The telomeric DNA sequences (TTTGGG)(n) and (TTAGGG)(n) where n = 12 and 24 form ordered structures which show large variations in stability with the nature of the alkali metal counterion, as do the CD spectra of the complexes. Concentration dependence of T(m) suggests that for n = 12 the complexes are intermolecular, whereas for n = 24 they are composed of G quartets with intramolecular loops. Hoogsteen base pairing at N7 of purines is known in the third strand of triple helices. We have now obtained """"""""unconstrained"""""""" Hoogsteen TC.AG double helices with parallel strands. The only Hoogsteen duplexes reported previously were those we had prepared by use of blocking groups at C2 of A to force use of the alternative binding site. The duplexes exhibit strong inverse dependence of T(m) on pH because of the need to protonate C for binding to N7 of G. The triple helix poly d(A).2poly d(T) was assigned the A conformation by Arnott and Selsing. We have now observed in the infrared spectrum in D(2)0 solution a band at 838 cm(-1) which is characteristic of B form DNA and no A form band at 810-1. We conclude that the helix has the B and not the A conformation and from modeling studies that the third strand has the reverse Hoogsteen structure and is antiparallel to the A strand. Similar conclusions were reached for the triple helix poly d(C-T) poly d(A-G).poly dC+-T). AG and AI mispairs introduced in dGGTACGCGTACC cause little change in the UV thermal transition in the central 6,7 position but a large depression when moved out one position to 5,8. Further movement to the 4,9 and 3,10 positions results in a return of T(m) to approximately that of the basic sequence. The anomalously low value for 5,8 AG or AI is due to formation of a four-base loop rather than a two-base loop for the 6,7 mispairs of the basic sequence. The cause of the return of T(m) values to """"""""baseline"""""""" is not yet clear, but it is possible that duplexes with single base loops rather than hairpinned monomers may be responsible.
