TC-NER E3 adaptor; ubiquitylation	Cockayne syndrome A
CSB	ERCC6	TC-NER Pol II-stalling sensor	Cockayne syndrome B
RPA	RPA1/2/3	ssDNA binding; template stabilization	—
PCNA	PCNA	Sliding clamp for resynthesis	—

Progeroid syndromes — causal link to aging

The NER field provides some of the most direct genetic evidence that DNA damage accumulation drives organismal aging. Three syndromes and one engineered mouse model are the pillars.

Xeroderma pigmentosum (XP)

XP arises from mutations in XPA through XPG (or the variant XPV/Pol η). Hallmarks: extreme UV hypersensitivity, ~1000-fold elevated skin cancer risk, and in XP-A, C, D, and G neurological degeneration. XP is primarily a cancer predisposition syndrome; patients do not display premature somatic aging of visceral organs. The neurodegeneration in severe XP-A is attributed to unrepaired oxidative DNA lesions in post-mitotic neurons that are processed through a TC-NER-like mechanism rather than classic UV photoproducts ¹. needs-replication — the neuronal oxidative damage hypothesis for XP-A neurodegeneration is supported by cell biology but not resolved in vivo.

Cockayne syndrome (CS)

CS is the classic progeroid NER syndrome. CSA (ERCC8) or CSB (ERCC6) mutations abolish TC-NER while preserving GG-NER. Clinical features: severe growth failure (cachectic dwarfism), progressive neurodegeneration, photosensitivity, sensorineural deafness, retinal degeneration, and premature death (median survival ~12 years for severe CS type I; range 12–30+ years for milder CS type II). Critically, CS patients do not show elevated cancer rates — consistent with intact GG-NER preventing mutagenesis while TC-NER loss causes cell death and senescence in transcriptionally active tissues ¹. Some XPG and XPD mutations produce combined XP/CS (both cancer and progeroid features), highlighting the modular contribution of each sub-pathway.

Trichothiodystrophy (TTD)

TTD (XPB/XPD/TTDN1 mutations) presents with brittle sulfur-deficient hair (trichorrhexis nodosa, hallmark of the name), ichthyosis, intellectual disability, short stature, and premature aging features including cachexia and reduced life expectancy. TTD is notable because XPD mutations in TTD reduce the stabilizing function of XPD within TFIIH — lowering total TFIIH levels and thereby reducing basal transcription in addition to NER — which explains the non-NER features (sulfur metabolism, developmental defects) not seen in XP ^{2 1}.

Syndrome	Defective proteins	GG-NER	TC-NER	Cancer risk	Progeroid?
XP (A/B/C/D/E/F/G)	XPA-G	Deficient	Deficient (B/D/G only TC-NER too)	High (skin, ocular)	Partial (neurodegen only)
Cockayne syndrome	CSA, CSB	Intact	Deficient	Not elevated	Yes — severe
TTD	XPD (+ XPB, TTDN1)	Partial	Partial	Not elevated	Yes — moderate
XFE progeroid	XPF/ERCC1 (partial)	Deficient	Deficient	Moderate	Yes — severe

Niedernhofer 2006: the ERCC1 progeroid mouse — strongest causal evidence

The most direct evidence that NER failure causes multi-system aging phenotypes comes from the ERCC1-null mouse ³. Ercc1^-/-^ mice (pure null; n=27 in the lifespan cohort) accumulated unrepaired DNA damage throughout life and developed within approximately 3–4 weeks a cluster of multi-organ aging phenotypes, including:

• Liver degeneration (polyploidy, apoptosis; senescent polyploid hepatocytes visible by BrdU/lysosome staining)
• Neurodegeneration (dystonia and progressive ataxia indicative of neurodegeneration; renal insufficiency also reported)
• Sarcopenia and kyphosis
• Cachexia and reduced adipose tissue
• Oxidative stress at the cellular level; premature replicative senescence in primary cells

Shortened lifespan: typically culminating in death by four weeks (n=27; median approximately 3 weeks from Fig. 2b Kaplan-Meier curve); wild-type littermates survive ~2–3 years.

Dimension	Status
Pathway conserved in humans?	yes — XFE progeroid syndrome (partial ERCC1-XPF deficiency) phenocopies elements in human patients
Phenotype conserved in humans?	partial — XFE patients show progeroid features; full ERCC1 null is presumably lethal in humans
Replicated in humans?	partial — XFE patients provide human genetic confirmation; no RCT-level evidence

Crucially, the authors showed that the Ercc1^-/-^ mouse liver transcriptional profile closely matched that of normally aged wild-type C57BL/6 animals (16-week vs 130-week old controls) — providing direct evidence that DNA damage accumulation is sufficient to drive the aging gene expression programme. The Spearman rank correlation between Ercc1^-/-^ and old-mouse liver transcriptomes was r=0.32 (P≤0.0001); correlation with young wild-type liver was r=−0.03 (non-significant) ³. Kidney trends were confirmed by qRT-PCR but a full genome-wide brain comparison was not reported in this paper. needs-replication — the transcriptional-overlap argument was based primarily on liver; genome-wide comparison in post-mitotic tissues (neurons, cardiomyocytes) requires further study.

Note on alleles: Niedernhofer 2006 used Ercc1^-/-^ pure null mice for the core lifespan and transcriptomic data. The related Ercc1^Δ/-^ compound heterozygote (one null allele, one hypomorphic Δ/truncation allele) with a longer lifespan of 4–6 months was the model used in Vermeij et al. 2016 (see below). These are distinct allelic combinations with meaningfully different severities.

Vermeij 2016: 30% dietary restriction triples lifespan of Ercc1^Δ/-^ mice

Vermeij et al. 2016 subjected Ercc1^Δ/-^ compound heterozygous mice (F1 C57BL6J/FVB hybrid background; lifespan 4–6 months) to 30% dietary restriction (DR), starting at 7 weeks with 10% reduction, reaching 30% from 9 weeks onward ⁴. Results:

• Median remaining lifespan extended ~250% in males (10 → 35 weeks; p<0.0001) and ~200% in females (13 → 39 weeks; p<0.0001) in the primary cohort
• A replication cohort in a different facility showed ~180% extension (p<0.0001), confirming robustness
• DR animals retained ~50% more neurons in the neocortex compared to ad libitum controls; locomotor function preserved well beyond the lifespan of ad libitum mice
• Genome-wide DNA damage foci (γH2AX) were reduced in DR mice, suggesting DR attenuates damage accumulation rather than compensating downstream
• ~2/3 of DR-induced transcriptional changes in Ercc1^Δ/-^ liver overlapped with wild-type DR responses, and 684/688 common DEGs changed in the same direction — indicating a strongly shared mechanism

The Xpg^-/-^ Cockayne syndrome-like mouse (lifespan ~18–22 weeks) showed a comparable ~80% median lifespan extension on 30% DR (p<0.0001), extending the finding beyond Ercc1.

Dimension	Status
Pathway conserved in humans?	partial — CR modulates IGF-1/mTOR in humans; whether the same DNA-damage-attenuating mechanism operates at equivalent DR levels is not established
Phenotype conserved in humans?	unknown — no human equivalent experiment is possible in NER-deficient patients
Replicated in humans?	no — mouse model only; authors propose DR(mimetics) as a therapeutic direction for human progeroid NER syndromes

needs-replication — Vermeij 2016 is a single study from one group (though two independent cohorts internally); the 180–250% figure should not be cited without noting the progeroid-model caveat. needs-human-replication

De Boer 2002: TTD mouse confirms transcription as the aging axis

Ercc1-null mice die too young to study aging trajectories in full. The TTD mouse (Xpd^TTD/TTD^; homozygous XPD R722W knockin mimicking a human TTD allele) provided a more tractable model: mice survive to 2 years with moderate premature aging features, including reduced bone density, skin atrophy, reduced subcutaneous fat, and shortened median lifespan (~14% reduction in females) ². Combined Xpd^TTD^ with an Xpa-null background further reduced NER and accelerated aging, confirming dose-response of DNA damage to the aging phenotype.

Dimension	Status
Pathway conserved in humans?	yes — ERCC2/XPD mutations cause TTD in humans with the same hypomorphic-TFIIH mechanism
Phenotype conserved in humans?	yes — TTD patients show comparable aging features (though variable expressivity)
Replicated in humans?	in-progress — human TTD cohort is small (~200 patients globally); longitudinal data limited

NER decline with normal aging

Beyond progeroid syndromes, NER capacity in normal individuals declines with age, contributing to the genomic instability hallmark in the absence of Mendelian defects ¹. Evidence for age-dependent NER decline:

• XPC and XPB protein levels decrease in skin fibroblasts from older donors needs-replication — most data from cross-sectional comparisons with small n
• Unscheduled DNA synthesis (UDS) assays show reduced NER activity in older donor cells
• CPD accumulation per unit UV exposure increases with donor age in human skin

Mechanistically, NER decline with age may reflect: (i) decreased expression of NER factors (XPC transcription declines in aged skin); (ii) chromatin compaction reducing access; (iii) accumulation of oxidized XPA/RPA reducing protein function. no-mechanism — the quantitative contribution of each mechanism is not resolved; longitudinal data in humans is absent.

Aging relevance and cross-links

NER intersects with aging biology at multiple levels:

• genomic-instability (Hallmark of Aging) — NER is a primary genome maintenance pathway; its decline contributes directly to genomic instability accumulation ¹
• cellular-senescence — unrepaired TC-NER substrates in active genes cause persistent DNA damage signalling (via atm / dna-damage-response) → p53/p21 activation → senescence, particularly in post-mitotic tissues
• stem-cell-exhaustion — NER deficiency preferentially depletes tissue stem cell compartments (demonstrated in the ERCC1 mouse ³); NER proficiency is disproportionately important in long-lived stem cells
• dna-damage-response — NER is initiated downstream of damage recognition by the DDR; ATR-mediated checkpoint signalling is triggered by the ssDNA gap intermediate during NER
• atm — ATM responds primarily to DSBs, but NER failure produces secondary DSBs (when replication forks collapse at unrepaired lesions); this ATM-mediated secondary signalling links NER deficiency to accelerated replicative senescence
• mus-musculus — the Ercc1^-/-^ null mouse (Niedernhofer 2006; background not specified in main text) and the Ercc1^Δ/-^ compound heterozygote (Vermeij 2016; F1 C57BL6J/FVB hybrid background) are the principal NER-aging mouse models; the Xpd^TTD/TTD^ model (de Boer 2002) is separately maintained; note these are distinct backgrounds and alleles

Limitations and gaps

• #gap/needs-human-replication — NER decline with aging is demonstrated in cross-sectional cell culture assays; no prospective human cohort has measured NER capacity longitudinally against aging biomarkers.
• #gap/dose-response-unclear — The relationship between quantitative NER activity (e.g., UDS assay output) and aging phenotype severity in non-progeroid humans is unknown. It is unclear how much NER decline is sufficient to contribute to accelerated aging.
• #gap/no-mechanism — Age-associated NER decline mechanism is unresolved: transcriptional silencing vs. post-translational modification of NER factors vs. chromatin accessibility changes.
• #gap/long-term-unknown — There are no validated pharmacological NER enhancers with human safety data; the translational route from NER insufficiency → intervention is undefined.
• #gap/needs-canonical-id — WikiPathways ID WP3528 is unverified; WikiPathways was inaccessible at page-creation time.
• The ERCC1-deficient mouse model uses a severely hypomorphic allele that produces disease within weeks; this extreme is likely not representative of the milder, decades-long NER decline in normal aging. Quantitative gap between mouse model and normal human aging trajectory is not bridged.

Footnotes

Footnotes

1. doi:10.1038/nrm3822 · Marteijn JA et al. · Nature Reviews Molecular Cell Biology 2014 · review · model: human/mouse NER genetics and cell biology · 1,127 citations · not_oa (no local PDF); archive status: not downloaded ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹

2. doi:10.1126/science.1070174 · de Boer J et al. · Science 2002 · in-vivo · n=cohort of Xpd^TTD/TTD^ mice · model: Mus musculus · premature aging features + lifespan reduction in TTD mouse · 539 citations · not_oa (no local PDF) ↩ ↩²

3. doi:10.1038/nature05456 · Niedernhofer LJ et al. · Nature 2006 · in-vivo · n=27 (Ercc1^-/-^ lifespan cohort; Fig. 2b) · model: Mus musculus Ercc1^-/-^ null (pure null; background not specified in main text; WT comparators for microarray were C57BL/6) · progeroid phenotype cluster within ~3 weeks; liver transcriptome Spearman r=0.32 (P≤0.0001) vs 130-wk-old WT liver · 653 citations · local PDF: ↩ ↩² ↩³

4. doi:10.1038/nature19329 · Vermeij WP et al. · Nature 2016 · in-vivo · n=cohort (exact per-group n in supplementary; primary and replication cohorts) · model: Mus musculus Ercc1^Δ/-^ compound heterozygote; F1 C57BL6J/FVB hybrid background; 30% dietary restriction from 9 weeks · median remaining lifespan extended ~250% males / ~200% females (primary cohort; p<0.0001); ~180% in replication cohort (p<0.0001) · local PDF: ↩