The fluorescence transition of 2-aminopurine in double- and single-stranded DNA

2-Aminopurine (2AP) is a widely used marker for nucleic acid structure because the fluorescence of 2AP in double-stranded DNA is approximately half that of 2AP in single-stranded DNA. The underlying photophysical mechanism responsible for the change in fluorescence with structure is not understood. We have performed CIS and TDB3LYP level calculations on double-stranded trinucleotide models, (X2APX)·(YTY), where X and Y represent the natural nucleobases. The results reveal that base stacking combined with hydrogen bonding reduces the oscillator strength for the fluorescence transition, and hence the fluorescence quantum yield. It is also shown that the electronic transitions of double-stranded trinucleotides are not simply those of the individual bases, but rather involve orbitals that are delocalized across several bases. Although the results obtained at the CIS level suffer from the neglect of electron correlation, the results obtained at the TDB3LYP level are not necessarily better because this method has problems describing long-range interactions and tends to underestimate the energies of charge-transfer states. Calculations on a full (rather than model) dinucleotide 5′-G2AP-3′ confirm that the inclusion of the sugar and phosphate groups is not necessary when studying the energetically low-lying excited states of nucleic acids containing 2AP.