The use of spectral decomposition via the convex constraint algorithm in interpreting the CD-observed unfolding transitions of coiled coils.

Decomposition of CD spectra for the unfolding of both coiled-coil and single-helical molecules is carried out via the convex constraint algorithm (CCA) [A. Perczel, M. Hollósi, G. Tusnády, and G. D. Fasman (1991) Protein Engineering, Vol. 4, pp. 669-679]. Examined are (1) our thermal unfolding data for rabbit alpha alpha-tropomyosin and chicken gizzard gamma gamma-tropomyosin coiled coils, and for a 35-residue, tropomyosin-model peptide that forms single helices, not coiled coils; (2) extent pH-induced unfolding data for 50- and 400-residue poly-L-glutamic acid. Each set of spectra shows a sharp isodichroic point near 203 nm. We find here that the CCA is of sharply limited use for analyzing such data. The component spectra obtained for a given substance not only depend on the particular experimental spectra included and on the chosen number of component spectra, but all pass through the experimental isodichroic point. The latter is physically unlikely for more than three component spectra, and physically impossible for conformers, such as beta structures, having known isodichroic points elsewhere. Our conclusions are in contrast to those of an extant decomposition via CCA of thermal spectra for rabbit alpha alpha-tropomyosin [N. J. Greenfield and S. E. Hitchcock-DeGregori (1993) Protein Science, Vol. 2, pp. 1263-1273] that postulates the existence of five conformers, including beta structures, in the unfolding. Moreover, an extant diagnostic based on the theta 222/theta 208 ratio and allegedly distinguishing between spectra for coiled coil and for single alpha-helix is shown here to be unreliable.