Doctor of Philosophy / Department of Chemistry / Duy H. Hua / The description of this thesis is divided into three chapters following the chronological events of my research development. In chapter one, a series of new 1,4-anthracenediones were synthesized via functionalizations of the methyl side chain of 6-methyl-1,4-anthracenedione. The new 1,4-anthracenediones were found to exhibit potent cytotoxic activities against human L1210 leukemic and HL-60 cell lines. A key intermediate, 6-bromomethyl-1,4- anthracenedione (1.44), was first synthesized through a sequence of reactions including a double Friedel-Crafts reaction, reductive quinone formation, and selective benzylic bromination. The bromide (1.44) was further converted to other 1,4-anthracenediones via hydrolysis, subsequent oxidation, and reductive amination or nucleophilic substitution. Chapter two deals with a continuous research project aiming at macropolycyclic cyclodecitycene or [10]beltene derivative using Diels-Alder reaction as the key strategy for cyclization. A tetraene, (4aR,5R,7S,7aS,11aR,12R,14S,14aR)-5,7,12,14-tetrahydroxy- 2,3,9,10-tetramethylene-1,4,4a,5,6,7,7a,8,11,11a,12,13,14,14a-tetradecahydro-6,13-obenzenopentacene (2.51), was synthesized by following previous work from this laboratory. Unfortunately, the Diels-Alder reaction of tetraene 2.51 with triptycene bisquinone showed predominantly polymerization over intramolecular cyclization. The use of double activated quinone such as 1,4,5,8-naphthodiquinone (2.64) and 1,2,4,5- tetraethoxycarbonyl-1,4-benzoquinone (2.70) as dienophiles gave monoadducts 2.67 and 2.71 respectively. However, they both failed to cyclize under high dilution conditions at elevated temperature, which may be rationalized by chair conformations adopted in six membered rings causing unfavorable twist for intramolecular cylization. Further study showed tetraene 2.51 underwent an unexpected furan ring forming reaction. In chapter three, an unprecedented substituted cyclododeciptycene, 2,4,6,8,10,12,14,16,18,20,22,24-dodecahydro-9,11,21,23-tetramethoxy-(2,14:4,16:6,18:8, 20:10,22:12,24)-hexa(o-benzeno)-[12]cyclacene-1,3,5,7,13,14,17,19-octaone (3.138), was successfully synthesized based on a successful intramolecular Diels-Alder reaction, which was developed from the above [10]beltene project and previously reported literature work. A series of all cis-iptycenequinones were synthesized as bisdienophile building blocks from a sequence of Diels-Alder reactions, separation of individual Diels- Alder adducts, enolization, and oxidative demethoxylation. It was found that each Diels- Alder adduct isomer shows distinguish [superscript]1HNMR signals inherent to its structure. The characteristic [superscript]1HNMR signals allow the identification of the structures of iptycenequinones derived from the above reactions. A bisdimethoxyanthracene, 6,8,15,17-tetramethoxy-7,16-dihydro-7,16-(o-benzeno)heptacene (3.56), was synthesized as bisdiene building block, which reacted with cis,cis-heptiptycene tetraquinone (3.23). The cycloadduct was transformed to cyclododeciptycene 3.138, whose structure was firmly established by a single-crystal X-ray analysis.

Sortases are cell-membrane associated cysteine transpeptidases that are essential for the assembly and covalent anchoring of certain surface proteins to the cell wall in Gram-positive bacteria. Thus, they play critical roles in virulence, infection and colonization by pathogens. Sortases have been classified as type A, B, C, D, E and F based on their phylogeny and the target-protein motifs that they recognize. Sortase A (SrtA) enzymes participate in cell wall anchoring of proteins involved in bacterial adhesion, immune evasion, internalization, and phage recognition and in some cases pili formation. SrtA substrates are characterised by the presence of a C-terminal cell wall sorting signal as LPXTG motif, followed by a stretch of hydrophobic residues and a positively charged tail. Experimental and bioinformatics studies show that class A sortases are housekeeping as well as virulence determining proteins. Hence, Sortase A enzymes are considered as promising antibacterial drug targets, particularly because many organisms are developing multi-drug resistance behaviour. SrtA adopts an eight-stranded β-barrel structure and the overall fold is conserved among the sortase isoforms, with some modifications. The thesis candidate has determined the three dimensional (3D) crystal structures of wild-type and active site mutant of Sortase A from Streptococcus pneumoniae R6 strain by using X-ray diffraction method. The wild-type enzyme crystallized in P21 space group whereas active site cysteine mutant crystallized in C2 space group. In both the cases, N-terminal 81 residue deletion constructs (ΔN81) were used for crystallization. Uncommonly, both the structures showed a phenomenon of domain-swapping which resulted in the protein adopting a domain-swapped dimeric form. Two such dimers in wild-type protein and three dimers in mutant protein were observed in the asymmetric unit. To the best of our knowledge, our work reveals for the first time the occurrence of domain-swapping in sortase superfamily. Experimental techniques like size-exclusion chromatography, native-PAGE, analytical centrifugation and thiol cross-linking (carried out in our collaborator’s laboratory at National Institute of Immunology (NII), New Delhi, India) of functionally active wild-type SrtA from S. pneumoniae showed dimerization as well as domain-swapping in solution state. These results support the possibility that the protein indeed exists in a domain-swapped dimeric form and the determined structure is not the result of crystal packing artifact but is physiologically relevant as well. The work done by the thesis candidate covering crystallization of both, the active and inactive protein constructs, their structure determination using molecular replacement method, detailed structural analyses, structural comparisons with known SrtA structures and new structural findings are described in from Chapter 2 to Chapter 4. Based on the SrtA crystal structure the author of the thesis has also proposed various point mutations which are likely to disrupt domain– swapping and result in loss of dimer formation. In addition, as a part of the ongoing project in our laboratory, molecular dynamics studies of these domain-swapped dimers containing two sets of active site residues facing each other in a very compact volume have been initiated to understand substrate binding, which in future could lead to inhibitor design. Apart from the crystal structure analyses of SrtA structures, the author of the thesis has also carried out systematic crystal structure investigation of dipeptides and pentapeptides containing non-standard amino acids (ΔPhe, Aib and β-amino acids) along with computational studies. Conformationally restricted α,β-dehydrophenylalanine residue (ΔF) and α-aminoisobutyric acid (Aib) have been incorporated in highly amyloidogenic human Islet Amyloid Polypeptide (hIAPP) fragments. Amyloid deposits, observed in a vast majority of Type 2 diabetic patients, are primarily on account of misfolding and aggregation into fibrils of hIAPP, a 37 residue endocrine hormone secreted by pancreatic β-cells. It has been suggested that intermediates produced in the process of fibrillization are toxic to insulin producing β-cells. Hence, the inhibition of misfolding of hIAPP that involves structural transition from its native state (coil and/or helical and/or transient helical conformation) to β-sheet conformation, could be a possible strategy to mitigate Type 2 Diabetes Mellitus (T2DM). All the peptides discussed in this thesis were synthesized in our collaborator, Prof. V. S. Chauhan’s laboratory at the International Centre for Genetic Engineering and Biotechnology (ICGEB), New Delhi, India. In this work, author of the thesis has designed short peptides containing helicogenic residue, α,β-dehydrophenylalanine (ΔF) and determined their 3D crystal structures. It was found that pentapeptides, FGA∆FL and FGA∆FI act as inhibitors of hIAPP fibrillization. As revealed by crystal structure analysis, both the peptides have similar backbone conformation consisting of a ‘nest’ motif, which is an anion receptor. Molecular docking suggested that both the pentapeptides interact with the hIAPP20-27 segment, stabilizing the hIAPP in helical form by shielding the core aggregation initiation region. This reduces the possibility of oligomerization, formation of toxic intermediates and subsequently the transition to β-structure and fibrillization. Thus, the crystal structures of pentapeptide inhibitors together with computational docking studies suggest an atomic level model of the possible mode of action by which the FGAΔF(L/I) peptides manifest their fibrillization inhibition activity and this could be of value in the design of a new class of amyloid inhibitors. In another peptide design, L→U (Aib) mutation was done in core fibrillization region ANFLV i.e. hIAPP13-17. The resulting mutant peptide ANFUV as well as native fragment ANFLV was crystallized and their 3D crystal structures were determined. ANFLV crystallized in two space groups C2 and P2 adopting extended conformation. Crystal packing of ANFLV in both the crystals shows parallel beta sheet arrangement which is favoured and strengthened by hydrogen bonding between asparagine side chains of Asn-Asn pair each located in neighbouring parallel beta-strands. Hydrogen bonded Asn-Asn residue pairing in parallel beta-strands suggests its significant contribution during hIAPP fibril formation. The substitution L→U abolished its fibrillization property and the structure of ANFUV was solved by direct methods in P21 space group. The occurrence of β-bulge in ANFUV induced by Aib, as observed in crystal packing, suggests that Aib acts as a β-breaker through β-bulge inducing property in the highly amyloidogenic hIAPP segment. β-bulge forming property, an attribute of Aib as β-breaker may be responsible for the curtailment of fibrillization potential of the peptide in which the residue was incorporated. The aim of the anti-amyloid work is to design potent anti-fibrillization peptides and the work is important to design peptide based drugs to fight type II diabetes. The utilization of ΔPhe in the molecular self-assembly offers an added benefit in terms of variety and stability. Taking advantage of the conformation constraining property of ΔPhe residue, its incorporation in dipeptide molecules has been probed. The author has studied nanotube formation through molecular self-assembly, involving two classes of non¬standard amino acids i.e. ΔF and β-amino acids. FΔF in D-form, L-form and DL-mixture crystallized in different space groups forming rectangular/hexagonal channels constituting different channel dimensions. Recently, the application of FΔF nanotubes have been demonstrated in controlled drug delivery, showing the relevance of the work in health care. Another class of dipeptides containing β-amino acids (β-FF, β-FΔF, β-AΔF, β-VΔF, β¬LΔF, β-IΔF, and β-LF) was also explored for the self-assembled nanotube formation. These β-peptides were crystallized and their 3D structures were determined solely by the author of the thesis. Except the β-AΔF & β-LΔF, these peptides self-assemble and form rectangular/ hexagonal channels. Structures of ΔF and β-amino acid containing dipeptides forming ordered nanotubes through self-assembly are detailed in Chapters 8 and 9 in the thesis. Overall, the author of the thesis has crystallized and determined structures of more than twenty peptides. Experimentally, β-peptide nanotubes were observed to encapsulate drug molecules and thus might be useful as a drug delivery system. In the present thesis crystal structures of the following designed peptide sequences (including one natural sequence ANFLV) are reported in detail. Table 1 Peptide sequence Representation Length Discussed in 1. Phe-Gly-Ala-ΔPhe-Leu FGAΔFL 5 Chapter 6 2. Phe-Gly-Ala-ΔPhe-Ile FGAΔFI 5 Chapter 6 3. Ala-Asn-Phe-Leu-Val (2 forms) ANFLV_P2, ANFLV_C2 5 Chapter 7 4. Ala-Asn-Phe-Aib-Val ANFUV 5 Chapter 7 5. LPhe-ΔPhe (2 forms) LFΔF1 , LFΔF2 2 Chapter 8 6. DPhe-ΔPhe DFΔF 2 Chapter 8 7. DLPhe-ΔPhe DLFΔF 2 Chapter 8 8. LTyr-ΔPhe LYΔF 2 Chapter 8 9. LSer-ΔPhe LSΔF 2 Chapter 8 10. Boc-D,LPhe-ΔPhe Boc-DLFΔF 2 Chapter 8 11. Cbz-D,LPhe-ΔPhe Z-DLFΔF 2 Chapter 8 12. D,LMet-ΔPhe DLMΔF 2 Chapter 8 13. β-Phe-ΔPhe β-FΔF 2 Chapter 9 14. β-Phe-Phe β-FF 2 Chapter 9 15. β-Val-ΔPhe β-VΔF 2 Chapter 9 16. β-Ile-ΔPhe β-IΔF 2 Chapter 9 17. β-Leu-ΔPhe β-LΔF 2 Chapter 9 18. β-Leu-Phe β-LF 2 Chapter 9 19. β-Ala-ΔPhe β-AΔF 2 Chapter 9 20. Cyclo(Phe-ΔPhe) DKP-FΔF 2 Appendix C 21. Cyclo(Ile-ΔPhe) DKP-IΔF 2 Appendix C 22. Cyclo(Cha-Cha) DKP-ChaCha 2 Appendix C 23. Cyclo(Cha-Phe) DKP-ChaF 2 Appendix C 24. Cyclo(Cha-ΔPhe) DKP-ChaΔF 2 Appendix C 25. Cyclo(S-tritylCys-ΔPhe) DKP-CΔF 2 Appendix C Most of the dipeptides, except the N-terminal protected dipeptides, cyclic dipeptides (i.e. DKPs) and LSΔF, were found in the zwitterionic conformation and out of these, ten dipeptides resulted in tubular structures of dimensions in the nanoscale range. The thesis is organized into nine chapters and five appendices. Chapter 1 is an introduction to the work presented in the thesis, while Chapter 2, Chapter 3 and Chapter 4 describe the crystallographic work on the protein Sortase A. Chapter 5 is an introduction to the non-standard amino acids used for peptide designs and Chapter 6, Chapter 7, Chapter 8, Chapter 9 and Appendix C describe the crystallographic work on peptides. Chapter 1 starts with a general introduction to the Gram-positive bacteria containing sortase enzymes, and the bacterial cell-wall where sortase catalyzed proteins get attached for implicating their virulence during host-pathogen interactions. Pneumococcal diseases mostly affect children and their count has been observed to be higher than the combined total cases of malaria, AIDS and tuberculosis in child population worldwide. The chapter describes different virulence factors of S. pneumoniae out of which many are proteins. Among these, LPXTG containing proteins, which are the prime substrates of the sortase enzymes, are discussed in detail. Sortase enzymes, their classification and their structural studies with conserved ‘Sortase fold’ are discussed elaborately. A brief mention is made about the enzymatic activity of Sortase A to understand the transpeptidation mechanism. To appreciate the biomedical and biotechnological importance of the sortase enzyme, some potential applications of Sortase A are detailed in this chapter. A section is dedicated to describe the protein in the present study 'Sortase A from Streptococcus pneumoniae'. At the end, the scope of the present work, comprising of both protein and peptide crystallography, is presented. Chapter 2 begins with a brief account of the sequence analysis of Sortase A from S. pneumoniae and phylogenetic analysis of the sortase superfamily enzymes, followed by the details of protein purification & crystallization of two different constructs, wild-type SrtA from S. pneumoniae (Spn-∆N59SrtAWT and Spn-∆N81SrtAWT) as well as that of an active site cysteine mutant (Spn-∆N81SrtAC207A). This chapter includes X-ray intensity data collection of both types of crystals and data processing. Sortases are membrane anchored enzymes and therefore their expression as a full-length protein is a difficult task. Hence, the deletion of N-terminal transmembrane region from the enzyme is crucial for expression in its soluble form and is important for its successful crystallization. Thus, two wild-type constructs of S. pneumoniae sortase A, ∆N59SrtAWT (N-terminal 59 residue deletion) and ∆N81SrtAWT (N-terminal 81 residue deletion), and one active site mutant ∆N81SrtAC207A (N-terminal 81 residue deletion & active site Cys207 to Ala mutation) were cloned, expressed and purified. Cloning, expression and purification of the protein were done at the laboratory of our collaborator Prof. Rajendra P. Roy, Cell biology lab-II, National Institute of Immunology (NII), New Delhi, India. Crystallization of Spn-∆N59SrtAWT (~23 kDa) construct was initiated by manual screening using sparse matrix conditions from Hampton research. Initial trials were set up by following hanging-drop vapour diffusion method. Spn-∆N59SrtAWT construct crystallized in diamond, needle, rod and wedge-shaped crystal forms in more than one crystallization condition but they failed to diffract. Further trials were set up in microbatch plates that resulted in diamond-shaped crystals again, which diffracted up to a maximum of 4.0 Å resolution. Sequence comparison of the present construct was performed to modify the construct to achieve better diffraction. Thus, we made modifications in the Spn¬∆N59SrtAWT construct by deleting additional 22 residues at the N-terminal (i.e. total 81 residues deletion in the original sequence from the N-terminal) similar to SrtA from S. pyogenes. Hence, Spn-∆N81SrtAWT construct was prepared. For further crystallization experiments, we used the new construct Spn-∆N81SrtAWT. Similar to Spn-∆N59SrtAWT construct, crystallization set up for Spn-∆N81SrtAWT were done in microbatch plates at 293 K by using the Hampton conditions. During the crystallization set up, protein concentration was varied from 6-30 mg/ml. Notably, the protein crystals grown with 25 mg/ml protein concentration diffracted very well. Thus increasing the protein concentration helped to improve diffraction quality. Crystals obtained in Index-88 condition (0.2 M tri-ammonium citrate and 20% (w/v) PEG 3350, pH 7.0) diffracted up to 2.9 Å. Additive screen was used to improve its diffraction quality. This time many diffracting crystals were obtained and the best rod-shaped crystals grown in additive screen-79 (40% v/v (±)-1,3-butanediol) diffracted well up to 2.70 Å at home source. Thus, Spn-ΔN81SrtAWT crystallized at protein concentration of 25 mg ml-1 (in 10 mM Tris buffer, pH 7.5; 2 mM β-mercaptoethanol) with a condition containing 0.2 M tri-ammonium citrate and 20% (w/v) PEG 3350, pH 7.0, along with 40% v/v (±)¬1,3-butanediol as an additive agent by using microbatch-under-oil crystallization method. The chapter also includes crystallization of active site mutant Cys207Ala of ∆N81SrtAWT from S. pneumoniae (Spn-∆N81SrtAC207A). Spn-∆N81SrtAC207A mutant crystallized as a beautiful rectangular block type crystal (with a diffraction up to 2.7 Å at home source and up to 2.48 Å at synchrotron) at protein concentration of 25 mg ml-1 (in 10 mM Tris buffer, pH 7.5; 2 mM β-mercaptoethanol) with a condition containing 0.2 M tri-ammonium citrate and 20% (w/v) PEG 3350, pH 7.0, along with 1.0 M guanidine hydrochloride as an additive agent by using microbatch-under-oil crystallization method. Data collection was done on home-source diffraction facility for both the crystals however; mutant data in better resolution was collected by the author of the thesis at BM-14 beamline at ESRF, Grenoble, France. Thus, two crystals of SrtA, wild-type (Spn-∆N81SrtAWT) and its C207A mutant (Spn-∆N81SrtAC207A) were indexed satisfactorily in two space groups and their cell parameters are given in the following table 2. Table 2 Protein Space group a (Å) b (Å) c (Å) β (°) X-ray source Spn-∆N81SrtAWT P21 66.94 103.45 74.87 115.65 Home source Spn-∆N81SrtAC207A C2 155.57 113.33 81.34 90.80 Synchrotron The quality of both the data sets was assessed by SFCHECK and none of them showed twinning. Thus, the data sets collected were found appropriate and useful for structure determination as discussed in Chapter 3. Chapter 3 details the structure determination of Sortase A from S. pneumoniae for a wild-type construct (Spn-ΔN81SrtAWT) and for an active site cysteine mutant construct (Spn-ΔN81SrtAC207A). Sortase A from S. pyogenes was used as a search model in the molecular replacement (MR) method and a single solution for each data set was obtained through PHASER program. It resulted in four-molecules in wild-type sortase structure and six-molecules in the mutant structure in the respective crystal asymmetric unit. Iterative model building and structure refinement revealed a clear case of domain-swapping as observed in the electron density map. Finally, in the asymmetric unit of wild-type structure and in mutant protein structure two and three domain-swapped dimers were located, respectively. Simulated annealing and TLS refinement resulted in the protein structure with best refinement statistics. All these are elaborately discussed in Chapter 3. The last round of refinement of Spn-ΔN81SrtAWT converged to Rwork = 18.10% and Rfree = 23.39 % for 25152 unique reflections in the resolution range 30.7-2.7 Å whereas for Spn¬ΔN81SrtAC207A structure these parameters converged to Rwork = 18.25% and Rfree = 22.39% for 50010 unique reflections in the resolution range 47.15-2.48 Å. Chapter 4 describes the wild-type (Spn-ΔN81SrtAWT) as well as mutant (Spn¬ΔN81SrtAC207A) structures of Sortase A. The structure of Sortase A is not found in its commonly observed monomeric form but occur in a domain-swapped dimeric form. There are two dimers in Spn-ΔN81SrtAWT and three in Spn-ΔN81SrtAC207A as observed in the asymmetric unit. Each dimer contains two characteristic 8-stranded beta-barrel folds i.e. ‘sortase fold’ which is unique to the sortase superfamily. Unlike the structure of SrtA from other organisms known so far, the monomer does not form the 8-stranded beta-barrel all by itself. One monomer exchanges the β7 and β8 strands with the other monomer having β1 to β6 strands, thereby forming a complete 8-stranded β-barrel fold and such kind of two complete folds are present in each dimer. Because of the mutual swapping of strands between two monomers in a dimer, the dimer thus formed is defined as a domain-swapped dimer. This is the first time we have observed Sortase A structure in the domain-swapped dimeric form and is also the first example of domain-swapping in the sortase superfamily. Interestingly, all the catalytic residues (His141, Cys207 and Arg215) in each sortase fold in the swapped dimer lie at the secondary interface (open interface) generated by domain-swapping. Catalytic R215 (in one fold) interacts with D209 residue (in other fold of same dimer) through salt bridge interactions. Each dimer contains two pairs of such residues at the secondary interface but only one pair shows this kind of interaction. R215 (B-chain) interacts with D209 (A-chain) in AB dimer whereas R215 (D-chain) interacts with D209 (C-chain) in CD dimer. Asymmetry in the catalytic residues for their orientations and observed interactions at the secondary interface was evidenced. These active site residues were seen buried to a great extent except Arg215 which is slightly better exposed. It was difficult to find the exact substrate-binding pocket to approach the catalytic Cys207. However, biochemical and biophysical analyses (done at NII, New Delhi) provided strong evidence for the existence of the swapped-dimeric form at physiological pH as well. The enzyme exists with an equilibrium between its monomeric and dimeric forms, and the dimeric population is the most active species of the functionally active enzyme. An important role of Glu208 (in all the chains of two dimers; e.g. Chain A) was seen in the catalytic site where its side chain wobbles between His141 and H142 (both in Chain B) residues for interaction. Due to such kind of interactions the backbone conformation between C207-E208 (Chain B) shows variability, and coordinates the distance between His141 (ND1, Chain A) and Cys207 (SG, Chain B) each belonging to opposite chains in a swapped-dimer. The nature of side chain conformations of Glu208 in all the four sets of active site residues (in wild-type as well as in cysteine mutant structure) indicates that its movement presumably regulates thiolate-imidazolium acid-base pair formation which is a crucial condition for the sortase function where cysteine thiolate acts as nucleophile. Based on the crystal structure, the thesis candidate has suggested several mutants which might disrupt domain-swapping pointing to future studies on the system. Domain movement analyses by using HingeProt and DynDom servers indicate that the two-sortase folds joined with hinge loops in each dimer may show twist movement around the hinge axis. Possibly, such motion will affect the secondary interface covering active site residues and may allow increasing the exposure of the catalytic residues to perform catalysis. Presumably, such kind of domain movements may play a key role for the unique kind of regulatory mechanism for transpeptidase activity in sortase enzymes. However, more study has to be done to explore the role of these possibilities, if any, in the enzyme function and its regulation. Chapter 5 provides an introduction to non-standard amino acids, their sources and their uses in de novo peptide design; this is followed by a description of outcomes of structural investigations of modified peptides and their applications in various fields of medical and material science. Specifically, α, β-dehydrophenylalanine (ΔPhe), α-aminoisobutyric acid (Aib) and β-amino acids are discussed and their structures and conformational preferences are highlighted for their use in naturally occurring peptides or peptide fragments. Chapter 6 begins with an introduction to the human Islet Amyloid Polypeptide (hIAPP), which is an amyloidogenic protein and considered to be an important protein constituent of the amyloid plaques in pancreatic beta-cells in Type 2 diabetes patients. Therefore, fibrillization inhibition of hIAPP is considered as an important therapeutic approach to combat Type 2 Diabetes Mellitus (T2DM). In this chapter, the author of the thesis describes an approach to design peptide based inhibitors of hIAPP fibrillization using non¬standard amino acid ΔPhe (α,β-dehydrophenylalanine) residue. The first designed inhibitor has the sequence origin from hIAPP23-27 and it was developed by replacing I→ΔF (i.e. β¬favouring residue to helical conformation favouring) which resulted in FGAΔFL peptide. Fibrillization inhibition studies were done by co-incubation of hIAPP and FGAΔFL in 1:5 molar ratio and monitored by electron microscopy and thioflavin T binding assay that showed ~75% fibrillization inhibition. It suggested that the inhibitor is working effectively and thus the author determined its crystal structure by X-ray diffraction method. Peptide synthesis and experimental studies like electron microscopy and Thioflavin T binding assay were done in our collaborator’s laboratory at ICGEB, New Delhi, India. Subsequently a sequence similar peptide FGAΔFI was also designed by mutating L→I in the first inhibitor sequence. The resulting peptide FGAΔFI showed ~70% fibrillization inhibition. Following this success, crystal structures of both peptides were determined. FGAΔFL crystallized in P212121 space group whereas FGAΔFI crystallized in P21 space group. Though it was not anticipated, crystal structure analysis revealed that FGAΔFL and its analogue FGAΔFI harbour the anion receptor ‘nest’ motif. Both peptides dock with the helical form of hIAPP which may contribute to the inhibitory function of the peptides through their interaction with hIAPP in the core fibrillization region. These peptides effectively inhibit hIAPP fibrillization in vitro and it seems that these are unique examples of ‘nest-motif’ containing peptides that inhibit fibrillization. We also propose a model for fibrillization inhibition by these peptides; this has been published in Chemical Communications, a journal published by the Royal Society of Chemistry (RSC) and its reprint is enclosed within the thesis. In general, the approach described in the chapter may be applicable to target helices or helical intermediates and could be utilized in developing inhibitors useful, apart from T2DM, in other amyloid diseases including Alzheimer’s disease and Parkinson’s disease. Table 3 Peptide Crystal system and space group Unit cell details X-ray data Structure solution and refinement Agreement factor FGAΔFL Orthorhombic, P212121 a=8.9951 (9) Åb=13.0144 (12) Åc=27.7521 (24) ÅV=3248.82 (5) Å3 Z=4 Mo Kα(λ=0.71073Å) 4703 Unique reflections 2581 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 5.95 % for [|Fo| > 4σ (|Fo|)] FGAΔFI Monoclinic, P21 a=8.9951 (9) Åb=13.0144 (12) Åc=27.7521 (24) Å β=92.637 (2)°V=935.59 (2) Å3 Z=2 Mo Kα(λ=0.71073Å) 4024 Unique reflections 2612 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 5.02 % for [|Fo| > 4σ (|Fo|)] Chapter 7 describes another important but less studied core fibrillization fragment of hIAPP (hIAPP13-17) different than the hIAPP23-27 discussed in the previous chapter. It also discusses the development of fibrillization inhibitor design from this segment. The fragment hIAPP13-17 i.e. ANFLV crystallized in two space groups; C2 with one molecule in the asymmetric unit and P2 with two molecules in the asymmetric unit. In these structures, ANFLV peptide shows fully extended conformation i.e. a β-conformation. Crystal packing shows parallel β-sheet arrangement with the involvement of dry ‘steric-zippers’. The peptide prefers cross-strand Asn-Asn residue pair by side chain hydrogen bonding and is discussed in comparison with a few crystal structures of hIAPP fragments, solved by Eisenberg’s group, containing Asn residue in their sequence. It is observed that if the Asn is located in the sequence between two terminal residues the peptide will arrange itself in parallel beta sheet. This supports a structural model of hIAPP fibril in parallel beta sheet arrangement as the hIAPP sequence contains several Asn residues. To develop an inhibitor from ANFLV, a partial success was achieved where the Leu → Aib mutant i.e. ANFUV was developed. ThT (Thioflavin T) and TEM (Transmission electron microscopy) results show that the mutant peptide does not fibrilize on its own. This strongly supports the fact that the native peptide (ANFLV) lost its inherent fibrillization characteristic with the introduction of Aib in place of Leu i.e. the resultant mutant ANFUV is a non-fibrillizing peptide. The logic behind the development was to retain ANF in the same extended conformation and then break the β-strand with β-breaker residues. The structure of ANFLV showed parallel beta-sheets along with the additional side chain-side chain hydrogen bonding in the same direction as the fibril axis. Thus, we retained the ANF region to keep the sticky segment in the design and then Leu was mutated to Aib, a known β-breaker, to alter backbone conformation. The crystal structure of the peptide ANFUV resulted in the similar ANF region in beta conformation and Aib in helical conformation. Interestingly, in this situation the conformation of Aib develops a beta-bulge observed in the crystal packing and this bulge structure probably turned the peptide to have non-fibrillizing characteristics. These results will be useful in designing peptide inhibitors by using U as a beta breaker to inhibit hIAPP fibrillization. Table 4 Peptide Crystal system and space group Unit cell details X-ray data Structure solution and refinement Agreement factor ANFLV1 Monoclinic, C2 a=36.1350 (20) Åb=4.8050 (10) Åc=19.4190 (20) Å β=98.644 (5)°V=3333.40 (27) Å3 Z=4 Synchrotron (λ=0.77490 Å) 1982 Unique reflections 1825 [|Fo| > 4σ (|Fo|)] Direct methods: Sir92 & SHELXL97 11.71% for [|Fo| > 4σ (|Fo|)] ANFLV2 Monoclinic, P2 a=18.7940 (80) Åb=4.7970 (10) Åc=35.4160 (50) Å β=103.929 (10)°V=3099.03 (81) Å3 Z=4 Synchrotron (λ=0.77490 Å) 2651 Unique reflections 2580 [|Fo| > 4σ (|Fo|)] Direct methods: Sir92 & SHELXL97 15.39% for [|Fo| > 4σ (|Fo|)] ANFUV Monoclinic, P21 a=10.8140 (22) Åb=9.1330 (18) Åc=16.7540 (34) Å β=107.960 (30)°V=1574.07 (161) Å3 Z=2 Synchrotron (λ=0.97918 Å) 1426 Unique reflections 1398 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 5.45% for [|Fo| > 4σ (|Fo|)] Chapter 8 elaborates the self-assembly of α-dipeptides containing conformationally constrained achiral amino acid, α,β-dehydrophenylalanine (ΔF). The structural polymorphism in LFΔF peptide and the resulting self-assembly are discussed. Its D-isomer (DF∆F) and its racemic mixture (DLF∆F) are also discussed as these peptides self-assemble to give channel-forming assemblies. In addition to LFΔF, crystal structures of LYΔF, DLMΔF and LSΔF peptides and their self-assemblies are presented as well. Except DLMΔF xi and N-terminal protected DLFΔF (Boc-DLF∆F and Z-DLF∆F) peptides, the other dipeptides discussed in this chapter resulted in tubular structures of nanoscale dimensions through molecular self-aggregation. Table 5 Peptide Crystal system and space group Unit cell details X-ray data Structure solution and refinement Agreement factor LFΔF1 Hexagonal, P65 a=23.1873(24) Åb=23.1873(24) Åc=5.5260(8) ÅV=2573.01(5) Å3 Z=6 Mo Kα(λ=0.71073Å) 3489 Unique reflections 2915 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 6.19% for [|Fo| > 4σ (|Fo|)] LFΔF2 Monoclinic, P21 a=5.5739(2) Åb=13.1383(4) Åc=13.5816(4) Å β=96.137(2)°V=988.90(2) Å3 Z=2 Mo Kα(λ=0.71073Å) 4865 Unique reflections 3402 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 4.35% for [|Fo| > 4σ (|Fo|)] DFΔF Orthorhombic, P21212 a=13.1690(21) Åb=25.3673(40) Åc=5.5622(9) ÅV=1858.12(5) Å3 Z=4 Mo Kα(λ=0.71073Å) 4370 Unique reflections 3426 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 4.44% for [|Fo| > 4σ (|Fo|)] DLFΔF Monoclinic, P21/c a=5.5392(14) Åb=26.0376(55) Åc=13.1839(27) Å β=90.278(16)°V=1901.46(8) Å3 Z=4 Mo Kα(λ=0.71073Å) 2051 Unique reflections 1264 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 7.08% for [|Fo| > 4σ (|Fo|)] LYΔF Hexagonal, P65 a=23.5523(4) Åb=23.5523(4) Åc=5.5183(1) ÅV=2650.96(1) Å3 Z=6 Mo Kα(λ=0.71073Å) 2746 Unique reflections 1871 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 3.91% for [|Fo| > 4σ (|Fo|)] LSΔF Monoclinic, P21 a=5.2998(20) Åb=9.6732(30) Åc=14.1827(57) Å β=95.604(27)°V=723.62(20) Å3 Z=2 Mo Kα(λ=0.71073Å) 1978 Unique reflections 1558 [|Fo| > 4σ (|Fo|)] Direct methods: SHELXS97 & SHELXL97 13.59% for [|Fo| > 4σ (|Fo|)] DLMΔF Monoclinic, P21/c a=9.9032(5) Åb=8.6675(4) Åc=34.0283(18) Å β=90.088(3)°V=29

My research work for PhD has focused on: (i) the development and application of new NMR methodologies to solve challenging problems in structural biology and (ii) studying important biological systems to correlate their structural and functional aspects. I have worked on diverse research projects ranging from NMR methodology development to the study of structure and dynamics of protein-based nanotubes. Chapter 1 of my thesis gives brief introduction to bio-molecular NMR spectroscopy and the different biological systems that I have studied. In recent years, several new methods have emerged for rapid NMR data collection. One class of methods is G-matrix Fourier transform (GFT) projection NMR spectroscopy. GFT NMR spectroscopy involves phase sensitive joint sampling of two or more chemical shifts in an indirect dimension of a multidimensional NMR experiment. Chapter 2 describes a new method based on the principle of GFT NMR for increasing further the speed of data collection. In the current implementations of the GFT method, cosine/sine modulation of all chemical shifts involved in the joint sampling are collected and stored as separate FIDs. A post-acquisition data processing step (application of G-matrix) then separates the different inter-modulations of chemical shifts. Thus, joint sampling of K+1 spins results in 2K combination of chemical shifts (also representing 2K projection angles). One limitation of this approach is that even if only a few of the 2K components of the multiplet (or projection angles) is desired, an entire data set containing information for all 2K shift combinations is collected. We have proposed a simple method which releases this restriction and allows one to selectively detect only the desired linear combination of chemical shifts/projection angles out of 2K combinations in a phase sensitive manner. The method involves selecting the appropriate cosine/sine modulations of chemical shifts and forming the desired linear combination by phase cycling of the radiofrequency pulses and receiver. This will benefit applications where only certain linear combination of shifts are desired or/and are sufficient. Further, G-matrix transformation required for forming the linear combination is performed within the pulse sequence. This avoids the need for any post-acquisition data processing. Taken together, this mode of data acquisition will foster new applications in projection NMR spectroscopy for rapid resonance assignment and structure determination. Chapter 3 describes another GFT NMR-based method for rapid estimation of secondary structure in proteins. This involves the detection of specific linear combination of backbone chemical shifts and facilitates a clear separation and estimation of residues in different secondary structures of a given protein. This methodology named as CSSI-PRO (Combination of Shifts for Secondary structure Identification in PROteins), involves detection of specific linear combination of backbone 1Hα and 13C’ chemical shifts in a two dimensional (2D) NMR experiment. Such linear combination of shifts facilitates editing of residue belonging to α-helical/ β-strand regions into distinct spectral regions nearly independent of the amino acid type. This helps in the estimation of overall secondary structure content of the protein. Comparison of the estimated secondary structure content with those obtained from the respective 3D structures and/or the method of Chemical Shift Index (CSI) was carried out for 254 proteins and gives a correlation of more than 90% and an overall RMSD of 6.5%. The method has high sensitivity and data can be acquired in a few minutes. This methodology has several applications such as for high-throughput screening of proteins in structural proteomics and for monitoring conformational changes during protein folding and/or ligand-binding events. Chapter 4 (Part-A and Part-B) describes an area of my research which involves the study of structure and function in the Insulin-like Growth Factor Binding Protein (IGFBP) family. IGFBPs (six in number; IGFBP1-6) belong to the IGF-system, which plays an important role in growth and development of the human body. This system is comprised of the following components: (i) Two peptide hormones, IGF-1 and -2, (ii) type 1 and type 2 IGF receptors, (iii) six IGF-binding proteins (IGFBP; numbered 1-6) and (iv) IGFBP proteases. IGF-1 and -2 are small signalling peptides (~7.5 kDa) that stimulate action by binding to specific cell surface receptors (IGF-1R) evoking subsequent response inside the cell. Six soluble IGF binding proteins, the IGFBPs, which range in 22-31 kDa in size and share overall sequence and structural homology with each other, regulate the activity of the IGFs. IGFBPs bind strongly to IGFs (KD ~ 300-700 pM) to ensure that all the circulating IGF in the blood stream is sequestered and inhibit the action of IGFs by blocking their access to the receptors. Proteolysis of the IGFBPs dissociates IGFs from the complex, enabling them to bind and activate the cell surface receptors. IGFBPs have been recently implicated in different cancers and HIV/AIDS. However, the nature of their interaction with the ligand: IGF-1 or IGF-2 at a molecular level poorly understood. This is due to the difficulty in over-expressing these proteins in large scale and in soluble amounts which is required for structural studies. We have for the first time developed an efficient method for bacterial expression of full-length human IGFBP-2, a 33 kDa system, in soluble (upto 30 mg/ml) and folded form. Using a single step purification protocol, hIGFBP-2 was obtained with >95% purity and structurally characterized using NMR spectroscopy. The protein was found to exist as a monomer at the high concentrations required for structural studies and to exist in a single conformation exhibiting a unique intra-molecular disulfide-bonding pattern. The protein retained full biologic activity as evident from its strong binding to IGF-1 and IGF-2 detected using surface plasmon resonance (SPR). This study represents the first high-yield expression of wild-type recombinant human IGFBP-2 in E. coli and first structural characterization by NMR. Using different NMR methods, we are now in the process of elucidating the 3D structure of this molecule. Chapter 5 (Part-A and Part-B) describes our discovery of nanotubular structures formed by spontaneous self-assembly of a small fragment from the C-terminal domain of hIGFBP-2. The nanotubular structures are several micrometers long and have a uniform outer diameter of ~35 nm. These structures were studied extensively by NMR and other techniques such as TEM, fluorescence and circular dichroism (CD). The water soluble nanotubes form through intermolecular disulphide bonds due to the presence of three cysteines in the polypeptide chain and exhibit enhanced tyrosine fluorescence. Based on different experimental evidences we have proposed a mechanism for the formation of the nanotubes. This was considered as a breakthrough by the journal ChemComm and featured on the cover-page of the journal. An article highlighting the discovery was also published in RSC news. In recent years, a number of novel polypeptide and DNA based nanotubes have been reported. Our study reveals intrinsically fluorescent self-assembling nanotubes made up of disulphide bonds having the following novel properties: (i) their formation/dissociation can be controlled by tuning the redox conditions, (ii) they do not require the support of any additional chemical agent for self-assembly, (iii) they have high stability due to the involvement of covalent interactions, (iv) the monomer is a small polypeptide chain which can be chemically synthesized or produced using simple recombinant methods and (v) they possess high inherent fluorescence and can thus be easily detected against a background of other proteins. In addition, the presence of an RGD motif in this polypeptide fragment offers avenues for novel biomedical applications. The RGD motif is known to be recognized by integrins. The design of such self-assembling polypeptide fragments containing an RGD motif can be utilized to enhance the efficacy of cancer therapeutics. Towards this end, we have investigated the structural basis of formation of these nanotubular structures by NMR spectroscopy and proposed its application for cancer cell imaging.