A multistep damage recognition mechanism for global genomic nucleotide excision repair

Abstract

A mammalian nucleotide excision repair (NER) factor, the XPC-HR23B complex, can specifically bind to certain DNA lesions and initiate the cell-free repair reaction. Here we describe a detailed analysis of its binding specificity using various DNA substrates, each containing a single defined lesion. A highly sensitive gel mobility shift assay revealed that XPC-HR23B specifically binds a small bubble structure with or without damaged bases, whereas dual incision takes place only when damage is present in the bubble. This is evidence that damage recognition for NER is accomplished through at least two steps; XPC-HR23B first binds to a site that has a DNA helix distortion, and then the presence of injured bases is verified prior to dual incision. Cyclobutane pyrimidine dimers (CPDs) were hardly recognized by XPC-HR23B, suggesting that additional factors may be required for CPD recognition. Although the presence of mismatched bases opposite a CPD potentiated XPC-HR23B binding, probably due to enhancement of the helix distortion, cell-free excision of such compound lesions was much more efficient than expected from the observed affinity for XPC-HR23B. This also suggests that additional factors and steps are required for the recognition of some types of lesions. A multistep mechanism of this sort may provide a molecular basis for ensuring the high level of damage discrimination that is required for global genomic NER.

Figure 1

Defined DNA substrates used in this study. Arrows indicate major incision sites in each AAF-damaged substrate, which were determined by the primer extension method.

Figure 2

Gel mobility shift analysis of XPC–HR23B binding to a single UV photolesion. (A,B) Indicated amounts of purified XPC–HR23B were incubated with 3.5 fmole each of the ³²P-labeled DNA fragment containing a 6-4PP or a CPD, or the cognate nondamaged (ND) control fragment. The binding reactions were performed in the absence (A) or presence (B) of a small amount (0.5 ng) of covalently closed circular plasmid DNA. The resulting DNA–protein complexes were fixed with glutaraldehyde and separated by PAGE. (C) A competition experiment using the labeled 6-4PP probe (3.5 fmole), which was incubated with (+) or without (−) 2 ng of XPC–HR23B and various amounts of cold competitor DNA fragments as indicated. (D) The amount of the labeled probe complexed with XPC–HR23B was calculated for each lane in C, and expressed as a percentage of the control (lane 2) without competitors. The mean values and standard errors were calculated from two or three independent experiments.(●) ND; (▴) 6-4PP; (▪) CPD.

Figure 3

Damage-independent binding of XPC–HR23B to a small bubble. (A) Gel mobility shift assay using the ³²P-labeled probe with or without the AAF adduct and a bubble. (B) A competition experiment involving the ³²P-labeled 6-4PP probe (3.5 fmole) and 2 ng of XPC–HR23B. Various amounts of nonlabeled competitor DNA fragments were included in the binding reactions as indicated. (C) Quantitative representation of the competition profiles. Complex formation in each lane of B was quantified as in Fig. 2D. The mean values and standard errors were calculated from at least two independent experiments. The competition profile by 6-4PP obtained in Fig. 2D is superimposed. (○) B0(AAF−); (●) B0(AAF+); (□) B3(AFF−); (▪) B3(AFF+); (▴) 6-4PP. (D) A competition experiment using the ³²P-labeled B5(AAF +) probe and 1 ng of XPC–HR23B.

Figure 4

DNase I footprinting analysis of XPC–HR23B binding to a 3-base bubble. (A,B) The indicated substrates were 5′-end-labeled for top strands (A) or bottom strands (B), and subjected to the DNase I footprinting assay using various amounts of XPC–HR23B as indicated. The digested DNA samples were subjected to denaturing PAGE followed by autoradiography. The Maxam-Gilbert G ladder prepared from each probe was loaded in parallel, where indicated, below the gels. The position of the bubble in each substrate is shown by an asterisk. (C) A schematic representation of the protection patterns for the B3 substrates. In each panel, strongly and weakly protected regions are shown by solid and shaded bars, respectively. The sites that became hypersensitive to DNase I upon binding are indicated by arrowheads (A,B) or arrows (C). Size of the arrows in C corresponds to the observed degree of hypersensitivity.

Figure 5

Binding of XPC–HR23B is not sufficient for the NER incision, but damage is also required. (A) The indicated substrates were internally labeled with ³²P and used for the dual incision assay involving the XP-C whole cell extract and various amounts of XPC–HR23B. The DNA samples were subjected to denaturing PAGE followed by autoradiography. (M) ³²P-labeled 25-bp ladder. (B) A map of HaeIII-cutting sites in the DNA substrates, where the size of each fragment (in base pairs) is indicated. (C) The nonlabeled closed circular DNA substrates indicated were incubated in the cell-free NER reactions including the XP-C whole cell extract, various amounts of XPC–HR23B, and aphidicolin. The purified DNA samples were subjected to gap-filling DNA synthesis with T4 DNA polymerase and radiolabeled dNTPs, digestion with HaeIII, and nondenaturing PAGE followed by autoradiography.

Figure 6

Presence of mismatched bases opposite CPD enhances the binding affinity for XPC–HR23B as well as the damage excision efficiency. (A) Gel mobility shift assay using the indicated probes and various amounts of XPC–HR23B. (B) A competition experiment involving the ³²P-labeled 6-4PP probe (3.5 fmole) and 2 ng of XPC–HR23B. Various amounts of nonlabeled competitor DNA fragments were included in the binding reactions as indicated. (C) Quantitative representation of the competition profiles. The mean values and standard errors were calculated from at least two independent experiments. The competition profile of 6-4PP is superimposed. (▪) CPD; (○) CPD(GA); (●) CPD(GG); (▴) 6-4PP. (D) The indicated, internally labeled substrates were assayed for the NER incision in the XP-C whole cell extract supplemented with various amounts of XPC–HR23B. A part of the autoradiograph showing the dual incision products is presented. (M) ³²P-labeled 25-bp ladder.

Figure 7

A model for two-step damage recognition in global genomic NER. XPC–HR23B first recognizes the site where helix distortion is induced. Other NER factors, involving TFIIH, XPA, XPG, and RPA, may be then recruited to the suspected site, where they somehow verify the presence of a lesion that is suitable to be handled by NER. If there is a lesion (left), the pre-incision complex containing fully opened DNA would be assembled, leading to dual incision by XPF–ERCC1 and XPG. If there is not a lesion (right), the process may be cancelled at a certain stage prior to the open complex assembly.