Human XPA is an essential component in the multienzyme nucleotide excision repair (NER) pathway. The solution structure of the minimal DNA binding domain of XPA (XPA-MBD: M98-F219) was recently determined [Buchko et al. (1998) Nucleic Acids Res. 26, 2779-2788, Ikegami et al. (1998) Nat. Struct. Biol. 5, 701-706] and shown to consist of a compact zinc-binding core and a loop-rich C-terminal subdomain connected by a linker sequence. Here, the solution structure of XPA-MBD was further refined using an entirely new class of restraints based on pseudocontact shifts measured in cobalt-substituted XPA-MBD. Using this structure, the surface of XPA-MBD which interacts with DNA and a fragment of the largest subunit of replication protein A (RPA70 Delta C327: M1-Y326) was determined using chemical shift mapping. DNA binding in XPA-MBD was highly localized in the loop-rich subdomain for DNA with or without a lesion [dihydrothymidine (dhT) or 6-4-thymidine-cytidine (64TC)], or with DNA in single- or double-stranded form, indicating that the character of the lesion itself is not the driving force for XPA binding DNA. RPA70 Delta C327 was found to contact regions in both the zinc-binding and loop-rich subdomains. Some overlap of the DNA and RPA70 Delta C327 binding regions was observed in the loop-rich subdomain, indicating a possible cooperative DNA-binding mode between XPA and RPA70 Delta C327. To complement the chemical shift mapping data, the backbone dynamics of free XPA-MBD and XPA-MBD bound to DNA oligomers containing dhT or 64TC lesions were investigated using 15N NMR relaxation data. The dynamic analyses for the XPA-MBD complexes with DNA revealed localized increases and decreases in S2 and an increase in the global correlation time. Regions of XPA-MBD with the largest increases in S2 overlapped regions having the largest chemical shifts changes upon binding DNA, indicating that the loop-rich subdomain becomes more rigid upon binding DNA. Interestingly, S2 decreased for some residues in the zinc-binding core upon DNA association, indicating a possible concerted structural rearrangement on binding DNA.

The isolated cytochrome (cyt) b(6)f complex from spinach is inhibited by Cu(2+) with a K(D) of about 1 microM at pH 7.6 in the presence of 1.6 microM decyl-plastoquinol (C(10)-PQH(2)) as a substrate. Inhibition was competitive with respect to C(10)-PQH(2) but noncompetitive with respect to horse heart cyt c or plastocyanin (PC). Inhibition was also pH-sensitive, with an apparent pK at about 7, above which inhibition was stronger, suggesting that binding occurred at or near a protonatable amino acid residue. Equilibrium binding titrations revealed ca. 1.4 tight Cu(2+) binding sites with a K(D) of about 0.5 microM and multiple (>8) weak (K(D) > 50 microM) binding sites per complex. Pulsed electron paramagnetic resonance (EPR) techniques were used to identify probable binding sites for inhibitory Cu(2+). A distinct enhancement of the relaxation time constant for the EPR signal from bound Cu(2+) was observed when the cyt f was paramagnetic. The magnitude and temperature-dependence of this relaxation enhancement were consistent with a dipole interaction between Cu(2+) and the cyt f (Fe(3+)) heme at a distance of between 30 and 54 A, depending upon the relative orientations of Cu(2+) and cyt f heme g-tensors. Two-pulse electron spin-echo envelope modulation (ESEEM) and 4-pulse 2-dimensional hyperfine sublevel correlation (2D HYSCORE) measurements of Cu(2+) bound to isolated cyt b(6)f complex indicated the presence of a weakly coupled nitrogen nucleus. The nuclear quadrupole interaction (NQI) and the hyperfine interaction (HFI) parameters identified one Cu(2+) ligand as an imidazole nitrogen of a His residue, and electron-nuclear double resonance (ENDOR) confirmed the presence of a directly coordinated nitrogen. A model of the 3-dimensional structure of the cytochrome b(6)f complex was constructed on the basis of sequences and structural similarities with the mitochondrial cyt bc(1) complex, for which X-ray structures have been solved. This model indicated three possible His residues as ligands to inhibitory Cu(2+). Two of these are located on the "Rieske" iron-sulfur protein protein (ISP) while the third is found on the cyt f protein. None of these potential ligands appear to interact directly with the quinol oxidase (Q(o)) binding pocket. A model is thus proposed wherein Cu(2+) interferes with the interaction of the ISP protein with the Q(o) site, preventing the binding and subsequent oxidation of plastoquinonol. Implications for the involvement of ISP "domain movement" in Q(o) site catalysis are discussed.