A unified theory of the B-Z transition of DNA in high and low concentrations of multivalent ions.

We showed recently that the high-salt transition of poly[d(G-C)]. poly[d(G-C)] between B-DNA and Z-DNA (at [NaCl] = 2.25 M or [MgCl(2)] = 0.7 M) can be ascribed to the lesser electrostatic free energy of the B form, due to better immersion of the phosphates in the solution. This property was incorporated in cylindrical DNA models that were analyzed by Poisson-Boltzmann theory. The results are insensitive to details of the models, and in fair agreement with experiment. In contrast, the Z form of the poly[d(G-m5C)] duplex is stabilized by very small concentrations of magnesium. We now show that this striking difference is accommodated quantitatively by the same electrostatic theory, without any adjustable parameter. The different responses to magnesium of the methylated and nonmethylated polymers do not come from stereospecific cation-DNA interactions: they stem from an experimentally derived, modest difference in the nonelectrostatic component of the free energy difference (or NFED) between the Z and B forms. The NFED is derived from circular DNA measurements. The differences between alkaline earth and transition metal ions are explained by weak coordination of the latter. The theory also explains the induction of the transition by micromolar concentrations of cobalt hexammine, again without specific binding or adjustable parameters. Hence, in the case of the B-Z transition as in others (e.g., the folding of tRNA and of ribozymes), the effect of multivalent cations on nucleic acid structure is mediated primarily by nonspecific ion-polyelectrolyte interactions. We propose this as a general rule for which convincing counter-examples are lacking.