La MAPK p38α est une protéine clé de l’inflammation, également impliquée dans de nombreux processus liés au cancer. Les petites molécules chimiques inhibitrices de p38α bloquent son activité kinase par un mécanisme de compétition à l’ATP. En raison de la grande conservation du domaine kinase, la spécificité de la majorité de ces inhibiteurs n’est pas restreinte à la p38α et des effets « off-targets » ont été rapportés. Dans ce contexte, le projet de ma thèse a porté sur l’utilisation de fragments d’anticorps au format scFv comme nouvel outil de ciblage pharmacologique afin de définir une/des voie(s) alternative(s) d’inhibition de la p38α. Les fragments d’anticorps lient un motif antigénique avec une affinité et une spécificité élevées, tout comme les immunoglobulines classiques. Leur expression intracellulaire permet également le ciblage de protéines cytoplasmiques et l’étude de leurs fonctions dans des processus physiologiques et pathologiques. Nous avons sélectionné par phage display, à partir d’une banque de fragments d’anticorps, 5 scFv spécifiques de la MAPK p38α. Alors que tous ces scFv empêchent l’activation par phosphorylation de la p38α par MKK6, l’un d’entre eux agit directement sur l’enzyme pour inhiber totalement son activité kinase in vitro. Ce scFv possède un site de liaison et un mécanisme d’inhibition distincts des inhibiteurs pharmacologiques déjà décrits : bien qu’il ne cible pas le domaine kinase et n’empêche pas la fixation de l’ATP, le scFv se comporte comme un inhibiteur compétitif de l’hydrolyse de l’ATP. Ces résultats suggèrent un effet allostérique du scFv sur l’activité de la p38α et permettent de le caractériser comme un inhibiteur compétitif non conventionnel. La détermination de son épitope d‘interaction ainsi que la confirmation de sa fonctionnalité une fois exprimé dans le cytosol des cellules nous permettra de définir une voie alternative d'inhibition de la p38α et de valider notre approche de ciblage par l’utilisation de fragments d’anticorps.Ces données ouvrent de nouvelles perspectives de design d’inhibiteurs chimiques de la p38α de meilleure spécificité que ceux actuellement disponibles. / MAPK p38α is a key protein in inflammation, but is also involved in many cancer-related processes. All the currently described chemical inhibitors of p38α inhibit its kinase activity by an ATP-competitive mechanism. Because of the high conservation of the ATP-binding pocket, the majority of these inhibitors are not specific to p38α and off-target effects have been reported. To identify alternative approaches to inhibit p38α MAPK, my thesis project focused on the use of scFv antibody fragments as a new highly specific pharmacological tool.Antibody fragments bind to an antigen with high affinity and specificity like conventional immunoglobulins. Their intracellular expression also allows to target cytoplasmic proteins and study the target functions in physiological and pathological processes. Using a naïve library of antibody fragments, we have selected by phage display five scFv specific of MAPK p38α isoform. While all these scFv inhibit the activation of p38α by MMK6, one of them also completely inhibits its kinase activity in vitro. This scFv has a binding site and a mechanism of inhibition distinct from the pharmacological inhibitors currently described: although it does not target the ATP-binding pocket and does not prevent ATP binding, it behaves like a competitive inhibitor of ATP hydrolysis. These results suggest an allosteric effect of this scFv on p38α activity and allow to characterize it as an unconventional competitive inhibitor. The determination of its epitope as well as the confirmation of its inhibitory activity once expressed in the cell cytosol will allow us to propose an alternative approach to target p38α function using antibody fragments.These data open up new perspectives for the design of more specific p38α chemical inhibitors than those currently available.

This thesis describes the development of chemical tools that inhibit the sialidases NanA and NanB from Streptococcus pneumonia. The primary focus was on the discovery of allosteric inhibitors of NanA and NanB, however, promising inhibitors that act by binding at the active site of these enzymes were also investigated. Chapter 1 gives an overview of the use of chemical tools in the field of chemical biology. It focuses in particular on chemical tools that function by the allosteric regulation of their target proteins. The uses, advantages and methods of discovery of allosteric tools are discussed. Finally this chapter introduces the use of serendipitous binders for the discovery of allosteric sites. In particular, the use of CHES to identify novel allosteric sites on the sialidase NanB is proposed. Chapter 2 describes how the ‘hits' from a series of high throughput screens were reanalysed using a wide range of secondary assays to eliminate any false positives that were contaminating the results. This process removed eight of the eleven ‘hits'. Two of the remaining three compounds were then analysed further in an attempt to characterise their binding mode to NanA and/or NanB using modelling and X-ray crystallographic studies. Whilst, it was not possible to confirm the binding mode by X-ray crystallography modelling studies using the modelling software GOLD generated possible binding modes for these inhibitors. A structure activity relationship study was conducted for both compounds in an attempt to generate more potent inhibitors. Chapter 3 moves from the use of high throughput screens to identify hits against NanA and NanB to the use of the serendipitous binding of N-cyclohexyl-2-aminoethanesulfonic acid in the active site of NanB for the development of selective NanB inhibitors. First taurine was identified as the minimum unit of N-cyclohexyl-2-aminoethanesulfonic acid required to bind to the active site of NanB. Taurine was then used as the basis of an optimisation study. This chapter concludes with the identification of 2-(benzylammonio)ethanesulfonate as the next key intermediate in the development of N-cyclohexyl-2-aminoethanesulfonic acid based active site inhibitors of NanB. Chapter 4 follows on from Chapter 3 with the optimisation of 2-(benzylammonio)ethanesulfonate describing the design and synthesis of a wide range of analogues. From these compounds 2-[(3-chlorobenzyl)ammonio]ethanesulfonate was identified as the most potent and selective inhibitor. Detailed analysis of the binding of 2-[(3-chlorobenzyl)ammonio]ethanesulfonate to NanB gave a rationale for its improved inhibitory activity. The increase in inhibition occurred because on binding of 2-[(3-chlorobenzyl)ammonio]ethanesulfonate to the active site of NanB a well coordinated water molecule was displaced. The displacement of this water caused an increase in the flexibility of the enzyme's 352 loop. A detailed study of the flexibility of this loop in response to various N-cyclohexyl-2-aminoethanesulfonic acid based chemical tools was then conducted. The research in chapters 2 and 3 has recently been published. In Chapter 5 a molecule of N-cyclohexyl-2-aminoethanesulfonic acid that binds serendipitously in a previously unmentioned secondary site is elaborated into a ligand, known as Optactin, that binds strongly and selectively at this secondary site. It was then shown that Optactin inhibited NanB by binding at this secondary site. It was therefore concluded that this secondary site was in fact an allosteric site that could be used for the regulation of NanB. Chapter 6 describes the development of a rationalisation for the inhibition of NanB by Optactin. This study included the X-ray crystallographic analysis of the apo-NanB structure and the NanB-Optactin complex under a range of conditions. This was followed by mechanistic studies that identified the point in the catalytic cycle at which Optactin was inhibiting NanB. This chapter concludes with a hypothesis for the mechanism of inhibition of NanB by Optactin.

Glycerol kinase (GK) from Escherichia coli is allosterically controlled by fructose 1,6-bisphosphate (FBP) and the glucose-specific phosphocarrier protein IIAGlc of the phosphotransferase system. These controls allow glucose to regulate glycerol utilization. Fluorescence spectroscopic and enzyme kinetic methods are applied to investigate these allosteric controls in this study. The linkage between FBP binding and GK tetramer assembly is solved by observation of homo-fluorescence energy transfer of the fluorophore Oregon Green (OG) attached specifically to an engineered surface cysteine in GK. FBP binds to tetramer GK with an affinity 4000-fold higher than to dimeric GK. A region named the coupling locus that plays essential roles in the allosteric signal transmission from the IIAGlc binding site to the active site was identified in GK. The relationship between the coupling locus sequence in Escherichia coli or Haemophilus influenzae GK variants and the local flexibility of the IIAGlc binding site is established by fluorescence anisotropy determinations of the OG attached to the engineered surface cysteine in each variant. The local flexibility of the IIAGlc binding site is influenced by the coupling locus sequence, and in turn affects the binding affinity for IIAGlc. Furthermore, the local dynamics of each residue in the IIAGlc binding site of GK is studied systematically by the fluorescence anisotropy measurements of OG individually attached to each position of the IIAGlc binding site. The fluorescence steady-state anisotropy measurement provides a valid estimation of the local flexibility and correlates well with the crystallographic B-factors. Steady-state kinetics of FBP inhibition shows that the data are best described by a model in which the partial inhibition and FBP binding stoichiometry are taken into account. Kinetic viscosity effects show that the product-release step is not the purely rate-limiting step in the GK-catalyzed reaction. Viscosity effects on FBP inhibition are also discussed.

Alzheimers disease (AD) represents one of the greatest health problems in industrialized countries considering the ageing population. Only four drugs are currently approved for the treatment of this disease. As these drugs are characterized with a limited time efficacy, it has become urgent to develop additional innovative AD treatments. Amongst the approaches that are actively investigated, the one consisting in potentiating a subclass of glutamate receptors appears attractive. This well advanced pharmacological approach includes three major classes of compounds amongst which appear the benzothiadiazine 1,1-dioxides. The present thesis is a pursuit of the preliminary efforts that were published in 1998 and 2001 by our team. Based on promising in vitro results obtained with the lead compound 59, pharmacomodulations around 59s structure have been achieved in order to enhance its in vivo activity and to optimize its pharmacokinetic parameters. First efforts were devoted to exploratory synthesis where attention was paid to the impact of the substituent introduced at the 7-position. Moreover, some pyridothiadiazine dioxides as well as thienothiadiazine dioxides were prepared. The most important part of our pharmacomodulations was focused on the thiadiazine ring system. Considering that the poor in vivo results obtained with 59 could be due to a metabolic weakness of the latter, the introduction of fluorine atoms was tempted as a lead optimization strategy. This approach was successful, since it led to the synthesis of 95b which was selected for further pharmacological evaluations. This new lead compound was shown to exert significant cognitive-enhancing effects in vivo after oral administration to Wistar rats. Moreover, the study of the metabolic degradation of 95b allowed the assessment of the starting hypothesis that had dictated the pharmacomodulations philosophy. Finally, additional exploratory pharmacomodulations were achieved notably leading to the preparation of a quinazolinone series and 1,4-benzothiazine compounds. This research allowed to significantly improve the pharmacokinetic profile of our series and led to the identification of 95b as a new lead compound. However, many pharmacomodulations remain to be explored. The data collected during this thesis are appealing further studies. Efforts in the near future should lead to the design of novel drug candidates among which a future innovative AD treatment could emerge. Au vu du vieillissement de la population dans les pays industrialisés, la maladie dAlzheimer constitue un problème majeur en termes de santé publique. A lheure actuelle, seuls quatre médicaments sont utilisés pour traiter cette maladie. Sachant que ces substances nont quune efficacité limitée, la conception de nouveaux médicaments actifs contre cette pathologie apparaît comme une priorité. Dans cette optique, la potentialisation des récepteurs glutamatergiques de sous-type AMPA semble une approche thérapeutique intéressante. Parmi les composés étudiés dans cette voie, apparaissent les benzothiadiazine dioxydes. Cette thèse sinscrit dans la continuité des recherches publiées en 1998 et 2001 par notre équipe et poursuit les pharmacomodulations entamées autour du composé leader 59. Ce travail sest principalement focalisé sur lamélioration de lactivité in vivo de ce dérivé et sur loptimisation de ses paramètres pharmacocinétiques afin dobtenir un candidat médicament potentiel. Notre attention sest tout dabord portée sur la subsitution aromatique en position 7. Par ailleurs, plusieurs séries pyrido- et thiénothiadiazines ont été également préparées. Ensuite, en postulant que le manque dactivité in vivo du composé 59 pouvait être dû à une faiblesse métabolique de ce composé, lintroduction judicieuse datomes de fluor a été choisie comme stratégie doptimisation. Cette approche nous a notamment amené à lidentification du composé 95b. Ce dérivé sest montré particulièrement actif dans un test de reconnaissance dobjets chez le rat Wistar. Ce résultat encourageant a amené à réaliser toute une série dévaluations pharmacologiques sur ce produit, afin de caractériser son mécanisme daction. Les données récoltées à lissue de ces investigations suggèrent que le composé 95b présente un intérêt potentiel en tant quagent procognitif. Dautre part, létude du profil de dégradation métabolique de ce dérivé a permis de confirmer lhypothèse qui avait dicté lintroduction datomes de fluor. Enfin, des pharmacomodulations supplémentaires ont été réalisées de façon exploratoire et ont entre autres débouché sur une série de quinazolinones et une série de 1,4-benzothiazines. Les travaux réalisés au cours de cette thèse ont permis daméliorer grandement le profil pharmacocinétique des séries explorées. Néanmoins, de nombreuses pharmacomodulations restent à explorer. Des études complémentaires savèrent nécessaires et devraient déboucher sur de nouveaux candidats médicaments innovants, parmi lesquels pourraient figurer un futur traitement de la maladie dAlzheimer.

<p><i>Lactococcus lactis,</i> prolidase (<i>Lla</i>prol) hydrolyzes Xaa-Pro dipeptides. Since Xaa-Pro is known as bitter peptides, <i>Lla</i>prol is potentially applicable to reduce bitterness of fermented foods. <i>Lla</i>prol shows allosteric behaviour and substrate inhibition, which are not reported in other prolidases. Computer models of <i>Lla</i>prol based on an X-ray structure of non-allosteric <i>Pyrococcus furiosus</i> prolidase showed that a loop structure (Loop^32-43) is located at the interface of the protomers of this homodimeric metallodipeptidase. This study investigated roles of four charged residues (Asp³⁶, His³⁸, Glu³⁹, and Arg⁴⁰) of Loop^32-43 in <i>Lla</i>prol using a combination of kinetic examinations of ten mutant enzymes and their molecular models. Deletion of the loop structure by Î36-40 mutant resulted in a loss of activity, indicating Loop^32-43 is crucial for the activity of <i>Lla</i>prol. D36S and H38S exhibited 96.2 % and 10.3 % activity of WT, whereas little activities (less than 1.0 % of WT activity) were observed for mutants E39S, D36S/E39S, R40S, R40E, R40K and H38S/R40S. These results implied that Glu³⁹ and/or Arg⁴⁰ play critical role(s) in maintaining the catalytic activity of <i>Lla</i>prol. These observations suggested that the loop structure is flexible and this attribute, relying on charge-charge interactions contributed by Arg⁴⁰, Glu³⁹ and Lys¹⁰⁸, is important in maintaining the activity of <i>Lla</i>prol. When the loop takes a conformation close to the active site (closed state), Asp³⁶ and His³⁸ at the tip of the loop can be involved in the catalytic reaction of <i>Lla</i>prol. The two active mutant prolidases (D36S and H38S) resulted in modifications of the unique characteristics; the allosteric behaviour was not observed for D36S, and H38S <i>Lla</i>prol showed no substrate inhibition. D36E/R293K, maintaining the negative charge of position 36 and positive charge of position 293, still possessed the allosteric behaviour, whereas the loss of the charges at these positions (D36S of this study and R293S of a previous study (Zhang et al., 2009 BBA-Proteins Proteom 1794, 968-975) eliminated the allosteric behaviour. These results indicated the charge-charge attraction between Asp³⁶ and Arg²⁹³ is important for the allostery of <i>Lla</i>prol. In the presence of either zinc or manganese divalent cations as the metal catalytic centre, D36S and H38S enzymes also showed different substrate preferences from WT <i>Lla</i>prol, implying the influence of Asp³⁶ and His³⁸ on the substrate binding. D36S and H38S also showed higher activities at pH 5.0 to 6.0, in which range WT <i>Lla</i>prol steeply decreased its activity, indicating Asp³⁶ and His³⁸ are involved in the active centre and influence the microenvironment of catalytic His²⁹⁶. The above observations are attributed to modifications of their local structure in the active centre since the temperature dependency and thermal denaturing temperature indicated little effects on the overall structure of the <i>Lla</i>prol mutants.</p> <p>From these results, we concluded that the unique behaviours of <i>Lla</i>prol are correlated to Loop^32-43 and Asp³⁶ and His³⁸ on it. When Loop^32-43 takes a closed conformation, Asp³⁶ interacts with Arg²⁹³ via charge-charge attraction to form an allosteric subsite. The saturation of the allosteric site with substrates further allowed the communications of His³⁸ with S₁ site residues to complete the active site. When the substrate concentration becomes higher than it is required to saturated productive S₁' site, His³⁸, Phe¹⁹⁰ and Arg²⁹³ would resemble the residue arrangement of S₁' site residues (His²⁹², Tyr³²⁹, and Arg³³⁷) and bind to the proline residue of substrates. This non-productive binding would prevent the conformational change of Loop^32-43, which further results in the substrate inhibition. For further confirmation of this mechanism, crystallographic studies will be conducted. In this thesis, we have indentified the conditions to produce crystals of <i>Lla</i>prol proteins.</p>

Phosphofructokinase (PFK) from the lactic acid bacterium Lactobacillus delbrueckii subspecies bulgaricus (LbPFK) is a non-allosteric PFK with weak binding affinity for both the allosteric ligands phospho(enol)pyruvate (PEP) and magnesium adenosine diphosphate (MgADP). PEP and MgADP bind to the same allosteric binding site but exhibit opposite effects, PEP acting as an inhibitor and MgADP an activator. In 2005, Parichatttanakul, et al. solved the first crystal structure of LbPFK to 1.87 A resolution and allowed for a structural comparison of LbPFK to the allosteric forms of PFK from E. coli (EcPFK) and Bacillus stearothermophilus (BsPFK). Two additional structures of LbPFK have been determined with the first having phosphates bound at the four active sites and four allosteric sites solved to 2.20 A resolution. The second structure solved to 1.83 A resolution contains phosphates at all eight sites with the addition of the substrate fructose-6-phosphate (F6P) in the active sites. These structures are similar to the published sulfate-bound LbPFK structure. Overall, the secondary, tertiary and quaternary structure is conserved with the exception of the residues in the allosteric site. E55, H59, S211, D214, H215 and G216, as well as the long cassettes of residues 52-61 (PFKs1) and 206-218 (PFKs2) were mutated to the corresponding residue/residues in Thermus thermophilus PFK (TtPFK). PFKs1 and PFKs1 were also combined to form PFKs1s2. The single mutations along with PFKs1 and PFKs2 showed no enhancement in PEP binding, but PFKs1s2 enhanced PEP binding 10-fold with no change in MgADP binding compared to LbPFK. D12, located along the active site interface 15 A away from the allosteric site, was mutated to an alanine and exhibited enhanced binding 9-fold for both PEP and MgADP to the allosteric binding site. A crystal structure of D12A was solved to 2.30 A resolution with sulfate bound to all eight binding sites, and showed no major changes in secondary, tertiary or quaternary structure when compared to the sulfate-bound wild-type LbPFK structure. Combining D12A with PFKs1s2 (PFKs1s2/D12A) further enhanced PEP binding with a 21-fold tighter binding compared to LbPFK with MgADP binding being similar to D12A. PEP inhibition was also quantitated in PFKs1s2/D12A with a Q_ay = 0.007 plus/minus 0.0008. Coupling between PEP and F6P in PFKs1s2D12A is 2-fold stronger than the coupling measured in EcPFK and 7-fold stronger than the coupling measured in BsPFK. The coupling measured in PFKs1s2D12A is the first measured in any of the LbPFK variants.

<p><i>Lactococcus lactis,</i> prolidase (<i>Lla</i>prol) hydrolyzes Xaa-Pro dipeptides. Since Xaa-Pro is known as bitter peptides, <i>Lla</i>prol is potentially applicable to reduce bitterness of fermented foods. <i>Lla</i>prol shows allosteric behaviour and substrate inhibition, which are not reported in other prolidases. Computer models of <i>Lla</i>prol based on an X-ray structure of non-allosteric <i>Pyrococcus furiosus</i> prolidase showed that a loop structure (Loop^32-43) is located at the interface of the protomers of this homodimeric metallodipeptidase. This study investigated roles of four charged residues (Asp³⁶, His³⁸, Glu³⁹, and Arg⁴⁰) of Loop^32-43 in <i>Lla</i>prol using a combination of kinetic examinations of ten mutant enzymes and their molecular models. Deletion of the loop structure by Î36-40 mutant resulted in a loss of activity, indicating Loop^32-43 is crucial for the activity of <i>Lla</i>prol. D36S and H38S exhibited 96.2 % and 10.3 % activity of WT, whereas little activities (less than 1.0 % of WT activity) were observed for mutants E39S, D36S/E39S, R40S, R40E, R40K and H38S/R40S. These results implied that Glu³⁹ and/or Arg⁴⁰ play critical role(s) in maintaining the catalytic activity of <i>Lla</i>prol. These observations suggested that the loop structure is flexible and this attribute, relying on charge-charge interactions contributed by Arg⁴⁰, Glu³⁹ and Lys¹⁰⁸, is important in maintaining the activity of <i>Lla</i>prol. When the loop takes a conformation close to the active site (closed state), Asp³⁶ and His³⁸ at the tip of the loop can be involved in the catalytic reaction of <i>Lla</i>prol. The two active mutant prolidases (D36S and H38S) resulted in modifications of the unique characteristics; the allosteric behaviour was not observed for D36S, and H38S <i>Lla</i>prol showed no substrate inhibition. D36E/R293K, maintaining the negative charge of position 36 and positive charge of position 293, still possessed the allosteric behaviour, whereas the loss of the charges at these positions (D36S of this study and R293S of a previous study (Zhang et al., 2009 BBA-Proteins Proteom 1794, 968-975) eliminated the allosteric behaviour. These results indicated the charge-charge attraction between Asp³⁶ and Arg²⁹³ is important for the allostery of <i>Lla</i>prol. In the presence of either zinc or manganese divalent cations as the metal catalytic centre, D36S and H38S enzymes also showed different substrate preferences from WT <i>Lla</i>prol, implying the influence of Asp³⁶ and His³⁸ on the substrate binding. D36S and H38S also showed higher activities at pH 5.0 to 6.0, in which range WT <i>Lla</i>prol steeply decreased its activity, indicating Asp³⁶ and His³⁸ are involved in the active centre and influence the microenvironment of catalytic His²⁹⁶. The above observations are attributed to modifications of their local structure in the active centre since the temperature dependency and thermal denaturing temperature indicated little effects on the overall structure of the <i>Lla</i>prol mutants.</p> <p>From these results, we concluded that the unique behaviours of <i>Lla</i>prol are correlated to Loop^32-43 and Asp³⁶ and His³⁸ on it. When Loop^32-43 takes a closed conformation, Asp³⁶ interacts with Arg²⁹³ via charge-charge attraction to form an allosteric subsite. The saturation of the allosteric site with substrates further allowed the communications of His³⁸ with S₁ site residues to complete the active site. When the substrate concentration becomes higher than it is required to saturated productive S₁' site, His³⁸, Phe¹⁹⁰ and Arg²⁹³ would resemble the residue arrangement of S₁' site residues (His²⁹², Tyr³²⁹, and Arg³³⁷) and bind to the proline residue of substrates. This non-productive binding would prevent the conformational change of Loop^32-43, which further results in the substrate inhibition. For further confirmation of this mechanism, crystallographic studies will be conducted. In this thesis, we have indentified the conditions to produce crystals of <i>Lla</i>prol proteins.</p>

M. tuberculosis copper-sensitive operon repressor (Mtb CsoR) is the founding member of a new metalloregulatory protein family in prokaryotes that regulates the transcription of the cso operon in response to copper toxicity. Mtb CsoR tetramer binds 1 monomer mol equiv of Cu(I) with very high affinity (log KCu=18.0) via three conserved residues, Cys36, His61' and Cys65'. Binding of Cu(I) allosterically inhibits the CsoR binding to the DNA operator (CsoO) overlapping the cso promoter (DeltaGc=+3.6 kcal/mol, pH 7.0, 25 oC). These findings are consistent with a role of CsoR as a transcriptional repressor with Cu(I) binding inducing transcriptional derepression. To explore the mechanism of this regulation, His61 was substituted with 1-methylhistidine (MeH) or Beta- (2-thiazolyl)-alanine (Thz) using a native chemical ligation strategy. The CsoO binding affinities of the resultant H61MeH and H61Thz CsoRs are both refractory to inhibition by Cu(I) binding despite the fact that each forms a high affinity 3-coordinate complex with Cu(I). This suggests that while Cu(I) is coordinated by the N?11 atom of His61, the N?22 atom plays an critical role in driving this allosteric switch. Evidence in support of a formation of a hydrogen bonding network involving the N?1 face of His61 and two conserved "second coordination shell" residues, Glu81' and Tyr35, is presented. Remarkably, this mechanism is analogous to that proposed for the Zn(II) sensor CzrA from S. aureus. To test this, we employed the same native chemical ligation approach to substitute the key Zn(II) ligand His97 with 1-methylhistidine; with the preliminary findings fully consistent with an intersubunit allosteric switch involving the N?2 face of this key His97 residue in CzrA. Two predicted homologs of Mtb CsoR were also biochemically characterized to obtain additional support for the hypothesis that CsoR is a key Cu(I) regulatory protein in many bacterial species. B. subtilis CsoR, known to regulate the transcription of the copZA operon, was found to have biochemical properties similar to those of Mtb CsoR as to Cu(I) binding, DNA binding and Cu(I)-dependent allosteric regulation. Interestingly, Bsu CsoR also binds other divalent metal ions (Zn, Ni) with high affinity but with metal coordination geometries distinct from that of Cu(I). Binding of these divalent metal ions only weakly regulates copZA operator binding in vitro, suggesting that coordination number and geometry are most closely related to the allosteric regulation. Finally, a putative CsoR from the pathogenic S. aureus Newman strain was identified and characterized, and was found to exhibit biochemical properties similar to those of Mtb and Bsu CsoRs. Parallels between Cu(I)-sensing CsoRs and functional orthologs in the CsoR/RcnR family are further discussed in the context of the mechanism and evolutionary divergence of this new family of regulatory proteins.

Protein-ligand and protein-protein interactions are critical to cellular function. Most cellular metabolic and signal tranduction pathways are influenced by these interactions, consequently molecular level understanding of these associations is an important area of biochemical research. We have examined the thermodynamics of several protein-protein associations and the protein-ligand interactions that mediate them. Using Fluorescence Correlation Spectroscopy, we have examined the putative interaction between pig heart malate dehydrogenase (MDH) and citrate synthase (CTS). We demonstrate a specific, low-affinity interaction between these enzymes. The association is highly polyethylene glycol (PEG)-dependent, and at high concentrations of NaCl or PEG, non-specific aggregates are formed. We demonstrate that oxaloacetate, the intermediate common to both CTS and MDH, induces the association at concentrations below the Km of CTS, suggesting that the open conformation of CTS is involved in the association. Using several biophysical techniques, we have examined the subunit associations of B. stearothermophilus phosphofructokinase (PFK). We demonstrate that the inhibitor bound conformation of the enzyme has reduced subunit affinity. The kinetics and thermodynamics of the phosphoenolpyrvuate (PEP)-induced dissociation of PFK have been quantified. Binding substrate, fructose-6-phosphate (F6P), stabilizes the enzyme to inhibitor-induced dissociation by 132-fold. These data suggest that subunit associations may play a role in the allosteric inhibition of PFK by PEP. The thermodynamics of the protein-ligand associations and allosteric inhibition of E. coli phosphofructokinase have been examined using intrinsic fluorescence and hydrostatic pressure. Both ligand-binding affinity and PEP inhibition are diminished by pressure, whereas substrate-binding affinity for inhibitor-bound enzyme is pressure-insensitive. Larger entropic than enthalpic changes with pressure lead to the overall reduction in free energies. Using a fluorescence-based assay, we have developed a series of baroresistant buffer mixtures. By combining a buffer with acid dissociation of negative volume with a buffer of positive volume, a pressure-resistant mixture is produced. Alteration of the molar ratio of the two component buffers yields mixtures that are pressure-insensitive at pH values around neutrality.

Glycerol kinase (GK) from Escherichia coli is allosterically controlled by fructose 1,6-bisphosphate (FBP) and the glucose-specific phosphocarrier protein IIAGlc of the phosphotransferase system. These controls allow glucose to regulate glycerol utilization. Fluorescence spectroscopic and enzyme kinetic methods are applied to investigate these allosteric controls in this study. The linkage between FBP binding and GK tetramer assembly is solved by observation of homo-fluorescence energy transfer of the fluorophore Oregon Green (OG) attached specifically to an engineered surface cysteine in GK. FBP binds to tetramer GK with an affinity 4000-fold higher than to dimeric GK. A region named the coupling locus that plays essential roles in the allosteric signal transmission from the IIAGlc binding site to the active site was identified in GK. The relationship between the coupling locus sequence in Escherichia coli or Haemophilus influenzae GK variants and the local flexibility of the IIAGlc binding site is established by fluorescence anisotropy determinations of the OG attached to the engineered surface cysteine in each variant. The local flexibility of the IIAGlc binding site is influenced by the coupling locus sequence, and in turn affects the binding affinity for IIAGlc. Furthermore, the local dynamics of each residue in the IIAGlc binding site of GK is studied systematically by the fluorescence anisotropy measurements of OG individually attached to each position of the IIAGlc binding site. The fluorescence steady-state anisotropy measurement provides a valid estimation of the local flexibility and correlates well with the crystallographic B-factors. Steady-state kinetics of FBP inhibition shows that the data are best described by a model in which the partial inhibition and FBP binding stoichiometry are taken into account. Kinetic viscosity effects show that the product-release step is not the purely rate-limiting step in the GK-catalyzed reaction. Viscosity effects on FBP inhibition are also discussed.