A number of clinically important antitumor agents such as cisplatin, cyclophosphamide (a nitrogen mustard) or carmustine (BCNU, a chloro ethyl nitroso urea) form DNA interstrand crosslinks (ICLs) as key cytotoxic lesions. ICLs covalently link two strands of a DNA duplex and therefore provide a potent block to DNA replication and transcription. Despite the enormous success of ICL-forming agents in treating a large variety of tumors, the occurrence of resistance caused by the repair of ICLs (and other mechanisms) and the occurrence of secondary tumors remain significant problems. Studies aimed at understanding the biological responses triggered by ICLs formed by antitumor agents have been hampered by the limited availability of site-specific ICLs for biochemical and cell biological studies. We have developed new methodology for the synthesis of site-specific ICLs formed by nitrogen mustards and chloro ethyl nitroso ureas to overcome this limitation. This will enable us to synthesize structurally diverse ICLs and incorporate them into longer oligonucleotides and plasmids for the study of ICL repair. In collaboration with the laboratory of Johannes Walter (Harvard Medical School) these substrates were used to establish the first defined biochemical system for the study of replication- dependent ICL repair, revealing incisions around the ICL and translesion synthesis past an unhooked ICL as key steps. Along with preliminary studies exploring the reactions of translesion synthesis polymerases with ICL templates, these studies provide the foundation for the proposed studies of structure-function relationships in ICL repair. The guiding hypothesis of these studies is that differences in ICL structure will affect the translesion synthesis and nucleotide excision repair steps in ICL repair in particular, and that these differences have important implication for therapeutic outcomes in antitumor chemotherapy. In Aim 1 we propose to further our efforts to synthesize ICLs that link the DNA through the major groove or base-pairing surfaces, generating ICLs that induce severe, intermediate, mild or no distortion in the DNA double helix. We will furthermore synthesize ICLs in structures that represent intermediates in ICL repair to study how they are processed by DNA polymerases. In Aim 2, we will characterize the structures of these ICLs by NMR spectroscopy and molecular dynamics simulations to gain detailed insights into how the various ICLs affect DNA structure. In Aim 3, we will investigate how these structurally diverse ICLs are processed in replication-dependent ICL repair and how the structures of the ICLs influence how they are processed by translesion synthesis polymerases. We expect that these studies will reveal commonalities and also important differences of how structurally diverse ICLs are processed in human cells. Our studies should provide important insights into the mechanisms that underlie resistance of tumors to crosslinking agents used in cancer chemotherapy as well as the formation of secondary tumors. Since our studies involve ICLs formed by antitumor agents as well as ones with novel structures, they could lead to the development of antitumor agents with improved properties.