Theory of .alpha.-helix-to-random-coil transition of two-chain, coiled coils. Application of the augmented theory to thermal denaturation of .alpha.-tropomyosin

Abstract

The statistical mechanical theory of the helix-to-random-coil transition in two-chain, a-helical coiled coils has recently been augmented by inclusion of the effects of loop entropy and out-of-register (mismatched) structures. This theory is applied to experimental data on non-cross-linked a-tropomyosin at nearly neutral and at acidic pH, using extant values of helix initiation (u) and propagation (s) parameters for each amino acid in the sequence. A semiquantitative fit of the helix content (from circular dichroism measurements) vs. temperature (0-80 C) is obtained at each pH, covering a 1000-fold range of protein concentration. The algorithms for the mean interhelix interaction free energy per mole of turn pairs (RT In w(T)) needed to produce the fit at each pH provide curves of RT In w(T) vs. T that are similar in range and in shape, each showing a minimum near room temperature. Theory is also compared with independent experiments, in particular light scattering and cross-linkability studies at nearly neutral pH. The temperature dependence of the weight-average molecular weight at nearly neutral pH, as recently determined by light scattering, agrees well with the theoretical prediction. The observed high degree of cross-linkability of tropomyosin in the native state can be reconciled with the theoretically calculated fraction of in-register molecules under the benign conditions of the cross-linking experiments. Examination shows that the principal cause of the greater stability of a-tropomyosin at low pH lies in the augmented short-range (a,s) interactions of aspartic and glutamic residues over those of the aspartate and glutamate species which predominate near neutral pH. In fact, it is shown that, with small adjustment (within experimental error) in these parameters, the same interhelix interaction free energy algorithm can be used to explain the full range of data at both pHs. A discussion of the implications of this result is given, wherein it is shown that the interhelix salt bridges, while they may provide enough free energy at nearly neutral pH to ensure that the helices associate in parallel, make a contribution to the total interhelix interaction that is relatively small compared with the hydrophobic contribution. The statistical theory developed here is compared with the all-or-none-stages model brought forward elsewhere; it is suggested that the latter disagrees with the recent light scattering data and is difficult to reconcile with accepted ideas concerning loop entropy.