Computational methods of analyzing, simulating, and modeling proteins are essential towards understanding protein structure and its interactions. Computational methods are easier as not all protein structures can be determined experimentally due to the inherent difficultly of working with some proteins. In order to predict, design, analyze, simulate or model a protein, data from experimentally determined proteins such as those located in the repository of the Protein Data Bank (PDB) are essential. The assumption here is that we can use pieces of known proteins to piece together a "new" protein hence, de novo protein design. The analysis of the geometric relationships between secondary structure elements in proteins can be extremely useful to protein prediction, analysis, and de novo design. This thesis project involves creating a database of protein secondary structure elements and geometric information for rapid protein assembly, de novo protein design, prediction and analysis.

Threefoil is a small engineered protein of 141 amino acids which is a member of the beta-trefoil superfamily, with three-fold symmetry and high thermal and kinetic stability. Its primary sequence was designed based on a predicted beta-trefoil glycosidase from the halophilic Archaeon Haloarcula marismortui. Threefoil predominantly forms inclusion bodies when over-expressed in Escherichia coli at 37??C, with little to no protein soluble in the cytoplasm. Nevertheless, Threefoil is capable of adopting a native beta-trefoil structure when refolded from solubilized inclusion bodies. The focus of this thesis is on characterization of the folding and stability of Threefoil through thermodynamic and kinetic experiments for wild-type Threefoil, in addition to sugar- and metal-binding studies and characterization of Threefoil inclusion bodies. Various Threefoil mutants, designed to increase protein stability, are also characterized to probe the origins of, as well as to give insight into, the mechanism of inclusion body formation. The thermodynamic and kinetic stability of wild-type Threefoil was studied using spectral probes, mainly fluorescence, circular dichroism (CD) and dynamic light scattering (DLS). The major observed spectral changes in kinetic and thermodynamic experiments can be fit to a 2-state transition between the folded state and a denatured state containing extensive residual secondary structure. At high protein concentrations, the folding of wild-type Threefoil is complicated by protein misfolding and aggregation. As Threefoil is remarkably resistant to denaturation even at high concentrations of urea and guanidine hydrochloride (GuHCl), studies were also conducted in guanidine isothiocyanate (GuSCN), which is a much stronger denaturant than urea and GuHCl. Remarkably, the time that is required for Threefoil samples to reach equilibrium in renaturation curves is approximately 100 days, while equilibrium by denaturation in the stronger denaturant, GuSCN, requires more than two years. The expression levels of Threefoil mutants A62V, Q78I, D85P and D93P were also studied. None of the four mutants studied exhibited any pronounced increase in solubility compared to wild-type when expressed in E. coli.

Synthetic construction of protein molecules has been widely pursued over the last two decades. A primary goal behind de novo protein design has been to build minimal systems by capturing the essential features of protein structures. Such minimal models can be used to understand underlying principles governing folding, structure, and function of proteins molecules. Several approaches envisioning successful construction of synthetic proteins have been described over the years, some of them being admirably successful (DeGrado et al, 1999; Richardson et al> 1992; Baltzer, 1998). Specific patterning of polar and apolar residues in synthetic sequences has been widely used to achieve designed polypeptide structures like helix bundles (DeGrado et ah, 1999) and (3-sheets (Smith and Regan, 1997; Lacroix et a/., 1998), with reliance on hydrophobic driving forces for folding. Our laboratory has been pursuing a distinctly alternative approach, that employs stereochemically constrained amino acids to generate specific secondary structures which can then be assembled into composite structures by appropriately chosen linking segments. This approach, which involves linking prefabricated modules of secondary structures can be termed as a "Meccano set" approach to protein design (Balaram, 1992). The studies embodied in the present thesis describe attempts at construction of synthetic polypeptide motifs using the stereochemically directing influence of conformationally constrained amino acid residues, such as DPro or Aib (α-aminoisobutyric acid). This thesis is subdivided into 8 chapters, with Chapter 1 providing a perspective of the field of protein design. Subsequent chapters (2-8) describe studies directed towards the specific goal of construction of polypeptide motifs. Chapter 2 describes synthesis and conformational characterization of two octapeptides Boc-Leu-Val-Val-DPro-LAla-Leu-Val-Val-OMe (1) and Boc-Leu-Val-Val-DPro-DAla-Leu-Val-Val-OMe (2), designed to investigate the effect of specific β-turn stereochemistry on β-hairpin structures. 500 MHz NMR studies establish that both peptides 1 and 2 adopt predominantly β-hairpin conformations in chloroform and methanol solutions, with interstrand registry established by observation of long-range nuclear Overhauser effects (NOEs). Specific NOEs provide evidence for a type II' β-turn conformation for the DPro-LAla segment in 1, while the NMR data suggest that a type I' DPro-DAla β-turn conformation predominates in the peptide 2. The crystal structure of 1 reveals two independent molecules in the crystallographic asymmetric unit, both of which adopt β-hairpin conformations nucleated by a type II’ β-turn across DPro-LAla and stabilized by 3 cross strand hydrogen bonds. These designed β-hairpins with defined tight turns produce characteristic vibrational circular dichroism (VCD) patterns, demonstrating the utility of VCD as a probe for conformational analysis of β-hairpins. In Chapter 3, we present conformational analysis on designed β-hairpin sequences incorporating a 'Phe-Phe' residue pair at a non-hydrogen bonding position. Two octapeptides Boc-Leu-Phe-Val-DPro-Gly-Leu-Phe-Val-OMe and Boc-Leu-Phe-Val-DPro-Ala-Leu-Phe-Val-OMe were synthesized and conformationally characterized by 500 MHz NMR spectroscopy. Specific NOEs observed in solution provide conclusive evidence favoring specific orientation effects pertaining to the 'Phe-Phe' pair. The peptides exhibited anomalous electronic CD, which has been explained in terms of aromatic contributions by the side chain chromophores. Interestingly, the VCD patterns obtained for these peptides were almost identical to those obtained for other β-hairpins, described in Chapter 2. Chapter 4 describes the synthesis and conformational analysis of designed decapeptide sequences with centrally located DPro-Xxx β-trun segments. Two sequences Boc-Met-Leu»Phe-Val'DPro-Ala-Leu-Val-Val-Phe-OMe (1) and Boc-Met-Leu-Val-Val-^ro-Gly-Leu-Val-Val-Phe-OMe (2) were designed to study the effect of chain length elongation, of β-strands, on designed β-hairpin structures. 500 MHz NMR studies establish β-hairpin folds in both these sequences, with strand segments aligned even at the termini of the structures. Multi-stranded, antiparallel β-sheet structures can be generated by successive placement of β-hairpin sequences in a single polypeptide chain. The successful construction of three stranded β-sheet structures is described in Chapter 5 of this dissertation. A 14-residue peptide Boc-Leu-Phe-Val-DPro-Gly-Leu-Val-Leu-Ala-DPro-Gly-Phe-Val-Leu-OMe (LFV14) was designed such that it is composed of three strand segments linked by two DPro-Gly turn segments. The peptide showed excellent solubility in apolar media, permitting detailed conformational analysis by 500 MHz NMR spectroscopy in organic solvents. Observation of long-range, interstrand NOEs, diagnostic of multiple hairpin structures, provides conclusive evidence for a predominantly populated three stranded β-sheet structure in solution. Extension of this strategy has been described in which an 18-residue peptide, Arg-Gly-Thr-Ile-Lys-DPro-Gly-Val-Thr-Phe-Ala-DPro-Ala-Thr-Lys-Tyr-Gly-Arg, was designed with enhanced solutility in water to probe (β-sheet structure formation in aqueous and mixed aqueous-methanol systems. NMR data provided conclusive evidence in favor of the desired structure being significantly populated in methanol and methanol-water mixtures (50 %, v/v). In water, spectroscopic evidence suggests that the long-range order expected of a three-stranded structure is lost, possibly due to water invading the interstrand hydrogen bonds. Successful construction of a four-stranded antiparallel β-sheet structure has been demonstrated in Chapter 6. A 26-residue peptide Arg-Gly-Thr-Ile-Lys»DPro-Gly-Ile-Thr- Phe-Ala-DPro-Ala-Thr-Val-Leu-Phe-Ala-Val-DPro-Gly-Lys-Thr-Leu-Tyr-Arg was designed to have four strand segments linked by three DPro-Xxx turn segments. The peptide exhibited excellent NMR properties permitting structure determination by analysis of NOE data, which revealed that a four stranded β-sheet structure is indeed populated in methanol. Structural studies on this peptide in mixed methanol-water established that the four stranded β-sheet is appreciably populated at a composition of 50 % (v/v) methanol-water mixture, with the β-sheet structure still detectable even at a composition of 70 % water-30 % methanol. In a completely aqueous environment, the β-sheet structures is significantly disrupted, presumably due to solvent invasion. The nucleating β-turns, however, appear to have retained their structural integrity even in this competitive environment. Chapter 7 describes the insertion of L-Lactic acid (Lac), a hydroxy acid, into polypeptide helices stabilized by a-aminoisobutyricacid (Aib). This study was undertaken to investigate the effect of hydrogen bond deletion on peptide helices. Crystal structure determination of three oligopeptides containing Lac residues has been performed. Peptide 1, Boc-Val-Ala-Leu-Aib-Val-Lac-Leu-Aib-Val-Ala-Leu-OMe, and peptide 2, Boc-Val-Ala-Leu-Aib-Val-Lac-Leu-Aib-Val-Leu-OMe adopt completely helical conformations in the crystalline state, with the Lac(6) residue comfortably accommodated in the center of a helix. NMR studies of peptide 1 and its all amide analog 4, Boc-Val-Ala-Leu-Aib-Val-Ala-Leu-Aib-Val-Ala-Leu-OMe, provide firm evidence for a continuous helical segment in both the cases. In a 14-residue peptide 3, Boc-Val-Ala-Leu-Aib- Val- Ala-Leu- Val- Ala-Leu- Aib-Val-Lac-Leu-OMe, residues Val( 1 )-Leu( 10) adopt a helical conformation, which is terminated by formation of a Schellman motif, with Aib(ll) as the site of chiral reversal. The loss of the hydrogen bond at the C-terminus appears to facilitate the chiral reversal at Aib(l 1). In the final section of this thesis, Chapter 8, successful construction of a synthetic motif containing two distinct elements of secondary structure, a (β-hairpin and a helix, has been described. The design of a 17-residue peptide Boc-Val-Ala-Leu-Aib-Val-Ala-Leu-Gly-Gly-Leu-Phe-Val-DPro-Gly-Leu-Phe-Val-OMe, BH17, is based on a modular approach, in which previously characterized β-hairpin (Leu-Phe-Val-DPro-Gly-Leu-Phe-Val) and helix (Val-Ala-Leu-Aib-Val-Ala-Leu) modules are linked by a Gly-Gly linker. The positioning of the achiral Gly residue at position 8 facilitates termination of the potential helical segment (residues 1-7) by formation of a Schellman motif. Gly(9) is anticipated to be the sole conformationally flexible residue. NMR studies on BH17 indicated the presence of both the helix (residues 1-7) and the β-hairpin (residues 10-17) structures in the sequence, with four major conformational possibilities at the linking segment. Crystal structure determination of BH17 revealed that the two elements of structure are approximately arranged in an orthogonal fashion. The crystal structure validates the original premise that a modular assembly strategy may be viable for the construction of larger synthetic structures. Chapter 9 summarises the major results of this thesis. (For formulae, please refer "pdf" format)