Site-resolved energetics in DNA triple helices containing G*TA and T*CG triads

Recognition of specific sites in double-helical DNA by triplex-forming oligonucleotides has been limited until recently to sites containing homopurine-homopyrimidine sequences. G*TA and T*CG triads, in which TA and CG base pairs are specifically recognized by guanine or by thymine, have now extended this recognition code to DNA target sites of mixed base sequences. In the present work, we have obtained a characterization of the stabilities of G*TA and T*CG triads, and of the effects of these triads upon canonical triads, in triple-helical DNA. The three DNA triplexes investigated are formed by the folding of the 31-mers d(GAAXAGGT(5)CCTYTTCT(5)CTTZTCC) with X = G, T, or C, Y = C, A, or G, and Z = C, G, or T. We have measured the exchange rates of imino protons in each triad of the three triplexes using nuclear magnetic resonance spectroscopy. The exchange rates are used to map the local free energy of structural stabilization in each triplex. The results indicate that the stability of Watson-Crick base pairs in the G*TA and T*CG triads is comparable to that of Watson-Crick base pairs in canonical triads. The presence of G*TA and T*CG triads, however, destabilizes neighboring canonical triads, two or three positions removed from the G*TA/T*CG site. Moreover, the long-range destabilizing effects induced by the T*CG triad are larger than those induced by the G*TA triad. These findings reveal the molecular basis for the lower overall stability of G*TA- and T*CG-containing triplexes.