Abstract
The probability of DNA base‐pair opening was calculated at temperatures below the denaturation region. The helix–coil transition theory, modified to include different nearest‐neighbor interactions, was employed. Predictions of base‐pair opening employing DNA melting‐curve parameters differed considerably from predictions based on parameters evaluated from synthetic RNA oligomer data of Gralla and Crothers (G‐C) and formaldehyde–DNA binding experiments of McGhee and von Hippel (M‐vH). Calculations based on the latter parameters indicate a base‐pair‐opening probability of 10^−2^–10^−3^ at 35°C in 0.1__M__ NaCl. DNA melting‐curve parameters predict values about 10^3^ smaller. At a temperature 10°C below the transition midpoint of a specific DNA sequence, DNA melting parameters predict base‐pair opening of about 10^−3^, whereas the G‐C and M‐vH parameters predict that ≳0.16 of the DNA is melted. Experiments and theoretical assumptions relevant to the calculations are analyzed. Evidence suggests that DNA melting parameters are valid when the average loop size exceeds some minimum value, whereas the G‐C and M‐vH parameters appear more valid for single base‐pair loops. A reconciliation of the two sets of predictions can be made if interactions extending beyond neighboring base pairs are considered. Such interactions will make the parameters of the nearest‐neighbor model appear to change with the average loop size. Experiments that may provide further measurements of base‐pair opening are discussed.