Ming, Xun2011-07-132011-07-132011-05https://hdl.handle.net/11299/109187University of Minnesota M.S. thesis. May 2011. Major: Medicinal Chemistry. Advisor:Dr. Natalia Y. Tretyakova. 1 computer file (PDF); xix, 200 pages.1,1,2,2- Cis-diamminedichloroplatinum (II) (cisplatin) is a common anticancer drug used in treatment of solid tumors. The biological activity of cisplatin is generally attributed to its ability to form DNA-DNA cross-links by sequential platination of two nucleophilic sites within the DNA duplex. However, cisplatin also forms a variety of DNA-protein cross-links (DPCs) whose structures are not well characterized. While the biological outcomes of cisplatin-induced DPC lesions are not understood, they are hypothesized to interrupt important cellular processes such as DNA replication and transcription, potentially leading to toxicity. In the present work, a human DNA repair protein O6-alkylguanine DNA alkyltransferase (AGT) was used as a model to investigate cisplatin-induced DNA-protein cross-linking. The normal physiological function of AGT is to repair alkylation damage by transferring O6-alkylguanine groups from DNA to an active site cysteine residue of the protein (Cys145), restoring normal guanine. Incubation of recombinant AGT protein with 32P-labeled oligonucleotides duplexes in the presence of cisplatin resulted in concentration-dependent formation of DNA-protein conjugates as revealed by denaturing gel electrophoresis. Capillary HPLC-electrospray ionization mass spectrometry analysis (ESI-MS) of AGT protein treated with dG-Pt-Cl monoadduct as a model of monoplatinated DNA confirmed the ability of the protein to form multiple dG-AGT cross-links. Upon heating, dG-Pt-AGT complexes undergo platination migration from protein to the N7 position of guanine to form dG-Pt-dG cross-links. This can be explained by greater thermodynamic stability of the Pt-N bond as compared to the Pt-S bond. HPLC-ESI+-MS/MS sequencing of tryptic peptides derived from dG-Pt-AGT complexes revealed that cisplatin-mediated cross-linking involves six different sites within this protein: Glu110, Lys125, Cys145, His146, Arg147, and Cys150. Among these, Cys145, His146, Arg147, and Cys150 are located in the protein active site and directly or indirectly participate in alkyl transfer. Finally, HPLC-ESI+ - MS/MS analysis of total proteolytic digests detected 1,1-cis-diammine-2-(5-amino-5-carboxypentyl)amino-2-(2’-deoxyguanosine-7-yl)-platinum (II) (dG-Pt-Lys) conjugates produced via platination of lysine residues within AGT. To identify protein targets of cisplatin-induced cross-linking in nuclear protein extracts from human cervical carcinoma (HeLa) cells, an affinity capture methodology was combined with mass spectrometry-based proteomics and immunological detection. A total of 131 nuclear proteins were identified to form covalent DPCs in the presence of cisplatin. An estimated DNA-protein cross-linking efficiency following treatment with 50 μM cisplatin was 2-16%, depending on protein identity. HPLC-ESI+-MS/MS analysis of total proteolytic digests of cross-linked proteins revealed the presence of dG-Pt-Lys conjugates. We further extended this work to characterize DNA-protein cross-linking by cisplatin in human fibrosarcoma (HT1080) cells. Following drug treatment, DPCs were isolated by a modified phenol/chloroform DNA extraction incorporating proteasome inhibitors. Proteins were released from DNA by heating and identified by mass spectrometry-based proteomics and immunological detection. Over 250 nuclear proteins were found to be captured on chromosomal DNA following treatment with cisplatin. HPLC-ESI+-MS/MS analysis of total proteolytic digests revealed the formation of dG-Pt-Lys conjugates between the N7 guanine of DNA and the ε-amino group of lysine. Although cisplatin-induced DPCs spontaneously release proteins to form DNA-DNA cross-links upon heating, they appear to be stable enough under physiological conditions to inhibit DNA replication and transcription, contributing to the biological effects of cisplatin. These results indicate that clinically relevant concentrations of cisplatin induce covalent cross-links between chromosomal DNA and a large range of nuclear proteins. If not repaired, the resulting bulky DPC lesions are likely to contribute to both on-target and off-target toxicity of cisplatin. Reactive oxygen species produced as part of normal cellular metabolism and immune response can damage cellular DNA, giving rise to promutagenic nucleobase lesions. Since the biological impact of a given oxidative adduct is influenced by its position within gene sequence, previous studies have focused on determining the distribution of oxidative lesions along DNA sequences. However, since these studies have relied on gel electrophoresis to locate the sites of oxidative damage, they could not analyze the distribution of structurally defined nucleobase lesions and suffered from the high background of direct strand breaks induced by sugar oxidation. We now report the use of stable isotope labeling of DNA-mass spectrometry (ILD-MS) approach to map the formation of 8-oxo-7,8-dihydro-2’-deoxyguanosine (8-oxo-dG) and 2,2-diamino-4-[2-deoxy-β-D-erythro-pentofuranosyl)amino]-2,5-dihydrooxazol-5-one (oxazolone) lesions along DNA sequences derived from the p53 tumor suppressor gene. In each duplex, one of the guanine bases was labeled with [1,7,NH2-15N3-2-13C]-guanine which served as an isotope “tag” to enable specific quantification of guanine lesions originating from that position. Following photooxidation in the presence of riboflavin, DNA was enzymatically digested to 2’-deoxynucleosides, and the formation of 8-oxo-dG and oxazolone at each site of interest was quantified from isotope ratios obtained from capillary HPLC-ESI+-MS/MS. We found that in double stranded DNA, both oxidative lesions were generated non-randomly, and their distribution was strongly influenced by the local DNA sequence. In particular, the 5’ Gs in guanine repeats and guanines within MeCG dinucleotides were preferentially targeted for photooxidation in the presence of riboflavin. This can be explained by the low ionization potential of 5’-guanine at 5’-GG sites and/or the preferential intercalation of riboflavin at MeCG sites. Furthermore, the most frequently adducted position, G5 in exon 5, G4 and G7 in exon7, and G6 in exon 8, coincide withthe known p53 lung cancer mutational “hotspots” at p53 codons 158 (CGC), 245 (GGC), 248 (CGG), and 273 (CGT), respectively, suggesting that oxidative DNA damage may contribute to mutagenesis in the p53 gene.en-USMedicinal chemistryThesis or Dissertation