The relationships between XPC binding to conformationally diverse DNA adducts and their excision by the human NER system: is there a correlation?

Abstract

The first eukaryotic NER factor that recognizes NER substrates is the heterodimeric XPC-RAD23B protein. The currently accepted hypothesis is that this protein recognizes the distortions/destabilization caused by DNA lesions rather than the lesions themselves. The resulting XPC-RAD23B-DNA complexes serve as scaffolds for the recruitment of subsequent NER factors that lead to the excision of the oligonucleotide sequences containing the lesions. Based on several well-known examples of DNA lesions like the UV radiation-induced CPD and 6-4 photodimers, as well as cisplatin-derived intrastrand cross-linked lesions, it is generally believed that the differences in excision activities in human cell extracts is correlated with the binding affinities of XPC-RAD23B to these DNA lesions. However, using electrophoretic mobility shift assays, we have found that XPC-RAD23B binding affinities of certain bulky lesions derived from metabolically activated polycyclic aromatic hydrocarbon compounds such as benzo[a]pyrene and dibenzo[a,l]pyrene, are not directly, or necessarily correlated with NER excision activities observed in cell-free extracts. These findings point to features of XPC-RAD23B-bulky DNA adduct complexes that may involve the formation of NER-productive or unproductive forms of binding that depend on the structural and stereochemical properties of the DNA adducts studied. The pronounced differences in NER cleavage efficiencies observed in cell-free extracts may be due to differences in the successful recruitment of subsequent NER factors by the XPC-RAD23B-DNA adduct complexes, and/or in the verification step. These phenomena appear to depend on the structural and conformational properties of the class of bulky DNA adducts studied.

Keywords: Benzo[a]pyrene; DNA adduct; Dibenzo[a,l]pyrene; Electrophoretic mobility shift assay; Nucleotide excision repair (NER); XPC-RAD23B binding.

Figure 1

Chemical structures of DNA adducts investigated.

Figure 2

Structures of lesion-containing duplexes. NMR solution structures: S-trans-B[a]P-G [31], R-cis- B[a]P-G [30], R-trans-DB[a,l]P-G [35], S-trans-DB[a,l]P-G [33, 34]. Modeled structures: R-trans-DB[a,l]P-A, and S-trans-DB[a,l]P-A [36].

Figure 3

Relative dual incision efficiencies of 135-mer duplexes containing different DNA adducts determined in cell-free HeLa cell extracts. The data for the B[a]P-G adducts are from Hess et al. [22] and Mocquet et al [23], and for the DB[a,l]P-G/A data are from Kropachev et al. [21].

Figure 4

Typical EMSA experiments demonstrating the binding of XPC-RAD23B to 50-mer (A) unmodified duplexes or (B) with single R-cis-B[a]P-G or (C) S-trans-B[a]P-G adducts in native 4.5% polyacrylamide gels. (D) Quantitative analysis of fractions of DNA molecules in complexes with XPC-RAD23B. The results for the cis- and trans-adducts were indistinguishable from one another within experimental error, and thus only the averages are shown (black squares). Each point and associated error bar were derived from four data points in each case (including the results for the unmodified duplexes, open circles). Solid and dashed lines are the best fits to the Hill equation based on the Hill independent bimodal binding of XPC-RAD23B to DNA (Eq. 1) with K_D = (K_D1K_D2)^1/2 = 1.2 ± 0.6 nM (unmodified duplex), 0.67±0.20 nM R-cis- and S-trans-B[a]P-G duplexes

Figure 5

Typical EMSA experiments demonstrating the binding of XPC-RAD23B to 50-mer (A) unmodified duplexes or (B) with single S-trans-DB[a,l]P-G or (C) R-trans-DB[a,l]P-G adducts in native 4.5% polyacrylamide gels. (D) Quantitative analysis of fractions of DNA molecules in complexes with XPC-RAD23B as described in the case of Figure 4, but with K_D = (K_D1K_D2)^1/2 = 0.67±0.20 nM. The results obtained with the two stereoisomeric adducts were identical to one another within experimental error.

Figure 6

Typical EMSA experiments demonstrating the binding of XPC-RAD23B to 50-mer (A) unmodified duplexes or (B) with single S-trans-DB[a,l]P-A or (C) R-trans-DB[a,l]P-A adducts in native 4.5% polyacrylamide gels. (D) Quantitative analysis of fractions of DNA molecules in complexes with XPC-RAD23B as described in the case of Figure 4, with K_D = (K_D1K_D2)^1/2 = 1.2±0.75 nM (DB[a,l]P-A) nM. The results obtained with the two stereoisomeric adducts were identical to one another within experimental error.

Figure 7

Competition binding experiments. (A) – (H): EMSA experiments as in Figure 4, except that the unmodified and unlabeled 50-mer duplexes were first mixed with XPC-RAD23B at the concentrations indicated, allowed to equilibrate for 10 min; then the ³²P-endlabeled unmodified duplexes (panels A and D), or the same duplexes bearing single stereoisomeric S-trans- or R-trans-DB[a,l]P-G adducts (B and C, respectively), or S-trans- or R-trans-DB[a,l]P-A (panels D and E, respectively) were added at the indicated concentrations.

Figure 8

(A) Example of effects of a large excess of unlabeled and unmodified 50-mer duplexes (500 nM) on the binding of XPC-RAD23B to 32P-labeled 1 nM unmodified 50-mer duplexes (open squares), or identical duplexes but with single R-cis-B[a]P-G (black circles) or S-trans-B[a]P-G (black squares). (B) Effects of NaCl (225 mM) on the binding of XPC-RAD23B to ³²P-labeled unmodified 50-mer DNA duplexes (open squares), or to ³²P-labeled modified 50-mer DNA duplexes with R-cis-B[a]P-G (open circles) or with S-trans-B[a]P-G duplexes (black squares).

Figure 9

Structural models to illustrate the productive binding of the yeast Rad4/Rad23 apo enzyme to duplex DNA containing the R-trans-DB[a,l]P-G lesion intercalated from the minor groove; the latter is a very good NER substrate [21]. (A) NMR solution structure of the R-trans-DB[a,l]P-G lesion (PDB [53] ID: 2LZK, [35]) with the apo Rad4/Rad23 heterodimer before binding (PDB ID: 2QSF, [18]) positioned to initiate the insertion of the BHD3 β–hairpin from the major groove (red arrow), with extrusion of the lesion and its 3’ adjacent dC through the minor groove (pink arrow). The model was created by superposing, respectively, the TGD domain of the apo enzyme and the NMR structure of the R-trans-DB[a,l]P-G lesion-containing DNA, to the TGD domain and the DNA backbone of 6 base pairs around the lesion site in the complex crystal structure (PDB ID: 2QSG, [18]). (B) the BHD3 hairpin loop has inserted into the duplex from the major groove side and extruded the R-trans-DB[a,l]P-G stacked with its 3’ adjacent dC (purple in A and B) through the minor groove. The Rad4/Rad23 complex containing a cis-syn thymine dimer (PDB ID: 2QSG, [18]) was remodeled to replace the thymine dimer (whose coordinates were missing) with the R-trans-DB[a,l]P-G-C dinucleotide from the NMR structure (PDB ID:2LZK, [35]) followed by 10 ns of molecular dynamics simulation [11] The missing BHD3 hairpin loop in (A) was modeled based on its structure in (B). This model suggests how productive binding may take place for a well-repaired PAH-derived NER substrate.