Uma etapa essencial no ciclo de vida do vírus HIV é a integração do DNA viral no cromossomo hospedeiro. Essa etapa é catalisada pela enzima integrase (IN) de 32-kDa. HIV-1 IN é um importante e validado alvo, e as drogas que inibem seletivamente a enzima, quando utilizadas em combinação com os inibidores da transcriptase reversa (RT) e protease (PR), são consideradas altamente eficazes em suprimir a replicação viral. IN catalisa dois processos enzimáticos designados por 3\' processamento e transferência de DNA. Agentes ativos contra integrase, inibindo a etapa de transferência da vertente já estão em fase clínica. O fármaco Raltegravir® é o primeiro nesta nova classe. Os ensaios clínicos no tratamento em novos pacientes têm uma atividade anti-retroviral potente e bem tolerado. Dada a sua potência, segurança e novo mecanismo de ação, os inibidores da integrase representam um importante avanço terapêutico contra o HIV-1. Na presente tese de doutorado, foram realizados estudos quimiométricos utilizando descritores teóricos e QSAR bi- (2D) e tridimensionais (3D) empregando, respectivamente, as técnicas holograma QSAR (HQSAR) e a análise comparativa dos campos moleculares (CoMFA), visando à geração de modelos preditivos para um conjunto de inibidores da integrase do vírus HIV-1. Modelos de QSAR com boa consistência interna, habilidade preditiva e estabilidade foram obtidos em todos os casos. Os modelos gerados, associados às informações obtidas pelos mapas de contribuição 2D e de contorno 3D, são guias químico-medicinais úteis no planejamento de novos inibidores mais potentes e seletivos da integrase do HIV-1. / An essential step in the HIV life cycle is integration of the viral DNA into the host chromosome. This step is catalyzed by a 32-kDa viral enzyme HIV integrase (IN). HIV-1 IN is an important and validated target, and the drugs that selectively inhibit this enzyme, when used in combination with reverse transcriptase (RT) and protease (PR) inhibitors, are believed to be highly effective in suppressing the viral replication. IN catalyzes two discrete enzymatic processes referred as 3\' processing and DNA strand transfer. Agents active against HIV-1, which target the viral integrase by inhibiting the strand transfer step of integration, have now initialized the clinical trials. The Raltegravir® is the first drug in this new class. Clinical trials in treatment-experienced and in treatment-naive patients have shown that raltegravir-containing regimens have potent antiretroviral activity and are well tolerated. Given their potency, safety and novel mechanism of action, integrase inhibitors represent an important advance in HIV-1 therapy. In the present thesis, Bi- and Tridimensional Quantitative Structure-Activity Relationship (QSAR) studies were performed applying chemometric methods based on theoretical descriptors, Comparative Molecular Field Analysis (CoMFA) and Holograma QSAR (HQSAR) techniques, aiming to generate predictive models for a large set of HIV-1 IN inhibitors. QSAR models presenting good internal consistency, predictive power and stability were obtained in all cases. The final models along with the information resulted by 2D contribution and 3D contour maps should be useful in the design of new inhibitors with increased potency and selective within the chemical diversity of the data.

As metaloproteinases de matriz (MMP) são enzimas superexpressas em quase todos os tumores humanos, sendo que os subtipos MMP-2 e MMP-9 têm sido associados ao potencial metastático e prognóstico desfavorável em neoplasias malignas como, por exemplo, melanoma metastático e glioma. Compostos capazes de inibir a atividade destas enzimas podem representar potenciais agentes terapêuticos. O composto 4-nerolidilcatecol (4-NC), isolado de plantas do gênero Pothomorphe, apresentou resultados promissores para o tratamento do melanoma e glioma e foi capaz de atuar em várias etapas bioquímicas importantes envolvidas na progressão dessas patologias, inclusive inibindo MMP-2 e MMP-9. No entanto, o mecanismo de ação do 4-NC não está completamente elucidado. O presente estudo envolveu a aplicação de métodos de modelagem molecular e de formalismos do planejamento de novas moléculas auxiliado por computador, CAMD (Computer-Aided Molecular Design) a fim de explorar a interação entre esta molécula e as enzimas MMP-2 e MMP-9, além de planejar novos inibidores para estes alvos. Análise exploratória de dados, que compreende a análise de agrupamentos hierárquicos e de componentes principais. foi desenvolvida para um conjunto de hidroxamatos (N=64) descritos como inibidores de MMP-2 e MMP-9, a fim de identificar as propriedades moleculares que mais influenciavam o processo de discriminação dos compostos. As propriedades termodinâmicas, eletrônicas e estéricas foram importantes para descrever os compostos mais ativos no conjunto de dados da MMP-2. Para a MMP-9, o coeficiente de distribuição (ClogD) em pH 1,5 foi relevante no processo de discriminação do conjunto. A presença de substituintes volumosos na porção R3 parece ser crucial para o conjunto de inibidores investigados. Esta região está envolvida em interações moleculares com a cavidade S1 de ambas as enzimas, mas há um limite de volume a ser considerado para estes substituintes. O formalismo QSAR-4D independente do receptor (IR) foi aplicado ao mesmo conjunto de dados e permitiu estabelecer o mapeamento do farmacóforo, além de explorar diferentes alinhamentos para a obtenção da hipótese de conformação bioativa prevista pelo melhor modelo de QSAR. OS modelos QSAR apresentaram boa capacidade de previsão, auxiliaram na proposição de novos inibidores e estimaram a atividade do 4-NC. Com o melhor modelo QSAR para MMP-9 (N=64), a atividade prevista para o 4-NC foi classificada na faixa dos inibidores com atividade moderada. Entretanto, o melhor modelo QSAR obtido para MMP-2 (N=38) não foi capaz de prever, de forma adequada, a atividade de compostos com arcabouço químico diferente daqueles utilizados na construção dos modelos. Estudos de ancoramento molecular foram desenvolvidos para investigar a orientação do 4-NC no sitio catalítico das duas enzimas e as interações que poderiam ser estabelecidas nestes complexos. Duas conformações favoráveis foram encontradas. Simulações computacionais de dinâmica molecular foram desenvolvidas com os complexos mais promissores selecionados nos estudos de ancoramento, a fim de obter informações mais detalhadas e de maior confiabilidade. sobre suas interações intermoleculares. O 4-NC tende a se orientar no sítio de forma a acomodar sua cadeia lateral no bolso S1 adjacente ao sítio catalítico em ambas as enzimas. Ensaios de zimografia também foram realizados com o objetivo de elucidar possíveis contribuições da cadeia lateral e do núcleo catecólico do 4-NC na atividade inibitória frente às enzimas em estudo. O núcleo catecólico parece ser o responsável por sua atividade, pois o composto 1,2dimetoxibenzeno, que possui as hidroxilas bloqueadas por grupos metil, não foi capaz de exercer atividade inibitória significante frente à MMP-2 e MMP-9. Estudos de voltametria reforçaram a hipótese de que o 4-NC tem a capacidade de quelar os íons zinco presentes no tampão de incubação. / Matrix metalloproteinases (MMP) enzymes are overexpressed in almost all human tumors, and MMP-2 and MMP-9 subtypes have been associated with metastatic potential and poor prognosis in malignant tumors, such as metastatic melanoma and glioma. Compounds capable of inhibiting the activity of theses enzymes would be considered as potential therapeutic agents. The 4-nerolidylcatechol compound (4-NC), isolated from plants of genus Pothomorphe, has showed promising results in the treatment of melanoma and glioma, and was able to act in several important biochemical steps involved in the progression of these diseases, as well as inhibiting MMP-2 and MMP-9. However, the 4-NC mechanism of action is not completely understood. This study has involved the application of molecular modeling methods and formalisms of computer-aided molecular design (CAMD) in order to explore the interaction between 4-NC and MMP-2/MMP-9, and to design new inhibitors for these targets. Exploratory data analysis, which comprises hierarchical cluster analysis and principal components analysis, was performed to a set of hydroxamates (N=64). previously reported as MMP-2 and MMP-9 inhibitors, in order lo identify the molecular properties that is most critical for the discrimination process regarding the investigated compounds. The thermodynamic, electronic, and steric properties were: quite important to describe the highly active compounds in the data set of MMP-2, whereas the apparent partition coefficient (ClogD) at pH 1.5 was the property more relevant for MMP-9 data set. The presence of bulky substituents on the R3 moiety seems to be crucial for this set of inhibitors due to the molecular interaction with the S1 subsite of both enzymes. However, there is a limit regarding the substituents volume in this region. Receptor independent (RI) 4D-QSAR analysis was applied lo the same data set and it was possible to establish the pharmacophore mapping, besides to explore different alignments in order to generate the hypothesized bioactive conformation through the best QSAR model. The QSAR models have presented good predictability, assisted in proposing new inhibitors, and estimated the activity of 4-NC. Regarding the best QSAR model for MMP-9 (N=64), the 4-NC predicted activity was classified in the range of the moderate active inhibitors. The best QSAR model obtained for MMP-2 (N=38), however was not able to properly predict the activity for compounds with different chemical scaffold from those used to build up the QSAR model. Molecular docking studies have been developed to investigate the 4-NC binding mode into the catalytic site of the two enzymes and the interactions that could be established in those complexes. The results have shown two favorable conformers regarding the MMP inhibition. Molecular dynamics computational simulation were combined to molecular docking studies in order to obtain more detailed and reliable information regarding the intermolecular interactions of each complex. The 4-NC molecule tends to accommodate the side chain in the S1 pocket adjacent to the catalytic site in both enzymes. Experimental zymography assays were also performed to elucidate the possible contribution of the side chain and the catechol core in the 4-NC inhibitory activity against the MMP-2 and MMP-9 enzymes. The catechol core seems to be responsible for its activity, since the 1,2 dimethoxybenzene compound, which has the hydroxyl blocked by a methyl group, was not able to exert any significant inhibition on enzymes. Voltametric assays confirmed the hypothesis that 4-NC chelates zinc ions present in the incubation buffer.

Non-steroidal anti-inflammatory drugs (NSAIDs) are one of the most commonly used classes of drugs. They target the cyclooxygenases (COX) 1 and 2 to reduce the physiological responses of pain, fever, and inflammation. Due to their role in inducing angiogenesis, COX proteins have also been identified as targets in cancer therapies. In this thesis, I describe computational protocols of molecular docking, molecular dynamics simulations and free energy calculations. These methods were used in this thesis to determine structure-activity relationships of a diverse set of NSAIDs in binding to their target proteins COX-1 and 2. Binding affinities were calculated and used to predict the binding modes. Based on combinations of molecular dynamics simulations and free energy calculations, binding mechanisms of sub-classes of NSAIDs were also proposed. Two stable conformations of COX were probed to understand how they affect inhibitor affinities. Finally, a brief discussion on selectivity towards either COX isoform is discussed. These results will be useful in future de novo design and testing of third-generation NSAIDs.

Les antibiotiques sont les médicaments les plus utilisés dans la médecine moderne. Depuis leurs découvertes, ils ont drastiquement changé la façon dont les infections sont traitées. Toutefois, à travers le processus d’adaptation, les bactéries deviennent éventuellement résistantes aux antibiotiques. Malgré leur omniprésence dans la biosphère, l’émergence de souches résistantes est favorisée par le mauvais usage des antibiotiques, ce qui crée une menace importante pour la santé publique. Les antibiotiques actuelles perdent graduellement leur efficacité, et vue le faible nombre de nouvelles molécules développées, la priorité est donnée pour la découverte de nouvelles stratégies capables de combattre les pathogènes. Les nouvelles cibles thérapeutiques idéales doivent exercer une faible pression évolutive, diminuer la virulence et être unique aux microorganismes. Une façon d’atteindre cet objectif est d’interférer dans la régulation et l’homéostasie du Fer chez les bactéries. La biodisponibilité du Fer a fortement influencé l’émergence de la vie sur terre et les stratégies évolutives qu’elle a adoptée. Ce qui a mené à l’apparition d’un mécanisme central de détection du Fer assurant la régulation de cet élément de haute importance. Ce senseur est un point faible que nous pourrons exploiter dans notre combat contre les infections bactériennes. La protéine Fur, pour « Ferric Uptake Regulator », est un régulateur de transcription métal dépendant qui est impliqué dans vaste réseau de régulation contrôlant principalement l’homéostasie du Fer et l’expression de facteurs de virulence. Le travail présenté dans ce manuscrit complète les études précédentes sur des inhibiteurs de la protéine Fur en utilisant une approche combinée théorique et expérimentale grâce a des expériences de XAS, SAXS et MALLS associé a de la modélisation moléculaire. Nous décrivons pour la première fois la structure de Fur d’E. coli ainsi que la structure d’un tétramère de Fur d’un mutant de P. aeruginosa. Par ailleurs, les profils d’énergie libre des protéines Fur de différentes espèces ont été déterminé, pour des complexes tetramériques ou dans le cas de dimères liés à l’ADN, permettant une compréhension préliminaire de leur mécanistique. Les informations structurales obtenues grâce aux travaux présentés dans ce manuscrit permettront de mieux comprendre les mécanismes d’inhibition des protéines Fur ainsi que fournir de nouvelles opportunités pour le développement de molécules a visée thérapeutique. / The most commonly prescribed drugs in human medicine are antibiotics. Since their discovery, they have drastically impacted the way we treat infections. However, a bacterium eventually becomes resistant to antimicrobial treatment through the natural process of adaptative evolution. Even if resistant bacteria are omnipresent in the biosphere, their emergence rate is accelerated by the misuse of antimicrobial agents leading to the public health threat we are facing now. As currently available antimicrobial agents lose their effectiveness and very few new drugs are being developed, a breakthrough in new strategies to fight pathogens should be a priority. Ideal new therapeutic targets should exert weak evolutionary pressure, disarm or weaken the pathogen and be unique to microorganisms. One way to do so is by interfering with the iron regulation and its homeostasis within Bacteria. The bioavailability of iron strongly influenced early life and the metabolic strategies that sustained it. A central iron sensing mechanism evolved to ensure the regulation of such an important element. Sadly for bacteria this sensor became an exploitable weakness in our battle against infection. The “Ferric Uptake Regulator” is a metal dependent transcription regulator with a large regulatory network controlling iron homeostasis and bacterial virulence. This work continues previous investigations on Fur inhibitors using a combined experimental and theoretical approach by performing XAS, SAXS and MALLS experiments together with computer simulations. We describe for the first time the structures of Fur from E. coli in addition to a tetrameric Fur structure of a mutant from P. aeruginosa. Moreover, free energy profiles of Fur proteins, as tetramers or dimers bound to DNA, from different species were generated and key residues involved in the interactions determined, providing mechanistic insights into Fur complexes. The structural information gathered from this work will be used to better understand inhibition mechanisms of Fur proteins providing new opportunities to overcome drug development challenges.

The glucose transport protein, GLUT1, is highly expressed in rapidly proliferating cells, including cancer cells, while decreased GLUT1 levels are found in diseases such as GLUT1 deficiency syndrome and Alzheimer’s. There is increased interest in developing GLUT1 inhibitors as novel anticancer therapeutics, and the discovery of compounds that directly stimulate GLUT1 function. This work investigates how small molecules stimulate and/or inhibit GLUT1-mediated glucose transport, either directly or through the AMPK pathway. Using sugar transport assays and docking analyses to explore Ligand–GLUT1 interactions and specificity of binding, we show that: 1) Ligands inhibit GLUT1 by competing with glucose for binding to the exofacial or endofacial sugar binding sites; 2) Subsaturating inhibitor concentrations stimulate sugar uptake; 3) Ligands inhibit GLUT1–, GLUT3– and GLUT4–mediated sugar uptake in HEK293 cells; and 4) Inclusion of a benzonitrile head group on endofacial GLUT1 inhibitors confers greater inhibitory potency. Furthermore, we investigated AMPK-regulated GLUT1 trafficking in cultured blood-brain barrier endothelial cells, and show that inhibition of GLUT1 internalization is not responsible for increased cell surface levels of GLUT1 observed with AMPK activation in these cells. This study provides a framework for screening candidate GLUT1 inhibitors for specificity, and for optimizing drug design and delivery. Our data on transport stimulation at low inhibitor concentrations support the idea that GLUT1 functions as a cooperative oligomer of allosteric alternating access subunits.

Philosophiae Doctor - PhD / Given the significant disease burden caused by human coronaviruses, the discovery of an effective antiviral strategy is paramount, however there is still no effective therapy to combat infection. This thesis details the in silica exploration of ligand libraries to identify candidate lead compounds that, based on multiple criteria, have a high probability of inhibiting the 3 chymotrypsin-like protease (3CUro) of human coronaviruses. Atomistic models of the 3CUro were obtained from the Protein Data Bank or theoretical models were successfully generated by homology modelling. These structures served the basis of both structure- and ligand-based drug design studies. Consensus molecular docking and pharmacophore modelling protocols were adapted to explore the ZINC Drugs-Now dataset in a high throughput virtual screening strategy to identify ligands which computationally bound to the active site of the 3CUro . Molecular dynamics was further utilized to confirm the binding mode and interactions observed in the static structure- and ligand-based techniques were correct via analysis of various parameters in a IOns simulation. Molecular docking and pharmacophore models identified a total of 19 ligands which displayed the potential to computationally bind to all 3CUro included in the study. Strategies employed to identify these lead compounds also indicated that a known inhibitor of the SARS-Co V 3CUro also has potential as a broad spectrum lead compound. Further analysis by molecular dynamic simulations largely confirmed the binding mode and ligand orientations identified by the former techniques. The comprehensive approach used in this study improves the probability of identifying experimental actives and represents a cost effective pipeline for the often expensive and time consuming process of lead discovery. These identified lead compounds represent an ideal starting point for assays to confirm in vitro activity, where experimentally confirmed actives will be proceeded to subsequent studies on lead optimization.

PhD (Microbiology) / Department of Microbiology / High genetic diversity is a major contributory factor in the development of drug resistance, in addition to challenges in diagnosis and treatment monitoring in the therapeutics of human immunodeficiency virus (HIV) .Within the wide HIV-1 diversity, differences in mutational frequency, disease progression, drug response and transmission amongst HIV-1 subtypes have been shown. In spite HIV-1 subtype C (HIV-1C) being the most prevalent variant globally, none of the available drugs nor screening assays for inhibitory molecules have been developed targeting the genetics of this important subtype. This study therefore aimed to overexpress and biophysically characterize HIV-1C reverse transcriptase and protease to serve as reagents in the development of assays for routine screening of molecules inhibitory to HIV-1C. Heterologous expression of HIV-1C reverse transcriptase and protease isolates that are prevalent in South Africa was carried out in Escherichia coli (E. coli (BL21-DE3). The secondary and tertiary structures of the proteins were determined using, circular dichroism (CD) and fluorescence spectroscopy respectively. Thereafter, interaction studies to delineate interaction properties of natural products for possible inhibition of protease were conducted. Furthermore, in silico studies to determine binding interactions, further confirmed by in vitro binding assays of a pepsin inhibitor homolog (Bm-33) from Brugia malayi , against protease were also conducted. Expressed reverse transcriptase and protease from the globally prevalent HIV-1C were shown to be structurally and functionally intact for application in downstream HIV-1 inhibition assays. Interaction studies on the other hand revealed successful inhibition of the expressed HIV-1C PR with gallotanin. Furthermore, binding interactions of Bm-33 and HIV-1 PR revealed the first intermolecular interactions of the two molecules displaying possible inhibition of HIV-1 PR / NRF

Philosophiae Doctor - PhD / Tsetse flies are the biological vectors of human and animal trypanosomiasis and hence representant medical and veterinary importance. The sense of smell plays a significant role in tsetse and its ecological interaction, such as finding blood meal source, resting, and larvicidal sites and for mating. Tsetse olfactory behaviour can be exploited for their management; however, olfactory studies in tsetse flies are still fragmentary. Here in my PhD thesis, using scanning electron microscopy, electrophysiology, behaviour, bioinformatics and molecular biology techniques, I have investigated tsetse flies (Glossina fuscipes fuscipes) olfaction using behaviourally well studied odorants, tsetse repellent by comparing with attractant odour. Insect olfaction is mediated by olfactory sensory neurons (OSNs), located in olfactory sensilla, which are cuticular structures exposed to the environment through pore and create a platform for chemical communication. In the sensilla shaft the dendrite of OSNs are housed, which are protected by called the sensillum lymph produced by support cells and contains a variety of olfactory proteins, including the odorant binding protein (OBP) and chemosensory proteins (CSP). While on the dendrite of OSNs are expressed olfactory receptors. In my PhD, studies I tried to decipher the sense of smell in tsetse fly. In the second chapter, I demonstrated that G. f. fuscipes is equipped with diverse olfactory sensilla, that various from basiconic, trichoid and coeloconic. I also demonstrated, there is shape, length, number difference between sensilla types and sexual dimorphism. There is a major difference between male and female, while male has the unique basiconic sensilla, club shaped found in the pits, which is absent from female pits. In my third chapter, I investigated the odorant receptors which are expressed on the dendrite of the olfactory sensory neurons (OSNs). G. f. fuscipes has 42 ORs, which were not functionally characterised. I used behaviourally well studied odorants, tsetse repellents, composed of four components blend. I demonstrated that tsetse repellent is also a strong antifeedant for both G. pallidipes and G. f. fuscipes using feeding bioassays as compared to the attractant odour, adding the value of tsetse repellent. However, the attractant odour enhanced the feeding index. Using DREAM (deorphanization of receptors based on expression alterations of mRNA levels). I found that in G. f. fuscipes, following a short in vivo exposure to the individual tsetse repellent component as well as an attractant volatile chemical, OSNs that respond to these compounds altered their mRNA expression in two opposite direction, significant downregulation and upregulation in their number of transcripts corresponding to the OR that they expressed and interacted with odorant. Also, I found that the odorants with opposite valence already segregate distinctly at the cellular and molecular target at the periphery, which is the reception of odorants by OSNs, which is the basis of sophisticated olfactory behaviour. Deorphanization of ORs in none model insect is a challenge, here by combining DREAM with molecular dynamics, as docking score, physiology and homology modelling with Drosophila a well-studied model insects, I was able to predict putative receptors of the tsetse repellent components and an attractant odour. However, many ORs were neutral, showing they were not activated by the odorants, demonstrating the selectivity of the technique as well as the receptors. In my fourth chapter, I investigated the OBPs structures and their interaction with odorants molecules. I demonstrated that OBPs are expressed both in the antenna, as well as in other tissues, such as legs. I also demonstrated that there are variations in the expression of OBPs between tissues as well as sexes. I also demonstrated that odorants induced a fast alteration in OBP mRNA expression, some odorants induced a decrease in the transcription of genes corresponding to the activated OBP and others increased the expression by many fold in OBPs in live insect, others were neutral after 5 hours of exposure. Moreover, with subsequent behavioural data showed that the behavioural response of G. f. fuscipes toward 1-octen-3-ol decreased significantly when 1-octen-3-ol putative OBPs were silenced with feeding of double-stranded RNA (dsRNA). In summary, our finding whereby odorant exposure affects the OBPs mRNA, their physiochemical properties and the silencing of these OBPs affected the behavioural response demonstrate that the OBPs are involved in odour detection that affect the percept of the given odorant. The expression of OBPs in olfactory tissues, antenna and their interaction with odorant and their effect on behavioural response when silenced shows their direct involvement in odour detection and reception. Furthermore, their expression in other tissues such as legs indicates they might also have role in other physiological functions, such as taste.

Philosophiae Doctor - PhD / Tsetse flies are the biological vectors of human and animal trypanosomiasis and hence representant medical and veterinary importance. The sense of smell plays a significant role in tsetse and its ecological interaction, such as finding blood meal source, resting, and larvicidal sites and for mating. Tsetse olfactory behaviour can be exploited for their management; however, olfactory studies in tsetse flies are still fragmentary. Here in my PhD thesis, using scanning electron microscopy, electrophysiology, behaviour, bioinformatics and molecular biology techniques, I have investigated tsetse flies (Glossina fuscipes fuscipes) olfaction using behaviourally well studied odorants, tsetse repellent by comparing with attractant odour. Insect olfaction is mediated by olfactory sensory neurons (OSNs), located in olfactory sensilla, which are cuticular structures exposed to the environment through pore and create a platform for chemical communication. In the sensilla shaft the dendrite of OSNs are housed, which are protected by called the sensillum lymph produced by support cells and contains a variety of olfactory proteins, including the odorant binding protein (OBP) and chemosensory proteins (CSP). While on the dendrite of OSNs are expressed olfactory receptors. In my PhD, studies I tried to decipher the sense of smell in tsetse fly. In the second chapter, I demonstrated that G. f. fuscipes is equipped with diverse olfactory sensilla, that various from basiconic, trichoid and coeloconic. I also demonstrated, there is shape, length, number difference between sensilla types and sexual dimorphism. There is a major difference between male and female, while male has the unique basiconic sensilla, club shaped found in the pits, which is absent from female pits. In my third chapter, I investigated the odorant receptors which are expressed on the dendrite of the olfactory sensory neurons (OSNs). G. f. fuscipes has 42 ORs, which were not functionally characterised. I used behaviourally well studied odorants, tsetse repellents, composed of four components blend. I demonstrated that tsetse repellent is also a strong antifeedant for both G. pallidipes and G. f. fuscipes using feeding bioassays as compared to the attractant odour, adding the value of tsetse repellent. However, the attractant odour enhanced the feeding index. Using DREAM (deorphanization of receptors based on expression alterations of mRNA levels). I found that in G. f. fuscipes, following a short in vivo exposure to the individual tsetse repellent component as well as an attractant volatile chemical, OSNs that respond to these compounds altered their mRNA expression in two opposite direction, significant downregulation and upregulation in their number of transcripts corresponding to the OR that they expressed and interacted with odorant. Also, I found that the odorants with opposite valence already segregate distinctly at the cellular and molecular target at the periphery, which is the reception of odorants by OSNs, which is the basis of sophisticated olfactory behaviour. Deorphanization of ORs in none model insect is a challenge, here by combining DREAM with molecular dynamics, as docking score, physiology and homology modelling with Drosophila a well-studied model insects, I was able to predict putative receptors of the tsetse repellent components and an attractant odour. However, many ORs were neutral, showing they were not activated by the odorants, demonstrating the selectivity of the technique as well as the receptors. In my fourth chapter, I investigated the OBPs structures and their interaction with odorants molecules. I demonstrated that OBPs are expressed both in the antenna, as well as in other tissues, such as legs. I also demonstrated that there are variations in the expression of OBPs between tissues as well as sexes. I also demonstrated that odorants induced a fast alteration in OBP mRNA expression, some odorants induced a decrease in the transcription of genes corresponding to the activated OBP and others increased the expression by many fold in OBPs in live insect, others were neutral after 5 hours of exposure. Moreover, with subsequent behavioural data showed that the behavioural response of G. f. fuscipes toward 1-octen-3-ol decreased significantly when 1-octen-3-ol putative OBPs were silenced with feeding of double-stranded RNA (dsRNA). In summary, our finding whereby odorant exposure affects the OBPs mRNA, their physiochemical properties and the silencing of these OBPs affected the behavioural response demonstrate that the OBPs are involved in odour detection that affect the percept of the given odorant. The expression of OBPs in olfactory tissues, antenna and their interaction with odorant and their effect on behavioural response when silenced shows their direct involvement in odour detection and reception. Furthermore, their expression in other tissues such as legs indicates they might also have role in other physiological functions, such as taste.

In drug development, molecular docking methods aim at characterizing the binding of a drug-like molecule to a protein. In a typical drug development process, a docking task is repeated millions of time, which makes optimization efforts essential. In particular, modern heterogeneous architectures, such as GPUs, allow for significant acceleration opportunities. AutoDock-GPU, a state-of-the-art GPU-accelerated molecular docking software, estimates the geometrical conformation of a docked ligand-protein complex by minimizing an energy-based scoring function. Our profiling results indicated that a reduction operation, which is performed several millions times in a single docking run, limits performance in AutoDock-GPU. Thus, we proposed a method to accelerate the block-level sum reduction of four-element vectors by using matrix operations. We implemented our method to make use of the high throughput capabilities offered by NVIDIA Tensor Cores to perform matrix operations. We evaluated our approach by designing a simple benchmark, and achieved a 4 to 7-fold runtime improvement compared to the original method. We then integrated our reduction operation into AutoDock-GPU and evaluated it on multiple chemical complexes on three GPUs. This evaluation allowed to assess the possibility to use half-precision reduction operations in parts of AutoDock-GPU code, without detrimental effects on the simulation result. In addition, our implementation achieved an average 27% improvement on the overall docking time during a real-world docking run. / Vid läkemedelsutveckling syftar molekylär dockningsmetoder till att karakterisera bindningen av en läkemedelsliknande molekyl till ett protein. I en typisk läkemedelsutvecklingsprocess upprepas en dockinguppgift miljontals gånger, vilket gör optimeringsinsatser nödvändiga. Framför allt moderna heterogena arkitekturer som GPU:er ger betydande accelerationsmöjligheter. AutoDock-GPU, en modern GPU-accelererad programvara för molekylär dockning, uppskattar den geometriska konformationen hos ett ligand-protein-komplex genom att minimera en energibaserad poängsättningsfunktion. Våra profileringsresultat visade att en reduktionsoperation, som utförs flera miljoner gånger i en enda dockningskörning, begränsar prestandan i AutoDock-GPU. Vi har därför föreslagit en metod för att accelerera summareduktionen på blocknivå av vektorer med fyra element med hjälp av matrisoperationer. Vi implementerade vår metod för att utnyttja den höga genomströmningskapacitet som erbjuds av NVIDIA Tensor Cores för att utföra matrisoperationer. Vi utvärderade vårt tillvägagångssätt genom att utforma ett enkelt testfall och uppnådde en 4- till 7-faldig förbättring av körtiden jämfört med den ursprungliga metoden. Vi integrerade sedan vår reduktionsoperation i AutoDock-GPU och utvärderade den på flera kemiska komplex på tre GPU:er. Denna utvärdering lät oss bedöma möjligheten att använda reduktionsoperationer med halvprecision i delar av AutoDock-GPU-koden, utan negativa effekter på simuleringsresultatet. Dessutom uppnådde vår version en genomsnittlig förbättring på 27% av den totala dockningstiden under en riktig dockningskörning.