Strand- and site-specific DNA lesion demarcation by the xeroderma pigmentosum group D helicase

Abstract

The most detrimental responses of the UV-exposed skin are triggered by cyclobutane pyrimidine dimers (CPDs). Although placental mammals rely solely on nucleotide excision repair (NER) to eliminate CPDs, none of the core NER factors are apparently able to distinguish this hazardous lesion from native DNA, raising the question of how CPDs are circumscribed to define correct excision boundaries. A key NER intermediate involves unwinding of the damaged duplex by transcription factor TFIIH, a reaction that requires xeroderma pigmentosum group D (XPD) protein. This study was prompted by the observation that the ATPase/helicase activity of XPD is necessary for an effective anchoring of this subunit to UV lesions in mammalian nuclei. The underlying mechanism by which XPD impinges on damaged DNA has been probed with a monomeric archaeal homolog, thus revealing that the collision with a single CPD inhibits the helicase but stimulates its ATPase activity. Restriction and glycosylase protection assays show that the XPD helicase remains firmly bound to a CPD situated in the translocated strand along which the enzyme moves with 5'-3' polarity. Competition assays confirm that a stable complex is formed when the XPD helicase encounters a CPD in the translocated strand. Instead, the enzyme dissociates from the substrate after running into a CPD in the complementary 3'-5' strand. These results disclose a damage verification and demarcation process that takes place by strand-selective immobilization of the XPD helicase and its conversion to a site-specific ATPase at DNA lesions.

Fig. 1.

Interaction of human XPD with damaged sites in living cells. (A) Representative foci of UV damage in CHO cells detected by immunochemistry against CPDs 30 min after UV irradiation. The accumulation of XPD proteins (wild-type and K48R mutant) is visualized by measuring the fluorescence intensity in cells transfected with the respective GFP constructs. (B) Comparison between total fluorescence, reflecting the overall expression of XPD fusions, and local fluorescence intensity in UV foci (N = 30; ± SEM). See SI Text for quantification methods. (C) Dissociation of XPD-GFP proteins from UV foci determined by fluorescence recovery after photobleaching on local damage (FRAP-LD) (N = 13; ± SEM). See SI Text for a detailed description of data acquisition and analysis.

Fig. 2.

Immobilization of the XPD enzyme. (A) Electrophoretic analysis and Coomassie staining of FaXPD protein. (B) Oligonucleotide used for competition assays. The CPD is indicated by a triangle and the ³²P label by an asterisk. (C) Competition in the presence of 3 mM ATP showing the dissociation of FaXPD from undamaged oligomers (lanes 3–8) and the stability of radiolabeled complexes containing a single CPD (lanes 10–15). Competitor DNA (undamaged 51-mer) was added in a 50-fold excess. Lanes: 1, incubation without protein; 2 and 9, control incubations (60 nM FaXPD and 5 nM radiolabeled oligonucleotides) without competitor DNA. F, free probes; B, protein-bound fraction. (D) ATP-dependent dissociation of FaXPD from undamaged oligonucleotides. Lanes: 4 and 8, competition assays with nonhydrolyzable ATPγS. (E) Quantification of competition assays. FaXPD (60 nM) was incubated (15 min) with radiolabeled 51-mers (5 nM). A 50-fold molar excess of unlabeled 51-mers was then added in the presence of 3 mM ATP and, after varying competition periods, the samples were analyzed in mobility shift gels. The fraction of protein-bound DNA is represented as the percentage of total radioactivity (N = 3; ± SD).

Fig. 3.

Differential impact on XPD enzyme activity. (A) Schematic view of fork substrates. The CPD is located either in the 5′–3′ translocated or the 3′–5′ displaced strand. (B) Typical autoradiographs showing the inhibition of XPD helicase by a single CPD either in the translocated (Bottom) or displaced strand (Middle) of forked substrates (5 nM). (C) Dose dependence of helicase activity. The indicated concentrations of FaXPD were incubated (15 min) with forked DNA substrate (5 nM). The CPD is located either in the translocated or the displaced strand. Oligonucleotide displacement is expressed as the percentage of total radioactivity in each reaction (N = 3; ± SD). (D) Time course experiments. FaXPD (60 nM) was incubated with forked substrates (5 nM) for the indicated time periods (N = 3; ± SD). (E) Dose-dependent stimulation of ATPase activity. The indicated concentrations of FaXPD were incubated with forked DNA (5 nM) in helicase reaction buffer containing 3 mM ATP (N = 3; ± SD). (F) Time course of P_i release upon incubation of FaXPD protein (60 nM) with forked DNA substrates (5 nM) (N = 3; ± SD).

Fig. 4.

Protection assays showing that FaXPD forms a lesion demarcation complex. (A) Position of the HaeIII, SalI, and PstI recognition sequences in 125-mer forked substrates. (B) Protection from SalI cleavage. FaXPD (60 nM) was preincubated (15 min) with partial duplexes (5 nM) and ATP (3 mM), followed by treatment with SalI. The SalI site is occluded by the XPD helicase when the substrate contains a CPD in the 5′–3′ strand (Left) but not if the CPD is situated in the 3′–5′ strand (Right). Lanes: 6–9 and 15–18, control reactions with incomplete helicase mixtures. The arrows indicate the position of the displaced strand. (C) Glycosylase protection assay with single-stranded DNA (Left) and forked substrates (Right). Helicase reaction products were probed by incubation with T4 denV and resolved on denaturing polyacrylamide gels. Lanes: 4–7 and 12–15, control incubations with incomplete helicase mixtures.