Analysis of melting transitions of the DNA hairpins formed from the oligomer sequences d[GGATAC(X)4GTATCC] (X = A, T, G, C)

Abstract

Optical melting transitions of the short DNA hairpins formed from the self‐complementary DNA oligomers d[GGATACX₄GTATCC] where X = A, T, G, or C measured in 100 mM NaCl are presented. A significant dependence of the melting transitions on loop sequence is observed and transition temperatures, t_m, of the hairpins vary from 58.3°C for the T₄ loop hairpin to 55.3°C for the A₄ loop. A nearest‐neighbor sequence‐dependent theoretical algorithm for calculating melting curves of DNA hairpins is presented and employed to analyze the experimental melting transitions. Experimental melting curves were fit by adjustment of a single theoretical parameter, F_end(n), the weighting function for a hairpin loop comprised of n single‐strand bases. Empirically determined values of F_end(n) provide an evaluation of the free‐energy of hairpin loop formation and stability. Effects of heterogeneous nearest‐neighbor sequence interactions in the duplex stem on hairpin loop for mation were investigated by evaluating F_end(n) in individual fitting procedures using two of the published sets of nearest‐neighbor stacking interactions in DNA evaluated in 100 mM NaCl and given by Wartell and Benight, 1985. In all cases, evaluated values of F_end(n) were obtained that provided exact theoretical predictions of the experimental transitions. Results of the evaluations indicate: (1) Evaluated free‐energies of hairpin loop formation are only slightly dependent on loop sequences examined. At the transition temperature, T_m, the free‐energy of forming a loop of four bases is approximately equal for T₄, G₄, or C₄ loops and varies from 3.9 to 4.8 kcal/mole depending on the set of nearest‐neighbor interactions employed in the evaluations. This result suggests, in light of the observed differences in stability between the T₄, G₄, and C₄ loop hairpins, that sequence‐dependent interactions between base residues of the loop are most likely not the source of the enhanced stability of a T₄ loop. In contrast, the evaluated free‐energy of forming an A₄ loop is approximately 400 cal/mole higher for each nearest‐neighbor set indicating unfavorable interactions between A bases in a loop‐affect loop formation and overall hairpin stability, (2) The absolute value for the free‐energy of loop formation at the T_m of each hairpin varies by about 1 kcal/mole depending on the set of nearest‐neighbor interactions employed and the relative hierarchy of stability for each loop is conserved for different nearest neighbor sets, (3) The melting process of each hairpin deviates from strict two‐state behavior in the order according to loop sequence of T > A > G > C, (4) Results of our analysis are compared with the early work of Scheffler et al., 1970 on the hairpins formed from the copolymer sequences d(T – A)_q where q = 9–21. Comparisons with the more recent works a DNA dumbbell (Benight et al., 1988) and the very similar DNA hairpins studied by Senior et al., 1988 are presented.