Correlation of local stability of DNA during melting with environmental conditions

Abstract

DNA MELTING studies have revealed fine structural features in the melting profiles of many natural DNAs^1–4. Restriction fragments have been used to assign modes showing up in melting profiles to the denaturation of well delineated segments inside the DNA molecule^1,3 and possible correlations of these segments with biological functions have been suggested^1,5,6. However, when the exact sequence of natural DNAs became available, known genetic or regulatory features extending over tens to hundreds of base pairs did not appear to be characterized systematically by simple physical parameters. For example, the base pair composition fluctuation of ΦX174 DNA is almost smooth^6,7, whereas several distinct modes are seen in experimental melting profiles^8,9, each reflecting separate melting of segments of hundreds of base pairs. Attempts have been made to calculate the melting profiles of DNA from its sequence^10–13, and two calculated melting profiles have been published for ΦX174 DNA. Both calculated curves show fine structure, but comparison with the experimental profiles showed ‘no obvious systematic relationship’ (ref. 8). The very different profiles observed for the same ΦX174 DNA in two different ionic strengths⁸ could not be accounted for by the theory used. Thus the melting profile is not determined by the DNA sequence alone. However, recently good agreement between the calculated and experimental melting profile of ΦX174 DNA was obtained by Azbel^14–16. Azbel's model is based on the observation that nucleation of the helix–coil transition requires much higher energies than that exchanged between base pairs, so that behaviour of DNA is mainly correlated to its ground state at a given temperature inside the melting range (see ref. 16). We have applied Azbel's model to another DNA of known sequence (a restriction fragment of SV40 DNA) and ΦX174 DNA in two different environmental conditions, and we obtained good agreement between predictions and experimental observations.