Adenosine deaminase is a key enzyme in purine catabolism. The elucidation of the crystal structure and recent analysis of site-directed mutants have provided details on the structure and function of the enzyme. The studies described in this thesis provide new insights into ADA. The ADA mechanism is well defined except for the identification of the proton acceptor. Aspartate 295 is important to metal coordination (Wilson, 1991). This residue may be the catalytic base because of its position in the original crystal structure. Mutation of this residue to asparagine resulted in a 100,000 fold reduction in activity but retention of binding to substrates and inhibitors. UV difference spectra confirm that the tetrahedral intermediate is not formed during catalysis. These results suggest that Asp295 functions as the proton acceptor and in metal coordination. The crystal structure of mouse ADA provides a static picture of the enzyme in a single conformation. The sequestered nature of the active site indicates that the motion of one or more side chains is necessary for the proper functioning of the enzyme. Several mutants were created involving a flexible loop at residues 109-124. Complete excision, mutation to a hinge, or the replacement of the loop with the smaller loop found in E. coli ADA resulted in decreases in activity. Binding was not affected by any of these mutations. The 109-124 loop most likely acts in sequestering the active site from solvent during catalysis. An alanine to valine point mutation has been located in codon 329 of patients with SCID. Site directed mutagenesis was used to generate this mutation. Protein containing this change was generated and was found to be misfolded and catalytically inactive. The loop mutations, A329V, and D295N affect the rate of catalysis of ADA for three different reasons. The loop serves to shield the active site from solvent. Asp295 is the catalytic base. Ala329 is important to protein folding. These studies show how slight alterations in primary structure can affect the overall function of an enzyme.

LacI and PurR are members of the extended LacI/GalR family of bacterial proteins that regulate genetic expression. Detailed kinetic and thermodynamic studies on PurR were compared with results for homologous LacI. Operator binding affinity was increased by the presence of guanine as demonstrated previously, and conversely guanine binding affinity was increased by the presence of operator. Guanine enhanced operator affinity by increasing the association rate constant and decreasing the dissociation rate constant for binding. Operator had minimal effect on the association rate constant for guanine binding; however, this DNA decreased the dissociation rate constant for corepressor by ∼10-fold. Despite significant sequence and structural similarity between PurR and LacI proteins, PurR binds to its corepressor ligand with a lower association rate constant than LacI binds to its inducer ligand. However, the rate constant for PurR-guanine binding to operator is ∼3-fold higher than for LacI binding to its cognate operator under the same solution conditions. The distinct metabolic roles of the enzymes under regulation by these two repressor proteins provide a rationale for the observed functional differences. Addition of the LacI C-terminal tetramerization domain to the C-terminus of PurR resulted in the tetramerization of PurR. This tetrameric mutant of PurR exhibited very similar corepressor and operator binding affinity to wild-type PurR. These results establish that the C-terminal assembly motif from LacI can elicit tetramer formation in a naturally folded homologous dimer without interference with its biological function. Studies of the N-terminal DNA binding domain of LacI with variant hinge regions linked by a disulfide bond confirmed that the hinge region is important for operator recognition and interaction. DNA binding domains containing the hinge region from LacI (LH and VH) have similar affinity to lacO 1 operator, even though they form the disulfide bond at different positions in the hinge region. In addition, they bind to lacO 1 operator with ∼10-fold tighter affinity than that derived from the LacI DNA binding domain with the hinge region from PurR (PH). Deletion of the hinge region from the DNA binding domain (nH) resulted in a mutant without detectable operator binding even when linked by a disulfide bond.

The intricate temporal and spatial regulation of membrane fusion is critical for all living organisms. Fusion of two opposing membranes occurs in a wide range of processes. These include intracellular transportation, cell-to-cell fusion and viral fusion. In all known cases, the SNARE proteins (Soluble NSF attachment protein receptors) (Sollner et al., 1993; Whiteheart et al., 1993) have been shown to be required for vesicular membrane fusion within cells and sufficient to drive membrane fusion in vitro (Nickel et al., 1999; Parlati et al., 1999; Weber et al., 1998). While SNAREs combine in specific combinations to drive highly specific membrane fusion, it is clear that SNARE proteins do not act independently to regulate the entire fusion process. Many regulatory proteins from different families have been identified that interact with individual SNARE proteins and SNARE complexes, yet the precise role of many of these remains unclear. One such group of regulatory proteins is the Sec1/Munc18 (SM) family. Sec1 proteins are likely to be critical players in membrane trafficking. My work has focused on the role of the yeast Sec1p in post-Golgi secretion in Saccharomyces cerevisiae. To analyze Sec1p function in vitro, I have utilized a well-characterized SNARE-mediated membrane fusion assay. For this application, conditions were optimized to allow for specific protein-protein interactions to be tested. Conditions for expression and purification of the previously elusive recombinant Sec1p are documented. In addition, an overexpressing Sec1p yeast strain was generated. Sec1p interactions with SNARE proteins that mediate post-Golgi secretion were then tested. I found that recombinant Sec1p binds strongly to the t-SNARE complex (Sso1p;Sec9c) as well as to the fully assembled ternary-SNARE complex (Sso1p;Sec9c/Snc2p), and also weakly to free Sso1p. I tested the ability of Sec1p to regulate fusion in the fusion assay. Concentration dependent stimulation of membrane fusion is observed when Sec1p is associated with the SNARE proteins. The binding and fusion data strongly argue that Sec1p directly stimulates SNARE-mediated membrane fusion. With this new information, specific binding modes of neuronal-Sec1 are currently being investigated further in yeast, Drosophila and mammalian SNARE systems.

Tropomyosin plays a central role in the regulation of skeletal, cardiac, and smooth muscle regulation. The regulatory properties of tropomyosin are mediated by its interactions with muscle-specific tropomyosin binding proteins. In skeletal and cardiac muscle the regulatory protein is troponin, while in smooth muscle the primary protein is caldesmon. The structures and interactions of tropomyosin with caldesmon, skeletal troponin, and cardiac troponin have been studied using X-ray crystallography and optical biosensors. Only whole caldesmon and the carboxyl-terminal domain of caldesmon bound tightly to tropomyosin. X-ray studies showed that whole caldesmon bound to tropomyosin in several places. Experiments with the carboxyl-terminal domain of caldesmon revealed that this region corresponded to the strongest binding site seen for whole caldesmon. Weaker association of other regions of caldesmon to tropomyosin was also observed. The structure of cocrystals of skeletal and cardiac troponin subunit T revealed that the two isoforms interacted with tropomyosin in the same general area but that cardiac troponin T bound to tropomyosin over a more extended region. The longer amino-terminal domain of the cardiac protein bound further along the carboxyl region of tropomyosin than skeletal troponin T, and the carboxyl-terminal domain of cardiac troponin T bound to tropomyosin more tightly than its skeletal counterpart. Biosensors studies of tropomyosin interacting with caldesmon and troponin measured association rate, dissociation rate, and equilibrium rate constants of these proteins for the first time. Caldesmon bound with similar affinity to several tropomyosin isoforms while troponin bound most tightly to striated muscle tropomyosin. An atomic model of tropomyosin at 5 A resolution has been constructed using a simulated annealing procedure and X-ray diffraction data from the spermine crystal form of tropomyosin. During these refinements the R-free and R were monitored. However, failure to lower the R-free suggests that this model does not accurately describe the structure of tropomyosin within these crystals. These results define interactions and structures within thin filaments of cardiac, skeletal, and smooth muscle which will be useful in elucidating the exact role of these proteins in the unique regulation of each type of muscle.

The Escherichia coli RecA protein is a remarkable protein, playing key roles in initiating the SOS response to DNA damage, restarting stalled replication forks, and participating in recombinational DNA repair by binding single-stranded DNA and pairing it with homologous duplex DNA. We investigated the ligand binding specificity of the RecA nucleoprotein filament, specifically its interaction with homologous DNA strands, divalent metal ions, and nucleotide cofactors. We probed the kinetics of the mechanism of homology recognition by RecA using substituted oligonucleotide substrates to disrupt complete homology. Using this system, we were able to lend support to a base-pairing model proposed in the literature for the process of homology recognition and to provide unique structural and mechanistic insights not previously noted in investigations of the strand exchange mechanism of RecA. We also determined that RecA is inhibited by divalent metals in a novel fashion, as zinc(II), copper(II), and mercury(II) all inactivate RecA in vitro by initiating aggregation of the protein. A mechanistic hypothesis was developed for the action of metal-ligand complexes on RecA, and the potential use for metal-ligand complexes as inhibitors of RecA activities is discussed herein. We demonstrated the first proof of small-molecule inhibition of RecA activities. Using negative design, we developed several putative nucleotide analog inhibitors of RecA, and found that N6-(1-naphthyl)-ADP is a potent and highly specific inhibitor of ReCA function. In this report, we discuss the inhibitory properties of this compound on RecA and the implications for the development of novel antibiotic therapies. Finally, we investigated the role of the nucleotide cofactor binding site of RecA in the determination of substrate specificity and also activation of the RecA NPF. The D100R RecA protein was found to have reduced activity relative to wild-type in SOS induction, DNA damage repair, and ATPase assays, while neutrally charged residues at position 100 generated hyperactive RecA proteins. We postulated that RecA uses the Asp100 residue to discriminate against GTP activation and maintain selectivity for ATP binding, and that RecA uses Coulombic forces to communicate NTP binding cooperatively to both the protein conformational changes and DNA binding required for active NPF formation.

Nitric oxide (NO) has been implicated as mediator in a variety of physiological functions, including neurotransmission, platelet aggregation, macrophage function, and vasodilation. Mild hypertensive events have been reported during in vivo trials of extracellular protein-based blood substitutes and have been attributed to consumption of NO by hemoglobin and subsequent vasoconstriction due to lack of activation of guanylate cyclase. Thus, understanding the reactions of NO with hemoglobin is crucial to the development of a risk-free extracellular oxygen carrier. The reactions of NO with recombinant oxy-, deoxy-, and metmyoglobin have been examined in order to determine the chemical mechanisms involved in these processes. Sperm whale myoglobin was chosen as a simple model system for determining the structural and chemical factors that regulate NO reactivity. These myoglobin studies have led to four major conclusions. (1) NO-induced oxidation of MbO$\sb2$ takes place by direct reaction between NO and bound O$\sb2$, and the rate-limiting step is NO diffusion into the distal pocket. (2) In wild-type ferrous myoglobin, the strength of hydrogen bonding between His64(E7) and bound NO is several-fold stronger than hydrogen bonding to bound CO but is $\sim$100-fold weaker than hydrogen bonding to bound O$\sb2$. (3) The pathway for NO entry into the distal pocket of either deoxy- or metmyoglobin appears to involve a channel between Phe43(CD1), the heme-7-propionate, and Thr67(E10) which is opened by outward rotation of the His64(E7) side chain. (4) The rate-limiting step for aerobic oxidation of MbNO is dissociation of NO followed by NO-induced oxidation of newly-formed MbO$\sb2$. These results have already proved to be applicable to other heme proteins and have led to the design of recombinant hemoglobins which can function as extracellular oxygen-carriers with minimal interference with NO messenger pathways.

A cataract is an opacification of the ocular lens, which is suspended behind the pupil by filaments called zonules. While cataracts are surgically treatable, limited economic resources currently hinder the treatment of this debilitating disease in the developing world. Couching of the cataractous lens by a proteolytic enzyme could provide a satisfactory and economical alternative to surgical lens removal. The HIV protease, because of its small size, could potentially be delivered non-invasively into the eye to effect release of the lens. Unfortunately, our results suggest that the wild-type protease is unable to cleave the zonules. Directed molecular evolution may provide a viable means of modifying the HIV protease to recognize and efficiently cleave the zonular fibers. Utilizing this approach, however, requires the availability of a method for screening very large protease mutant libraries. Accordingly, we have developed a novel two-tiered system for the rapid screening of site-specific protease libraries with, in some cases, as many as one billion members. The first tier is an in vivo genetic selection that links protease activity to antibiotic resistance. The activity of proteases isolated in this first tier is then verified in a more stringent second tier based on the in vitro cleavage of a recombinant fluorescent substrate bound to a solid support. This system has been adapted and optimized for screening the HIV, HCV NS3/4a, and TEV proteases, and has been demonstrated to work well in a model directed evolution experiment. With the ability to now screen large mutant libraries of the HIV protease, it should be possible in the future to isolate a mutant with cleavage activity toward the zonules. Additionally, the broad flexibility of the system for screening additional site-specific proteases should be of great assistance to the scientific community by providing a means of gathering important structure-function information and by allowing for the isolation of protease mutants with novel properties such as altered substrate specificity, stability, and/or reaction chemistry.

The lanosterol synthases from Trypanosoma cruzi, Trypanosoma brucei and Pneumocystis carinii, the lanosterol 14-demethylases from Trypanosoma brucei and Trypanosoma cruzi, and the cytochrome P450 reductase from Trypanosoma brucei were cloned and characterized. The two trypanosome lanosterol synthases showed a novel difference in protein sequence identity from that of other lanosterol synthases, which could be exploited for the development of specific antitrypanosome inhibitors. Yeast strains for expressing lanosterol 14alpha-demethylases were also developed. The first strains developed by tetrad dissection were time-consuming to produce, therefore another expression system was developed. The new system involved transforming the existing yeast strains, BJY1[pTb14DM] or BJY5[pTb14DM], and selection on FOA medium. With the addition of the T. brucei P450 reductase characterized in this study, the trypanosome lanosterol 14alpha-demethylases were able to regenerate their catalytic activity more efficiently than the strains containing only the native yeast P450 reductase. The various yeast strains developed in this study should be useful for screening antiparasite drugs. The BJY1[pTb14DM] (without reductase) and BJY5[pTb14DM] (with reductase) strains would also be useful in creating expression strains for other 14DM genes. Hopefully, these will be cloned in the near future.

The electrostatic theory for ligand discrimination in myoglobin is based on the physiological importance of hydrogen bonding between the distal histidine and bound ligands in myoglobin. A quantitative estimation of this theory in hemoglobin had not been established. A series of 'electrostatic' mutations was created to examine for the existence of electrostatic stabilization of bound O2 in both the alpha and beta subunit of R-state recombinant human hemoglobin. A set of HisE7 to Ala, Leu, and Gln replacements was used to investigate the role of the distal histidine. The aliphatic mutations, AlaE7 and LeuE7, abolish all hydrogen bonding capabilities near bound ligands and cause dramatic increases in ligand association rate constants, particularly for O2. The GlnE7 substitution retains hydrogen bonding capabilities and results in wild-type-like ligand binding behavior. A second set of mutations were created to alter and examine the electrostatic potential in the distal pocket: PheCD4 → Val, ValE11 → Asn, and the double mutant HisE7 → Leu/ValE11 → Asn. The ValCD4 mutation increases the flexibility of the distal histidine side chain, weakening hydrogen bonding interactions with bound ligands. The addition of an asparagine into the distal pocket at position E11 provides a second hydrogen bonding donor in both subunits. The LeuE7/AsnE11 double mutation maintains 'wild-type' electrostatic behavior, due to the compensatory effects of the addition of a hydrogen bond donor at position E11 and removal of one at position E7. Finally, the O2 and CO properties of the double mutant HisE7 → Gln/LeuB10 → Trp were examined as a blood substitute prototype with reduced rates of ligand capture but normal or increased rates of ligand dissociation. This combination of amino acid replacements causes marked decreases in O2 affinity in both subunits and maintains high rates of dissociation, both of which are favorable for efficient O2 transport. The rate of association is decreased by the large size of the Tip side chain and the weakening of hydrogen bonding by the GlnE7 replacement markedly enhances O2 dissociation compared to that measured for the single TrpB10 mutants.

Transcription regulation and RNA metabolism, two essential cellular processes, are linked by Disconnected Interacting Protein 1 (DIP1) in Drosophila melanogaster. This linkage occurs through DIP1 interactions with double-stranded RNA (dsRNA) and its association with other proteins, such as Disconnected (zinc-finger transcription factor), Ultrabithorax (Ubx) (Hox transcription factor), Suppressor of variegation 3--9 (chromatin methyltransferase), and Adat (tRNA deaminase). This thesis explores the molecular mechanisms of DIP1 function by elucidating its interactions with dsRNA and two protein partners, Adat and Ubx. Proteins with dsRNA-binding domains bind dsRNA with high affinity and no sequence specificity in vitro. Nonetheless, these proteins are involved in many important facets of cellular metabolism in vivo that encompass splicing, deamination, and RNA interference. Analytical ultracentrifugation suggests that DIP1 is a homodimer in solution, consistent with data for other proteins with dsRNA-binding domains. Gel retardation demonstrates that DIP1 prefers dsRNA over dsRNA. Additionally, DIP1 binds the Adenovirus VA1 RNA with the highest affinity we have been able to identify for this class of proteins. DIP1 also binds to a stem-loop structure similar to those involved in long-range enhancer effects of the Bithorax complex as well as in the generation of two microRNAs that potentially regulate Ubx. Adat·DIP1 interaction was established in yeast two-hybrid assays, and the Adat interaction with DIP1 was confirmed with GST pulldown experiments. DIP1 binds with high affinity to the pre-tRNAala, a potential in vivo target of Adat. We hypothesize that Adat is recruited by the DIP1·RNA complex in order to properly deaminate the pre-tRNA ala. DIP1 also binds Ubx, a Hox transcription factor necessary for haltere formation during embryogenesis. All Hox proteins rely on protein-protein interactions to define their cellular function, although few such interactions have been identified. The Ubx·DIP1 interaction was demonstrated using phage display, immunoprecipitation, GST pulldown assays, and gel supershifts in complex with DNA. Yeast one-hybrid assays establish DIP1 can block Ubx activation, and a functional consequence of this interaction may be to differentiate Hox function in vivo. Elucidating these molecular mechanisms via DIP1 interactions potentially provide insight into cancer regulation and a physical bridge that links transcription and RNA metabolism.