Les intégrines, protéines transmembranaires hétérodimériques de type αβ, sont impliquées dans un grand nombre de phénomènes physiologiques et pathologiques. L’intégrine α5β1 est considérée à l’heure actuelle comme une cible thérapeutique pertinente en oncologie, en particulier pour le traitement des glioblastomes. Ces tumeurs cérébrales très agressives résistent aux traitements actuels, en partie par leur capacité à envahir le tissu cérébral sain. Nos résultats mettent en évidence, in vitro, le rôle de l’intégrine α5β1 dans la migration de cellules de glioblastome. Ils ont permis également de caractériser les effets inhibiteurs de la migration d’antagonistes sélectifs de l’intégrine α5β1 non reproduits par des antagonistes de l’intégrine αvβ3. Pour caractériser des antagonistes originaux de l’intégrine α5β1, nous avons combiné des techniques in silico et un test fonctionnel de migration in vitro. Cette démarche a permis la sélection de 3 molécules intéressantes, antagonistes allostériques de l’intégrine α5β1, se démarquant des antagonistes de référence par leur capacité à inhiber la migration cellulaire sans affecter la liaison du ligand endogène de l’intégrine, la fibronectine. / Integrins are αβ heterodimeric transmembrane proteins implicated in various physiological and pathological processes. Currently, α5β1 integrin is considered as a relevant therapeutic target in oncology, particularly for the treatment of glioblastomas. These highly aggressive brain tumours are resistant to current therapies, notably by their ability to invade healthy brain tissues. Our results highlight the role of the α5β1 integrin in the in vitro migration of glioblastoma cells. We characterized the inhibitory effects of selective α5β1 integrin antagonists in cell migration, which are not reproduced by αVβ3 integrin antagonists. To identify original and selective α5β1 integrin antagonists, we combined in silico screening and in vitro functional cell migration assays. This allowed the selection of 3 interesting molecules, behaving as allosteric α5β1 integrin antagonists. Contrarily to known α5β1 antagonists, our three hits inhibit cell migration without interfering with the binding of fibronectin, the endogenous ligand of this integrin.

Thesis advisor: Evan R. Kantrowitz / A full understanding of an enzyme's catalytic mechanism and a crystal structure representative of its in vivo form are powerful tools in computational drug screening and design. In the case of aspartate transcarbamoylase (ATCase), an allosteric enzyme, the mechanism and allosteric transition are still being explored. The crystallization of the ATCase mutant Asp236 to alanine, a T-state destabilized mutant, in the presence of phosphonoacetamide (PAM) by microdialysis was successful at pH 5.7. The enzyme crystallized in the R-state in the presence of only one substrate analogue. Globally the enzyme had converted to R, but the active site domains are more open than previously observed. Due to the ordered nature of the reaction, the R-state active site exists with a variety of small molecules bound at different times through out the course of the reaction. This structure shows an R-state active site with only one substrate analogue bound, and may therefore represent the R active site after catalysis has occurred and the active site is binding new substrates to perform its reaction again. Docking studies of small molecules can be conducted using this more open, emptier active site as it may be more representative of an in vivo conformation of the enzyme just before catalysis. Additionally, Arg296, previously unobserved as part of the active site, makes a hydrogen bonding interaction with the PAM molecule. The role of this residue will require further investigation. / Thesis (BS) — Boston College, 2005. / Submitted to: Boston College. College of Arts and Sciences. / Discipline: Chemistry. / Discipline: College Honors Program.

Thesis advisor: Evan R. Kantrowitz / The enzyme aspartate transcarbamoylase (ATCase) is an enzyme in the pyrimidine nucleotide biosynthetic pathway. It was once an attractive target for anti-proliferation drugs but has since become a teaching model due to kinetic properties such as cooperativity and allostery exhibited by the Escherichia coli form of the enzyme. ATCase from E. coli has been extensively studied over that last 60 years and is the textbook example of allosteric enzymes. Through this past research it is understood that ATCase is allosterically inhibited by CTP, the end product of pyrimidine biosynthesis, and allosterically activated by ATP, the end product of the parallel purine biosynthetic pathway. Part of the work discussed in this dissertation involves further understanding the catalytic properties of ATCase by examining an unregulated trimeric form from Bacillus subtilis, a bacterial ATCase that more closely resembles the mammalian form than E. coli ATCase. Through X-ray crystallography and molecular modeling, the complete catalytic cycle of B. subtilis ATCase was visualized, which provided new insights into the manifestation of properties such as cooperativity and allostery in forms of ATCase that are regulated. Most of the work described in the following chapters involves understanding allostery in E. coli ATCase. The work here progressively builds a new model of allostery through new X-ray structures of ATCase*NTP complexes. Throughout these studies it has been determined that the allosteric site is bigger than previously thought and that metal ions play a significant role in the kinetic response of the enzyme to nucleotide effectors. This work proves that what is known about ATCase regulation is inaccurate and that currently accepted, and taught, models of allostery are wrong. This new model of allostery for E. coli ATCase unifies all old and current data for ATCase regulation, and has clarified many previously unexplainable results. / Thesis (PhD) — Boston College, 2013. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Chemistry.

Intracellular polymerization of deoxygenated sickle hemoglobin (Hb S) remains the principal cause of the pathophysiology associated with sickle cell disease (SCD). Naturally occurring and synthetic allosteric effectors of hemoglobin (AEH) have been investigated as potential therapeutic agents for the treatment of SCD. Several aromatic aldehydes, including vanillin, have been studied previously as AEHs for their antisickling activities. Specifically, these compounds form Schiff- base adduct with Hb to stabilize the oxygenated (R) state, increase Hb affinity for O2 and concomitantly prevent the hypoxia-induced primary pathophysiology of Hb S polymerization and RBC sickling, in turn, ameliorating several of the cascading secondary adverse effects. These compounds, however, undergo significant metabolism leading to suboptimal pharmacokinetic properties, e.g. short duration of pharmacologic action and low bioavailability. These drawbacks have severely hampered the use of aromatic aldehydes as AEHs to treat SCD. To counter these challenges, we designed and synthesized 14 novel compounds (PP- compounds) based on previously studied pyridyl derivative of vanillin. These modifications were expected to increase binding interactions with the protein and thus stabilize the Schiff-base adduct, as well as lead to perturbation of the surface-located F-helix that would stereospecifically destabilize polymer contacts. We investigated the in vitro pharmacokinetic/pharmacodynamic properties of these newly synthesized compounds to ascertain sustained binding and modification of normal human Hb. Subsequently, we conducted in vitro screening assays to test for inhibition of sickling, modification of Hb to the high-affinity form, as well as for a direct left-shift in oxygen equilibrium curves (OEC). Three selected compounds, PP6, PP10, and PP14, that demonstrated not only high antisickling activity but also showed sustained pharmacologic action were chosen for preliminary in vivo studies. Our results showed maximal levels of Hb modification, which were sustained for the entire 24 h experimental period. In contrast, TD-7 after reaching maximum effect at 1 h gradually decreased in potency and at 24 h has lost 45% of its activity, consistent with the low bioavailability of this compound. These findings suggested that our modifications appeared to successfully limit drug metabolism in red blood cells. Most of these compounds showed almost complete inhibition of sickling at 2 mM concentration; with significant modification of Hb to a higher affinity Hb as well as an increase in O2 affinity of Hb. Interestingly, while TD-7 demonstrated a clear linear correlation between its ability to increase Hb-O2 affinity and antisickling activity, PP2, PP3, PP6, PP9, PP10, and PP14, showed a weak correlation between these parameters. In fact, these compounds demonstrated high antisickling effect despite only marginally increasing Hb affinity for O2. This observation indicated that these compounds possibly exhibit the dual mechanism of antisickling activity. We hypothesize that their secondary mechanism of action is due to interactions with the surface located αF-helix that would lead to direct or stereospecific inhibition of polymer formation. To establish the mode of interaction with Hb, we further conducted x-ray crystallography studies and co-crystallized PP2, PP6, PP9 and PP11 with CO-liganded Hb. Our studies showed that these compounds bind in symmetry-related fashion at the α-cleft of Hb to form Schiff-base adducts with the N-terminal Val1 and showed direct interactions with the F-helix which overall enhanced interactions with Hb. PP6, PP10, and PP14 demonstrated significant in vivo modification of intracellular Hb in mice after IP administration, with increasing levels from 1 h to the 6 h experimental period. They also showed corresponding changes in O2 affinity observed at 3 h and 6 h, compared to 0 h pre-treatment samples in vivo. Thus, our results establish these compounds as a novel, promising group of potent anti-sickling agents, demonstrate their proposed mechanism of action and provide proof-of-concept justifications for our structure-based approach to developing potent therapeutics for SCD.

The small- and intermediate-conductance Ca2+ activated K + (SK/IK) channels play a fundamental role in the regulation of neurons in the central nervous system. In animal models, SK/IK channel positive modulators have been shown to be effective in reducing the symptoms of neurological diseases such as ataxia. Ataxia is a lethal neurological rare disease characterized by lack of balance and incoordination of muscle movements, often as a result of cerebellar or spinocerebellar neurodegeneration. SK/IK channel modulators have been developed over the past few decades. Currently available modulators are often weak in potency. Lack of knowledge about the binding site for the compounds is the main reason hindering the development of more potent and effective therapeutics targeting SK channels. Dr. Zhang and his colleagues recently discovered the binding pocket for these positive modulators of SK/IK channels. This pocket is located at the interface between the channel and calmodulin. Dr. Zhang and his colleagues performed screening of a large number of compounds in silico, to find those fitting into the binding pocket. I performed electrophysiological recordings to evaluate the efficacy and the potency of these modulators on SK2 channels. We discovered a correlation between the total binding energy values calculated from the structures and the potencies determined from electrophysiological recording.

In order to maintain genetic stability in eukaryotes, tight regulation of the relative sizes of deoxyribonucleoside triphosphate (dNTP) levels inside the cell is essential for optimal fidelity of DNA replication. Ribonucleotide reductase (RNR) is the enzyme responsible for proportional production of DNA precursors. Studies on regulation of this enzyme, the focus of this thesis, are important because mutations affecting RNR control mechanisms result in dNTP pool imbalance, thus promoting mutagenesis. By using mouse RNR as a model for mammalian forms of the enzyme, three major factors--allosteric effectors, rNDP substrate concentrations, and hypoxic conditions--that influence the substrate specificity of RNR have been investigated. Allosteric regulation has been studied by the four-substrate assay, which permits simultaneous monitoring of the four reactions catalyzed by this enzyme in one reaction mixture. Individual dNTPs affect the four activities differentially in a concentration-dependent manner with discrete effects of dTTP and dGTP on reduction of ADP and GDP, respectively. Ribonucleoside diphosphate (rNDP) substrate concentrations are equally important, as their variations lead to different product ratios. Results from nucleotide binding assays indicate that rNDPs directly influence binding of dNTP effectors at the specificity site, one of the two classes of allosteric sites, whereas ADP has an indirect effect, displacing other substrates at the catalytic site and consequently removing effects of those substrates upon dNTP binding. Hence, this is the first evidence of a two-way communication between the catalytic site and the specificity site. Oxygen limitation also plays an important role in controlling the enzyme specificity. Reactivation of the enzyme at different oxygen tensions, after treatment of the enzyme with hydroxyurea (HU) followed by removal of HU, reveals a distinct sensitivity of GDP reductase to low 0��� levels. Although the basis for specific inhibition of GDP reduction remains to be determined, some possibilities have been ruled out. This research proves that in addition to allosteric regulation by nucleoside triphosphates, mouse RNR is also controlled by other factors. Since these components can simultaneously exert their effects upon enzyme specificity, complex regulatory patterns of RNR to provide a proportional supply of the DNA building blocks in vivo are suggested. / Graduation date: 2003

Degradation of mRNA is a highly regulated step important for proper gene expression. Degradation of eukaryotic mRNA is initiated by shortening of the 3’ end located poly(A) tail. Poly(A)-specific ribonuclease (PARN) is an oligomeric enzyme that degrades the poly(A) tail with high processivity. A unique property of PARN is its ability to interact not only with the poly(A) tail but also with the 5’ end located mRNA cap structure. A regulatory role in protein synthesis has been proposed for PARN based on its ability to bind the cap that is required for efficient initiation of eukaryotic mRNA translation. Here we have investigated how the cap structure influences PARN activity and how PARN binds the cap. We show that the cap activates PARN and enhances the processivity of PARN. Further we show that the cap binding complex (CBC) inhibits PARN activity through a protein-protein interaction. To investigate the cap binding property of PARN, we identified the cap binding site at the molecular level using site-directed mutagenesis and fluorescence spectroscopy. We identified tryptophan 475, located within the RNA recognition motif (RRM) of PARN, as crucial for cap binding. A crystal structure of PARN bound to cap revealed that cap binding is mediated by the nuclease domain and the RRM of PARN. Tryptophan 475 binds the inverted 7-Me-guanosine residue through a stacking interaction. Involvement of the nuclease domain in cap binding suggests that the cap site and the active site overlap. Mutational analysis showed that indeed amino acids involved in cap binding are crucial for hydrolytic activity of PARN. Taken together, we show that PARN is an allosteric enzyme that is activated by the cap structure and that the allosteric cap binding site in one PARN subunit corresponds to the active site in the other PARN subunit.

Phosphofructokinase from Escherichia coli (EcPFK) is allosterically regulated by MgADP and phospho(enol)pyruvate (PEP). Both molecules compete for binding to the same allosteric site, however, MgADP activates and PEP inhibits the binding of fructose-6-phosphate (F6P) to the active site. The mode by which this enzyme can differentiate between the two ligands and cause the appropriate response is important for the understanding of the basis of allosteric regulation. We studied the interactions between an active site and an allosteric site (heterotropic interactions) within the protein, and found that each of the four unique heterotropic interactions is unique and the magnitudes of the coupling free energies for MgADP activation sum up to 100% that of wildtype EcPFK without homotropic cooperativity in F6P binding. We took on the kinetic and structural characterization of phosphofructokinase from Lactobacillus bulgaricus (LbPFK) to reveal an enzyme that exhibits allosteric properties in spite of previous kinetic studies performed by Le Bras et al. (1991). We have identified residues in EcPFK (Asp59, Gly184 and Asp273), which are important for the allosteric responses to both MgADP and PEP. Interestingly, Lys214 is only important in PEP inhibition and not MgADP activation. We can also differentially disrupt the MgADP heterotropic interactions with the introduction of G184C within the protein. These results suggest that there are different pathways for allosteric communication within the enzyme: different paths for MgADP activation and PEP inhibition, and different paths for each heterotropic interaction with Gly184 being important for the 33Å MgADP heterotropic interaction.

The ATP-PRTase enzyme catalyses the first committed step of histidine biosynthesis in archaea, bacteria, fungi and plants.1 As the catalyst of an energetically expensive pathway, ATP-PRTase is subject to a sophisticated, multilevel regulatory system.2 There are two families of this enzyme, the long form (HisGL) and the short form (HisGS) that differ in their molecular architecture. A single HisGL chain comprises three domains. Domains I and II house the active site of HisGL while domain III, a regulatory domain, forms the binding site for histidine as an allosteric inhibitor. The long form ATP-PRTase adopts a homo-hexameric quaternary structure.3,4 HisGS comprises a similar catalytic core to HisGL but is devoid of the regulatory domain and associates with a second protein, HisZ, to form a hetero-octameric assembly.5 This thesis explores the allosteric regulation of the short form ATP-PRTase, as well as the functional and evolutionary relationship between the two families. New insight into the mode allosteric inhibition of the short form ATP-PRTase from Lactococcus lactis is reported in chapter two. A conformational change upon histidine binding was revealed by small angle X-ray scattering, illuminating a potential mechanism for the allosteric inhibition of the enzyme. Additionally, characterisation of histidine binding to HisZ by isothermal titration calorimetry, in the presence and absence of HisGS, provided evidence toward the location of the functional allosteric binding site within the HisZ subunit. Chapter three details the extensive effort towards the purification of the short form ATP-PRTase from Neisseria menigitidis, the causative agent of bacterial meningitis. This enzyme is of particular interest as a potential target for novel, potent inhibitors to combat this disease. The attempts to purify the long form ATP-PRTase from E. coli, in order to clarify earlier research on the functional multimeric state of the enzyme, are also discussed. Chapter four reports the investigation of a third ATP-PRTase sequence architecture, in which hisZ and hisGS comprise a single open reading frame, forming a putative fusion enzyme. The engineering of two covalent linkers between HisZ and HisGS from L. lactis and the transfer of the regulatory domain from HisGL to HisGS, is also discussed, in an attempt to delineate the evolutionary pathway of the ATP-PRTase enzymes. Finally, the in vivo activity of each functional and putative ATP-PRTase was assessed by E. coli BW25113∆hisG complementation assays.

Antibiotic resistance is a worsening threat to human health. Increasing our understanding of the mechanisms causing this resistance will be of great benefit in designing methods to evade resistance and in developing new classes of antibiotics. In this thesis, I have used the TetR Family Transcriptional Regulators (TFRs), which constitute one of the largest antibiotic resistance regulator families, as a model system to study the structure, function and evolution of antibiotic resistance determinants. I performed a thorough examination of the variation and conservation seen in TFR sequences and structures using computational approaches. Through structure comparison, I have identified the most conserved features shared by the TFR family that are crucial for their stability and function. Based on my findings on conserved TFR structural features, a quantitative assay of binding affinity determination was developed. Through sequence comparison and a residue contact map method, I discovered the existence of a conserved residue network that correlates well with the known allostery pathway of TetR. This predicted allosteric communication network was experimentally tested in TtgR. I have also developed methods to identify TFR operator sequences through genomic comparisons and validated my prediction through experiments. In addition, I have developed an in vivo system that can be used to identify and characterize proteins that mediate resistance to almost any antibiotic. This system is simple, fast, and scalable for high-throughput applications, and could be used to discover a wide range of novel antibiotic resistance mechanisms. The principles that I applied to the TFR family could also be applied to other protein families.