<p>Stabilization of proteins is of great interest for the biotechnological society, industrial as well as research areas. Proteins with high stability are more suitable as reagents, easier to handle, store, transport and use in industrial processes. One way to stabilize a protein is to introduce a disulfide bridge into the structure by protein engineering. In this report the formation of a disulfide bridge between engineered cysteines in non ideal conformations in human carbonic anhydrase has been investigated. The disulfide bridge is not formed when the protein is in its native state. It is shown that when the protein is exposed to mild concentrations of urea in the presence of DTTox the disulfide bridge is formed. Also upon refolding in vitro, in a non oxidative environment, disulfide bridges are formed. This observation is worth to notice, since the disulfide bridge does not form to any appreciable extent when the protein is expressed and folded in vivo in Escherichia coli.</p>

Les matériaux en contact avec des fluides biologiques peuvent être colonisés par de nombreux microorganismes (bactéries, levures...) et des macromolécules telles que les protéines. Lorsqu'il s'agit de bactéries pathogènes, l'adhésion bactérienne devient un problème, en particulier dans les milieux agroalimentaire et biomédical, car elle se poursuit jusqu'à la formation de biofilms bactériens, des bio-structures plus résistantes à l'action des antibiotiques que les bactéries isolées. Malgré une littérature abondante sur les processus d'adhésion bactérienne, l'interface surface - bactérie est encore mal comprise principalement à cause du manque de caractérisation à l'échelle moléculaire. Dans ce travail, nous utilisons une technique d'optique non linéaire du second ordre, la spectroscopie vibrationnelle de Génération de Fréquence Somme (SFG) à large bande, pour sonder spécifiquement des interfaces ordonnées à l'échelle moléculaire. Le principe consiste à envoyer un faisceau picoseconde visible et un faisceau femtoseconde infrarouge (accordé aux longueurs d'onde des vibrations des molécules de la surface) sur le substrat en contact avec des biomolécules en milieu aqueux. L'optimisation de la déconvolution et la modélisation du spectre SFG expérimental, réalisées dans le cadre de travail, permettent d'obtenir quantitativement la conformation des molécules de la surface du substrat. Nous nous sommes intéressés à la colonisation d'une surface structurée en " brosse " composée de monocouches autoassemblées (SAM) hydrophobes d'OctaDécaneThiol (ODT) par des bactéries Lactococcus lactis. Afin de reproduire les conditions naturelles de la colonisation bactérienne, nous avons aussi étudié le rôle de la présence de protéines, en l'occurrence l'albumine de sérum bovin. L'étude par spectroscopie SFG couplée avec des mesures de microscopie confocale de fluorescence a permis de proposer un mécanisme de l'adhésion bactérienne sur la SAM d'ODT, qui dépend de la présence des protéines. Nous avons démontré que les bactéries seules en suspension avaient un impact sur la conformation du support pouvant conduire à une augmentation ou à une diminution de la colonisation bactérienne selon le caractère hydrophobe / hydrophile de la paroi bactérienne. La présence des protéines avant ou pendant la colonisation bactérienne conduit à de nouveaux changements structuraux de la SAM d'ODT et à une importante diminution de l'adhésion bactérienne et du biofilm résultant (indépendamment du caractère hydrophobe / hydrophile de la paroi bactérienne). Cette étude démontre d'une part la faisabilité et l'intérêt de la spectroscopie vibrationnelle SFG pour l'étude de l'adhésion bactérienne in situ, et d'autre part que l'effet des bactéries et des protéines sur la conformation des surfaces est à prendre en compte lors de l'ingénierie de nouveaux matériaux à effet antiadhésif et/ou bactéricide.

In order to understand biological processes, to treat and diagnose diseases, find appropriate vaccines and to prevent the outbreak of epidemics, it is essential to obtain more knowledge about carbohydrate structures. This thesis deals with structure and conformation of carbohydrates, analysed by NMR spectroscopy and MD simulations.In the first two papers, the structures of O-antigen polysaccharides (PS) from two different E. coli bacteria were determined using NMR spectroscopy. The O-antigenic PS from E. coli O152 (paper I) consists of branched pentasaccharide repeating units, built up of three different carbohydrate residues and a phosphodiester, whilst the repeating unit of the O-antigen from E. coli O176 (paper II) is built up of a linear tetrasaccharide consisting of two different monosaccharides. In papers III and IV, the conformational analysis of different disaccharides is described. Conformational analysis was performed using NMR spectroscopy and MD simulations (paper IV). In paper III four different glucobiosides were studied using coupling constants and Karplus-type relationships. By use of specific 13C isotopically labelled derivatives, additional coupling constants were obtained and the number of possible torsion angles was reduced by half. In paper IV, we examine the conformations of two disaccharides that are part of an epitope of malignant cells. From NOE and T-ROE experiments, short proton-proton distances around the glycosidic linkage were estimated. Furthermore, interpretation of the extracted coupling constants using Kaplus relationships gave the values of the torsion angles. As in paper III, isotopically labelled compounds were synthesised in order to enhance the sensitivity of the analysis. Finally, MD simulations were performed and the results were compared with results from NMR data.

Stabilization of proteins is of great interest for the biotechnological society, industrial as well as research areas. Proteins with high stability are more suitable as reagents, easier to handle, store, transport and use in industrial processes. One way to stabilize a protein is to introduce a disulfide bridge into the structure by protein engineering. In this report the formation of a disulfide bridge between engineered cysteines in non ideal conformations in human carbonic anhydrase has been investigated. The disulfide bridge is not formed when the protein is in its native state. It is shown that when the protein is exposed to mild concentrations of urea in the presence of DTTox the disulfide bridge is formed. Also upon refolding in vitro, in a non oxidative environment, disulfide bridges are formed. This observation is worth to notice, since the disulfide bridge does not form to any appreciable extent when the protein is expressed and folded in vivo in Escherichia coli.

Carbohydrates play important roles in biological processes. Their function is closely related to their conformation. In this thesis, conformational studies of carbohydrates by NMR spectroscopy and molecular dynamics computer simulations are described. The first two papers discuss the anomalous solubility of β-cyclodextrin compared to other cyclodextrins. Time correlation functions revealed flexibility in all cyclodextrins. Molecular dynamics computer simulations showed that the glycosidic linkages were rather rigid and the flexibility was suggested to be macrocyclic. From spatial distribution functions β-cyclodextrin was found to have greater ability to order the surrounding water than the other cyclodextrins. Paper III deals with some of the difficulties of conformational studies. In Paper IV, a new method, Additative Potential Maximum Entropy, APME, is applied to a disaccharide. Conformational distribution functions are derived from NOEs, J-couplings and residual dipolar couplings and calculated from computer simulations. All distribution functions were found to be in good agreement. In papers V and VI oligosaccharides from human milk are studied. Residual dipolar coupling, J-couplings and cross relaxation rates were measured by NMR spectroscopy and molecular dynamics computer simulations were carried out. Both oligosaccharides showed high flexibility for the β-D-GlcpNAc-(1→3)-β-D-Galp linkage.

Soluble intermediates of the amyloid-β (Aβ) aggregation process are suggested to play a central role in the pathogenesis of Alzheimer’s disease (AD) by causing synaptic dysfunction and neuronal loss. In this thesis, soluble Aβ aggregates have been studied with a particular focus on the Aβ protofibril, which has served as the antigen for developing conformation dependent monoclonal antibodies. Antibodies generated from mice immunized with Aβ protofibrils were characterized regarding Aβ binding properties and the amino acid sequences of their antigen binding sites. A conformation dependent IgG antibody, mAb158, was further characterized and found to bind to Aβ protofibrils with a 200-fold higher affinity than to monomeric Aβ without affinity for soluble amyloid-β precursor protein (AβPP) or other amyloidogenic proteins. A sandwich enzyme-linked immunosorbent assay (ELISA) based on mAb158 was used to measure soluble Aβ protofibrils in brain extracts from AβPP-transgenic mice. Low levels of protofibrils could also be detected in human AD brain. However, positive signals generated from measurements in AD and control CSF samples were attributed to interference from heterophilic antibodies (HA), generating false positive signals by cross-binding the assay antibodies; consequently, a study on HA interference in Aβ oligomer ELISAs was initiated. A large set of plasma and CSF samples from AD and non-AD subjects were analyzed with and without measures taken to block HA interference, revealing that virtually all signals above the assay limit of detection were false and generated by HA interference. Many types of soluble Aβ aggregates have been described and suggested to impair neuron and synapse function. To investigate the soluble Aβ pool, synthetic Aβ and brain extracts from AβPP-transgenic mice and AD patients were ultracentrifuged on a density gradient to separate Aβ by size under native conditions. Four distinct gradient fractions were defined based on the appearance of synthetic Aβ in atomic force microscopy (AFM) and immunoreactivity in our protofibril specific sandwich ELISA. Interestingly, most Aβ from AD patients and AβPP-transgenic mice separated in the same fraction as toxic synthetic protofibrils.

The biological function of a protein is determined by its amino acid sequence, structure, and internal dynamics. In turn the prediction of a protein structure from its folding pathway involves the characterization of the dynamics of the polypeptide backbone. This study addresses how the internal dynamics of arborescent polypeptides are affected by increased crowding of the interior of these branched polymer molecules. Linear, comb-branched, and arborescent poly(L-glutamic acid) (PGA) samples were analyzed by 1H NMR spectroscopy to determine their chain conformation. The PGA chains of these constructs were shown to adopt α-helical and random coil conformations in N,N-dimethylformamide (DMF) and in dimethyl sulfoxide (DMSO), respectively. The hydrodynamic diameter (Dh) of the arborescent PGAs, determined using dynamic light scattering measurements, increased with increasing generation number and when the side-chains adopted random coil instead of α-helical conformations. The PGA samples were labelled with 1-pyrenemethylamine to determine how their structure affected the internal dynamics of the arborescent polymers in solution, from the analysis of their fluorescence spectra and decays. For each pyrene-labelled polymeric construct excimer formation increased with increasing pyrene content, and the efficiency of excimer formation increased with the generation number due to the increased density of the macromolecules. Comparison of the time-resolved fluorescence results acquired in DMF and in DMSO demonstrated that the helical conformation led to slower chain dynamics in DMF and that despite the higher viscosity of DMSO, the polypeptide side-chains were more mobile as a consequence of the random coil conformation of the linear PGA segments. These results suggest that the formation of structural motives inside a polypeptide slows down its internal dynamics.

M-DNA is a complex formed between normal double-stranded DNA and the transition metal ions Zn2+, Ni2+, and Co2+ that is favoured by an alkaline pH. Previous studies have suggested that M-DNA formation involves replacement of the imino protons of G and T bases by the transition metal ions involved in forming the complex. Owing to the conductive properties of this unique DNA conformation, it has potential applications in nanotechnology and biosensing. This work was aimed at improving existing methods and developing new methods of characterizing M-DNA. The effects of base substitutions, particularly those of G and T, were evaluated in light of the proposed structure. Differences between M-DNA conformations induced by Zn2+ and Ni2+ were also investigated with a variety of techniques and compared to the effects of Cd2+ and Mg2+ on double-stranded DNA. M-DNA formation and stability were studied with an ethidium bromide (EtBr) based assay, M-DNA induced fluorescence quenching of DNA labelled with fluorescein and a compatible quenching molecule, isothermal titration calorimetry (ITC), and surface plasmon resonance (SPR). Production of monoclonal antibodies against the conformation was also attempted but was unsuccessful. The EtBr-based assay showed Ni(II) M-DNA to be much more stable than Zn(II) M-DNA as a function of pH and in the presence of ethylenediaminetetraacetic acid. Sequence-dependency and the effect of base substitutions were measured as a function of pH. With regards to sequence, d(G)nd(C)n tracts were found to form the conformation most easily. Base substitutions with G and T analogues that lowered the pKa of these bases were found to stabilize M-DNA more strongly than other base substitutions. A combination of temperature-dependant EtBr and ITC assays showed M-DNA formation to be endothermic, and therefore entropy driven. The SPR studies demonstrated many qualitative differences between Zn(II) and Ni(II) M-DNA formation, allowed characterization of Zn2+, Ni2+, Cd2+, and Mg2+ complexes with single-stranded DNA, and provided unambiguous evidence that M-DNA formation results in very little denaturation of double-stranded DNA. Specifically, the SPR study showed Ni(II) M-DNA to be more stable than Zn(II) M-DNA in the absence of transition metal ions, but also showed that Ni(II) M-DNA required higher concentrations of Ni2+ than Zn2+ to fully form the respective M-DNA conformations. Finally, quenching studies demonstrated Zn(II) M-DNA formation over a pH range from 6.5 to 8.5 provided that a Zn2+:H+ ratio of roughly 105 was maintained. The Keq for this interaction was 1.3 x 10-8 with 1.4 H+ being liberated per base bair of M-DNA formed. These results support the proposed structural model of M-DNA, as lowering the pKa of the bases having titratable protons over the pH range studied facilitated M-DNA formation. The fact that Zn(II) M-DNA formation was observed by fluorescence quenching at any pH provided that a constant ratio of Zn2+:H+ was maintained was consistent with a simple mass-action interaction for M-DNA formation. The differences between Zn(II) and Ni(II) M-DNA formation show that although it requires a higher pH or transition metal ion concentration, Ni(II) M-DNA is more stable than Zn(II) M-DNA once formed. This difference could play an important role in applications of M-DNA which required modulation in the stability of the M-DNA conformation.

M-DNA is a complex formed between normal double-stranded DNA and the transition metal ions Zn2+, Ni2+, and Co2+ that is favoured by an alkaline pH. Previous studies have suggested that M-DNA formation involves replacement of the imino protons of G and T bases by the transition metal ions involved in forming the complex. Owing to the conductive properties of this unique DNA conformation, it has potential applications in nanotechnology and biosensing. This work was aimed at improving existing methods and developing new methods of characterizing M-DNA. The effects of base substitutions, particularly those of G and T, were evaluated in light of the proposed structure. Differences between M-DNA conformations induced by Zn2+ and Ni2+ were also investigated with a variety of techniques and compared to the effects of Cd2+ and Mg2+ on double-stranded DNA. M-DNA formation and stability were studied with an ethidium bromide (EtBr) based assay, M-DNA induced fluorescence quenching of DNA labelled with fluorescein and a compatible quenching molecule, isothermal titration calorimetry (ITC), and surface plasmon resonance (SPR). Production of monoclonal antibodies against the conformation was also attempted but was unsuccessful. The EtBr-based assay showed Ni(II) M-DNA to be much more stable than Zn(II) M-DNA as a function of pH and in the presence of ethylenediaminetetraacetic acid. Sequence-dependency and the effect of base substitutions were measured as a function of pH. With regards to sequence, d(G)nd(C)n tracts were found to form the conformation most easily. Base substitutions with G and T analogues that lowered the pKa of these bases were found to stabilize M-DNA more strongly than other base substitutions. A combination of temperature-dependant EtBr and ITC assays showed M-DNA formation to be endothermic, and therefore entropy driven. The SPR studies demonstrated many qualitative differences between Zn(II) and Ni(II) M-DNA formation, allowed characterization of Zn2+, Ni2+, Cd2+, and Mg2+ complexes with single-stranded DNA, and provided unambiguous evidence that M-DNA formation results in very little denaturation of double-stranded DNA. Specifically, the SPR study showed Ni(II) M-DNA to be more stable than Zn(II) M-DNA in the absence of transition metal ions, but also showed that Ni(II) M-DNA required higher concentrations of Ni2+ than Zn2+ to fully form the respective M-DNA conformations. Finally, quenching studies demonstrated Zn(II) M-DNA formation over a pH range from 6.5 to 8.5 provided that a Zn2+:H+ ratio of roughly 105 was maintained. The Keq for this interaction was 1.3 x 10-8 with 1.4 H+ being liberated per base bair of M-DNA formed. These results support the proposed structural model of M-DNA, as lowering the pKa of the bases having titratable protons over the pH range studied facilitated M-DNA formation. The fact that Zn(II) M-DNA formation was observed by fluorescence quenching at any pH provided that a constant ratio of Zn2+:H+ was maintained was consistent with a simple mass-action interaction for M-DNA formation. The differences between Zn(II) and Ni(II) M-DNA formation show that although it requires a higher pH or transition metal ion concentration, Ni(II) M-DNA is more stable than Zn(II) M-DNA once formed. This difference could play an important role in applications of M-DNA which required modulation in the stability of the M-DNA conformation.

none