A quantitative reconstruction of the amide I contour in the IR spectra of globular proteins: from structure to spectrum.

The Amide I contours of six globular proteins of varied secondary structure content along with a peptide model for collagen and pulmonary surfactant protein C have been simulated very closely by using a modified GF matrix method. The starting point for the method uses the three-dimensional structure as obtained from the Protein Data Bank. Elements of the interactions between peptide groups (e.g., transition dipole coupling) are very sensitive to tertiary structure, thus the current formalism demonstrates that the Amide I contour may be useful for a more detailed probe of 3-D conformation that goes beyond the traditional use of this band to probe the percentages of particular elements of secondary structure. For example, postulated changes to a known structure can be tested by comparing the new simulated band to the experimental band. A number of refinements to the transition dipole interaction calculation have been made. Most of the important interactions between the C=O oscillators that define the Amide I mode appear to have been identified, including through space transition dipole coupling, through valence bond and through hydrogen bond coupling. The eigenvector matrix produced by the method permits the contribution of each peptide group to the spectrum to be precisely determined. Analysis of the results shows that the often-used structure-frequency correlations are at best approximate and at worst misleading. The subbands from helices, sheets, turns, and loops are much broader and more overlapped than has been commonly assumed. Furthermore, the traditional alpha-helical marker band may be substantially distorted in short segments. Difference spectra based on isotope editing, a technique thought capable of revealing the spectral contributions of individual peptide groups, are shown to be prone to misinterpretation.