- The function of a protein is defined by its three-dimensional structure, and in understanding the three-dimensional structure of a protein, we gain an understanding of its function and mechanism. Protein structures, especially at high resolution, can provide detailed insights into many elements of enzyme function and catalysis – identifying residues directly involved in binding ligand or in carrying out catalysis, illuminating factors involved in promoting catalysis, and revealing subtleties which allow functionally and/or structurally similar enzymes to carry out distinct chemistries.
This dissertation presents work aimed at the functional and structural characterization of two types of proteins: sedoheptulose 7-phosphate cyclases (SH7PCs) and flavoenzymes. Six chapters of original work are presented in this dissertation and include one review and five primary research reports. The first three chapters (Chapters 2-4) focus on SH7PCs, and the second three chapters (Chapters 5-7) focus on flavoenzymes. All but the last of these chapters are published; the final original research chapter on lactate monooxygenase is at the stage of a manuscript in preparation for submission. These central chapters are bookended by a brief introduction to SH7PCs and flavoenzymes (Chapter 1) and concluding remarks on the main highlights and impacts of this work as well as future directions (Chapter 8).
With regard to the work on SH7PCs, I report the first structures of any SH7PC (Chapters 3 and 4). The crystal structures show these enzymes are structurally similar to each other and to other, closely related sugar phosphate cyclases, in particular dehydroquinate synthase (DHQS). These structures reveal subtle but informative differences between the three members of the SH7PC family – 2-epi-5-epi-valiolone synthase (EEVS), desmethyl-4-deoxygadusol synthase (DDGS), and 2-epi-valiolone synthase (EVS) – and suggest how these enzymes all utilize the same substrate to specifically generate one of three distinct products. The review (Chapter 2) provides an overview of the state of our understanding of SH7PCs, including the structural work reported in this thesis, and gives a broader context for the roles, evolution, and prevalence of these enzymes in nature. Based on structural insights, I propose a novel anomer selection hypothesis to differentiate two of these enzymes, EEVS and DDGS, from the third, EVS, based on which anomer of the substrate, sedoheptulose 7-phosphate, the enzyme selectively binds (Chapter 3). Furthermore, these structures allowed for corrections and additions to the “fingerprint” used to identify and differentiate SH7PCs and DHQS, which further guided genome mining (Chapter 4).
With regard to the work on flavoenzymes, enzymes utilizing either FAD or FMN to carry out their unique chemistries, I report studies that have advanced our understanding of enzyme families represented by three different proteins: glycerol 3-phosphate oxidase from Mycoplasma pneumoniae (MpGlpO), ferredoxin-NADP+ reductase from corn root (FNR), and lactate monooxygenase from Mycobacterium smegmatis (LMO).
In facilitating glycerol metabolism and producing hydrogen peroxide as a byproduct, glycerol 3-phosphate oxidase (GlpO), which converts glycerol 3-phosphate to dihydroxyacetone phosphate, is implicated as a pathogenicity factor and a potential drug target in pathogens such as Mycoplasma pneuomniae. I report the structure of MpGlpO (Chapter 5). Sequence and structural comparisons of MpGlpO to other structurally known glycerol 3-phosphate oxidases and dehydrogenases (GlpO/DHs) reveal there are two distinct types of GlpO/DHs: Type I GlpO/DHs, including mitochondrial GlpDHs and Streptococcus GlpO, and Type II GlpO/DHs, of which MpGlpO is representative. Guided by a liganded structure of a close homolog for which the structure was solved as part of the Protein Structure Initiative, I proposed the first plausible binding mode of the glycerol 3-phosphate substrate and a detailed catalytic mechanism that arguably will apply to all GlpO/DHs, despite the distinct differences in the composition of their active sites.
The FAD-dependent enzyme FNR catalyzes the transfer of electrons from photoreduced ferredoxin to NADP+ during photosynthesis and serves as a model for a broad superfamily of enzymes including NO synthase, cytochrome P450 reductase, and NADPH oxidases. Using a variant replacing a conserved aromatic amino acid to capture the productive binding mode of the nicotinamide portion of NADP(H), we analyzed a suite of high resolution structures (~1.5 Å) in complex with nicotinamide, NADP+, and NADPH. A reinterpretation of previous kinetic data also supports the relevance of these complexes to catalysis. Based on these high resolution structures, we further report insights into factors promoting hydride transfer in FNR and other FNR-like superfamily members. Specifically, we infer that higher anisotropic mobility of the C4 atom of NADP+ compared to NADPH, distortion of FAD geometry from planarity, and a tightly packed active site implicate significant active site compression as a factor promoting hydride transfer. A broadly relevant conclusion of this work is the recognition of active site compression as an important and general – although often overlooked and underappreciated – factor that can promote catalysis.
LMO is an FMN-dependent enzyme that catalyzes the conversion of lactate to acetate, carbon dioxide, and water. LMO is part of a family of α-hydroxy acid oxidases, all of which carry out the same oxidation chemistry, but unlike LMO, proceed through an uncoupled pathway. I report the first structure of an LMO (Chapter 7). This structure reveals details of the LMO active site and provides new insights as to how LMO kinetically and functionally deviates from the other family members by proceeding along a coupled reaction pathway. A highly mobile, variable loop (“loop 4”) known to seal the active site in these α-hydroxy acid oxidases is significantly larger in LMO than in the other enzymes in the family, and has both more compact folding and greater buried surface area. We suggest it is the dynamics of this loop that governs the kinetics of intermediate release (or lack thereof) in these enzymes.