Les introns de groupe II forment une classe d'ARN connus avant tout pour leur activité ribozymique, qui leur permet de catalyser leur propre réaction d'épissage. Sous certaines conditions, ces introns peuvent s'exciser des ARN précurseurs dont ils font partie et assurer la ligation des exons qui les bordent sans l'aide d'aucune protéine. Les introns de groupe II sont généralement excisés sous forme d'un lariat, semblable à celui formé par les introns des prémessagers nucléaires, dont l'épissage est assurée par le spliceosome. De telles similarités dans le mécanisme d'épissage suggèrent que les introns de groupe II et les introns des prémessagers nucléaires pourraient avoir un ancêtre évolutif commun.Malgré leurs séquences très diverses, les introns de groupe II peuvent être définis par une structure secondaire commune, hautement conservée. Celle-ci est formée de six domaines (domaine I à domaine VI ; D1-D6), émergeant d'une roue centrale. L'épissage des introns de groupe II comprend deux étapes, et autant de réactions de transestérification, qui produisent les exons liés et l'intron excisé sous forme lariat. Il est généralement admis que la structure du ribozyme subit des changements conformationnels entre les deux étapes de l'épissage et que le domaine VI est un acteur clé dans ce phénomène. Cependant, malgré l'identification d'un certain nombre d'interactions tertiaires entre domaines, ni la RMN, ni les études faisant appel à des modifications chimiques ne sont parvenues à déterminer l'environnement immédiat, au niveau du site actif du ribozyme, de l'adénosine qui sert de point de branchement de la structure en lariat, ainsi que des nucléotides qui entourent cette adénosine au sein du domaine VI. A l'aide d'analyses phylogénétiques et d'une modélisation moléculaire tridimensionnelle, nous avons identifié plusieurs sections du ribozyme susceptibles de constituer le site de fixation du domaine VI au cours de l'étape de branchement. Des mutations ont été introduites dans ces sites de fixation potentiels et la cinétique de réaction des ARN mutants résultants a été déterminée. Afin de démontrer formellement l'interaction du domaine VI avec le site récepteur le plus probable, une molécule de ribozyme dont la réaction de branchement est assurée par l'addition d'oligonucléotides ADN ou ARN qui positionnent correctement le domaine VI vis-à-vis de son partenaire a été construite. En combinant l'information apportée par différentes expériences de ce type, nous avons pu générer un modèle à résolution atomique du complexe formé par le domaine VI, son site de branchement et le reste de l'intron au moment où l'épissage est initié.

OmniMerge performs a systematic search to enumerate all conformations of a molecule (at a given level of torsion-angle resolution) that satisfy a set of local geometric constraints. Constraints would typically come from NMR experiments, but applications such as docking or homology modeling could also give rise to similar constraints. The molecule to be searched is partitioned into small subchains so that the set of possible conformations for the whole molecule may be constructed by merging the feasible conformations for the subchain parts. However, instead of using a binary tree for straightforward divide-and-conquer, OmniMerge defines a sub-problem for every possible subchain of the molecule. Searching every subchain provides a counter-intuitive advantage: with every possible subdivision available for merging, one may choose the most favorable merge for each subchain, particularly for the bottleneck chain(s). Improving the bottleneck step may therefore cause the whole search to be completed more quickly. Finally, to discard infeasible conformations more rapidly, OmniMerge filters the solution set of each subchain based on compatibility with the solutions sets of all overlapping subchains. These two innovations—choosing the most favorable merges and enforcing consistency between overlapping subchains—yield significant improvements in run time. By determining the extent of structural variability permitted by a set of constraints, OmniMerge offers the potential to aid error analysis and improve confidence for NMR results on peptides and moderate-sized molecules. / Singapore-MIT Alliance (SMA)

The kinetic mechanism of an α/β hydrolase fold epoxide hydrolase from potato, StEH1, has been studied with the aims of explaining the underlying causes for enantio- and regioselectivity, both being important for product purity. Further effort has been laid upon understanding the causes of a hysteretic behavior discovered in the measurements leading to Paper I. The enantioselectivity was investigated with substrates differing only in substituent size at one carbon of the oxirane ring structure. In catalysis with trans-stilbene oxide and styrene oxide, enantioselectivity is the result of differences in alkylation rates. In pre-steady state measurement with trans-2-methylstyrene oxide (2-MeSO), a rate-limiting step involving slow transitions, referred to as hysteresis, was discovered. With this substrate enantioselectivity is proposed to be a consequence of the catalytic rate of (1R,2R)-enantiomer being more influenced by the hysteretic behavior than was the rate of the other enantiomer. In steady-state measurements with (1R,2R)-2-MeSO, at different temperatures and pH, hysteretic cooperativity was displayed. It can be concluded that this behavior is dependent on the relationship between kcat and the rate of transition between two Michaelis complexes. From the differences in pH dependence of kcat/KM in formation of the two diols resulting from low regioselectivity in catalysis of (1R,2R)-2-MeSO, it is suggested that hysteresis is a result of the substrates placed in different conformational modes within the active site cavity. Regioselectivity is proposed to be the result of specific interactions between the catalytically important Tyr and the substrate, with a link between KM-values and degree of regioselectivity. Furthermore, the hysteretic kinetic model proposed can explain hysteresis, cooperativity and regioselectivity resulting from StEH1 catalyzed hydrolysis of (1R,2R)-2-MeSO.

A deeper understanding of the diversity of carbohydrates and the many applications of oligo- and polysaccharides found in nature are of high interest. Many of the processes involving carbohydrates affect our everyday life. This thesis is based on six papers all contributing to an extended perspective of carbohydrate property and functionality. An introduction to carbohydrate chemistry together with a presentation of selected carbohydrate synthesis and analysis methods introduces the reader to the research field. The first paper is an NMR spectroscopy reinvestigation of the structures of the O-antigens from the lipopolysaccharides (LPS) of Shigella dysenteriae type 3 and Escherichia coli O124. The repeating units were concluded to be built of identical branched pentasaccharides now with the correct anomeric configurations. Paper II is a structural investigation of the O-antigen from the LPS of E. coli O74 which is built of branched tetrasaccharide repeating units including the uncommon monosaccharide d-Fuc3NAc. Paper III is a conformational study of a rhamnose derivative, using NMR spectroscopy and X-ray crystallography. The benzoyl ester group positioned at C4 prefers an “eclipsed” conformation in the crystal as well as in solution. The use of site-specifically 13C-labeled compounds in conformational studies is discussed in Papers IV and V. The disaccharide α-L-Rhap-(1→2)-α-L-Rhap-OMe was synthesized together with two 13C-isotopologues and studied with NMR spectroscopy to give seven J-couplings related to torsion angles φ and ψ. The trisaccharide α-L-Rhap-(1→2)[α-L-Rhap-(1→3)]-α-L-Rhap-OMe was synthesized with 13C-labeling at two positions which presented a solution to a problem of overlapping signals in the 1H NMR spectrum. The site-specific labeling also facilitated the measurement of two 3JCC and two 2JCH coupling constants. Finally, chapter 6 gives a short introduction to glycosynthase chemistry and discusses the synthesis of α-glycosyl fluorides. A novel cyclic heptasaccharide was synthesized from α-laminariheptaosyl fluoride using a mutant of the enzyme laminarase 16A and subsequently analyzed by NMR spectroscopy. / At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 4: Manuscript. Paper 5: Manuscript.

Choline oxidase catalyzes the flavin-dependent, two-step oxidation of choline to glycine betaine via the formation of an aldehyde intermediate. The oxidation of choline includes two reductive half-reactions followed by oxidative half-reactions. In the first oxidation reaction, the alcohol substrate is activated to its alkoxide via proton abstraction and oxidized via transfer of a hydride from the alkoxide α-carbon to the N(5) atom of the enzyme-bound flavin. In the wild-type enzyme, proton and hydride transfers are mechanistically and kinetically uncoupled. The role of Ser101 was investigated in this dissertation. Replacement of Ser101 with threonine, alanine, cysteine, or valine demonstrated the importance of the hydroxyl group of Ser101 in proton abstraction and in hydride transfer. Moreover, the kinetic studies on the Ser101Ala variant have revealed the importance of a specific residue for the optimization of the overall turnover of choline oxidase. The UV-visbible absorbance of Ser101Cys suggests Cys101 can form an adduct with the C4a atom of the flavin. The mechanism of formation of the C4a-cysteinyl adduct has been elucidated. D-arginine dehydrogenase (DADH) catalyzes the oxidation of D-amino acids to the corresponding imino acids, which are non-enzymatically hydrolyzed to α-keto acids and ammonia. The enzyme is strick dehrogenase and deoesnot react with molecular oxygen. Steady state kinetic studies wirh D-arginine and D-histidine as a substrate and PMS as the electron acceptor has been investigated. The enzyme has broad substrate specificity for D-amino acids except aspartate, glutamate and glycine, with preference for arginine and lysine. Leucine is the slowest substrate in which steady state kinetic parameters can be obtained. The chemical mechanism of leucine dehydrogenation catalyzed by DADH was explored with a combination of pH, substrate and solvent kinetic isotope effects (KIE) and proton inventories by using rapid kinetics in a stopped-flow spectrophotometer. The data are discussed in the context of the crystallographic structures at high resolutions (<1.3 Å) of the enzyme in complex with iminoarginine or iminohistidine.

Caspases are important players in programmed cell death. Normal activities of caspases are critical for the cell life cycle and dysfunction of caspases may lead to the development of cancer and neurodegenerative diseases. They have become a popular target for drug design against abnormal cell death. In this study, the recognition of P5 position in substrates by caspase-3, -6 and -7 has been investigated by kinetics, modeling and crystallography. Crystal structures of caspase-3 and -7 in complexes with substrate analog inhibitor Ac-LDESD-CHO have been determined at resolutions of 1.61 and 2.45 Å, respectively, while a model of caspase-6/LDESD is constructed. Enzymatic study and structural analysis have revealed that Caspase-3 and -6 recognize P5 in pentapeptides, while caspase-7 lacks P5-binding residues. D-arginine dehydrogenase catalyzes the flavin-dependent oxidative deamination of D-amino acids to the corresponding imino acids and ammonia. The X-ray crystal structures of DADH and its complexes with several imino acids were determined at 1.03-1.30 Å resolution. The DADH crystal structure comprises a product-free conformation and a product-bound conformation. A flexible loop near the active site forms an “active site lid” and may play an essential role in substrate recognition. The DADH Glu87 forms an ionic interaction with the side chain of iminoarginine, suggesting its importance for DADH preference of positively charged D-amino acids. Comparison of the kinetic data of DADH activity on different D-amino acids and the crystal structures demonstrated that this enzyme is characterized by relatively broad substrate specificity, being able to oxidize positively charged and large hydrophobic D-amino acids bound within a flask-like cavity. Understanding biology at the system level has gained much more attention recently due to the rapid development in genome sequencing and high-throughput measurements. Current simulation methods include deterministic method and stochastic method. Both have their own advantages and disadvantages. Our group has developed a deterministic-stochastic crossover algorithm for simulating biochemical networks. Simulation studies have been performed on biological systems like auto-regulatory gene network and glycolysis system. The new system retains the high efficiency of deterministic method while still reflects the random fluctuations at lower concentration.

The exquisite frequency selectivity and amplification characteristics of mammalian hearing intimately depend on the fast electromechanical motion of the outer hair cells in the cochlea. This membrane based process, termed electromotility, is driven by the protein prestin which is uniquely present in the OHC lateral wall. Voltage dependent motility, in OHCs and mammalian cells expressing prestin, is accompanied by intramembranous charge movement which is widely considered a signature of electromotility and prestin function. How prestin converts changes in membrane potential into axial length changes of OHCs is currently not understood at the molecular level. Many electromotility models predict that prestin conformational changes are the underlying mechanism connecting charge movement and motility. Currently, however, only indirect evidence for a prestin conformational change is available. Various experiments have indicated that the oligomeric states of prestin may be an important determinant of function. Numerous reports have provided varying estimates of prestin oligomeric state. However, estimates have been based on measurements performed outside the membrane making, firm biophysical conclusions difficult. Biophysical studies of prestin function have demonstrated its dependence on membrane properties. Alterations of membrane cholesterol affect voltage dependence of charge movement and motility. In addition cholesterol manipulations cause spatial redistribution of prestin and possibly change prestin oligomeric state. However, the underlying cause for prestin sensitivity to cholesterol and its relation to membrane distribution is unknown. We have applied single molecule fluorescence (SMF) imaging, single particle tracking (SPT), and Förster resonance energy transfer (FRET) to investigate prestin interactions at the molecular level. The results of our SMF experiments have suggested that prestin forms mainly tetramers and dimers in the cell membrane. Using SPT to map the trajectories of prestin in the membrane, we have found that prestin undergoes diffusion in and hops between membrane confinements of varying size. In addition, we have found that cholesterol affects the size and confinement strength of the compartments but does not affect the diffusivity within the compartments. Finally, using a combination of electrophysiology and FRET we have demonstrated that prestin undergoes voltage dependent structural changes. In total, our results refine our molecular understanding of prestin function.

<p>The thesis contains two directions in the simulations of biomolecular systems. The first part (Chapter 2 - Chapter 4) mainly focuses on the simulations of electron transfer processes in condensed phase; the second part (Chapter 5 - Chapter 6) investigates the conformational sampling of polysaccharides and proteins. Electron transfer (ET) reaction is one of the most fundamental processes in chemistry and biology. Because of the quantum nature of the processes and the complicated roles of the solvent, calculating the accurate kinetic and dynamic properties of ET reactions is challenging but extremely useful. Based on the Marcus theory for thermal ET in weak coupling limit, we combined the rigorous ab initio quantum mechanical (QM) method and well-established molecular mechanical (MM) force field and developed an approach to directly calculate a key factor that affects the ET kinetics: the redox free energy. A novel reaction order parameter fractional number of electrons (FNE) was used to characterize the ET progress and to drive the QM/MMMD sampling of the nonadiabatic free energy surface. This method was used for two aqueous metal cations, iron and ruthenium in solution, and generated satisfactory results compared to experiments. In order to further reduce the computational cost, a QM/MM-minimum free energy path (MFEP) method is implemented and combined with the FNE in the calculation of redox free energies. The calculation results using QM/MM-MFEP+FNE generated identical results as the direct QM/MM-MD method for the two metal cations, demonstrating the consistency of the two different sampling strategy. Furthermore, this new method was applied to the calculation of organic molecules and enhanced the computational efficiency 15-30 times than the direct QM/MM-MD method, while maintaining high accuracy. Finally, I successfully extended the QM/MM-MFEP+FNE method to a series of redox proteins, azurin and its mutants, and obtained very accurate redox free energy differences with relative error less than 0.1 eV. The new method demonstrated its excellent transferability, reliability and accuracy among various conditions from aqueous solutions to complex protein systems. Therefore, it shows great promises for applications of the studies on redox reactions in biochemistry. In the studies of force-induced conformational transitions of biomolecules, the large time-scale difference from experiments presents the challenge of obtaining convergent sampling for molecular dynamics simulations. To circumvent this fundamental problem, an approach combining the replica-exchange method and umbrella sampling (REM-US) is developed to simulate mechanical stretching of biomolecules under equilibrium conditions. Equilibrium properties of conformational transitions can be obtained directly from simulations without further assumptions. To test the performance, we carried out REM-US simulations of atomic force microscope (AFM) stretching and relaxing measurements on the polysaccharide pustulan, a (1&rarr;6)-&beta;-D-glucan, which undergoes well-characterized rotameric transitions in the backbone bonds. With significantly enhanced sampling convergence and efficiency, the REMUS approach closely reproduced the equilibrium force-extension curves measured in AFM experiments. Consistent with the reversibility in the AFM measurements, the new approach generated identical force-extension curves in both stretching and relaxing simulations, an outcome not reported in previous studies, proving that equilibrium conditions were achieved in the simulations. In addition, simulations of nine different polysaccharides were performed and the conformational transitions were reexamined using the REM-US approach. The new approach demonstrated consistent and reliable performance among various systems. With fully converged samplings and minimized statistical errors, both the agreement and the deviations between the simulation results and the AFM data were clearly presented. REM-US may provide a robust approach to modeling of mechanical stretching on polysaccharides and even nucleic acids. However, the performance of the REM-US in protein systems, especially with explicit solvent model, is limited by the large system size and the complex interactions. Therefore, a Go-like model is employed to simulate the protein folding/unfolding processes controlled by AFM. The simulations exquisitely reproduced the experimental unfolding and refolding force extension relationships and led to the full reconstruction of the vectorial folding pathway of a large polypeptide, the 253-residue consensus ankyrin repeat protein, NI6C. The trajectories obtained in the simulation captured the critical conformational transitions and the rate-limiting nucleation event. Together with the AFM experiments, the coarse-grained simulations revealed the protein folding and unfolding pathways under the mechanical tension.</p> / Dissertation

The aim of this study is to further our understanding of the forces that contribute to protein stability and to investigate how site-directed mutagenesis might be used for increasing protein stability. Eleven proteins ranging from 36 to 370 residues have been studied here. A 36-residue VHP and a 337-residue VlsE were used as model systems for studying the contribution of the hydrophobic effect on protein stability. Mutations were made in both proteins which replaced bulky hydrophobic side chains with smaller ones. All variants were less stable than their wild-type proteins. For VHP, the destabilizing effects of mutations were smaller when compared with similar mutations reported in the literature. For VlsE, a similarity was observed. This different behavior was investigated and reconciled by the difference in hydrophobicity and cavity modeling for both proteins. Therefore, the stabilizing mechanism of the hydrophobic effect appears to be similar for both proteins. Eight proteins were used as model systems for studying the effects of mutating non-proline and non-glycine residues to statistically favored proline and glycine residues in ?-turns. The results suggest that proline mutations generally increase protein stability, provided that the replaced residues are solvent exposed. The glycine mutations, however, only have a stabilizing effect when the wild-type residues have ?, ? angles in the L? region of Ramachandran plot. Nevertheless, this strategy still proves to be a simple and efficient way for increasing protein stability. Finally, using a combination of eight previously identified stabilizing mutations; we successfully designed two RNase Sa variants (7S, 8S) that have both much higher Tms and conformational stabilities than wild-type protein over the entire pH range studied. Further studies of the heat capacity change upon unfolding (?Cps) for both proteins and their variants suggest that residual structure may exist in the denatured state of the 8S variant. An analysis of stability curves for both variants suggests that they achieve their stabilization through different mechanisms, partly attributed to the different role of their denatured states. The 7S variants may have a more rigid denatured state and the 8S variant may have a compact denatured state in comparison with that of wild-type RNase Sa.

A thermodynamic study of five homologous HPr proteins derived from organisms inhabiting diverse environments has been undertaken. The aim of this study was to further our understanding of protein stabilization in extremes of environment. Two of the proteins were derived from moderate thermophiles (Streptococcus thermophilus and Bacillus staerothermophilus) and two from haloalkaliphilic organisms (Bacillus halodurans and Oceanobacillus iheyensis); these proteins were compared with HPr from the mesophile Bacillus subtilus. Genes for three of these homologous HPr proteins were for the first time cloned from their respective organisms into expression vectors and they were over-expressed and purified in Escherichia coli. Stability measurements were performed on these proteins under a variety of solution conditions (varying pH, salinity and temperature) by thermal and solvent induced denaturation experiments. Stability curves were determined for every homologue and these reveal very similar conformational stability for these homologues at their habitat temperatures. The BstHPr homologue is the most thermostable and also has the highest G25; the stability of other homologues was ranked as Bst>Bh>St>Bs>OiHPr. Other key thermodynamic parameters, like Cp, have been estimated for all the homologues and it was found that these values are identical within errors of estimation. Also, it was found that the values of TS are very similar for these homologues. Together these observations allow us to propose a thermodynamic mechanism toward achieving higher Tm. The crystal structures of the BstHPr and a single tryptophan-containing variant (BstF29W) of this homologue are also reported here. Also reported is a domain-swapped dimeric structure for the BstF29W variant, together with a detailed investigation into the solution oligomeric nature of this protein. The crystal structure of BstHPr is analyzed to enumerate various stabilizating interactions like hydrogen bonds and salt-bridges and these were compared with those for the mesophilic homologue BsHPr. Finally, an analysis of sequence alignments together with structural information for these homologues has allowed design of numerous variants of both Bs and BstHPr. A detailed thermodynamic study of these variants is presented in an attempt to understand the origins of the differences in stability of the HPr homologues.