Modularität ist ein wiederkehrendes Thema in biologischen Netzwerken, in denen sich unabhängige Elemente wiederholen und neu zusammensetzen, um neue Funktionen zu erzeugen. In der angeborenen Immunität sind supramolekular organisierende Zentren (SMOCs) ein Beispiel für Modularität, die durch eine ligandeninduzierte komplexe Selbstassemblierung gekennzeichnet sind und deren Signalkomponenten in ihrer lokalen Konzentration zunehmen. Das Myddosom ist ein SMOC, das sich als Reaktion auf den Liganden IL-1, ein wichtiges proinflammatorisches Zytokin, zusammensetzt. Der Myddosom-vermittelte IL-1-Signalweg wurde mit biochemischen Methoden umfassend charakterisiert, doch Informationen zur Dynamik sind weiterhin begrenzt. Es gibt widersprüchliche Berichte darüber, wie die Proteine des IL-1-Signalwegs zusammengesetzt werden, und es ist nicht bekannt, ob die Komplexe des IL-1-Signalwegs wieder abgebaut werden. In dieser Arbeit wurde die Dynamik der Myddosom-vermittelten IL-1-Signaltransduktion untersucht, wobei eine völlig neue Hochdurchsatz-Bildanalyse-Pipeline, eine Phasenporträtanalyse, ein neuartiger Mikroskopie-Assay und CRISPR/Cas9-editierte lebende Lymphomzelllinien kombiniert wurden. Hier zeige ich, dass der Aufbau des IL-1-Signalwegs sequenzielle Schritte, eine ligandeninduzierte De-novo-Oligomerisierung, Proteine, die die Oligomergröße regulieren, und zwei Module umfasst: das Myddosom und die NF-κB-Signalosome. Das stromaufwärts gelegene Myddosom-Signalosom (MyD88, IRAK4, IRAK1, TRAF6, TAB2) erscheint zuerst, zeigt keine Erholung in FRAP-Experimenten und zeigt eine positive interne Rückkopplung. Das nachgeschaltete NF-κB-Signalosom (HOIL1, NEMO, RelA, A20) erscheint später, erholt sich nach FRAP und reguliert das Myddosom-Signalosom negativ. Ich zeige auch zum ersten Mal, dass sich der IL-1-Signalweg nach Überschreiten eines kritischen stöchiometrischen Verhältnisses auflöst und dass dynamische Gleichungen den Zusammenbau und den Abbau nachbilden können. Diese Ergebnisse zeigen, wie ein SMOC-abhängiger Signalweg Modularität zur Regulierung einsetzt. Das Verständnis von Signalisierungsmodulen ist von großer Bedeutung für die Aufklärung der regulatorischen Schaltkreise in biologischen Netzwerken. Meine Hochdurchsatz-Pipeline bietet einen vielversprechenden Ansatz, um dieses Ziel zu erreichen. / Modularity is a recurring theme in biological networks where independent elements repeat and shuffle to produce novel functions. In innate immunity, supramolecular organizing centers (SMOCs) are an example of modularity, characterized by ligand-induced complex self-assembly, and signaling components increase in local concentration. The Myddosome is a SMOC that assembles in response to the ligand IL-1, a vital proinflammatory cytokine. The Myddosome-mediated IL-1 signaling pathway has been extensively characterized using biochemical methods, yet dynamic information remains limited. There are conflicting reports of how the IL-1 pathway proteins assemble, and it is unknown whether IL-1 pathway complexes disassemble. This thesis investigated the Myddosome-mediated IL-1 signal transduction dynamics, combining a completely new high-throughput image analysis pipeline, phase portrait analysis, a novel microscopy assay, and CRISPR/Cas9-edited live lymphoma cell lines. Here, I show that the IL-1 pathway assembly has sequential steps, ligand-induced de novo oligomerization, proteins regulating oligomer size, and two modules: the Myddosome and NF-κB signalosomes. The upstream Myddosome signalosome (MyD88, IRAK4, IRAK1, TRAF6, TAB2) appears first, has no FRAP recovery, and shows positive internal feedback. The downstream NF-κB signalosome (HOIL1, NEMO, RelA, A20) appears later, recovers after FRAP, and negatively regulates the Myddosome signalosome. I also show, for the first time, that the IL-1 pathway disassembles after crossing a critical stoichiometric ratio, and that dynamical equations can recapitulate assembly-disassembly. These results demonstrate how a SMOC-dependent pathway uses modularity to achieve regulation. Understanding signaling modules holds significant implications for elucidating the regulatory circuitry in biological networks. My high-throughput pipeline offers a promising approach to achieving this objective.

Um mit ihrer Umgebung zu kommunizieren verfügen lebende Zellen über Rezeptoren, welche die umschließende Membran überbrücken. Die vorherrschende G-Protein-gekoppelte Rezeptoren (GPCR) erhalten Informationen von Außerhalb durch Bindung eines Liganden, wodurch der Rezeptor aktiviert wird. Während der Aktivierung bildet sich innerzellulär ein offener Spalt, in den ein G-Protein (Gαβγ, G) mit seinem C-terminalen Ende koppeln kann. Die Bindung an einen GPCR führt in der Gα-Untereinheit vom Gαβγ zu einen GDP/GTP-Austausch, welcher für die weitere Signalübertragung ins Zellinnere notwendig ist. Die Kopplung von Rezeptor und Gαβγ umfasst eine Reihe von dynamischen strukturellen Änderungen, die Geschwindigkeit und Spezifität der Interaktion regeln. Hier haben wir MD-Simulationen (Molekulardynamik) verwendet, um die molekularen Details der GPCR Gαβγ Kopplung vor und während der GPCR-Gαβγ-Komplexbildung bis hin zum GDP/GTP-Austausch zu untersuchen. / To communicate with their environment, living cells feature receptors that provide a bridge across the enclosing membrane. The prevalent G protein-coupled receptors (GPCR) receive outside information through the binding of a ligand, which activates the receptor. During activation, an open intracellular crevice forms, to which a G protein (Gαβγ, G) can couple with its Gα C-terminus. Binding to GPCRs triggers GDP/GTP exchange in the Gα subunit of Gαβγ, necessary for further signal transfer within the cell. The coupling between receptor and Gαβγ involves a series of dynamic structural changes that govern speed and specificity of the interaction. Here we used molecular dynamics (MD) simulations to elucidate molecular details of the GPCR Gαβγ coupling process before and during GPCR Gαβγ complex formation up to the GDP/GTP exchange.

Aminoacyl-tRNA synthetases (aaRSs) are at the center of the question of the origin of life and are essential proteins found in all living organisms. AARSs arose early in evolution to interpret genetic code and are believed to be a group of ancient proteins. They constitute a family of enzymes integrating the two levels of cellular organization: nucleic acids and proteins. These enzymes ensure the fidelity of transfer of genetic information from the DNA to the protein. They are responsible for attaching amino acid residues to their cognate tRNA molecules by virtue of matching the nucleotide triplet, which is the first step in the protein synthesis. The translation of genetic code into protein sequence is mediated by tRNA, which accurately picks up the cognate amino acids. The attachment of the cognate amino acid to tRNA is catalyzed by aaRSs, which have binding sites for the anticodon region of tRNA and for the amino acid to be attached. The two binding sites are separated by ≈ 76 Å and experiments have shown that the communication does not go through tRNA (Gale et al., 1996). The problem addressed here is how the information of binding of tRNA anticodon near the anticodon binding site is communicated to the active site through the protein structure. These enzymes are modular with distinct domains on which extensive kinetic and mutational experiments and supported by structural data are available, highlighting the role of inter-domain communication (Alexander and Schimmel, 2001). Hence these proteins present themselves as excellent systems for in-silico studies. Various methods involved for the construction of protein structure networks are well established and analyzed in a variety of ways to gain insights into different aspects of protein structure, stability and function (Kannan and Vishveshwara, 1999; Brinda and Vishveshwara, 2005). In the present study, we have incorporated network parameters for the analysis of molecular dynamics (MD) simulation data, representing the global dynamic behavior of protein in a more elegant way. MD simulations have been performed on the available (and modeled) structures of aaRSs bound to a variety of ligands, and the protein structure networks (PSN) of non-covalent interactions have been characterized in dynamical equilibrium. The changes in the structure networks are used to understand the mode of communication, and the paths between the two sites of interest identified by the analysis of the shortest path. The allosteric concept has played a key role in understanding the biological functions of aaRSs. The rigidity/plasticity and the conformational population are the two important ideas invoked in explaining the allosteric effect. We have explored the conformational changes in the complexes of aaRSs through novel parameters such as cliques and communities (Palla et al., 2005), which identify the rigid regions in the protein structure networks (PSNs) constructed from the non-covalent interactions of amino acid side chains. The thesis consists of 7 chapters. The first chapter constitutes the survey of the literature and also provides suitable background for this study. The aims of the thesis are presented in this chapter. Chapter 2 describes various techniques employed and the new techniques developed for the analysis of PSNs. It includes a brief description of well -known methods of molecular dynamics simulations, essential dynamics, and cross correlation maps. The method used for the construction of graphs and networks is also described in detail. The incorporation of network parameters for the analysis of MD simulation data are done for the first time and has been applied on a well studied protein lysozyme, as described in chapter 3. Chapter 3 focuses on the dynamical behavior of protein structure networks, examined by considering the example of T4-lysozyme. The equilibrium dynamics and the process of unfolding are followed by simulating the protein with explicit water molecules at 300K and at higher temperatures (400K, 500K) respectively. Three simulations of 10ns duration have been performed at 500K to ensure the validity of the results. The snapshots of the protein structure from the simulations are represented as Protein Structure Networks (PSN) of non-covalent interactions. The strength of the non-covalent interaction is evaluated and used as an important criterion in the construction of edges. The profiles of the network parameters such as the degree distribution and the size of the largest cluster (giant component) have been examined as a function of interaction strength (Ghosh et al., 2007). We observe a critical strength of interaction (Icritical) at which there is a transition in the size of the largest cluster. Although the transition profiles at all temperatures show behavior similar to those found in the crystal structures, the 500K simulations show that the non-native structures have lower Icritical values. Based on the interactions evaluated at Icritical value, the folding/unfolding transition region has been identified from the 500K simulation trajectories. Furthermore, the residues in the largest cluster obtained at interaction strength higher than Icritical have been identified to be important for folding. Thus, the compositions of the top largest clusters in the 500K simulations have been monitored to understand the dynamical processes such as folding/unfolding and domain formation/disruption. The results correlate well with experimental findings. In addition, the highly connected residues in the network have been identified from the 300K and 400K simulations and have been correlated with the protein stability as determined from mutation experiments. Based on these analyses, certain residues, on which experimental data is not available, have been predicted to be important for the folding and the stability of the protein. The method can also be employed as a valuable tool in the analysis of MD simulation data, since it captures the details at a global level, which may elude conventional pair-wise interaction analysis. After standardizing the concept of dynamical network analysis using Lysozyme, it was applied to our system of interest, the aaRSs. The investigations carried out on Methionyl-tRNA synthetases (MetRS) are presented in chapter 4. This chapter is divided into three parts: Chapter 4A deals with the introduction to aminoacyl tRNA synthetases (aaRS). Classification and functional insights of aaRSs obtained through various studies are presented. Chapter 4B is again divided into parts: BI and BII. Chapter 4BI elucidates a new technique developed for finding communication pathways essential for proper functioning of aaRS. The enzymes of the family of tRNA synthetases perform their functions with high precision, by synchronously recognizing the anticodon region and the amino acylation region, which is separated by about 70Å in space. This precision in function is brought about by establishing good communication paths between the two regions. We have modelled the structure of E.coli Methionyl tRNA synthetase, which is complexed with tRNA and activated methionine. Molecular dynamics simulations have been performed on the modeled structure to obtain the equilibrated structure of the complex and the cross correlations between the residues in MetRS. Furthermore, the network analysis on these structures has been carried out to elucidate the paths of communication between the aminoacyl activation site and the anticodon recognition site (Ghosh and Vishveshwara, 2007). This study has provided the detailed paths of communication, which are consistent with experimental results. A similar study on the (MetRS + activated methionine) and (MetRS+tRNA) complexes along with ligand free-native enzyme has also been carried out. A comparison of the paths derived from the four simulations has clearly shown that the communication path is strongly correlated and unique to the enzyme complex, which is bound to both the tRNA and the activated methionine. The method developed here could also be utilized to investigate any protein system where the function takes place through long distance communication. The details of the method of our investigation and the biological implications of the results are presented in this chapter. In chapter 4BII, we have explored the conformational changes in the complexes of E.coli Methionyl tRNA synthetase (MetRS) through novel parameters such as cliques and communities, which identify the rigid regions in the protein structure networks (PSNs). The rigidity/plasticity and the conformational population are the two important ideas invoked in explaining the allosteric effect. MetRS belongs to the aminoacyl tRNA Synthetases (aaRSs) family that play a crucial role in initiating the protein synthesis process. The network parameters evaluated here on the conformational ensembles of MetRS complexes, generated from molecular dynamics simulations, have enabled us to understand the inter-domain communication in detail. Additionally, the characterization of conformational changes in terms of cliques/communities has also become possible, which had eluded conventional analyses. Furthermore, we find that most of the residues participating in clique/communities are strikingly different from those that take part in long-range communication. The cliques/communities evaluated here for the first time on PSNs have beautifully captured the local geometries in their detail within the framework of global topology. Here the allosteric effect is revealed at the residue level by identifying the important residues specific for structural rigidity and functional flexibility in MetRS. Chapter 4C focuses on MD simulations of Methionyl tRNA synthetase (AmetRS) from a thermophilic bacterium, Aquifex aeolicus. As describe in Chapter 4B, we have explored the communication pathways between the anticodon binding region and the aminoacylation site, and the conformational changes in the complexes through cliques and communities. The two MetRSs from E.coli and Aquifex aeolicus are structurally and sequentially very close to each other. But the communication pathways between anticodon binding region and the aminoacylation site from A. aeolicus have differed significantly with the communication paths obtained from E.coli. The residue composition and cliques/communities structure participating in communication are not similar in the MetRSs of both these organisms. Furthermore the formation of cliques/communities and hubs in the communication paths are more in A. aeolicus compared to E.coli. The participation of structurally homologous linker peptide, essential for orienting the two domains for efficient communication is same in both the organisms although, the residues composition near domain interface regions including the linker peptide is different. Thus, the diversity in the functioning of two different MetRS has been brought out, by comparing the E.coli and Aquifex aeolicus systems. Protein Structure network analysis of MD simulated trajectories of various ligand bound complexes of Escherichia coli Cysteinyl-tRNA synthetase (CysRS) have been discussed in Chapter 5. The modeling of the complex is done by docking the ligand CysAMP into the tRNA bound structure of E.coli Cysteinyl tRNA synthetase. Molecular dynamics simulations have been performed on the modeled structure and the paths of communications were evaluated using a similar method as used in finding communication paths for MetRS enzymes. Compared to MetRS the evaluation of communication paths in CysRS is complicated due to presence of both direct and indirect readouts. The direct and indirect readouts (DR/IR) involve interaction of protein residues with base-specific functional group and sugar-phosphate backbone of nucleic acids respectively. Two paths of communication between the anticodon region and the activation site has been identified by combining the cross correlation information with the protein structure network constructed on the basis of non-covalent interaction. The complete paths include DR/IR interactions with tRNA. Cliques/communities of non-covalently interacting residues imparting structural rigidity are present along the paths. The reduction of cooperative fluctuation due to the presence of community is compensated by IR/DR interaction and thus plays a crucial role in communication of CysRS. Chapter 6 focuses on free energy calculations of aminoacyl tRNA synthetases with various ligands. The free energy contributions to the binding of the substrates are calculated using a method called MM-PBSA (Massova and Kollman, 2000). The binding free energies were calculated as the difference between the free energy of the enzyme-ligand complex, and the free ligand and protein. The ligand unbinding energy values obtained from the umbrella sampling MD correlates well with the ligand binding energies obtained from MM-PBSA method. Furthermore the essential dynamics was captured from MD simulations trajectories performed on E.coli MetRS, A. aeolius MetRS and E.coli CysRS in terms of the eigenvalues. The top two modes account for more than 50% of the motion in essential space for systems E.coli MetRS, A. aeolius MetRS and E.coli CysRS. Population distribution of protein conformation states are looked at the essential plane defined by the two principal components with highest eigenvalues. This shows how aaRSs existed as a population of conformational states and the variation with the addition of ligands. The population of conformational states is converted into Free energy contour surface. From free energy surfaces, it is evident that the E.coli tRNAMet bound MetRS conformational fluctuations are more, which attributes to less rigidity in the complex. Whereas E.coli tRNACys bound CysRS conformational fluctuations are less and this is reflected in the increase in rigidity of the complex as confirmed by its entropic contribution. Future directions have been discussed in the final chapter (Chapter 7). Specifically, it deals with the ab-initio QM/MM study of the enzymatic reaction involved in the active site of E.coli Methionyl tRNA synthetase. To achieve this, two softwares are integrated: the Quantum Mechanics (QM) part includes small ligands and the Molecular Mechanics (MM) part as protein MetRS are handled using CPMD and Gromacs respectively. The inputs for two reactions pathways are prepared. First reaction involves cyclization reaction of homocysteine in the active site of MetRS and the second reaction deals with charging of methionine in the presence of ATP and magnesium ion. These simulations require very high power computing systems and also time of computation is also very large. With the available computational power we could simulate up to 10ps and it is insufficient for analysis. The future direction will involve the simulations of these systems for longer time, followed by the analysis for reaction pathways.

Cette thèse comporte quatre fichiers vidéo. This thesis comes with four video files. / L'industrie chimique mondiale est en pleine mutation, cherchant des solutions pour rendre la synthèse organique classique plus durable. Une telle solution consiste à passer de la catalyse chimique classique à la biocatalyse. Bien que les avantages des enzymes incluent leur stéréo, régio et chimiosélectivité, cette sélectivité réduit souvent leur promiscuité. Les efforts requis pour adapter la fonction enzymatique aux réactions désirées se sont révélés d'une efficacité modérée, de sorte que des méthodes rapides et rentables sont nécessaires pour générer des biocatalyseurs qui rendront la production chimique plus efficace. Dans l’ère de la bioinformatique et des outils de calcul pour soutenir l'ingénierie des enzymes, le développement rapide de nouvelles fonctions enzymatiques devient une réalité. Cette thèse commence par un examen des développements récents sur les outils de calcul pour l’ingénierie des enzymes. Ceci est suivi par un exemple de l’ingénierie des enzymes purement expérimental ainsi que de l’évolution des protéines. Nous avons exploré l’espace mutationnel d'une enzyme primitive, la dihydrofolate réductase R67 (DHFR R67), en utilisant l’ingénierie semi-rationnelle des protéines. La conception rationnelle d’une librarie de mutants, ou «Smart library design», impliquait l’association covalente de monomères de l’homotétramère DHFR R67 en dimères afin d’augmenter la diversité de la librairie d’enzymes mutées. Le criblage par activité enzymatique a révélé un fort biais pour le maintien de la séquence native dans un des protomères tout en tolérant une variation de séquence élevée pour le deuxième. Il est plausible que les protomères natifs procurent l’activité observée, de sorte que nos efforts pour modifier le site actif de la DHFR R67 peuvent n’avoir été que modérément fructueux. Les limites des méthodes expérimentales sont ensuite abordées par le développement d’outils qui facilitent la prédiction des points chauds mutationnels, c’est-à-dire les sites privilégiés à muter afin de moduler la fonction. Le développement de ces techniques est intensif en termes de calcul, car les protéines sont de grandes molécules complexes dans un environnement à base d’eau, l’un des solvants les plus difficiles à modéliser. Nous présentons l’identification rapide des points chauds mutationnels spécifiques au substrat en utilisant l'exemple d’une enzyme cytochrome P450 industriellement pertinente, la CYP102A1. En appliquant la technique de simulation de la dynamique moléculaire par la force de polarisation adaptative, ou «ABF», nous confirmons les points chauds mutationnels connus pour l’hydroxylation des acides gras tout en identifiant de nouveaux points chauds mutationnels. Nous prédisons également la conformation du substrat naturel, l’acide palmitique, dans le site actif et nous appliquons ces connaissances pour effectuer un criblage virtuel d'autres substrats de cette enzyme. Nous effectuons ensuite des simulations de dynamique moléculaire pour traiter l’impact potentiel de la dynamique des protéines sur la catalyse enzymatique, qui est le sujet de discussions animées entre les experts du domaine. Avec la disponibilité accrue de structures cristallines dans la banque de données de protéines (PDB), il devient clair qu’une seule structure de protéine n’est pas suffisante pour élucider la fonction enzymatique. Nous le démontrons en analysant quatre structures cristallines que nous avons obtenues d’une enzyme β-lactamase, parmi lesquelles un réarrangement important des résidus clés du site actif est observable. Nous avons réalisé de longues simulations de dynamique moléculaire pour générer un ensemble de structures suggérant que les structures cristallines ne reflètent pas nécessairement la conformation de plus basse énergie. Enfin, nous étudions la nécessité de compléter de manière informatisée un hémisphère où l’expérimental n’est actuellement pas possible, à savoir la prédiction de la migration des gaz dans les enzymes. À titre d'exemple, la réactivité des enzymes cytochrome P450 dépend de la disponibilité des molécules d’oxygène envers l’hème du site actif. Par le biais de simulations de la dynamique moléculaire de type Simulation Implicite du Ligand (ILS), nous dérivons le paysage de l’énergie libre de petites molécules neutres de gaz pour cartographier les canaux potentiels empruntés par les gaz dans les cytochromes P450 : CYP102A1 et CYP102A5. La comparaison pour les gaz CO, N2 et O2 suggère que ces enzymes évoluent vers l’exclusion du CO inhibiteur. De plus, nous prédisons que les canaux empruntés par les gaz sont distincts des canaux empruntés par le substrat connu et que ces canaux peuvent donc être modifiés indépendamment les uns des autres. / The chemical industry worldwide is at a turning point, seeking solutions to make classical organic synthesis more sustainable. One such solution is to shift from classical catalysis to biocatalysis. Although the advantages of enzymes include their stereo-, regio-, and chemoselectivity, their selectivity often reduces versatility. Past efforts to tailor enzymatic function towards desired reactions have met with moderate effectiveness, such that fast and cost-effective methods are in demand to generate biocatalysts that will render the production of fine and bulk chemical production more benign. In the wake of bioinformatics and computational tools to support enzyme engineering, the fast development of new enzyme functions is becoming a reality. This thesis begins with a review of recent developments on computational tools for enzyme engineering. This is followed by an example of purely experimental enzyme engineering and protein evolution. We explored the mutational space of a primitive enzyme, the R67 dihydrofolate reductase (DHFR), using semi-rational protein engineering. ‘Smart library design’ involved fusing monomers of the homotetrameric R67 DHFR into dimers, to increase the diversity in the resulting mutated enzyme libraries. Activity-based screening revealed a strong bias for maintenance of the native sequence in one protomer with tolerance for high sequence variation in the second. It is plausible that the native protomers procure the observed activity, such that our efforts to modify the enzyme active site may have been only moderately fruitful. The limitations of experimental methods are then addressed by developing tools that facilitate computational mutational hotspot prediction. Developing these techniques is computationally intensive, as proteins are large molecular objects and work in aqueous media, one of the most complex solvents to model. We present the rapid, substrate-specific identification of mutational hotspots using the example of the industrially relevant P450 cytochrome CYP102A1. Applying the adaptive biasing force (ABF) molecular dynamics simulation technique, we confirm the known mutational hotspots for fatty acid hydroxylation and identify a new one. We also predict a catalytic binding pose for the natural substrate, palmitic acid, and apply that knowledge to perform virtual screening for further substrates for this enzyme. We then perform molecular dynamics simulations to address the potential impact of protein dynamics on enzyme catalysis, which is the topic of heated discussions among experts in the field. With the availability of more crystal structures in the Protein Data Bank, it is becoming clear that a single protein structure is not sufficient to elucidate enzyme function. We demonstrate this by analyzing four crystal structures we obtained of a β-lactamase enzyme, among which a striking rearrangement of key active site residues was observed. We performed long molecular dynamics simulations to generate a structural ensemble that suggests that crystal structures do not necessarily reflect the conformation of lowest energy. Finally, we address the need to computationally complement an area where experimentation is not currently possible, namely the prediction of gas migration into enzymes. As an example, the reactivity of P450 cytochrome enzymes depends on the availability of molecular oxygen at the active-site heme. Using the Implicit Ligand Sampling (ILS) molecular dynamics simulation technique, we derive the free energy landscape of small neutral gas molecules to map potential gas channels in cytochrome P450 CYP102A1 and CYP102A5. Comparison of CO, N2 and O2 suggests that those enzymes evolved towards exclusion of the inhibiting CO. In addition, we predict that gas channels are distinct from known substrate channels and therefore can be engineered independently from one another.

No description available.

L'ingénierie enzymatique est un outil très avantageux dans l'industrie biotechnologique. Elle permet d'adapter les enzymes à une activité ou à une condition de réaction spécifique. En outre, elle peut permettre de déchiffrer les éléments clés qui ont facilité leur modification. Bien que l'ingénierie enzymatique soit largement pratiquée, elle comporte encore plusieurs goulets d'étranglement. Certains de ces goulets d'étranglement sont techniques, comme le développement de méthodologies pour la création de banques de mutations ciblées ou la réalisation de criblages à haut débit, et d'autres sont conceptuels, comme le déchiffrage des caractéristiques clés pertinentes d'une protéine cible pour la réussite d'un projet d'ingénierie. Parmi ces défis, l'épistasie intra-génique, ou la non-additivité des effets phénotypiques des mutations, est une caractéristique qui entrave grandement la prévisibilité. L'amélioration de l'ingénierie enzymatique nécessite une approche multidisciplinaire qui inclut une meilleure compréhension des relations structure-fonction-évolution. Cette thèse vise à contribuer à l'avancement de l'ingénierie enzymatique en étudiant deux systèmes modèles. Premièrement, des variantes dynamiques de la ß-lactamase TEM-1 ont été choisies pour étudier le lien entre la dynamique des protéines et l'évolution. La ß-lactamase TEM-1 a été largement caractérisée dans la littérature, ce qui s'est traduit par des connaissances approfondies sur son mécanisme de réaction, ses caractéristiques structurelles et son évolution. Les variantes de la ß-lactamase TEM-1 utilisées comme système modèle dans cette thèse ont été largement caractérisées, montrant une dynamique accrue à l'échelle temporelle pertinente pour la catalyse (µs à ms) mais maintenant la reconnaissance du substrat. Dans cette thèse, l'évolution in vitro de ces variantes dynamiques a été réalisée par des cycles itératifs de mutagenèse et de sélection aléatoires pour permettre une exploration impartiale du paysage de ‘fitness’. Nous démontrons que la présence de ces mouvements particuliers au début de l'évolution a permis d'accéder à des voies de mutations connues. De plus, des interactions épistatiques connues ont été introduites dans les variantes dynamiques. Leur caractérisation in silico et cinétique a révélé que les mouvements supplémentaires sur l'échelle de temps de la catalyse ont permis d'accéder à des conformations conduisant à une fonction améliorée, comme dans le TEM-1 natif. Dans l'ensemble, nous démontrons que l'évolution de la b-lactamase TEM-1 vers une nouvelle fonction est compatible avec divers mouvements à l'échelle de temps µs à ms. Il reste à savoir si cela peut se traduire par d'autres enzymes ayant un potentiel biotechnologique. Deuxièmement, la lipase Cal-A, pertinente sur le plan industriel, a été choisie pour identifier les caractéristiques qui pourraient faciliter son ingénierie. La lipase Cal-A présente des caractéristiques telles que la polyvalence du substrat et une grande stabilité thermique et réactivité qui la rendent attrayante pour la modification des triglycérides ou la synthèse de molécules pertinentes dans les industries alimentaire et pharmaceutique. Contrairement à TEM-1, la plupart des études d'évolution in vitro de la lipase Cal-A ont été réalisées dans un but industriel, avec une exploration limitée de l'espace de mutation. Par conséquent, les caractéristiques qui définissent la fonction de la lipase Cal-A restent insaisissables. Dans cette thèse, nous faisons état de la mutagenèse ciblée de la lipase Cal-A, confirmant l'existence d'une région clé pour la reconnaissance du substrat. Cela a été fait en combinant une nouvelle méthodologie de création de bibliothèque basée sur l'assemblage Golden-gate avec une visualisation structurelle basée sur des scripts pour identifier et cartographier les mutations sélectionnées dans la structure 3D. La caractérisation et la déconvolution de deux des plus aptes ont révélé l'existence d'une épistasie dans l'évolution de la lipase Cal-A vers une nouvelle fonction. Dans l'ensemble, nous démontrons que l’identification d'une variété de propriétés suite à la mutagenèse ciblée peut grandement améliorer la connaissance d'une enzyme. Cette information peut être appliquée pour améliorer l'efficacité de l'ingénierie dirigée. / Enzyme engineering is a tool with great utility in the biotechnological industry. It allows to tailor enzymes to a specific activity or reaction condition. In addition, it can allow to decipher key elements that facilitated their modification. While enzyme engineering is extensively practised, it still entails several bottlenecks. Some of these bottlenecks are technical such as the development of methodologies for creating targeted mutational libraries or performing high-throughput screening and some are conceptual such as deciphering the key relevant features in a target protein for a successful engineering project. Among these challenges, intragenic epistasis, or the non-additivity of the phenotypic effects of mutations, is a feature that greatly hinders predictability. Improving enzyme engineering needs a multidisciplinary approach that includes gaining a better understanding of structure-function-evolution relations. This thesis seeks to contribute in the advancement of enzyme engineering by investigating two model systems. First, dynamic variants of TEM-1 ß-lactamase were chosen to investigate the link between protein dynamics and evolution. TEM-1 ß-lactamase has been extensively characterized in the literature, which has translated into extensive knowledge on its reaction mechanism, structural features and evolution. The variants of TEM-1 ß-lactamase used as model system in this thesis had been extensively characterized, showing increased dynamics at the timescale relevant to catalysis (µs to ms) but maintaining substrate recognition. In this thesis, in vitro evolution of these dynamic variants was done by iterative rounds of random mutagenesis and selection to allow an unbiased exploration of the fitness landscape. We demonstrate that the presence of these particular motions at the outset of evolution allowed access to known mutational pathways. In addition, known epistatic interactions were introduced in the dynamic variants. Their in silico and kinetic characterization revealed that the additional motions on the timescale of catalysis allowed access to conformations leading to enhanced function, as in native TEM-1. Overall, we demonstrate that the evolution of TEM-1 b-lactamase toward new function is compatible with diverse motions at the µs to ms timescale. Whether this can be translated to other enzymes with biotechnological potential remains to be explored. Secondly, the industrially relevant Cal-A lipase was chosen to identify features that could facilitate its engineering. Cal-A lipase presents characteristics such as substrate versatility and high thermal stability and reactivity that make it attractive for modification of triglycerides or synthesis of relevant molecules in the food and pharmaceutical industries. Contrary to TEM-1, most in vitro evolution studies of Cal-A lipase have been done towards an industrially-specified goal, with limited exploration of mutational space. As a result, features that define function in Cal-A lipase remain elusive. In this thesis, we report on focused mutagenesis of Cal-A lipase, confirming the existence of a key region for substrate recognition. This was done by combining a novel library creation methodology based on Golden-gate assembly with script-based structural visualization to identify and map the selected mutations into the 3D structure. The characterization and deconvolution of two of the fittest revealed the existence of epistasis in the evolution of Cal-A lipase towards new function. Overall, we demonstrate that mapping a variety of properties following mutagenesis targeted to specific regions can greatly improve knowledge of an enzyme that can be applied to improve the efficiency of directed engineering.

La dynamique structurelle des biomolécules est intimement liée à leur fonction, mais très coûteuse à étudier expériementalement. Pour cette raison, de nombreuses méthodologies computationnelles ont été développées afin de simuler la dynamique structurelle biomoléculaire. Toutefois, lorsque l'on s'intéresse à la modélisation des effects de milliers de mutations, les méthodes de simulations classiques comme la dynamique moléculaire, que ce soit à l'échelle atomique ou gros-grain, sont trop coûteuses pour la majorité des applications. D'autre part, les méthodes d'analyse de modes normaux de modèles de réseaux élastiques gros-grain (ENM pour "elastic network model") sont très rapides et procurent des solutions analytiques comprenant toutes les échelles de temps. Par contre, la majorité des ENMs considèrent seulement la géométrie du squelette biomoléculaire, ce qui en fait de mauvais choix pour étudier les effets de mutations qui ne changeraient pas cette géométrie. Le "Elastic Network Contact Model" (ENCoM) est le premier ENM sensible à la séquence de la biomolécule à l'étude, ce qui rend possible son utilisation pour l'exploration efficace d'espaces conformationnels complets de variants de séquence. La présente thèse introduit le pipeline computationel ENCoM-DynaSig-ML, qui réduit les espaces conformationnels prédits par ENCoM à des Signatures Dynamiques qui sont ensuite utilisées pour entraîner des modèles d'apprentissage machine simples. ENCoM-DynaSig-ML est capable de prédire la fonction de variants de séquence avec une précision significative, est complémentaire à toutes les méthodes existantes, et peut générer de nouvelles hypothèses à propos des éléments importants de dynamique structurelle pour une fonction moléculaire donnée. Nous présentons trois exemples d'étude de relations séquence-dynamique-fonction: la maturation des microARN, le potentiel d'activation de ligands du récepteur mu-opioïde et l'efficacité enzymatique de l'enzyme VIM-2 lactamase. Cette application novatrice de l'analyse des modes normaux est rapide, demandant seulement quelques secondes de temps de calcul par variant de séquence, et est généralisable à toute biomolécule pour laquelle des données expérimentale de mutagénèse sont disponibles. / The dynamics of biomolecules are intimately tied to their functions but experimentally elusive, making their computational study attractive. When modelling the effects of thousands of mutations, time-stepping methods such as classical or enhanced sampling molecular dynamics are too costly for most applications. On the other hand, normal mode analysis of coarse-grained elastic network models (ENMs) provides fast analytical dynamics spanning all timescales. However, the vast majority of ENMs consider backbone geometry alone, making them a poor choice to study point mutations which do not affect the equilibrium structure. The Elastic Network Contact Model (ENCoM) is the first sequence-sensitive ENM, enabling its use for the efficient exploration of full conformational spaces from sequence variants. The present work introduces the ENCoM-DynaSig-ML computational pipeline, in which the ENCoM conformational spaces are reduced to Dynamical Signatures and coupled to simple machine learning algorithms. ENCoM-DynaSig-ML predicts the function of sequence variants with significant accuracy, is complementary to all existing methods, and can generate new hypotheses about which dynamical features are important for the studied biomolecule's function. Examples given are the maturation efficiency of microRNA variants, the activation potential of mu-opioid receptor ligands and the effect of point mutations on VIM-2 lactamase's enzymatic efficiency. This novel application of normal mode analysis is very fast, taking a few seconds CPU time per variant, and is generalizable to any biomolecule on which experimental mutagenesis data exist.