Malaria is an infectious disease caused by parasites of the Plasmodium genus with mortality rates of more than a million annually, hence a major global public health concern. Plasmodium falciparum (P. falciparum) accounts for over 90% of malaria incidence. Increased resistance to antimalarial drugs by the Plasmodium parasite, coupled with the lack of an effective malaria vaccine necessitates the urgent need for new research avenues to develop novel and more potent antimalarial drugs. This study focused on falcipains, a group of P. falciparum cysteine proteases that belong to the clan CA and papain family C1, that have emerged as potential drug targets due to their involvement in a range of crucial functions in the P. falciparum life cycle. Recently, falcipain-2 has been validated as a drug target but little is known of its Plasmodium orthologs. Currently, there are several falcipain inhibitors that have been identified, most of which are peptide based but none has proceeded to drug development due to associated poor pharmacological profiles and susceptibility to degradation by host cysteine proteases. Non-peptides inhibitors have been shown to be more stable in vivo but limited information exists. In vivo studies on falcipain-2 and falcipain-3 inhibitors have also been complicated by varying outcomes, thus a good understanding of the structural variations of falcipain Plasmodium orthologs at the active site could go a long way to ease in vivo results interpretation and effective inhibitor design. In this study, we use bioinformatics approaches to perform comparative sequence and structural analysis and molecular docking to characterize protein-inhibitor interactions of falcipain homologs at the active site. Known FP-2 and FP-3 small molecule nonpeptide inhibitors were used to identify residue variations and their effect on inhibitor binding. This was done with the aim of screening a collection of selected non-peptide compounds of South African natural origin to identify possible new inhibitor leads. Natural compounds with high binding affinities across all Plasmodium orthologs were identified. These compounds were then used to search the ZINC database for similar compounds which could have better binding affinities across all selected falcipain homologs. Compounds with high binding affinities across all Plasmodium orthologs were found.

The papain-like cysteine protease family (C1 proteases) is highly important because of its involvement in research and industrial applications and its role in various human diseases. Protein inhibitors are an important aspect of C1 protease biology and are relevant to its clinical, industrial and research importance. To study the interaction between the proteases and the inhibitors it is very useful to have accurate structural models of the protease-inhibitor complexes. To this end, a high-throughput pipeline for modelling complexes of papain-like cysteine proteases and protein inhibitors was implemented and tested (Tastan Bishop & Kroon, 2011). The pipeline utilizes a novel technique for obtaining modelling templates by using superpositioning to combine coordinates from separate experimental structures. To test the pipeline, models of complexes with known structures (test set) were modelled using many different templates and the resultant models evaluated to compare the quality of the different templates. It was found that use of the new technique to obtain templates did not introduce significant errors, while allowing closer homologs to be used for modelling - leading to more accurate models. The test set models were also used to evaluate certain steps of the modelling protocol. The effect of Rosetta energy minimization on model accuracy and the use of Rosetta energy and DOPE Z-score values to identify accurate models were investigated. Several complexes were then modelled using the best available templates according to criteria informed by the previous results. A website was built that allows a user to download any of the metrics or models produced in the study. This website is accessible at http://rubi.ru.ac.za/cpmdb

In insects, the cuticle provides protection against physical injury and water loss, rigidness for muscle attachment and mechanical support, and flexibility in inter-segmental and joint areas for mobility. As most insects undergo metamorphosis, they need to shred off old cuticle and synthesize new cuticle to fit the body shape and size throughout their life cycles. The newly formed cuticle, mainly composed of cuticular proteins, chitin, and sclerotizing reagents, needs to be hardened through the crosslinks between cuticular proteins and sclerotizing reagents. This dissertation concerns the biochemical activities of several pyridoxal 5-phosphate (PLP)-dependent decarboxylases with most of them involved in insect cuticle hardening. Herein, we first present a detailed overview of topics in reactions and enzymes involved in insect cuticle hardening. Aspartate 1-decarboxylase (ADC) is at the center of this dissertation. beta-alanine, the product of ADC-catalyzed reaction from aspartate, is the component of an important sclerotizing reagent, N-beta-alanyldopamine; the levels of beta-alanine in insects regulate the concentrations of dopamine, therefore affecting insect sclerotization and tanning (collectively referred as cuticle hardening in this dissertation). Biochemical characterization of insect ADC has revealed that this enzyme has typical mammalian cysteine sulfinic acid decarboxylase (CSADC) activity, able to generate hypotaurine and taurine. The result throws lights on research in the physiological roles of insect ADC and the pathway of insect taurine biosynthesis. Cysteine was found to be  an inactivator of several PLP-dependent decarboxylases, such as ADC, glutamate decarboxylase (GAD) and CSADC. This study helps to understand symptoms associated with the abnormal cysteine concentrations in several neurodegenerative diseases. A mammalian enzyme, glutamate decarboxylase like-1 (GADL1), has been shown to have the same substrate usage as insect ADC does, potentially contributing to the biosynthesis of taurine and/or beta-alanine in mammalian species. Finally, the metabolic engineering work of L-3, 4-dihydroxyphenylalanine decarboxylase (DDC) and 3, 4-dihydroxylphenylacetaldehyde (DHPAA) synthase has revealed that the reactions of these enzymes could be determined by a few conserved residues at their active site. As both enzymes have been implicated in the biosynthesis of sclerotizing reagents, it is of great scientific and practical importance to understand the similarity and difference in their reaction mechanisms. The results of this dissertation provide valuable biochemical information of ADC, DDC, DHPAA synthase, and GADL1, all of which are PLP-dependent decarboxylases. ADC, DDC, DHPAA synthase are important enzymes in insect cuticle hardening by contributing to the biosynthesis of sclerotizing reagents. Knowledge towards understanding of these enzymes will promote the comprehension of insect cuticle hardening and help scientists to search for ideal insecticide targets. The characterization of GADL1 lays groundwork for future research of its potential role in taurine and beta-alanine metabolism. / Ph. D.

Master of Science / Department of Food Science / Karen A. Schmidt / Milk proteins are excellent sources of sulfur-containing amino acids methionine and cysteine, in particular whey proteins. Cysteine is synthesized from methionine by γ-cystathionase. However, cysteine has to be included in the diets of certain subpopulations due to diminished γ-cystathionase activity. Cysteine, a heat- liable amino acid, may lose bioavailability during thermal processing. The objective of this research was to enhance cysteine content in yogurt while maintaining its quality. First, yogurt mixes were formulated to a total solids content of 12.5% with nonfat dry milk (NDM) (N) or a combination of NDM (10%) and whey protein isolate (WPI) (2.5%) (W), and processed at 70°C (20 min) (70) or 90°C (7 min) (90). Yogurt was prepared and maintained at 4oC for 60 days. Three replications were performed and data were analyzed using SAS®. The W mixes had 65%, 32% and 190% more cysteine, true protein and whey protein contents respectively, compared to N mixes prior to processing. However in day 1 yogurt, the highest cysteine content (398.3 mg/L) was found in the W70 yogurt and its gel quality was comparable to the N90 yogurt except for firmness. During a 60 day storage period the W70 and N90 were similar in gel quality except for firmness. Secondly, a hedonic test was done on the W70 (HC) and N90 (LC) yogurts which had been reformulated to contain sugar and vanillin. One replication was performed and data were analyzed using SAS®. The LC and HC yogurts did not vary in liking of flavor (6.1), aftertaste (6.1) and overall acceptability (6.3) corresponding to the words of “like slightly” when compared. However, the appearance of the LC yogurt was liked more than the HC yogurt (6.7 vs. 6.1) whereas the thickness of HC yogurt was liked more than the LC yogurt (6.4 vs. 5.8). These results suggest that addition of WPI along with lower process treatment resulted in yogurt with enhanced cysteine; however, further studies may be needed to optimize the WPI addition to improve the visual characteristics of the yogurt for consumer acceptance.

In 2007, a gene homolog of CDO encoded by the gene Gm237 in the DUF164 family was identified as cysteamine dioxygenase (ADO). ADO is one of the only known thiol dioxygenases found in mammals. Both ADO and its partner cysteine dioxygenase (CDO) are non-heme iron dependent enzymes that play a crucial role in the biosynthesis of taruine/hypotaurine by insertion of a dioxygen molecule. However, ADO has been overshadowed by CDO as heavy research focus on CDO over the past decade has led to the elucidation of its structure and possible mechanistic properties. In an effort to further understand ADO’s mechanism and regulating role in vivo, this work will be focused on the mammalian hADO and trying to gain further insight on hADO’s structural features via crystallography work. Investigation of the crystallization parameters for hADO has elucidated several potential conditions. Detailed work on these crystallization parameters will be presented.

A partial cystatin cDNA from rainbow trout was generated by reverse transcription polymerase chain reaction with two degenerate primers. The partial cystatin PCR product was 168 bp and used to screen trout liver λgt 11 cDNA library. Four positive clones were isolated and designated as cstl, cst2, cst3 and cst4. Only cst2 contained the full-length cystatin cDNA which was 674 bp and included 5' untranslated region and the polyadenylation signal sequence AATAAA in the 3' region. Translation of the cDNA contains 132 amino acid residues. Comparison of the amino acid sequence with those of family II cystatin indicated that the 21 amino acids at N-terminal end is a signal peptide that leads to cystatin secretion, and the 111 amino acids are mature cystatin. Four cysteine residues in the cystatin may form two disulfide bonds for the secondary structure. Cst2 was subcloned into pGEM-3z for Northern and Southern blot experiments. Northern blot indicated that trout cystatin mRNA is about 750 bp. Cystatin is expressed in all tissues examined but at various levels. This difference may reflect the regulation of cysteine proteinase activities. Southern blot of trout genomic DNA showed that the copy number of the trout cystatin gene is probably one per haploid genome. / Graduation date: 1997

African animal trypanosomosis or nagana is a disease in livestock caused by various species of protozoan parasites belonging to the genus Trypanosoma particularly T. congolense, T. vivax and T. b. brucei. Nagana is the most important constraint to livestock and mixed crop-livestock farming in tropical Africa. Trypanosomes undergo part of their developmental life in their insect vector, the tsetse fly and part in their mammalian host. Measures for eradicating the continent of the tsetse fly vector include insecticidal spraying, targeting and trapping. Vaccine development has been hampered by the generation of an inexhaustible collection of variant surface glycoproteins that trypanosomes possess and allow for evasion of the host immune system. Anti-disease vaccines aimed at reducing the symptoms of the disease rather than killing the parasite itself have been demonstrated as an alternative approach. Trypanotolerant cattle are able to protect themselves from the disease-associated symptoms. They are able to mount a better antibody response to the CATL-like cysteine peptidase, TcoCATL, compared to trypanosusceptible breeds. Bovine trypanosomosis, however, continues to be controlled primarily by trypanocidal compounds such as isometamidium chloride, homidium and diaminazene that have been developed more than 50 years ago and consequently drug resistance is widespread. Trypanosomal cysteine peptidases have also been proven to be effective targets for chemotherapeutics. TcrCATL, inhibited by the vinyl sulfone pseudopeptide inhibitor K11777, was effective in curing or alleviating T. cruzi infection in preclinical proof-of-concept studies and has now entered formal preclinical drug development investigation. Understanding enzymatic as well as structural characteristics of pathogenic peptidases is the first step towards successful control of the disease. To date no such characterisation of the major cysteine peptidases from T. vivax has been conducted. Although the major cysteine peptidase from T. vivax, TviCATL, has not been proven as a pathogenic factor yet, its high sequence identity with the pathogenic counterparts such as TcrCATL and TcoCATL hold much speculation for TviCATLs role in pathogenocity. In the present study, native TviCATL was isolated from T. vivax Y486, purified and characterised. TviCATL showed to have a general sensitivity to E-64 and cystatin and has a substrate specificity defined by the S2 pocket. TviCATL exhibited no activity towards the CATB-like substrate, Z-Arg-Arg-AMC but was able to hydrolyse Z-Phe-Arg-AMC, the CATL-like substrate. Leu was preferred in the P2 position and basic and non-bulky hydrophobic residues were accepted in the P1 and P3 positions respectively. Similar findings were reported for TcoCATL. The substrate specificity of TviCATL and TcoCATL does argue for a more restricted specificity compared to TcrCATL. This was based on the Glu333 in TcrCATL substituted with Leu333 in TviCATL and TcoCATL. In the case of TcrCATL, the Glu333 allows for the accommodation of Arg in the P2 position. Like other trypanosomal cysteine peptidases, TviCATL was inhibited by both chloromethyl ketones, Z-Gly-Leu-Phe-CMK and H-D-Val-Phe-Lys-CMK. Determining further structural and functional characteristics as well as whether TviCATL, like the T. congolense homolog, TcoCATL, acts as a pathogenic factor, would be important information to the designing of specific chemotherapeutic agents. To date, TcrCATL and TbrCATL (from T. b. rhodesiense) are the only trypanosomal CATL-like cysteine peptidases been crystallised and their tructures solved. This advantage has allowed for the directed design of synthetic peptidase inhibitors. The crystal structure of TcoCATL will be of major significance to the design of specific chemotherapeutic agents. Furtherrmore, understanding the dimeric conformation of TcoCATL is important for vaccine design as immune responses are likely to recognise the dimer specific epitopes. In the current study, the catalytic domain of TcoCATL and TviCATL, were recombinantly expressed in Pichia pastoris and purified to homogeneity. The T. congolense cysteine peptidase pyroglutamyl peptidase (PGP), also proven to be pathogenic in T. b. brucei, was recombinantly expressed in E. coli BL21 (DE3) cells and also purified to homogeneity. Purified cysteine peptidases along with previously purified TcoCATL dimerisation mutants, TcoCATL (H43W) and TcoCATL (K39F; E44P), possessing mutated residues involved in TcoCATL dimerisation, as well as the mutant proenzyme TcoCATL (C25A), were screened for crystallisation conditions using the Rigaku robotic crystallisation suite. One-dimensional needle-like crystals were found for TcoCATL (K39F; E44P). Optimisation of the TcoCATL (K39F; E44P) crystals were analysed for X-ray diffraction. The poor diffraction pattern prompted further optimisations for better crystal quality, which is presently underway. The crystal structure of TcoCATL, with some of the residues involved in dimerisation mutated, will be pivotal in understanding the dimerisation model. Furthermore, the information about the structure will be valuable for vaccine design and chemotherapeutics development. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2011.

<p>The Ras family of small GTPases, comprised of the KRAS, NRAS, and HRAS genes, are mutated to encode constitutively-active, GTP-bound, oncogenic proteins in upwards of one quarter or more of all human cancers, which is well established to promote tumorigenesis. Despite the prominent role these genes play in human cancer, the encoded proteins have proven difficult to pharmacologically inhibit. Therefore, it is important to understand how Ras proteins are activated. </p><p>RAS proteins cycle between a GDP-bound inactive state and a GTP-bound active state through guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs). GEFs facilitate the GDP-to-GTP exchange of RAS and promote RAS activation. Similar to GEFs, reactive oxygen/nitrogen species can also promote RAS activation through reactions with the thiol residue of cysteine 118 (C118). This residue may therefore play a role in RAS activation in cancer. To test this possibility, I investigated the effect of mutating C118 to serine (C118S) in Kras on (1) carcinogen-induced lung tumorigenesis, and (2) xenograft tumor growth of HRAS12V-transformed cells.</p><p>To explore the impact of the C118S mutation in Kras on carcinogen-induced lung tumorigenesis, I introduced a C118S mutation into the endogenous murine Kras allele and exposed the resultant mice to the carcinogen urethane, which induces Kras mutation-positive lung tumors. Kras+/C118S and KrasC118S/C118S mice developed fewer and smaller lung tumors than Kras+/+ mice. Although the KrasC118S allele did not appear to affect tumorigenesis when the remaining Kras allele was conditionally oncogenic (KrasG12D), there was a moderate imbalance of oncogenic mutations favoring the native Kras allele in tumors from Kras+/C118S mice treated with urethane. Therefore, mutating C118 of Kras impedes urethane-induced lung tumorigenesis.</p><p>To explore the the impact of the C118S mutation in Kras on xenograft tumor growth of HRAS12V-transformed cells, I tested and found that redox-dependent reactions with cysteine 118 (C118) and activation of wild type KRAS are critical for oncogenic HRAS-driven tumorigenesis. Such redox-dependent activation of KRAS affected both PI3K-AKT and RAF-MEK-ERK pathways. These findings were confirmed in the endogenous mouse Kras gene. Speicfically, oncogenic HRAS-transformed KrasC118S/C118S MEFs grew in soft agar and as xenograft tumors more slowly than similarly transformed Kras+/+ MEFs, suggesting that redox-dependent reactions with C118 of Kras promotes transformation and tumorigenesis. </p><p>Taken together, I have demonstrated a critical role of redox-dependent reactions with Kras C118 in tumorigenesis.</p> / Dissertation

Autosomal dominant adult onset neuronal ceroid lipofuscinoses (ANCL/CLN4) is a rare neurodegenerative disorder caused by mutations in the human gene DNAJC5 which encodes cysteine string protein alpha (CSPα). ANCL is characterized by the appearance of aberrant lysosomal storage material in the post-mortem brains of patients, who usually die from widespread neuronal loss within 10 years from the onset of symptoms. CSPα is a neuroprotective co-chaperone specifically localized to synaptic vesicles (SVs) and is evolutionarily conserved in all animals. CSPα forms a chaperone complex with HSC70 to properly fold a limited number of synaptic proteins. Complete loss of CSP leads to neurodegeneration and reduced lifespans in flies and mice. However, the mechanism of degeneration induced by ANCL mutations is currently unknown and there are no available animal models to study the dysfunctional proteins in situ. In this thesis, I describe the generation and subsequent characterization of the first animal model of ANCL, using the fruit fly Drosophila melanogaster. First, I show that human CSPα (hCSPα) is conserved functionally from humans to flies. Wildtype hCSPα expressed in flies localizes properly to SVs and is able to rescue lifespan defects in CSP null mutant flies. Overexpression of hCSPα proteins with the ANCL causing L115R and L116Δ mutations recapitulates numerous phenotypes consistent with human disease pathology. This includes the appearance of high molecular weight (HMW) SDS-resistant aggregates on western blots, accumulation of aberrant osmophilic membrane structures observed via electron microscopy, and a dose-dependent reduction in adult viability. Mutant hCSPα is mislocalized from SVs to enlarged abnormal endosomes, which accumulate in neuronal axons and somata. These endosomes strongly co-localize with the endosomal sorting required for transport (ESCRT) complex protein HRS, contain large amounts of ubiquitinated proteins, and lack markers of lysosomal maturation. This suggests that the ANCL causing mutations may cause disruptions in endo-lysosomal trafficking via an ESCRT related mechanism. To probe the genetic nature of the mutant alleles I expressed the mutant hCSPα transgenes with various doses of endogenous Drosophila CSP (dCSP). I show that loss of dCSP suppresses toxicity, as well as the aberrant endosomal accumulations and HMW aggregates induced by overexpression of mutant hCSPα. Additionally, expression of a combination of the wildtype and mutant hCSPα showed a super-additive effect on viability and HMW aggregates. This suggests that the disease-causing mutations may act as hypermorphic gain of function alleles, contrary to existing models, which suggest a dominant-negative mechanism. I also performed an F1 candidate screen for genetic modifiers of toxicity, using a robust and easy-to-score adult eye morphology and pigmentation phenotype. Using this approach, I discovered several strong interactors, both enhancers and suppressors, including member of the ESCRT trafficking pathway and other known CSP-interacting proteins. Of particular interest was the CSP co-chaperone Hsc70, which had several loss of function alleles among the strongest observed suppressors. Loss of Hsc70 also greatly reduces toxicity and endosomal accumulations of overexpressed mutant hCSPα but interestingly does not have a significant effect on the levels of HMW CSPα aggregates. This further supports the model that ANCL mutations act as hypermorphs, with a toxic mechanism involving CSP’s endogenous interactions with HSC70. Finally, I discuss the implications of these findings in relation to previous studies of the ANCL causing mutations and endogenous CSPα/HSC70 function and propose a novel mechanistic disease model. This model postulates that mutant CSP is properly trafficked to synapses but, after a brief lifespan as a properly functioning HSC70 co-chaperone, is then ubiquitinated and localized onto endosomes. Ubiquitinated mutant CSP is then clustered by HRS but is unable to mature properly through an ESCRT dependent degradation pathway. These endosomes are retrogradely trafficked through the axon to the soma where they fuse, accumulate, and persist, eventually leading to cellular toxicity via an unknown mechanism. The hypermorphic nature of the mutants can be explained by the novel observation that normal endogenous CSP also traffics through a retrograde ESCRT dependent pathway, where it intersects and co-accumulates with mutant CSP, potentially contributing to toxicity.

The present study has resulted in the development of a procedure for the specific chemical fragmentation of human phosphoglucose isomerase into a minimal number of peptides. A two-cycle procedure for cleaving the protein with 2-nitro-5- thiocyanobenzoic acid results in four primary peptides and three overlap peptides. The peptides can be readily separated on the basis of their size by using sodium dodecyl sulfate polyacrylamide gel electrophoresis. Preliminary peptide alignments have been considered, and amino acid analyses have been performed. End-terminal analyses of the enzyme revealed a carboxyl terminal sequence of Asp-Val-Gln and a blocked amino terminus. The cysteine cleavage procedure provides an excellent method for the identification and location of specific genetic mutations of human phosphoglucose isomerase.