A New Method for Analysis of Temperature Dependent IR Amide I Spectra of Peptides and Proteins

Abstract

Changes in the amide I' IR band with temperature are widely used for studies of structural transitions in peptides and proteins. Since amide I' exhibits inherent temperature dependent frequency shifts, standard mixture analysis methods are not applicable. To reliably extract the true thermodynamic states, frequency shifts of the component spectra must be explicitly taken into account. New methods termed Shifted Multivariate Spectra Analysis (SMSA) and parametric SMSA (pSMSA) were developed. SMSA uses no specific functional form for the transition (soft modeling), while the parametric variant (pSMSA) assumes a thermodynamic model (hard modeling). The implementation is optimized specifically for amide I' IR in that it takes advantage of known, linear dependence of the frequencies, as well as intensities, on temperature. The methods are first tested on sets of synthetic data with varying amounts of noise as well as on a real experimental amide I' data for the thermal unfolding of an α-helical peptide. The synthetic data tests demonstrate that the methods very reliably recover the correct parameters, although the non-parametric SMSA is subject to the rotational ambiguity. Application to the peptide experimental amide I' data illustrates additional complications encountered with the analysis of real systems, namely the correction for the side-chain spectral bands and interference of spectral shape changes. Finally the pSMSA is applied to the analysis of site-specific thermal unfolding of two small α-helical proteins from sets of multiple 13C isotopically edited amide I' spectra.