The disease malaria afflicts more than a billion people and kills almost one to three million of them every year. Of the four species of Plasmodium affecting man viz., P. falciparum, P. vivax, P. ovale and P. malariae, Plasmodium falciparum is the deadliest as it causes cerebral malaria. The situation has become worse with the continuous emergence of drug resistance in the parasite. Therefore, improving existing drugs and deciphering new pathways for drug development are the need of the hour. The discovery of the type II fatty acid biosynthesis pathway in Plasmodium falciparum (Surolia and Surolia, 2001) has opened up new avenues for the development of new antimalarials as this pathway is entirely different from the human host in which type I pathway exists. Although many biochemical pathways such as the purine, pyrimidine and carbohydrate metabolic pathways, and the phospholipid, folate and heme biosynthetic pathways operate in the malaria parasite and are being investigated for their amenability as antimalarial therapeutic targets, no antimalarial of commercial use based on the direct use of these biochemical pathways as targets has emerged so far. This is due to the fact that the structure and function of the targets of these drugs overlaps with that of the human host. A description of such pathways forms the Chapter 1 of the thesis. This is followed by a description of the discovery and the importance of fatty acid biosynthesis pathway and the available literature on the various enzymes that are targets of potential antimalarials. Three isoforms are known for condensing enzymes - FabH which functions in initiation, and FabB and FabF which function in elongation. These isoforms differ in their biochemical properties and have unique roles to play in deciding the membrane composition of any organism. This aspect is also discussed in this chapter. Cloning and expression of -ketoacyl-ACP synthase, FabB/F from Plasmodium falciparum is described in Chapter 2. PfFabB/F is coded by the nuclear genome and is targeted to the apicoplast. The gene is coded by the locus MAL6P1.165 and the putative amino acid sequence of the protein exists in PlasmoDB. All apicoplast targeted proteins have a characteristic bipartite leader sequence consisting of a signal and a transit peptide sequence (Waller et al., 1998). Since the mature protein start site was not known and none of the software packages could predict the site, I aligned the PfFabB/F sequence with the sequences of other -ketoacyl-ACP synthases. On the basis of similarity with E. coli synthases and the mature protein start site of plant synthases, I cloned the first construct of PfFabB/F. The sequence was amplified by PCR and ligated in pET as well as pGEX vector. Expression in various hosts under different temperature and induction conditions could not solubilize the protein in significant quantities and most of the protein was found in inclusion bodies. Next I expressed the sequence with five more amino acids towards the N-terminal and expressed it as an N- terminal NusA fusion. The protein was purified by single step Ni-NTA affinity chromatography. Along with the full length protein (108 kDa), a truncated version of the protein was also obtained. The identity of the protein was confirmed by western blotting using anti-His antibody and anti-FabB/F antibody. In Chapter 3, the substrate specificity of PfFabB/F has been elucidated. PfFabB/F condenses malonyl-ACP with a range of acyl-ACPs. In vivo, acyl carrier protein (ACP) shuttles the acyl substrates between various enzymes of the fatty acid biosynthesis pathway. Enzymes of the pathway other than synthases can accept substrate analogs like acyl-CoA and acyl-NAC’s also in vitro. Acyl-ACPs are not very stable species and thus are not commercially available. Therefore, they have to be synthesized. Since malonyl-ACP could not be synthesized by chemical means, enzymatic synthesis of acyl-ACPs was done. Acyl-ACP synthetase (Aas) or holo-ACP synthase (ACPS) can be used for enzymatic synthesis. Aas is specific only for longer chain substrates; therefore, I decided to use holo-ACP synthase, an enzyme responsible for converting apo-ACP to holo-ACP in the presence of CoA in vivo (Lambalot and Walsch, 1995). When acyl-CoAs are supplied in place of CoA, acyl-ACP is produced. Malonyl-ACP and acyl-ACPs (C4-C16:1) were thus synthesized using holo-ACP synthase from E. coli. The reaction went to almost 95% completion, indicating broad substrate specificity of this enzyme. Bacterial or plant acyl-ACPs of different chain lengths can be resolved by Conformation Sensitive PAGE (Heath and Rock, 1995, Post- Beittenmiller et al., 1991). However, Pfacyl-ACPs synthesized using ACPS did not show any significant shift on CS-PAGE. Therefore I resorted to MALDI-TOF (Matrix Assisted Laser Desorption and Ionization- Time Of Flight) for monitoring the PfFabB/F condensation reactions. PfFabB/F condensed C4-C12-ACPs with malonyl-ACP to their corresponding -ketoacyl-ACP products, with C6, C8 and C10-ACPs being most readily elongated. C14-ACP was very sluggishly elongated, and C16 and C16:1-ACPs were not elongated at all. The condensation reaction was also followed by autoradiography using14C labeled malonyl-ACP, exploiting the clear mobility shift between malonyl-ACP and the other acyl-ACPs. The inhibitory effect of cerulenin, a known inhibitor of condensing enzymes was also checked. PfFabB/F also exhibited malonyl decarboxylase activity resulting in the production of acetyl-ACP in the absence of any significant condensation activity. All the enzymes of fatty acid synthesis pathway required to complete a cycle were assembled together for the in vitro reconstitution of Plasmodium fatty acid synthesis cycle which is described in Chapter 4. Earlier studies of Surolia & Surolia have shown that C12 and C14 fatty acids are the major constituents of Plasmodium lipids. One of my objectives was to determine the maximum length of the acyl ACP product that is synthesized when all the functionally active enzymes of fatty acid synthesis are put together (Kapoor et. al, 2001, Sharma et al., 2003, Karmodiya and Surolia, 2006). Condensing enzymes have a deterministic role in the fatty acid composition as they catalyze the only irreversible step in fatty acid biosynthesis. By analyzing products of the elongation cycle by mass spectrometry it was apparent that C14-ACP is the longest species formed. As already mentioned, PfFabB/F readily elongates C12-ACP but C14-ACP is weakly elongated. Thus the end product of the Plasmodium FAB pathway is influenced by the substrate specificity of PfFabB/F. This confirms the role of PfFabB/F as a decisive enzyme in determining the length of fatty acids synthesized. The inhibition of the cycle by cerulenin and triclosan is also described in this chapter. Chapter 5 describes the ability of the PffabB/F gene to complement for the mutation of condensing enzymes in CY244 cells (fabBtsfabF-, Yasuno et al., 2004). CY244 cells were transformed with pBAD alone or PfFabB/F cloned in pBAD vector (pBADPffabB/F) and the growth was monitored at non-permissive temperature. The product of PfFabB/F could rescue the growth of mutant cells at high temperature but only in the presence of oleic acid. FabB and FabF are the isoforms of condensing enzymes involved in elongation of the fatty acid synthesis cycle but they have a unique role to play (Garwin et al., 1980). FabB is responsible for unsaturated fatty acid synthesis, and fabB-mutants require oleic acid supplementation for growth. FabF is utilized in temperature regulation of membrane fluidity and E. coli FabF elevates the level of C18:1 or cis-vaccenic acid at lower growth temperature but FabF-mutants have no growth phenotype (Ulrich et al., 1983). Rescue of CY244 cells in the presence of oleic acid supplementation indicated that the PffabB/F gene behaves like FabF and not FabB. Analysis of the fatty acid composition of membrane lipids of CY244 cells transformed with pBAD vector or pBADPffabB/F by GC-MS demonstrated no elevated levels of cis-vaccenic acid in transformed cells. This observation is in agreement with the in vitro determined substrate specificity data which shows that PfFabB/F does not elongate C16:1ACP. The thesis ends with a summary of the findings in Chapter 6 in the context of FabB and FabF enzymes known from other sources. 2, 4, 4’-Trichloro-2’hydroxydiphenylether, commonly known as triclosan, has been used as a topical antibacterial agent for decades. I determined its efficacy in treating acute systemic bacterial infection in mouse model. Triclosan, as compared to other well known antibiotics, could extend the survival time of mice by 48 hours. This work is described in Appendix I. (Sharma et al., 2003)

The thesis deals with X-ray crystallographic analysis of two enzymes involved in the fatty acid biosynthesis pathway, known as Fatty Acid Synthase or FAS, of the malarial parasite, Plasmodium falciparum, in order to understand their functions at the atomic level and to provide structural basis for the rational design of antimalarial compounds. Targeting highly specific and well-characterized biochemical pathways to develop effective therapeutic agents has the advantage of designing new drugs or modifying the existing ones based on the details of the known features of the processes. Knowledge of the three-dimensional structures of the molecules involved in the reactions will enhance the capabilities of this procedure. The recently identified fatty acid biosynthesis pathway in Plasmodium falciparum is undoubtedly an attractive target for drug development as it differs from that in humans. In the malarial parasite, each reaction of the pathway is catalyzed by a specific enzyme whereas in humans, the synthesis is carried out by a single multidomain enzyme. Essentially each step in the FAS of P. falciparum can be a potential target to prevent the growth of the parasite as the fatty acids are essential for the formation of the cell membrane which is vital for its survival. All the reactions of this pathway have been well-characterized. Nevertheless, there is a dearth of structural information of the proteins involved in performing various functions in this pathway. When this work was initiated, crystal structures of none of these proteins were reported. The current work on the plasmodial FAS proteins has been undertaken with a view to obtain precise structural details of their reaction and inhibition mechanisms. The introductory chapter of the thesis includes a discussion on malaria, the fatty acid biosynthesis in various organisms and an overview of the structural features of the enzymes involved in the pathway that have been characterized from other organisms.The second chapter includes the tools of X-ray crystallography that were used for structural studies of the present work. It also discusses the other computational and biophysical methods used to further characterize the enzymes under study. FabI, the enoyl acyl carrier protein reductase, that regulates the third step in FAS has been crystallized as a binary complex with its cofactor NADH and as a ternary complex with NAD+and triclosan. The crystal structures of the binary and the ternary complexes have been determined at 2.5 and 2.2 ˚A, respectively. The mode of binding of the cofactor and the inhibitor triclosan to the enzyme with details of the interactions between them could be clearly obtained from these structures. Each subunit of the tetrameric FabI has the classical Rossmann fold. We carried out a thorough analysis of this structure and compared it with the FabI structures from various sources, four microbial (Escherichia coli, Mycobacterium tuberculosis and Helicobacter pylori) and one plant (Brassica napus), and arrived at a number of significant conclusions: Though the tertiary and the quaternary structures of the enzymes from different sources are similar, the substrate binding loop shows significant changes. The position and nature of the loop are strongly correlated to the affinity of triclosan to the enzyme. Small to major changes in the structure of the enzyme occur to enhance ligand binding. Water molecules play an important role in enzyme-ligand interactions. The crystal structure has also confirmed our previous prediction based on modeling studies of the enzyme that the introduction of bulkier groups at carbon 4’ of triclosan is likely to improve its efficacy as an inhibitor of FabI of P. falciparum. It has also provided the structural basis for its preference to bind to the coenzyme NADH but not to NADPH which was also predicted by our modeling studies. Chapters 3 and 4 discuss the structure solution and a comparative analysis of the crystal structures of FabIs from various sources. The crystal structure of FabZ, the β-hydroxyacyl acyl carrier protein dehydratase of P. falciparum, has been determined at a resolution of 2.4 ˚A. Each subunit of FabZ has a hotdog fold with one long central α-helix surrounded by a six-stranded antiparallel β-sheet. FabZ has been found to exist as a homodimer in the crystals of the present study in contrast to the hexameric form which is a trimer of dimers crystallized in a different condition, reported while we completed the structure of the dimeric form. In the dimeric form, the architecture of the catalytic site has been drastically altered with two catalytic histidine residues moving away from the catalytic site caused by two cis to trans peptide flips compared to the hexameric form. These alterations not only prevent the formation of a hexamer but also distort the active site geometry resulting in a dimeric form of FabZ that is incapable of substrate-binding. The dimeric state and an altered catalytic site architecture make the dimeric FabZ presented in the thesis distinctly different from the FabZ structures described so far. This is the first observation and report of the existence of an inactive form of the enzyme and its unique structural features. Further analysis using dynamic light scattering and size exclusion chromatographic studies have shown that a pH-related conformational switching occurs between the inactive dimers and active hexamers of FabZ in P. falciparum. These findings open alternate and entirely new strategies to design inhibitors to make FabZ inactive. FabZ crystals show polymorphism with varying length of its longest cell axis. We could collect X-ray diffraction data for three of these forms. An analysis of these forms revealed that three flexible loops of the structure including those containing the peptide flips compete for the space between two symmetry-related molecules. In the form with the longest cell axis, the loops have the highest stability resulting in a better diffraction from the crystal. We also performed docking studies with two previously characterized inhibitors of FabZ. The docking showed that the inhibitors bind strongly at the active site each one making a number of different interactions with the catalytic residues. Observations from our docking studies are in excellent agreement with and strongly supported by chemical modification and fluorimetric analysis of the wild type enzyme and its mutants. Chapters 5 and 6 explain in detail about the structure solution of dimeric form of PfFabZ, the pH induced conformational flipping of two cis-trans peptide flips that lead to different oligomeric states, and the structural basis for these oligomeric shifts. The mechanism of action of PfFabZ inhibitors NAS-21 and NAS-91 are also discussed in detail. Intrigued by the hot dog fold of the Fab enzyme, we have analyzed and compared proteins having this fold in their structures. It has been observed that the fold is often associated with fatty acids. However, the sequences, the quaternary structures, substrate specificities and the reactions that the proteins catalyze are quite diverse. The consensus sequence motifs lining the interface of quaternary association and at active site clearly indicated that the information for different modes of quaternary associations is embedded in the sequences itself. The diversity in function and quaternary association of hot dog fold proteins and their structure-function relationships are discussed in chapter 7. Malaria affects hundreds of millions of people worldwide causing about two million deaths every year. In spite of the commendable success of the available antimalarials, it has not been possible to contain the disease completely as the protozoan has become resistant to a majority of frontline drugs. The structural studies presented here should enhance the current biochemical knowledge to develop selective and more potent inhibitors of the pathway and contribute to the ongoing efforts to fight the disease.