The work presented in this thesis focusses primarily on the determination of protein structure at atomic resolution, with NMR spectroscopy as the principle investigative tool. The thesis is divided into four parts. Part I consists of Chapter 1 which provides an introduction to protein structure, folding and NMR spectroscopy. Part II, consisting of Chapters 2 and 3, describes the effects of aromatic interactions on nucleating structure in disordered regions of proteins, using variants of apo-cytochrome b5 as a model system. Part III consists of Chapter 4, which describes structural effects of introducing cross-strand disulfide bonds using variants of Thioredoxin. Part IV of this thesis consists of the Appendices A, B and C. Appendix A describes the purification and characterization of ilvM, the regulatory subunit of the E.coli enzyme AHAS II. Appendices B and C contain chemical shift information corresponding to Chapter 3 and Chapter 4 respectively. Part I : Introduction to protein structure, folding and solution structure studies Chapter 1 first gives a brief overview of protein structure followed by an introduction to protein folding, focussing on the forces involved in determining the final three-dimensional shape of the protein as well as the experimental and computational techniques involved in studying or predicting the fold of a given protein. The second section of this chapter details the methodology followed to obtain solution structures of proteins using NMR spectroscopy. Part II : Engineering aromatic interactions to nucleate folding in intrinsically disordered regions of proteins Chapter 2 describes site-specific mutagenesis, recombinant over-expression, purifica-tion and preliminary biophysical characterization of two aromatic mutants of the molten globule apo-cytochrome b5 (apocytb5) : H43F H67F cytochrome b5 (FFcytb5) and H43W H67F cytochrome b5 (WFcytb5). Analysis of the structure of wild-type apo - cytochrome b5 was done to introduce surface mutations and avoid perturbation of the interior pack-ing of the protein. The bacterial host E.coli BL21(DE3) was used for recombinant over-expression, and both mutant proteins were purified by anion-exchange chromatography followed by size-exclusion chromatography. Biophysical studies show a decrease in the hydrodynamic radii and surface hydropho-bicity of FFcytb5 and WFcytb5 compared to wt -apo cytb5. An increase in protein stability was also seen from the wt apocytb5 to WFcytb5 and FFcytb5 in the presence of the chemical denaturant Urea. Proton 1D NMR spectra exhibited sharp lines and good spectral dispersion in the amide region, indicating that both mutant proteins are well folded. In addition, conservation of two distinctive up field and downfield shifted resonances present in apocytb5 indicated that structural changes upon mutation accrued on the upon the scaffold of apocytb5. Chapter 3 describes solution structure studies to determine secondary and tertiary structure of FFcytb5 and WFcytb5. Structural studies were carried out using homonu-clear and heteronuclear NMR methods, for which isotopically enriched 15N- and 13C, 15N samples were prepared for each protein. Additionally a 2H, 13C, 15N ILV methyl labeled sample was prepared for FFcytb5 to obtain unambiguous NOE correlation data. The hydrogen bond network for WFcytb5 was determined using hydrogen/deuterium exchange data. The restraints required to define the orientations and interactions of the aromatic groups were obtained from 15N-edited NOESY HSQC, 13C -edited NOESY HSQC and 2D 1H - 1H NOE spectra. These correlations were crucial in determining the aromatic interactions present within each protein. The structure of FFcytb5 was calculated using 1163 NOE distance restraints, 179 φ and ψ dihedral angle restraints, along with 40 hydrogen bond restraints. Similarly the structure of WFcytb5 was calculated using 1282 NOE distance restraints, 177 φ and ψ dihedral angle restraints and 40 hydrogen bond restraints. The ensemble of structures obtained for FFcytb5 showed a root mean square deviation of 1.01±0.21 Å . The ensemble of structures obtained for WFcytb5 showed a root mean square deviation of 0.58±0.09 Å . In both cases, ≈ 80% of backbone dihedral angles were found to be in the allowed regions and ≈ 20% in the additionally allowed regions of the Ramachandran map. The final tertiary structure of both FFcytb5 and WFcytb5 consisted of a mixed four strand β -sheet with a four helix bundle resting on top and were seen to align well, with an RMSD of 0.6 Å. A comparison of the solution structures of apocytb5 with FFcytb5 and WFcytb5 convincingly showed the nucleation secondary and tertiary structure well beyond the site of mutation. The presence of aromatic trimers, non-canonical in context of the wt apoc-ytb5, was confirmed upon analysis of the structures of FFcytb5 and WFcytb5, with NOE correlations assigned to verify these interactions. The reduction in the hydrodynamic radii of FFcytb5 and WFcytb5 in relation to apocytb5 was also verified from tsuperscript15N-NMR relaxometry studies. The nucleation of long-range structure using aromatic interactions has been demonstrated in proteins for the first time, and can in principle be used to incorporate aromatic residues and interactions in protein design. Structural data, chemical shift data and restraints lists used for structure calculation of WFcytb5 and FFcytb5 were deposited with the PDB (accession numbers 5XE4 and 5XEE) and BMRB(accession numbers 36070, 36071) respectively1. Part III : Structural consequences of introducing disulfide bonds into β - sheets Chapter 3 describes the solution structure studies on two mutants of E.coli Thiore-doxin which were designed to incorporate a disulfide bond between two anti-parallel β-strands at the edge of the β-sheet. One mutant was designed with a disulfide bond at the hydrogen bonding position (HB, 78c90cTrx) and the other with the disulfide bond at the non-hydrogen bonding position (NHB, 77c91cTrx). Here we study the structural changes that accompany the introduction of a cross-strand disulfide and whether such structural changes could be correlated with the previously seen thermodynamic and catalytic changes. Solution structure studies were conducted using a suite of multidimensional heteronu-clear NMR experiments, for which isotopically enriched 15N and 13C, 15N labelled samples were used. The solution structure for 77c91cTrx was calculated using 1190 NOE distance restraints, 199 φ and ψ dihedral angle restraints and 48 hydrogen bond restraints. The solution structure for 78c90cTrx was calculated using 1123 NOE distance restraints, 197 φ and ψ dihedral angle restraints and 50 hydrogen bond restraints. The ensemble of structures for 77c91cTrx showed an RMSD of 0.78± 0.13 Å while the RMSD for the ensemble of structures of 78c90cTrx was seen to be 0.76±0.09 Å . In both cases, ≈ 80% of backbone dihedral angles were seen to be in the allowed regions and ≈ 20% in the additionally allowed regions of the Ramachandran map. The tertiary structures of both proteins were seen to have a 5-strand mixed β-sheet and 4 helices surrounding it. . A comparison of the solution structures of mutant and wt -Trx showed significant changes in secondary and tertiary structure. For example, an α helix was reduced from 3 turns to a single turn, and of the β-strands containing the mutation was elongated by 3 residues. A ≈ 50% loss of hydrogen bonds, primarily from the β -sheet, was seen for both mutants. The secondary and tertiary structure for both 77c91cTrx and 78c90cTrx was seen to be near identical, despite the greater strain of the disulfide bond at the hydrogen bonding position. In addition to this, the Ile75-Pro76 peptide bond is now seen to be in the trans conformation in 78c90cTrx, while in wt -Trx the Ile75-Pro76 peptide bond is in the cis conformation. This cis peptide bond is known to play a role in both folding and catalysis, and the solution structures were analyzed in the context of observed changes in folding and catalysis. The study shows that introducing disulfide bonds even at the edge of β sheets have long-range structural effects, and the structural effects cannot be directly correlated with the changes in stability. Part III: Appendix Appendix A describes the expression, purification and preliminary characterization of ilvM, the regulatory subunit of E.coliAHAS II, one of three enzyme isomers that catal-yse the first step in the synthesis of all branched chain amino acids. AHAS II is known to be insensitive to feedback regulation, but our studies showed that the presence of Ile, Leu and Val causes structural changes and increases the stability of ilvM. However we were not able to purify ilvM in sufficient quantities to proceed with solution structure studies. Appendices B and C contain chemical shift information for the structural studies carried out on FFcytb5, WFcytb5, 77c91cTrx and 78c90cTrx.

Design of complex protein folds requires complete understanding of the stereochemical principles that govern polypeptide chain folding. Extensive studies on design and synthesis of specific secondary structures like β-helices, β -sheets and hairpins have taught us that the unnatural amino acid aminoisobutyric acid (Aib) can be successfully employed for helix nucleation and tight turns of appropriate stereochemistry are facilitated by the use of DPro-Xxx sequences. Availability of such rigid secondary structure scaffolds therefore permits the design of synthetic peptides that can be used as models for investigation of tertiary interactions, primarily that of aromatic residues. Chapter 1 summarizes the present knowledge of peptide design using non-protein amino acids. The chapter also details the unique features of aromatic amino acids, especially tryptophan, and their employment as secondary structure stabilizing elements. Chapters 2-7 contain detailed descriptions of the work carried out on design, synthesis, and structural characterization of designed peptides containing aromatic amino acids. In Chapter 2, the use of aromatic pairs in strand segments of peptide hairpins has been discussed with the results clearly indicating that aromatic interactions at the non-hydrogen bonding position of peptide hairpins contribute to structure stability. In Chapter 3, accommodation of the Leu-Trp-Val segment in helical scaffolds the role of Trp residues in crystallization has been discussed. Chapter 4 outlines the influence of a large number of Trp residues on the preferred backbone conformation, with the studies clearly indicating a preference for helical scaffolds in small peptides. The role of Trp residues at turn regions of peptide hairpins has been discussed in Chapter 5, using examples from both synthetic peptides and from natural peptides containing Pro-Trp segments. The studies suggest that the Pro-Trp segments serve as helix nucleators and disrupt formation of peptide hairpins. The results of this study have been further extended to Conus monile peptides, discussed in Chapter 6. The studies also suggest the role of an aromatic-Pro segment on the cis-trans isomerization of the Xxx-Pro tertiary amide unit. Chapter 7 discusses the contribution of a Cys-His vs Tyr-His pair on strand segment stability in diproline nucleated peptide hairpins. Chapter 8 summarizes the key findings of the work. Chapter 9 lists the references cited in the thesis and the Appendix chapter provides details of experimental techniques used in the study.β

Les matériaux agroressourcés à base de gluten de blé présentent des propriétés mécaniques qui ne leurs permettent pas de concurrencer celles des plastiques usuels issus de la pétrochimie. Les objectifs de ce travail de thèse visent (i) à améliorer les propriétés d'élongation et de résistance des matériaux gluten pour atteindre celles des polymères courants et (ii) à maîtriser la réactivité du gluten au cours de l'élaboration des matériaux afin de pouvoir utiliser les procédés connus de la plasturgie. L'enjeu scientifique est de comprendre la réactivité du gluten sous l'effet des traitements thermomécaniques et les mécanismes régissant les propriétés mécaniques des matériaux. Les fonctions réactives visées sont les thiols/disulfures qui assurent la réticulation des protéines du gluten. Nous avons testé l'effet de bloqueurs de thiols de type maléimide mono- et bifonctionnels, de nature, de taille et d'hydrophobicité variées. Ces derniers ont éventuellement été bloqués par réaction de Diels-Alder. L'ajout d'additifs de type bismaléimide bloqué par réaction de Diels-Alder permet de différer la réticulation à l'étape de thermoformage et de substituer aux liens covalents habituels des liens thioéthers. L'ajout de cet additif permet de doubler l'élongation à la rupture du matériau gluten mais entraîne la chute de la rigidité. L'effet de l'ajout de molécules bis- et tétrathiols a également été testé. Ces additifs ont permis d'augmenter par plus de 1,5 fois l'élasticité des matériaux gluten. Une analyse multi-échelle (moléculaire par FTIR, macromoléculaire par SEC et macroscopique par test de traction ; le tout complété par une analyse DMTA) de la structure et des propriétés a montré que l'absence de gain en élasticité était due au maintien d'une organisation structurale majoritaire en hélices-α, qui est le propre du gluten natif. La création d'interactions interprotéiques par feuillets-β a été identifiée comme seule responsable du gain d'élasticité des matériaux, la formation d'agrégats protéiques par le biais de liaisons disulfures ou thioéthers ne jouant qu'un rôle secondaire. Un mécanisme réactionnel mettant en avant les conditions qui assurent la participation de toutes les classes de protéines du gluten à la constitution du réseau protéique est discuté. Deux nouvelles voies prometteuses de mélange avec du caoutchouc et copolymérisation par « grafting from » ont été explorées et restent à approfondir. / Wheat gluten can be used to make biomaterials. Nevertheless their mechanical properties are not competitive with commonly used petroleum-based plastics. The purposes of this work aim at (i) improving strain and strength properties of gluten-based materials in order to reach those of common polymers and (ii) controlling gluten reactivity during material process in order to use already well-known processes for manufacturing plastics. The scientific stakes are to understand gluten reactivity during thermo-mechanical treatments and the mechanisms which govern mechanical properties of materials. The reactive functions of gluten are thiols/disulfides which are responsible of gluten proteins crosslinking. The effect of thiol blocker molecules such as mono and bismaleimide of various nature, sizes and hydrophobicity was tested. These molecules were eventually blocked by Diels-Alder reactions. The addition of bismaleimide blocked by Diels-Alder reaction enables to postpone the crosslinking to the thermo-molding step and also to substitute disulfides bonds for thioether bonds. The addition of this additive succeeds in doubling the strain at break of gluten-based material but leads to a decrease of the stiffness. The effect of addition of bis- and tetrathiol molecules was also considered as tests. These additives lead to increase by 1.5 times the elasticity of gluten-based materials. A multi-scale study (molecular scale by FTIR, macromolecular scale by SEC and macroscopic scale by tensile test, all supported by DTMA analysis) of the structure and properties showed that a predominant conformation with α-helices which is the case of native gluten, leads to a decrease of elasticity. The formation of β-sheets interproteic interactions was identified as the only responsible of elasticity increases of the material. The formation of proteic aggregates with disulfide and thioeter bonds only plays a secondary role. A reaction mechanism highlighting the conditions that ensure the participation of all types of gluten proteins in the gluten network upbuilding is discussed. Two new promising ways of rubber melt and copolymerization by “grafting from” technique were explored and need to be further developed.

Disulfides are the primary covalent interactions within a protein molecule that connect residues which are sequentially distant. Naturally occurring disulfides enhance the stability of the protein by destabilization of the unfolded state. Previous attempts to introduce disulfide bridges as a means to enhance protein stability have met with mixed results. Tools have been developed to predict potential sites for disulfide introduction. However, it must be noted that engineering disulfides is not a trivial task. The effect of the engineered disulfide on protein stability is difficult to predict. There have been few systematic studies carried out to study disulfides in the context of secondary structures. The work in this thesis is aimed at studying disulfides in two kinds of secondary structures- antiparallel β-sheets and helices. In particular, the focus in this thesis is on cross-strand disulfides in antiparallel β-sheets and intrahelical disulfides. The analysis of naturally occurring disulfides in these structural elements coupled with protein engineering studies in model proteins were used to understand the effects of introducing disulfides in helices and antiparallel β-sheets. Synopsis This thesis also includes studies carried out on molten globules of four periplasmic binding proteins of E.coli- Maltose binding protein (MBP), Leucine, isoleucine, valine binding protein (LIVBP), Leucine binding protein (LBP) and Ribose binding protein (RBP). Work carried out in the lab previously had shown that these molten globules can bind the ligands that the proteins do in their corresponding native states. The analysis of the thermodynamic data obtained for these molten globules by differential scanning calorimetry (DSC) studies and isothermal titration calorimetry (ITC) to characterize stability and ligand binding respectively are described in this thesis. To further study the structural features of molten globules by fluorescence resonance energy transfer (FRET), double cysteine mutants of MBP were constructed and characterized. The rationale behind the construction of these mutants and their characterization is reported. Chapter 1 gives an introduction to disulfides in proteins. Previous attempts at cataloguing and characterizing naturally occurring disulfides are described. An overview of studies carried out to determine the effects of removal of naturally occurring disulfides in proteins and the effect of engineered disulfides in different proteins is given. The various tools developed to predict potential disulfide sites are described. Chapter 1 also briefly discusses various aspects of molten globules and FRET. Chapters 2 and 3 involve studies with cross-strand disulfides occurring in antiparallel β-sheets. A detailed analysis on various stereochemical aspects of naturally occurring cross-strand disulfides is described in Chapter 2. The reasons for these disulfides to almost exclusively occur at non-hydrogen-bonded registered pairs have been explored with conformational analysis, modeling studies and energy calculations. In Chapter 3, the effect of engineering cross-strand disulfides in four model proteins- LBP, LIVBP, MBP and Top7 are described. The ease of formation of the introduced disulfides and their effects on protein stability are described. The proteins with engineered cross-strand disulfides at exposed positions were also examined for redox activity. Our studies have shown that in antiparallel strands, engineered disulfides at exposed NHB registered pairs provide a robust means of increasing protein stability. In Chapters 4 and 5, studies about intrahelical disulfides are described. In Chapter 4, the various conformational aspects of intrahelical disulfides occurring naturally are studied. Analysis of structures of proteins in conjunction with modeling studies show that all naturally occurring intrahelical residues bridge cysteines occurring between the N-Cap and 3rd residue of helices. To further explore conformational requirements for intrahelical disulfides, Cys pairs were introduced at N-terminal and interior of helices in a E.coli thioredoxin mutant lacking its active site disulfide. The ease of formation of the engineered disulfides, and their effects on protein stability were studied. The redox activity of the engineered disulfides was also examined. The studies demonstrated that intrahelical disulfides can only occur at the N-terminus of an α-helix and that the N-terminal CYS residue must adopt a non-helical backbone conformation. Although none of the engineered intrahelical disulfides increased the stability of the protein, they conferred mild redox activity. In Chapter 5, the ability of an engineered CXXC motif to bind Zn(II) is also explored. The effect of Zn(II) on the stability of the reduced and oxidized states of an engineered protein with a N-terminal intrahelical CXXC was ascertained. I have also shown that iminodiacetate (IDA) and nitrilotriacetate (NTA) resins charged with zinc can bind the protein CGPC 95-98 in reduced state. These Synopsis preliminary experiments on metal binding show that this property of CXXC motif could be exploited to develop a protein purification method. In Chapter 6, thermodynamic characterization of molten globules of four periplasmic binding proteins (LBP, LIVBP, MBP and RBP) is described. Studies had been previously carried out in the lab to characterize the stability and ligand binding of these molten globules. All four molten globules were found to bind their corresponding ligands without conversion to the native state. In Chapter 6, the estimation of ΔCp of unfolding and ligand binding from the DSC and ITC data is described. The binding of molten globules to their ligands and the ability to undego cooperative thermal unfolding indicated the presence of native protein-like tertiary contacts. To study the molten globule structure, we decided to construct double cysteine mutants of MBP for FRET studies. We decided to employ a strategy for differential labeling of the two cysteines with two different fluorophores based on the conformational differences between MBP in the ligand bound and free forms. Seven double cysteine mutants of MBP were made. The rationale behind the construction of these mutants and their preliminary characterization is described in the appendix to Chapter 6. The optimization of the differential labeling procedure of the MBP double mutants needs to be fine-tuned before further studies through FRET. The work described in this thesis has resulted in the following publications: 1.Prajapati RS, Indu S, Varadarajan R. Identification and thermodynamic characterization of molten globule states of periplasmic binding proteins. Biochemistry. 2007 (46):10339-52. 1 Indu S, Kumar ST, Thakurela S, Gupta M, Bhaskara RM, Ramakrishnan C, Varadarajan R. Disulfide conformation and design at helix N-termini. Proteins.2010 (78):1228-42. 2 Indu S, Kochat V, Thakurela S, Ramakrishnan C, Varadarajan R. Conformational analysis and design of cross-strand disulfides in antiparallel β-sheets. (Manuscript submitted)

Yaku’amide A (YA) is a linear anticancer peptide that is rich in bulky dehydroamino acids (ΔAAs) and β-hydroxyamino acids (β-OHAAs). In our recent total synthesis of YA, we featured a one-pot anti dehydration–azide reduction–O→N acyl transfer process for the stereospecific construction of Z- and E- ∆Ile residues. Despite previous total syntheses and our efforts, the synthesis of YA remains lengthy. Via computational studies, we identified two analogue peptides that closely resemble the conformation of YA. The use of simpler and symmetrical bulky ΔAAs such as dehydrovaline (ΔVal) and dehydroethylnorvaline (ΔEnv) as surrogates of ∆Ile, along with azlactone chemistry for their incorporation, significantly decreased the overall number of synthetic steps. Biological studies revealed that our analogues exhibited very similar activity to that of the natural product YA, demonstrating their suitability as mimics and consistency with our computational model. Despite its utility in the construction of YA analogues, azlactone chemistry is sluggish and moderate to low yielding. For this reason, we have explored strategies to streamline the synthesis of peptides containing Z-dehydroaminobutyric acid (∆Abu), ∆Val, and Z-dehydrophenylalanine (∆Phe). The key process is to form the alkene moiety via elimination of a β-sulfonium or β-OHAA embedded within a peptide, avoiding the need to form the alkene moiety via azlactone-dipeptide dehydration and bypassing sluggish amidation/ring opening steps. β-sheet disruption of Tau-model hexapeptides is a key type of inhibition for modulating Alzheimer’s disease progression. Previous studies replaced key residues with proline, due to its rigidity and lack of amide proton, to inhibit β-sheet formation. Similar to proline, ∆AAs are also known for their rigidity and ability to favor other conformations (e.g. β-hairpin, 310-helix) along with increasing peptide half-life. We have incorporated ∆Abu, ∆Val and dehydrocyclohexylglycine (∆Chg) in a highly aggregative hexapeptide sequence, using previously studied methods, to assess their capabilities as putative β-sheet breakers and to stabilize against proteolysis. Studies are continuing.