Structural and functional motifs in synthetic peptides

Abstract

The construction of complex protein folds relies on the precise conversion of a linear polypeptide chain into a compact 3-dimensional structure. A variety of approaches can be employed in the understanding of the forces underlying such a process, namely (a) de novo design of secondary structural elements and their assembly into mini-proteins and (b) study of structures adopted by appropriate protein fragments in isolation. De novo design of various secondary structural motifs such as isolated helices (Rohl & Baldwin, 1998), ?-hairpins (Gellman, 1998; Balaram, 1999), and small ?-sheets (Gellman, 1998; Lacroix et al., 1999, Balaram, 1999) has seen much success in the recent past; with some spectacular examples of well-defined helical bundles (Harbury et al., 1998; Walsh et al., 1999; Lombardi et al., 2000) and synthetic peptides possessing functional characteristics (Baltzer, 1998; Jarvo et al., 1999). The above attempts at peptide/protein design have relied on the use of residues with appropriate secondary structural propensities and polar/nonpolar patterning of residues within the designed sequence. An alternate approach exploits the ability of conformationally constrained amino acid residues to nucleate desired structure. Well-defined elements of secondary structure, like ?-hairpins and helices, can be stabilized in acyclic peptides by incorporating stereochemically constrained residues like ?-aminoisobutyric acid (Aib) (for nucleation of helical structures) and DPro (for nucleation of ?-hairpins). This thesis describes specific attempts at the construction of such modules, which can then be assembled to form compact folds, or used as templates for the grafting of functional sites. In a parallel approach to understanding the principles underlying protein folding, attempts at studying membrane protein structure have also been detailed. This thesis consists of six chapters: Chapter 1 reviews the current paradigms in the field of peptide/protein design in order to define the context of the present studies, while Chapters 2-5 describe specific attempts to construct and study polypeptide structural and functional motifs.

Chapter 2 describes the design, synthesis, and characterization of a five-stranded ?-sheet, B5, with the aim of constructing well-defined ?-sheets that can act as precursors for the rational design of ?-sandwiches and ?-barrels. Positioning of DPro-Xxx sequences at the i/i+1 positions of ?-turns leads to the nucleation of tight turns, which serve to stabilize ?-hairpins and in consequence, multi-stranded ?-sheets. In addition, a metal ion binding site was created on the extended backbone of B5, by juxtaposing histidine residues on adjacent strands of the ?-sheet, to permit metal ion-mediated assembly of B5 into ?-sandwich structures. B5, a 35 residue peptide [RGIKVDPGETNTDPSVQFHTIDPGYKT LHEDPARIVLK], was synthesized by conventional solid-phase techniques and conformationally characterized by 500 MHz 1H NMR spectroscopy. B5 was demonstrated to fold into a five-stranded ?-sheet in methanol, with the DPro-Xxx motifs forming type II' turns. Observation of key cross-strand NH/NH and CaH/CaH NOEs have established the correct registry of strand segments. Chelation of both Ni2+ and Zn2+ ions by the imidazole groups of His 17 and His26 of B5 (in methanol) was established by ESI-MS experiments.

Chapter 3 details attempts at the assembly of multi-stranded ?-sheets into compact ?-sandwiches and ?-barrels utilizing the disulfide bond as a structural template. Four [B4cys: RGEC(Acm)KFTVDPGRTALNTDPAVQKWHFVLDPGYKCEILA] and five-stranded (B5cys: GEIKVDPGLTNTDPSVQWATIDPGFTLHVVNFDPAYECRIVLR) amphipathic ?-sheets, bearing single cysteine residues in their terminal strands were synthesized and homo-dimers covalently linked through disulfide bonds were created in an attempt to form extended ?-sheets with eight (B4cys2) or ten (B5cys2) contiguous strands. Folding into a compact ?-sandwich structure could, in principle, be achieved by manipulating solvent conditions. CD studies suggest that both B4cys2 and B5cys2 adopt ?-sheet conformations in methanol and introduction of a disulfide bond appears to have stabilized the ?-sheet structure in the homodimers as compared to the monomers. Interestingly, introduction of a disulfide bond in B4cys enhanced structure formation in aqueous solvents. Two-dimensional NMR experiments performed on B4cys2 in water indicate the presence of a ?-sheet. Observation of the requisite NOEs in NOESY experiments indicate the formation of a well-defined, central, 4-stranded region with the desired strand registry, although the structural integrity of the terminal strands has been compromised due to solvent invasion. The observation of many aromatic-methyl interactions across the strands of the interface validates the use of a disulfide bridge as a conduit to designing extended sheets even in aqueous solvents.

Chapter 4 describes the use of rigid helical modules as templates onto which catalytic sites can be grafted. A family of 15-residue helical peptides have been designed that can serve as model systems in the study of the reaction mechanisms and intermediates characterizing the early stages of the protein glycation reaction. The reaction of glucose with the ?-amino groups of lysine residues in proteins, with the concomitant formation of a Schiff base, is a consequence of prolonged or increased exposure to glucose, relevant to situations in diabetes and aging disorders. The Schiff base undergoes rearrangement to form a ketoamine product (the Amadori rearrangement), which, after a series of oxidative rearrangements, results in the formation of extensive protein cross-links and other advanced glycation endproducts (AGEs), which have been implicated in various complications associated with diabetes and aging. The committing step in protein glycation is the formation of the ketoamine product from reversibly formed Schiff base. We have utilized helical peptides bearing a lysine residue, the site of glycation, and other potentially catalytic residues implicated in the rearrangement reaction (such as aspartic acid and histidine) to assay their relative roles in Schiff base formation and the Amadori rearrangement. The reaction site lysine and the catalysts are brought into spatial juxtaposition by means of the helix periodicity, which dictates that residues at i and i+4 positions are close together, but residues at i and i+2 positions are far apart. The results demonstrate that catalysis of the Amadori rearrangement by a proximal Asp residue may be important in determining the rate of irreversible glycation. The reactions have been monitored using electrospray ionization mass spectrometry.

Chapter 5 deals with alternate approaches to investigating membrane protein structure. Traditional methods of studying protein structure have often failed in the case of membrane proteins. We have chosen a small multidrug resistance protein (MDR) from E. coli, EmrE, as a model system. Multidrug resistance pumps (MDRs) are membrane translocases that have the ability to extrude a variety of unrelated cytotoxic drugs from the cell (Lewis, 1994; Van Bambeke et al., 2000). MDRs are present in a wide range of organisms and are very large proteins ranging between 400-1500 amino acids in length (Lewis, 1994). But there exist a family of MDRs, the miniTEXANS, which renders bacteria resistant to a variety of toxic, lipophilic cations (Saier et al., 1997; Paulsen et al., 1996) but are merely 100-120 amino acids in size. In consequence, the miniTEXANs have often been used as model systems to study MDRs and EmrE is one member of this family of translocases. Two alternate structural models have been proposed for EmrE, a four-helix bundle model and a ?-barrel model. Preliminary structural characterization of the protein has been performed by other groups (Arkin et al., 1996; Schwaiger et al., 1998; Tate et al., 2001) and favors the helical bundle model. The validity of either model has been probed experimentally by synthesizing overlapping peptides, ranging in length from 19-27 residues, derived from the sequence of EmrE and studying their conformation in solution. The choice of peptides was done to provide sequences of two complete, predicted transmembrane helices [the transmembrane segments (TMS) 1 and 3: peptides HI and H3] and two helix-turn-helix motifs (TMS1-loop-TMS2, peptide A and TMS2-turn-TMS3, peptide B). One such helix-turn-helix peptide (B) would also correspond to a hairpin in the ?-barrel model. CD and two-dimensional NMR experiments were performed on the peptides in SDS micelles, in an attempt to mimic the membrane environment. NMR experiments indicate that peptide HI adopts a helical conformation in SDS micelles, although H3 proved insoluble and hence recalcitrant to purification and structural analysis. Peptides A and B both demonstrated helical structure in parts of the sequence. Interestingly, in peptide A, NOE data suggest the presence of two individual helical domains, separated by an unstructured region, a helix-loop-helix motif. Further, the limits of the helix in peptides HI and B, as defined by NMR, appears to correspond to those predicted in the four-helix bundle model. This study implies that study of judiciously selected peptide fragments can prove useful in the structural elucidation of membrane proteins, and such understanding might further lead to the ability to design peptide modules that adopt definite structures in membrane environments.

Chapter 6 summarizes the main results and conclusions derived from these chapters.