DNA quality control by a lesion sensor pocket of the xeroderma pigmentosum group D helicase subunit of TFIIH

Background: Nucleotide excision repair is a versatile DNA repair reaction that removes bulky adducts generated by environmental mutagens such as the UV spectrum of sunlight or chemical carcinogens. Current multistep models of this excision repair pathway accommodate its broad substrate repertoire but fail to explain the stringent selectivity toward damaged nucleotides among excess native DNA. To understand the mechanism of bulky lesion recognition, we postulated that it is necessary to analyze the function of xeroderma pigmentosum group D (XPD) protein beyond its well-known role in the unwinding of double-stranded DNA. Results: We engineered two new XPD mutants (Y192A and R196E), involving amino acid substitutions near its central protein pore, that confer defective DNA repair despite normal transcription. In situ fluorescence-based protein dynamics studies in living cells demonstrated that both new mutants were unable to recognize DNA damage and failed to form stable associations with lesion sites. However, when their biochemical properties were tested in the framework of an archaeal protein homolog, they both retained ATPase and DNA-unwinding activity. The outstanding difference versus the wild-type control was that their directional 5?-3? translocation along DNA was not stopped by a bulky lesion, and moreover, they were unable to build long-lived demarcation complexes at damaged sites. Conclusions: By uncoupling for the first time the unwinding and damage sensor activities of XPD, we describe an unprecedented genome quality control process whereby a recognition pocket near the central DNA helicase pore scans individual substrate strands to capture base adducts