The work contained in this thesis is focused on the optical spectroscopic characterization and thermodynamic analysis of selected 13C isotopically labeled -helical and -hairpin peptide models. Studies of such small peptide systems provide valuable insight toward understanding -helix and -sheet folding and stability. With the aid of isotopic substitutions, we can gain useful information about the nature of vibrational coupling in these two common secondary structures. Coupling provides the fundamental interaction that leads to the spectral variations characteristic of different secondary structures. In addition, we can study the local secondary structure of these peptides with residue- or regio- specificity by substituting 13C labeled residues at different positions of the peptide sequence. The primary optical spectroscopic methods utilized in this thesis are infrared (IR) absorption and vibrational circular dichroism (VCD), both of which are sensitive to isotope substitutions. Theoretical calculations on both ideal -helical and -hairpin models have successfully predicted the IR and VCD amide I spectra of the peptides studied, especially for the 13C band.
First, the nature of vibrational coupling in a helix was studied using a series of isotopically labeled (13C on two or more amide C=Os), 25 residue, -helical peptides with the sequence Ac-(AAAAK)4AAAAY-NH2. Theoretical simulations on Ac-A24-NHCH3 using density functional theory (DFT) parameters predicted that the vibrational coupling between i, i+1 (sequential labeling) and i, i+2 residues (alternate labeling) differ in sign, thus leading to a different shift in the 13C amide I frequency and a reversal of the 13C VCD pattern. IR and VCD spectra of the sequential and alternate labeled peptides confirmed these predictions, providing an excellent example of our ability to correctly model coupling in small peptides with well-defined secondary structures. Thermal denaturation of these labeled peptides was used to study the effect of unfolding on vibrational coupling. Site differentiation is shown to be lost in the disordered peptides.
Next, the role of cross-strand vibrational coupling in understanding -sheet folding was studied using a turn stabilized 12-residue -hairpin peptide model based on a design from Gellman. Use of a non-stereogenic -aminoisobutyryl-glycyl (Aib-Gly) dipeptidyl sequence in the i+1 and i+2 positions nucleates a stable type I’ -turn (confirmed by NMR), and promotes a stable, highly soluble -hairpin conformation in water. Theoretical simulations suggest the cross-strand vibrational couplings are different for a large H-bond ring and a small H-bond ring in anti-parallel strands, as confirmed by experimental IR spectra. Two 13C labels forming a large H-bond ring yield an amide I band with a higher intensity at a higher frequency than do two 13C labels forming a small H-bond ring. However, labeling on positions near the -turn region, did not generate a useful 13C probe. Intensities of the 13C amide I band decayed with increasing temperature resulting in a broad thermal transition for all the labeled hairpins. The large H-bond ring case, showed the most dramatic effect, which suggests that this labeling pattern could be useful in probing local structural change.
Finally, a further study of vibrational coupling in -hairpin peptides used another -hairpin template called Trpzip2 designed by Cochran. This hairpin has four tryptophans forming a strong hydrophobic core, a very different stabilization mechanism from the previous -hairpin peptide. We designed three variants (with Ala mutations at selected positions) of the original Trpzip2 to facilitate isotopic labeling. IR and electronic circular dichroism (ECD) confirmed these variants have very similar conformation. Isotopic substitutions on the 3 Ala variant using several pairwise cross-strand labeling patterns showed that the vibrational coupling in this -hairpin peptide has the same 13C features in their IR spectra as did those with the same labeling pattern in the Gellman variant. The cross-strand coupling is shown to be characteristic of the fundamental H-bonding patterns and not specific to peptide sequences. The thermal unfolding transitions of the Trpzip peptides are more sigmodial than for the Gellman variant, yielding more reliable Tm values but derivation of a full set of thermal parameters was not achieved with a two-state model due to unsatisfactory properties of the fit which may suggest an insufficient physical model.