Complex I (NADH:ubiquinone oxidoreductase) couples electron transfer to proton translocation across the inner mitochondrial membrane, to drive the synthesis of ATP. Its distinctive L-shaped structure comprises 45 subunits, encoded by both the mitochondrial and nuclear genomes, which are assembled by a complicated modular pathway. Complex I genetic defects are the most common cause of mitochondrial disorders and often present in early childhood, with high mortality rates. Recent high-resolution electron cryo-microscopy structures of mammalian complex I provide a foundation for both interpreting biochemical and biomedical data and understanding the catalytic mechanism. First, this thesis explores how the flavin cofactor is inserted into the NADH-binding (N-) domain of complex I. Genetic manipulation of cultured human cells, to starve them of flavin, revealed a hierarchal impact on the mitochondrial flavoproteome. High riboflavin content in the growth media ameliorated observed phenotypes, requiring cell conditioning in low riboflavin conditions. CRISPR knockout of the putative mitochondrial flavin transporter SLC25A32 demonstrated the severe impact of decreased flavin on complexes I and II, and mass spectrometry 'complexome' analyses suggest that the N-domain is still assembled onto complex I in the absence of the flavin. Second, the model organism Yarrowia lipolytica was used to assess the importance of residues in the quinone-binding site of complex I. Three residues with proposed roles in binding the quinone head-group were targeted. One variant was catalytically inactive, while two retained some activity. They showed decreased ability to reduce physiologically-relevant, long chain quinones, but their ability to reduce short-chain analogues was affected less severely. The results suggest a complicated picture in which interactions between the protein and both the hydrophilic quinone head-group and hydrophobic isoprenoid chain contribute to quinone-binding affinity and catalysis. Finally, a model for human complex I, generated from a recent high-resolution structure of mouse complex I, was used to investigate whether the pathogenicity of human variants could be predicted. Structural information on variant residues, including their secondary structure, proximity to key features and surface exposure, was collated and the power of each property to predict pathogenicity investigated. The analysis was then extended to the whole structure, to identify potential pathogenic hotspots in the enzyme, inform future studies of functionally important regions in complex I, and aid the diagnosis of clinically relevant pathogenic variants.

The mitochondrial folate transport protein (MFT) functions to transport folates into the mitochondrial matrix. The MFT is a member of a mitochondrial carrier family (MCF) of proteins that have a high degree of sequence and structural similarities, yet they transport vastly different substrates at high specificities. In this dissertation research, the folate-specific transport mechanism of the MFT was explored using experimental and computational techniques. MFT residues that differed from MCF consensus residues in conserved PxD/ExxK/R motifs and at a predicted substrate-binding site common to all MCF proteins were investigated. Site-directed mutagenesis of these anomalous residues in the MFT revealed that these residues were adapted for optimal folate transport, and that the MCF consensus residues at these positions were incompatible with folate transport. The structure of the MFT was predicted by homology modeling using the solved crystallographic structure of the ADP/ATP carrier as a template and this model was subjected to ~75 ns of molecular dynamics simulations. These simulations predicted a stepwise descent for the folate substrate into the MFT transport cavity and implicated several aromatic and basic residues in folate recognition and orientation. A predicted set of interactions at the base of the transport cavity between the MCF PxD/ExxK/R conserved motif residues did not appear static as previously hypothesized; these interactions appeared to be induced in the presence of the folate substrate. Therefore, we believe it is unlikely that these interactions form a barrier at the base of the transport cavity. We also investigated the role of the MFT in the compartmentalization of folate metabolism. Cell lines were created that could be induced with doxycycline to express either the cytosolic or mitochondrial isoform of the enzyme folylpoly-γ-glutamate synthetase (FPGS). The constructed cell lines were used to study the flux of folylpolyglutamates across the mitochondrial membrane. It appeared that cellular folylpolyglutamates are not transported across the mitochondrial membrane in either direction. We also demonstrated that many antifolates, including methotrexate and pemetrexed, impaired mitochondrial folate uptake. We believe that these folate analogs competitively inhibit the MFT and have purified the MFT protein for future analysis in reconstituted transport systems.