A number of dihydroxyacetone kinases (DAK) have been described that can utilize either ATP (in animals, plants and some bacteria) or phosphoenolpyruvate (PEP) (in most bacteria) as the energy source to convert dihydroxyacetone (Dha) to dihydroxyacetone phosphate (DhaP), which plays critical roles in many metabolic pathways. It has also been described that Homo sapiens DAK is able to regulate the innate immune system via its interaction with Melanoma differentiation-associated protein 5 (MDA5), which is recognized as an RNA sensor during virus infection. These findings make H. sapiens DAK a very noteworthy research target due to its multiple functions in many fields. Therefore, structural studies of DAK from H. sapiens are presented. The initial structure of the wild type (WT) H. sapiens DAK was determined to 2.5Å resolution and solved by molecular replacement. The structure forms a homodimer and four dimers were shown to be present in each asymmetric unit. However, within each monomer, most regions of the C-terminal domain were disordered, and therefore in order to improve structure quality, multiple site-directed mutagenesis was used. The mutated protein was then crystallized and the structure was determined to 1.4Å. The N-terminal Dha binding domain consists of two α/β regions and the C-terminal ATP binding domain is comprised of eight anti-parallel α-helices, which forms a deep pocket and is filled with a phospholipid molecule. In addition, the structures of mutated DAK in complex with ATP analogues and Dha are also described in the current study. The second part of the project concerned sialidases, which are glycoside hydrolases that specifically hydrolyse terminal sialic acid from various glycans. Streptococcus pneumoniae is one of the most common pathogenic bacteria of humans, and is reported to encode three sialidases that act as virulence factors in bacterial colonization and infection. One of these sialidases, NanA, was reported to be present in all clinical strains and plays a vital role during the bacterial infection. Consequently, the structure of N-terminal Carbohydrate-binding module (CBM) domain of NanA has been determined to 1.8Å, and reveals a β-sandwich fold. The apo form of NanA-CBM is present as a dimer in the asymmetric unit, whereas a monomer was detected when it is bound to sialic acid or its derivatives. Structural comparisons between NanA-CBM and other structures of the CBM40 family were also performed. The substrate binding sites of NanA-CBM forms a cavity that is able to accommodate the substrates. A potential molecular binding site located beside the sialic acid binding site was revealed, and is occupied by the side chain of a lysine from a symmetry- related molecule. Heteronuclear single quantum coherence (HSQC) NMR spectroscopy and fluorescent-based thermal shift assays were also carried out to further characterise the protein. The current results reveal the structure of both DAK from H. sapiens and NanA-CBM from S. pneumoniae, which may contribute to a better understanding towards cell metabolism and bacterial colonization.

The primary focus of this dissertation is the metabolic enzyme pyruvate carboxylase (PC). The structure and function of this fascinating enzyme has been studies and characterized by many laboratories over many decades. This extensive background is reviewed in Chapter 1, with an overview of the biotin-dependent carboxylase family and a particular focus on PC. In this dissertation, we primarily use X-ray crystallography to study PC at a structural level. This dissertation is divided into two overarching sections, with the first section (Chapters 2-5) focusing on the bacterial second messenger cyclic-di-AMP (c-di-AMP). This project was initiated by our collaborators in the laboratory of Josh Woodward at the University of Washington, who performed the first screen to identify c-di-AMP binding proteins in the bacterium Listeria monocytogenes. In Chapters 2 and 3, the regulation of PC by c-di-AMP in L. monocytogenes as well as the bacterium Lactococcus lactis is discussed. Crystal structures of the PC from each of these species in complex with cyclic-di-AMP reveal the binding site and give insights into the molecular mechanisms of this regulation. In Chapters 4 and 5, structural studies of other c-di-AMP binding proteins identified in the screen are discussed. The second section (Chapters 6) focuses on a second class of PC enzymes called the two-subunit PCs, which are found in a subset of Gram-negative bacteria. In Chapter 6, the first crystal structure of a two-subunit PC from the bacterium Methylobacillus flagellatus is determined. In collaboration with the Lars Dietrich laboratory at Columbia University, we investigate the physiological function of the two-subunit PC in the pathogen Pseudomonas aeruginosa. A theme which emerges from these studies is that PC is an incredibly diverse enzyme which has been adapted for the peculiar physiological needs of each organism it inhabits. Because PC is found throughout nature in every kingdom of life, further studies of its unique properties and role in each organism are sure to provide more surprising insights in the years to come.

Determination of the three-dimensional structure of enzymes at atomic resolution is a key prerequisite for elucidation of molecular mechanisms of catalysis and catalysis mechanism prediction. X-ray protein crystallography is the most widely used method today for determining protein structures. In this thesis we describe the expression, purification, crystallization and structure solution of two new enzymes: PyrH and PrnB. PyrH is a member of the new emerging family of FADH dependent tryptophan halogenases. It catalyzes the regioselective halogenation of tryptophan at the C-5 position of the indole ring. Elucidation of its structure (Chapter 2) and comparison with PrnA, aregioselective 7th tryptophan halogenase whose structure has already been solved confirmed the proposed mechanism of action for this class of enzymes. PrnB is the only enzyme known to perform exquisite and peculiar ring rearrangement chemistry: it converts 7-Cl-tryptophan and tryptophan into respectively monodechloroaminopyrrolnitrin and aminophenylpyrrole. We developed a method for expression and purification of milligrams of pure and homogeneous recombinant PrnB (Chapter 3). We identified suitable crystallization conditions and determined PrnB structure (Chapter 4). Analysis of the PrnB structure helped us to propose a reaction mechanism for this unique enzyme.

A quantitative understanding of metabolic control depends on a knowledge of the enzymes involved. The extrapolation of studies in vitro to the intact cell is controversial because the intracellular environment is relatively poorly characterised, particularly with respect to the interactions between weakly-associated enzymes. There is a clear need to study enzymes directly in the cell, yet there are few suitable techniques. Metabolites have been very successfully studied in cells by the non-invasive technique of nuclear magnetic resonance (NMR). NMR studies of enzymes in the cell have, however, been prevented by difficulties in assigning the resonances from the many proteins within the cell. A method for studying a specific enzyme in the cell has been developed, using Saccharomyces cerevisiae and phosphoglycerate kinase (PGK) as a model system. Using an inducible expression system, PGK was synthesised in the cell without significant synthesis of other proteins. With 5-fluorotryptophan in the growth medium, fluorine-labelled PGK was formed in situ. Fluorine is an excellent label for NMR since it is absent from most cells and has a high receptivity to NMR detection. ¹⁹ F NMR was used to study PGK in the intact cell. Comparisons with measurements in vitro showed that PGK was exposed to only a small fraction of the total intracellular [ADP], implying some form of compartmentalisation. The NMR relaxation properties observed in vivo and in vitro were compared with theoretical predictions. This showed that PGK was not part of a complex in the cell and that the viscosity of the cytoplasm, relative to water, was c. 4 at 30 °C. Fluorine-labelled pyruvate kinase and hexokinase have also been prepared; the spectra of these proteins in vitro are responsive to their ligands, and further work will study these proteins in vivo. NMR techniques were also applied to study the kinetics of PGK in the cell. PGK and GAPDH catalyse an ATP↔P_i exchange which is near-equilibrium in wild-type cells. ³¹P magnetisation transfer experiments in genetically manipulated cells showed that the reaction becomes unidirectional if the PGK activity is reduced by 95 %. Net flux is reduced by less than 30 %. In low-PGK cells, the ATP↔P_i exchange from oxidative phosphorylation can be isolated from that of glycolysis, facilitating direct measurements of the P:O ratio. In the cells studied, the P:O ratio was 2 to 3.

This thesis describes studies directed towards understanding structure-function relationships of triosephosphate isomerase (TIM), from a protozoan parasite Plasmodium falciparum and a thermophilic archaea Methanocaldococcus jannaschii. Triosephosphate isomerase, a ubiquitous glycolytic enzyme, has been the subject of biochemical, enzymatic and structural studies for the last five decades. Studies on TIM have been central to the development of mechanistic enzymology. The present study investigates the role of specific residues in the structure and function of Plasmodium falciparum triosephosphate isomerase (PfTIM). The structure and stability of a tetrameric triosephosphate isomerase from Methanocaldococcus jannaschii (MjTIM) is also presented. Chapter 1 provides a general introduction to the glycolytic enzyme triosephosphate isomerase, conservation of TIM sequences, its fold and three dimensional organization. The isomerisation reaction interconverting dihydroxyacetone phosphate and glyceraldehyde 3phosphate catalyzed by triosephosphate isomerase is an example of a highly stereospecific proton transfer process (Hall & Knowles, 1975; Rieder & Rose, 1959). This chapter briefly reviews mechanistic features and discusses the role of active site residues and the functional flexible loop 6. Triosephosphate isomerase adopts the widely occurring ( β/ α)8 barrel fold and mostly occurs as a dimer (Banner et al., 1975). Protein engineering studies, related to folding, stability and design of monomeric TIM are also addressed. A brief introduction to thermophilic TIMs and higher oligomeric TIMs is given. The role of this enzyme in disease states like hemolytic anemia and neuromuscular dysfunction is surveyed. The production of methylglyoxal, a toxic metabolite, as a byproduct of the TIM reaction is also considered. Many proteins utilize segmental motions to catalyze a specific reaction. The omega loop (loop 6) of triosephosphate isomerase is important for preventing the ene-diol intermediate from forming the cytotoxic byproduct, methylglyoxal. The active site loop-6 of triosephosphate isomerase moves about 7Ǻ on ligand binding. It exhibits a hinged lid motion alternating between two well defined, “open” and “closed”, conformations (Joseph et al., 1990). Though the movement of loop 6 is not ligand gated, in crystals the ligand bound forms invariably reveal a closed loop conformation. Plasmodium falciparum TIM is an exception which predominantly exhibits “open” loop conformations, even in the ligand bound state (Parthasarathy et al., 2002). Phe 96 is a key residue that is involved in contacts between the flexible loop-6 and the protein body in PfTIM. Notably, in all TIM sequences determined thus far, with the exception of plasmodial sequences, this residue is Ser 96. In Chapter 2 the mutants F96S, F96H and F96W are reported. The crystal structures of the mutant enzymes with or without bound ligand are described. In all the ligand free cases, loop-6 adopts an “open” conformation. Kinetic parameters for all the mutants establish that residue 96 does not play an essential role in modulating the loop conformation but may be important for ligand binding. Structural analysis of the mutants along with WT enzyme reveals the presence of a water network which can modulate ligand binding. Subunit interfaces of oligomeric proteins provide an opportunity to understand protein- protein interactions. Chapter 3 describes biochemical and biophysical studies on two separate dimer-interface destabilizing mutants C13E and W11F/W168F/Y74W of PfTIM. The intention was to generate a stable monomer by disrupting the interaction hubs. C13 is a part of a large hydrophobic patch (Maithal et al., 2002a) at the dimer interface. Introduction of a negative charge at position 13 destabilizes the interface and reduces activity. Y74 is a part of an aromatic cluster of the interface (Maithal et al., 2002b). The Y74W triple mutant was designed to disrupt the aromatic cluster by introducing additional atoms. Tryptophan is also a fluorophore, allowing studies of the dimer disruption by fluorescence, after mutating the two inherent tryptophan residues, W11 and W168 to phenylalanine. The mutants showed reduced activity and were more sensitive than the wild type enzyme to chemical denaturants as well as thermal denaturation. Evidenced for monomer formation is presented. These studies together with previous work reveal that the interface is important for both activity and stability. In order to develop a model for understanding the relationship between protein stabilization and oligomeric status, characterization of the TIM from Methanocaldococcus jannaschii (MjTIM) has been undertaken. Chapter 4 describes the purification and characterization of MjTIM. The MjTIM gene was cloned and expressed in pTrc99A and protein was isolated from AA200 E. coli cells. Hyperexpressed protein was purified to homogeneity and relevant kinetic parameters have been determined. The tetrameric nature of MjTIM is established by gel filtration studies. Circular dichroism (CD) studies establish the stability of the overall fold, even at temperatures as high as 95ºC. A surprising loss of enzyme activity upon prolonged incubation at high temperature was observed. ESI-MS studies establish that oxidation of thiol groups of the protein may be responsible for the thermal inactivation. Chapter 5 describes the molecular structure of MjTIM, determined in collaboration with Prof. MRN Murthy’s group at the Indian Institute of Science (Gayathri et al., 2007). The crystal structure of the recombinant triosephosphate isomerase (TIM) from the archaeabacteria Methanocaldococcus jannaschii has been determined at a resolution of 2.3 Å. MjTIM is tetrameric, as suggested by solution studies and from the crystal structure, as in the case of two other structurally characterised archaeal TIMs. The archaeabacterial TIMs are shorter compared to the dimeric TIMs, with the insertions in the dimeric TIMs occurring in the vicinity of the putative tetramer interface, resulting in a hindrance to tetramerization in the dimeric TIMs. The charge distribution on the surface of archaeal TIMs also facilitates tetramerization. Analysis of the barrel interactions in TIMs suggests that these interactions are unlikely to account for the thermal stability of archaeal TIMs. A feature of the unliganded structure of MjTIM is the complete absence of electron density for the loop 6 residues. The disorder of the loop may be ascribed to a missing salt bridge between residues at the N- and C- terminal ends of the loop in MjTIM. Chapter 6 is a follow up of an interesting observation made by Vogel and Chmielewski (1994), who noticed that subtilisin cleaved rabbit muscle triosephosphate isomerase religated spontaneously upon addition of organic solvents. Further extension of this nicking and religation process with PfTIM emphasizes the importance of tertiary interactions in contributing to the stability of the (β/α)8 barrel folds (Ray et al., 1999). This chapter establishes that subtilisin nicking and religation is also facile in thermophilic MjTIM. Fragments generated by subtilisin nicking were identified using MALDI mass spectrometry at early and late stages of the cleavage for both the dimeric PfTIM and tetrameric MjTIM. This chapter also describes the comparative thermal and denaturant stability of both the enzymes. The accessibility of the Cys residues of MjTIM has been probed by examining the rates of labeling of thiol groups by iodoacetamide. The differential labeling of Cys residues has been demonstrated by mass spectrometry. Chapter 7 summarizes the main results and conclusions of the studies described in this thesis.

Metabolism of D-amino acids is of considerable interest due to their key importance in cellular functions. The enzymes D-serine dehydratase (DSD) and D-cysteine desulfhydrase (DCyD) are involved in the degradation of D-Ser and D-Cys, respectively. We determined the crystal structure of Salmonella typhimurium DSD (StDSD) by multiple anomalous dispersion method of phasing using selenomethione incorporated protein crystals. The structure revealed a fold typical of fold type II PLP-dependent enzymes. Although holoenzyme was used for crystallization of both wild type StDSD (WtDSD) and selenomethionine labeled StDSD (SeMetDSD), significant electron density was not observed for the co-factor, indicating that the enzyme has a low affinity for the cofactor under crystallization conditions. Interestingly, unexpected conformational differences were observed between the two structures. The WtDSD was in an open conformation while SeMetDSD, crystallized in the presence of isoserine, was in a closed conformation suggesting that the enzyme is likely to undergo conformational changes upon binding of substrate as observed in other fold type II PLP-dependent enzymes. Electron density corresponding to a plausible sodium ion was found near the active site of the closed but not in the open state of the enzyme. Examination of the active site and substrate modeling suggested that Thr166 may be involved in abstraction of proton from the Cα atom of the substrate. Apart from the physiological reaction, StDSD catalyses α, β-elimination of D-Thr, D-Allothr and L-Ser to the corresponding α-keto acids and ammonia. The structure of StDSD provides a molecular framework necessary for understanding differences in the rate of reaction with these substrates. Salmonella typhimurium DCyD (StDCyD) is a fold type II PLP-dependent enzyme that catalyzes the degradation of D-Cys to H2S and pyruvate. We determined the crystal structure of StDCyD using molecular replacement method in two different crystal forms. The better diffracting crystal form obtained in presence of benzamidine illustrated the influence a small molecule in altering protein interfaces and crystal packing. The polypeptide fold of StDCyD consists of a small domain (residues 48-161) and a large domain (residues 1-47 and 162-328) which resemble other fold type II PLP-dependent enzymes. X-ray crystal structures of StDCyD were also obtained in the presence of substrates, D-Cys and βCDA, and substrate analogs, ACC, D-Ser, L-Ser, D-cycloserine (DCS) and L-cycloserine (LCS). The structures obtained in the presence of D-Cys and βCDA show the product, pyruvate, bound at a site 4.0-6.0 Å away from the active site. ACC forms an external aldimine complex while D and L-Ser bind non-covalently suggesting that the reaction with these ligands is arrested at Cα proton abstraction and transimination steps, respectively. In the active site of StDCyD cocrystallized with DCS or LCS, electron density for a pyridoxamine phosphate (PMP) was observed. Crystals soaked in cocktail containing these ligands show density for PLP-cycloserine. Spectroscopic observations also suggested formation of PMP by the hydrolysis of cycloserines. Mutational studies suggested that Ser78 and Gln77 are key determinants of enzyme specificity and the phenolate of Tyr287 is responsible for Cα proton abstraction from D-Cys. Based on these studies, we proposed a probable mechanism for the degradation of D-Cys by StDCyD. The acid-induced arginine decarboxylase (ADC) is part of an enzymatic system in Salmonella typhimurium that contributes to making this organism acid resistant. ADC is a PLP-dependent enzyme that is active at acidic pH. It consumes a proton in the decarboxylation of arginine to agmatine, and by working in tandem with an arginine-agmatine antiporter, this enzymatic cycle protects the organism by preventing the accumulation of protons inside the cell. We have determined the structure of the acid-induced StADC to 3.1 Å resolution. StADC structure revealed an 800 kDa decamer composed as a pentamer of five homodimers. Each homodimer has an abundance of acidic surface residues, which at neutral pH prevent inactive homodimers from associating into active decamers. Conversely, acidic conditions favor the assembly of active decamers. Therefore, the structure of arginine decarboxylase presents a mechanism by which its activity is modulated by external pH.