Elucidating the Role of [4Fe4S] Clusters in DNA Replication and Repair Proteins

Citation

Bartels, Phillip Leon (2018) Elucidating the Role of [4Fe4S] Clusters in DNA Replication and Repair Proteins. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z9H1307G. https://resolver.caltech.edu/CaltechTHESIS:03012018-094939210

Abstract

[4Fe4S] clusters, redox cofactors, have been discovered in DNA processing enzymes ranging from bacterial base excision repair glycosylases to eukaryotic DNA polymerases. Bacterial repair proteins are activated toward redox activity when bound to DNA and can take advantage of DNA-mediated charge transport (DNA CT) to search the genome for lesions. DNA CT involves the rapid transport of charges through the π-stacked base pairs and is sharply attenuated in the presence of lesions, mismatches, or other stacking perturbations. Thus, [4Fe4S] repair proteins use this chemistry to rapidly redistribute to target lesions and communicate with one another over long distances.

The general function of [4Fe4S] clusters in bacterial DNA repair has received much attention, but previous efforts have left several critical questions unanswered. First, while the redox potential of these proteins is affected by DNA binding, the relative importance of the negatively-charged DNA, the protein environment surrounding the cluster, and solvent has remained unclear. Second, the importance of [4Fe4S] clusters and DNA CT to human disease has never been directly addressed. The biological consequences of this chemistry are certainly a pressing issue, as numerous disease-relevant mutations in the human homologues of well-studied repair proteins have been recorded. Finally, the existence of [4Fe4S] clusters in eukaryotic DNA replication proteins in general, and in the B-family DNA polymerases in particular, was entirely unexpected. The function of the [4Fe4S] cluster in replication proteins was far from obvious, and the functional differences from repair proteins made them difficult to explain even in the context of CT signaling. Herein, these questions have been addressed using a combination of electrochemical, spectroscopic, and biochemical approaches.

First, we describe the use of pyrolytic graphite edge electrodes (PGE) and S K-edge X-ray absorption spectroscopy (XAS) to address the influence of protein environment, DNA, and solvation on the [4Fe4S] cluster redox potential in the bacterial base excision repair glycosylases endonuclease III (EndoIII) and MutY. The PGE surface is rough and favorable for protein binding; electron transfer can be further enhanced in the presence of carbon nanotubes. Electrochemical signals for EndoIII and MutY in the absence of DNA are large and reproducible, and a potential shift upon DNA binding is observed. With respect to studying proteins in the absence of DNA, the PGE electrode represents a significant advance over previously used highly-oriented pyrolytic graphite (HOPG), which is hydrophobic and difficult to prepare. To test the effect of protein environment on redox potential, a series of EndoIII point mutants were prepared in which the charge within 5 Å of the cluster was reversed or added in. None of these mutations induced a significant shift in redox potential relative to wild type, arguing that DNA electrostatics are the dominant factor in potential modulation. In parallel, XAS studies were performed on EndoIII and MutY in the presence and absence of DNA, and in the presence and absence of solvent. Ligating cysteinyl thiols and inorganic S atoms in the [4Fe4S] cluster absorb at different intensities in XAS depending on solvent environment and local electrostatics; these changes, in turn, directly correlate to redox potential. By XAS, DNA was found to induce a significant shift in absorbance, and thus potential; the removal of solvent had a smaller effect. Together, these studies provide new approaches for the study of DNA-binding [4Fe4S] proteins and reveal the critical role of DNA in tuning the redox potential.

Second, we report on a novel mutation in human MUTYH identified from a colorectal cancer patient and confirmed to be pathological. MUTYH is responsible for repairing certain lesions induced by oxidative stress and is thus frequently implicated in cancer. This new variant, C306W, contains a mutation in one of the cysteines that ligates the [4Fe4S] cluster. Electrochemistry, activity and DNA binding assays, and spectroscopic analyses were performed for C306W alongside wild type MUTYH and two other disease-relevant mutants, Y179C and G396D, with an unaltered cluster environment. From this work, it is now clear that C306W can still bind a cluster, but it is susceptible to oxidative degradation to the [3Fe4S]⁺ state upon redox signaling in an aerobic environment. Consequently, enzymatic activity is very low, and DNA binding is poor. Overall, this represents the first complete characterization of the [4Fe4S] cluster in a human homologue of MutY, and the first demonstration of pathology resulting from a mutation that primarily affects the [4Fe4S] cluster.

Moving into DNA replication proteins, we report on the characterization of the [4Fe4S] cluster in yeast DNA polymerase (Pol) δ, the eukaryotic lagging strand polymerase. Pol δ shows reversible electrochemical signals at a midpoint potential indistinguishable from EndoIII under the same conditions, and EPR spectroscopy confirms use of the [4Fe4S]^3+/2+ couple. The electrochemical signal is attenuated on DNA containing an abasic site or a CA mismatch, confirming that Pol δ is capable of DNA-mediated signaling. Bulk electrolysis and photooxidation were used to oxidize Pol δ under anaerobic conditions, and activity assays were carried out using oxidized or untreated protein. Oxidation stalls replication activity, while electrochemical reduction of oxidized samples restores activity to untreated levels. These results thus reveal that cluster oxidation serves as a reversible switch regulating Pol δ activity, suggesting an in vivo role in responding to replication stress, especially oxidative stress. In an effort to address these possibilities, we have carried out preliminary efforts in the characterization of two potentially CT-deficient mutants, W1053A and Y1078A. Both mutants were found to be too structurally unstable to proceed with in vivo experiments, but they can serve to guide future efforts in this direction.

Finally, a strategy to examine charge transport through RPA-bound single-stranded DNA is reported. RPA is the eukaryotic single-stranded binding protein and forms a protective coat around vulnerable unwound DNA at replication forks. Given the importance of redox signaling in replication proteins, we aimed to use photooxidation experiments to determine if CT through RPA is a viable pathway; if so, this would open up a large set of long-range transfer pathways to [4Fe4S] proteins in replication. These efforts are ongoing, but the experimental strategy and initial efforts are discussed.

Thesis Files