Structural and Biochemical Dissection of the Trehalose Biosynthetic Complex in Pathogenic Fungi

Trehalose is a non-reducing disaccharide essential for pathogenic fungal survival and virulence. The biosynthesis of trehalose requires the trehalose-6-phosphate synthase, Tps1, and trehalose-6-phosphate phosphatase, Tps2. More importantly, the trehalose biosynthetic pathway is absent in mammals, conferring this pathway as an ideal target for antifungal drug design. However, lack of germane biochemical and structural information hinders antifungal drug design against these targets.

In this dissertation, macromolecular X-ray crystallography and biochemical assays were employed to understand the structures and functions of proteins involved in the trehalose biosynthetic pathway. I report here the first eukaryotic Tps1 structures from Candida albicans (C. albicans) and Aspergillus fumigatus (A. fumigatus) with substrates or substrate analogs. These structures reveal the key residues involved in substrate binding and catalysis. Subsequent enzymatic assays and cellular assays highlight the significance of these key Tps1 residues in enzyme function and fungal stress response. The Tps1 structure captured in its transition-state with a non-hydrolysable inhibitor demonstrates that Tps1 adopts an “internal return like” mechanism for catalysis. Furthermore, disruption of the trehalose biosynthetic complex formation through abolishing Tps1 dimerization reveals that complex formation has regulatory function in addition to trehalose production, providing additional targets for antifungal drug intervention.

I also present here the structure of the Tps2 N-terminal domain (Tps2NTD) from C. albicans, which may be involved in the proper formation of the trehalose biosynthetic complex. Deletion of the Tps2NTD results in a temperature sensitive phenotype. Further, I describe in this dissertation the structures of the Tps2 phosphatase domain (Tps2PD) from C. albicans, A. fumigatus and Cryptococcus neoformans (C. neoformans) in multiple conformational states. The structures of the C. albicans Tps2PD -BeF3-trehalose complex and C. neoformans Tps2PD(D24N)-T6P complex reveal extensive interactions between both glucose moieties of the trehalose involving all eight hydroxyl groups and multiple residues of both the cap and core domains of Tps2PD. These structures also reveal that steric hindrance is a key underlying factor for the exquisite substrate specificity of Tps2PD. In addition, the structures of Tps2PD in the open conformation provide direct visualization of the conformational changes of this domain that are effected by substrate binding and product release.

Last, I present the structure of the C. albicans trehalose synthase regulatory protein (Tps3) pseudo-phosphatase domain (Tps3PPD) structure. Tps3PPD adopts a haloacid dehydrogenase superfamily (HADSF) phosphatase fold with a core Rossmann-fold domain and a α/β fold cap domain. Despite lack of phosphatase activity, the cleft between the Tps3PPD core domain and cap domain presents a binding pocket for a yet uncharacterized ligand. Identification of this ligand could reveal the cellular function of Tps3 and any interconnection of the trehalose biosynthetic pathway with other cellular metabolic pathways.

Combined, these structures together with significant biochemical analyses advance our understanding of the proteins responsible for trehalose biosynthesis. These structures are ready to be exploited to rationally design or optimize inhibitors of the trehalose biosynthetic pathway enzymes. Hence, the work described in this thesis has laid the groundwork for the design of Tps1 and Tps2 specific inhibitors, which ultimately could lead to novel therapeutics to treat fungal infections.

Structural Basis of RNA Recognition by Mycobacterium tuberculosis MazF Homologues

The MazEF toxin-antitoxin (TA) system consists of the antitoxin MazE and the toxin MazF. MazF is a sequence-specific endoribonuclease that upon activation causes cellular growth arrest and increass the level of persisters. Moreover, MazF-induced cells are in a quasi-dormant state that cells remain metabolically active while stop dividing. The quasi-dormancy is similar to the nonreplicating state of M. tuberculosis during latent tuberculosis, thus suggesting the role of mazEF in M. tuberculosis dormancy and persistence. M. tuberculosis has nine mazEF TA modules, each with different RNA cleavage specificities and implicated in selective gene expression during stress conditions. To date only the Bacillus subtilis MazF-RNA complex structure has been determined. As M. tuberculosis MazF homologues recognize distinct RNA sequences, their molecular mechanisms of substrate specificity remain unclear. By taking advantage of X-ray crystallography, we have determined structures of two M. tuberculosis MazF-RNA complexes, MazF-mt1 (Rv2801c) and MazF-mt3 (Rv1991c) in complex with an uncleavable RNA substrate. These structures have provided the molecular basis of sequence-specific RNA recognition and cleavage by MazF toxins.

Both MazF-mt1-RNA and MazF-mt3-RNA complexes showed similar structural organization with one molecule of RNA bound to a MazF-mt1 or MazF-mt3 dimer and occupying the same pocket within the MazF dimer interface. Similar to B. subtilis MazF-RNA complex, MazF-mt1 and MazF-mt3 displayed a conserved active site architecture, where two highly conserved residues, Arg and Thr, form hydrogen bonds with the scissile phosphate group in the cleavage site of the bound RNA. The MazF-mt1-RNA complex also showed specific interactions with its three-base RNA recognition element. Compared with the B. subtilis MazF-RNA complex, our structures showed that residues involved in sequence-specific recognition of target RNA vary between the MazF homologues, therefore explaining the molecular basis for their different RNA recognition sequences. In addition, local conformational changes of the loops in the RNA binding site of MazF-mt1 appear to play a role in MazF targeting different RNA lengths and sequences. In contrast, the MazF-mt3-RNA complex is in a non-optimal RNA binding state with a symmetry-related MazF-mt3 molecule found to make interactions with the bound RNA in the crystal. The crystal-packing interactions were further examined by isothermal titration calorimetry (ITC) studies on selected MazF-mt3 mutants. Our attempts to utilize a MazF-mt3 mutant bearing mutations involved in crystal contacts all crystallized with few nucleotides, which are still found to interact with a symmetry mate. However, these different crystal forms revealed the conformational flexibility of loops in the RNA binding interface of MazF-mt3, suggesting their role in RNA binding and recognition, which will require further studies on additional MazF-mt3-RNA complex interactions.

In conclusion, the structures of the MazF-mt1-RNA and MazF-mt3-RNA complexes provide the first structural information on any M. tuberculosis MazF homologues. Supplemented with structure-guided mutational studies on MazF toxicity in vivo, this study has addressed the structural basis of different RNA cleavage specificities among MazF homologues. Our work will guide future studies on the function of other M. tuberculosis MazF and MazE-MazF homologues, and will help delineate their physiological roles in M. tuberculosis stress responses and pathogenesis.

Bayesian variable selection in structured high-dimensional covariate spaces with applications in genomics

We consider the problem of variable selection in regression modeling in high-dimensional spaces where there is known structure among the covariates. This is an unconventional variable selection problem for two reasons: (1) The dimension of the covariate space is comparable, and often much larger, than the number of subjects in the study, and (2) the covariate space is highly structured, and in some cases it is desirable to incorporate this structural information in to the model building process. We approach this problem through the Bayesian variable selection framework, where we assume that the covariates lie on an undirected graph and formulate an Ising prior on the model space for incorporating structural information. Certain computational and statistical problems arise that are unique to such high-dimensional, structured settings, the most interesting being the phenomenon of phase transitions. We propose theoretical and computational schemes to mitigate these problems. We illustrate our methods on two different graph structures: the linear chain and the regular graph of degree k. Finally, we use our methods to study a specific application in genomics: the modeling of transcription factor binding sites in DNA sequences. © 2010 American Statistical Association.

Statistical analysis of crystallization database links protein physico-chemical features with crystallization mechanisms.

X-ray crystallography is the predominant method for obtaining atomic-scale information about biological macromolecules. Despite the success of the technique, obtaining well diffracting crystals still critically limits going from protein to structure. In practice, the crystallization process proceeds through knowledge-informed empiricism. Better physico-chemical understanding remains elusive because of the large number of variables involved, hence little guidance is available to systematically identify solution conditions that promote crystallization. To help determine relationships between macromolecular properties and their crystallization propensity, we have trained statistical models on samples for 182 proteins supplied by the Northeast Structural Genomics consortium. Gaussian processes, which capture trends beyond the reach of linear statistical models, distinguish between two main physico-chemical mechanisms driving crystallization. One is characterized by low levels of side chain entropy and has been extensively reported in the literature. The other identifies specific electrostatic interactions not previously described in the crystallization context. Because evidence for two distinct mechanisms can be gleaned both from crystal contacts and from solution conditions leading to successful crystallization, the model offers future avenues for optimizing crystallization screens based on partial structural information. The availability of crystallization data coupled with structural outcomes analyzed through state-of-the-art statistical models may thus guide macromolecular crystallization toward a more rational basis.