A model for DNA binding by the human XPF and ERCC1 complexes

Abstract

Cellular DNA is the carrier of genetic information. Therefore, maintaining genome stability is essential for the existence of all living organisms. Integrity of our genome is constantly threatened by a large variety of endogenous and exogenous agents, which generate a plethora of DNA lesions. For example, there can be deamination of cytosine to uracil, formation of 8-oxoguanine, single and double strand DNA breaks, aberrant methylation of guanines, mismatch base insertion and interstrand crosslink. These alterations modify the structure and the function of genomic DNA leading to cellular death and long term genomic instability and aging of multi-cellular organism. In order to survive genomic assaults, cells are equipped with several DNA repair mechanisms. One of the most versatile and conserved DNA repair mechanisms is nucleotide excision repair (NER). Nucleotide excision repair eliminates helixdistorting DNA lesions, which mainly arise due to UV light. The molecular mechanism of NER is complex and it involves multiple proteins with distinct functions and interacting partners. One of the vital NER factors is the XPF/ERCC1 complex, which makes an incision at the 5 side of the UV damaged DNA. The formation of an obligate XPF/ERCC1 heterodimer is indispensable for the in vivo stability and function of both XPF and ERCC1 proteins. The structural data on XPF/ERCC1 complexes is vital to understand the in vivo and in vitro function of the heterodimeric complex. In this thesis I provide detailed structural insight on the human XPF and ERCC1 complexes. Using mass spectrometry, CD and NMR spectroscopy, I have studied the protein-protein and protein-DNA complexes of the C-terminal double helix-hairpin-helix (HhH) domains of the human XPF and ERCC1 proteins. Chapter 1 of this thesis provides a concise introduction on DNA repair mechanisms and NMR methods to study proteins, nucleic acids and protein-nucleic acid complexes. In chapter 2, we have compared the in vitro stabilities of the homodimeric and heterodimeric complexes of the C-terminal helix-hairpin-helix (HhH) domains of XPF and ERCC1. We find that XPF homodimers were significantly more stable than XPF/ERCC1 heterodimers. The observed differences in the stability can be explained by the presence of a larger protein-protein interface, more hydrogen bonds and additional aromatic stacking interactions found in the structure of XPF homodimers as compared to the structure of XPF/ERCC1 heterodimers. In chapter 3 we have determined the structure of the complex of the XPF HhH domains bound to 10nt single strand DNA. In agreement with the findings of Tsodikov et al. that suggested a role for the XPF HhH domain in single strand DNA binding we find that the positively charged surface of the XPF interacts with the phosphate backbone of the single strand DNA. Furthermore, we noted that a guanine base flips out and binds in a cavity formed by residues from the conserved first HhH motif and non-canonical second HhH motif. Part of this cavity is already preformed, but it widens since some protein side chains reorient upon complex formation. The cavity in XPF, that is not present in double HhH motifs of ERCC1, is the result of the altered conformation of the second HhH motif due to the deletion of one residue in the second hairpin of human XPF. Based on our findings we propose a model to explain the binding of the XPF/ERCC1 heterodimer at the ss/ds DNA junction. In chapter 4 we have made use of the high affinity DNA binding by the double HhH domains of XPF homodimers to understand the substrate specificity of XPF/ERCC1 heterodimers. First, we demonstrate in biochemical assays that the HhH domain of the homodimeric XPF can bind cooperatively to ssDNA and to structure specific DNA substrates like fork, bubble and Holliday junctions. Next we show by NMR spectroscopy that in XPF/ERCC1 heterodimers the XPF HhH domains can bind single strand DNA using the same charged surface as found for XPF homodimers. Also the XPF that was mutated at the residues identified by NMR showed reduced DNA binding activity. The present DNA binding study with the wild type and mutant XPF HhH domains complement the structural findings of Chapter 3. In conclusion, now we have a model to understand how XPF/ERCC1 binds at the ss/ds DNA junction.