The DNA dodecamer D-GGTACGCGTACC and four derivatives containing GA and IA mispairs in the 6,7 and 5,8 positions were found to have van't Hoff enthalpies from the dependence of Tm on concentration that appeared to be much too high. Equilibrium centrifugation studies in collaboration with Dr. Ross revealed complex equilibria among duplex hairpin, and coil. Detailed analysis showed the dependence of the equilibria on concentration, ionic strength, and temperature and provided enthalpies and other thermodynamic parameters of the different transitions. The results suggest that many of the van't Hoff enthalpies reported in the literature for oligonucleotides may, be in error. UV melting of the oligomers in the above series showed a surprising dependence of Tm on sequence. The 6,7 mispairs had no effect on Tm, whereas the 5,8 exhibited large depressions. We attribute these observations to formation of two-base loop in the basic sequence and the 6,7 mispairs, leaving an identical five-base pair stem for the three oligomers. In the 5,8 mispairs, on the other hand, there is only a four-base pair stem, with resulting lower Tm. A two-base hairpin loop, postulated above, has been reported in the literature to be impossible in the ribo series and unstable or unfavorable in DNA. To evaluate the stereochemical feasibility of a two-base loop a theoretical modeling study was carried out in collaboration with Drs. Sagisekharan, Raghunathan, and Jernigan. The calculations showed that a stable two-base loop structure could be formed using standard values conformational torsion angles and having no short contacts. We have found that the drug distamycin binds as well to the homopolymer duplex d2NH(2)A*dT as to dA*dT and had previously found that both drugs bind well to the alternating copolymer of d2NH(2)A and dT. These results indicate that the accepted explanation of failure to bind GC (that the 2NH(2) group of G interferes sterically with 2 CH(2) residues of the drug) is not correct.
