To pursue my structural interests, I applied X-ray crystallography and small angle X-ray scattering (SAXS) as a graduate student in the laboratory of Dr. John Tainer (The Scripps Research Institute). I focused on proteins that defend against and repair DNA damage, including human catalase, uracil-DNA glycosylase (UDG), the DNA mimicking UDG inhibitor Ugi, and the RuvB DNA translocase. RuvB is part of the bacterial RuvABC Holliday junction migration and resolution that binds (RuvA), cleaves (RuvC), and migrates the crossover positions (RuvB) of four-stranded Holliday junctions. RuvB was revealed to be an AAA+-ATPase linked to a winged helix DNA-binding domain. The structure and engineered mutants revealed that composite ATPase sites at monomer-monomer interfaces whose nucleotide-binding state directly influenced the asymmetry of the hexameric ring and were the location of most of the dominant negative E. coli RuvB mutations. To determine the biologically relevant assembly, I introduced SAXS as a methodology into Dr. Tainer's laboratory and validated computational models of the RuvB hexameric ring generated by superimposing the RuvB monomer onto other hexameric AAA+-ATPases. These assemblies were used it derive models of RuvABC complexes with Holliday junctions that indicated that both RuvA and DNA were likely to introduce important asymmetries that underlie the branch migration by the RuvB molecular motor. The course of these experiments led Dr. Tainer, Dr. Ulrich Genick, and myself to specify the design and obtain initial funding for beamline 12.3.1 (SIBYLS) and the Advanced Light Source at the Lawrence Berkeley National Laboratories http://www.bl1231.als.lbl.gov.

To further explore genomic instability in vivo, I joined the laboratory of Professor Richard Kolodner (Ludwig Institute for Cancer Research and University of California San Diego School of Medicine) as a postdoctoral fellow to integrate structural biology, biochemistry, and genetics for understanding the roles of DNA damage surveillance and repair pathways in cancer initiation. I have primarily used a gross chromosomal rearrangement (GCR) assay developed in the Kolodner laboratory to understand genomic stability.

In a collaboration with Winfred Edelmann (Alfred Einstein College of Medicine), I have demonstrated that heterozygous defects in replication protein A that structurally disrupt the single-stranded DNA binding surfaces as revealed by molecular dynamics simulations causes genomic instability in the yeast Saccharomyces cerevisiae. The genetic interactions in yeast using the GCR assay revealed that a hypomorphic and partially dominant allele, rfa1-t48, both causes damage as well as prevents that damage from being properly repaired. This increased chromosomal instability is correlated with predisposition to specific invasive lymphomas with common chromosomal defects in heterozygous knock-in mice. Thus, RPA1 was identified as a novel tumor suppressor and provided validation that genes identified in the yeast GCR assay can be tumor suppressors in mammals.

I undertook the first bioinformatics analyses of rearrangements recovered in the GCR assay in order to understand the mechanisms behind their formation. The genotype of the strain directly controlled the types of rearrangements formed and likely dictated which sites were prone to breakage. In conjunction with the rfa1-t48 studies, these results suggest that early inactivation of genes involved in genome stability can have substantial impact on how progression of cancers can progress genetically and thus can dictate what types of cancer are formed.

My studies of de novo telomeres isolated in the GCR assay have provided the first evidence for the specificity of S. cerevisiae telomerase in vivo. De novo telomeres are clearly targeted by the terminal 3-4 bases of chromosomal DNA that serve as an annealing site for the telomerase guide RNA. These target sequences are similar to the much smaller number of de novo telomere additions observed in other non-ciliate organisms, including some human genetic diseases. The patterns revealed by the sequences of de novo telomeres provide evidence for how these ends serve to register the degenerate reveal the usage of the templating region of the guide and readily explained all common telomere addition sequences. The critical role for 3-4 bases of annealing resolved the apparent paradox of the S. cerevisiae possessing a non-processive telomerase that synthesizes telomeres containing full repeats of the templating region of the telomerase guide RNA.

My current studies of genomic stability using S. cerevisiae are in two areas. First, to understand the implications of duplications in genome rearrangements, I modified the GCR assay to include a natural 5 kilobase duplication. This new assay has identified pathways that suppress, mediate, or direct damage away from homology-driven rearrangements. These pathways have important differences from those in the standard assay. Moreover, the observed rearrangements identify genotype-specific types of damage that underlie the genomic rearrangements. Second, I am using a bioinformatics approach to identify new genome instability genes in S. cerevisiae and to evaluate homologs of validated targets for transcriptional or copy number dysregulation in human cancers. These studies have implicated new S. cerevisiae genes and pathways in controlling genome stability.

In addition to these studies, I have collaborated with members of Richard Kolodner's laboratory to understand the bacterial MutS homodimer and its eukaryotic homologs, the Msh2/Msh6 and Msh2/Msh3 heterodimers. I have designed an Msh6-Msh3 chimera that recognizes the insertion/deletion mutations with Msh3-like mispair recognition specificity, but still functions like Msh6 with regard to binding MutL homologs and requirements for the Msh2 mispair recognition domain. Using SAXS, I have determined that the N-terminus of yeast Msh6, including the PCNA interaction site, is unstructured and has two partially overlapping roles in recruiting the Msh2/Msh6 to newly replicated DNA with one binding partner being the proliferating cell nuclear antigen (PCNA). Computational analysis reveals that unstructured tethers are common for many PCNA binding proteins that have roles in DNA replication, DNA repair, and the cell cycle. As part of a larger project to determine the structure of the MutS/MutL/mispair ternary complex, I have determined the structure of the E. coli MutS tetramer by SAXS and the structure of the domain required both for tetramerization and stabilization of the dimer by X-ray crystallography. These have allowed the generation of specific MutS constructs that are only defective in tetramerization and are fully functional in vivo. I fully expect that these collaborations will continue in future. For example, a combination of genetic and structural modeling data have suggested a conserved, functional, and asymmetric metal-binding site in the C-termini of the eukaryotic MutL homologs that will be investigated using X-ray crystallography. I am particularly excited about the use of the tetramerization-defective MutS constructs in SAXS studies of MutS/MutL/mispair ternary complexes. In conjunction with ongoing deuterium-exchange mass spectrometry, these data could determine the low-resolution structure of the ternary complex that is one of the major problems in the mismatch-repair field.