Energy Transfer in DNA DOI: 10.1002/ange.201004732 Triplet–Triplet Energy Transfer in DNA: A Process that Occurs on the Nanosecond Timescale** Carles Curutchet* and Alexander A. Voityuk* In the last decade, much progress has been made in the understanding of DNA excited-state dynamics. [1] In this context, theoretical studies focused both on the photophysical properties of individual nucleobases as well as on the relevant interactions in assemblies of two or more bases have been a valuable tool for exploring decay mechanisms of excited states in DNA. In contrast to singlet excited states, our knowledge on the energetics and dynamics of triplet excited states is still largely limited to the properties of individual bases. [2] Thus, despite the fact that triplet–triplet electronic energy transfer (TET) can initiate phototoxic reactions in DNA, [3, 4] such as the formation of thymine cyclobutane dimers, [5] little is known about the strength of the electronic interactions and the timescales for TET in nucleobase p stacks: the factors that determine the fate of triplet states in native DNA. Therefore, the assignment of decay compo- nents measured through ultrafast spectroscopy experiments remains a difficult task owing to the fundamental uncertainty regarding the degree of delocalization of triplet excited states and the approximate timescales for their migration. [1] Herein, we present a study of TET between stacked adenine–adenine (A–A) and tymine–thymine (T–T) in polyA–polyT DNA sequences. We applied the semiempirical ZINDO method to investigate how DNA structural dynamics modulate the couplings for TET along a 15 ns classical molecular-dynamics (MD) trajectory. The suitability of the ZINDO method for describing the energetics and TET couplings of low-lying p !p* triplet states was validated by comparison with equation-of-motion coupled-cluster models with single and double substitutions (EOM-CCSD) and calculations of configuration interaction with single excita- tions (CIS). The couplings were calculated by using the method fragment excitation difference (FED) recently devel- oped by Hsu and et al. [6] This method extends the fragment- charge-difference scheme [7] to couplings of excited states and enabled us to estimate the electronic couplings for nonsym- metrical arrangements of the bases, while accurately account- ing for the short-range interactions between stacked bases that promote TET. Finally, we applied Marcus theory to predict TET rates between the base pairs. [8] We found that in both A–A and T–T stacks, triplet excitons are localized on single bases and can migrate along the DNA on the nano- second timescale. We explored the ability of the semiempirical ZINDO method to accurately estimate electronic couplings by com- parison with correlated EOM-CCSD calculations with the 6- 31G basis set for symmetrical A–A and T–T dimers. As TET couplings depend on wavefunction overlap, we explored the effect of polarization and diffuse functions on the results at the CIS level by using the 6-31G, 6-31G(d), and 6-311 ++ G(d,p) basis sets. This effect was also considered at the EOM- CCSD level for several model systems. Our results indicate that ZINDO underestimates the couplings by approximately 20–40 % (see the Supporting Information for a detailed discussion). Thus, given that the TET rate is proportional to V 2 , the predicted efficiency of TET is expected to be about 2– 3 times too low. To check the performance of ZINDO, we also estimated electronic couplings for 500 configurations of the p stack at the CIS/6-31G(d) level (see below). In Figure 1, we show the distribution of squared electronic couplings obtained for 15 000 structures extracted from the MD simulation, whereas in Table 1, we report the MD averages as well as the results obtained for A- and B-DNA reference structures. [9] The TET rate can be estimated by Table 1: Delocalization length of the lowest triplet excited state (L 1 ), squared electronic coupling (V 2 ), reorganization energy (l), and TET time (t TET ) computed for the A 7 –A 8 and T 23 –T 24 stacked base pairs along the 15 ns molecular-dynamics trajectory. For comparison, values obtained for reference DNA structures [9] are also listed. L 1 V 2 [eV] l [eV] t TET [ns] A 7 –A 8 MD average 1.03 2.12  10 À5 0.607 0.80 A-DNA [9] 1.03 6.39  10 À6 – 2.66 B-DNA [9] 1.03 1.19  10 À5 – 1.43 T 23 –T 24 MD average 1.01 1.59  10 À6 0.557 6.35 A-DNA [9] 1.08 4.83  10 À5 – 0.21 B-DNA [9] 1.02 2.91  10 À6 – 3.46 [*] Dr. C. Curutchet Institut de Química Computacional and Departament de Química Universitat de Girona, Campus Montilivi, 17071 Girona (Spain) Fax: (+ 34) 97-218-3241 E-mail: carles.curutchet@udg.edu Prof. Dr. A. A. Voityuk Institució Catalana de Recerca i Estudis Avançats 08010 Barcelona (Spain) and Institut de Química Computacional, Universitat de Girona Campus Montilivi, 17071 Girona (Spain) Fax: (+ 34) 97-241-8356 E-mail: alexander.voityuk@icrea.es [**] C.C. acknowledges support from the Departament d’Innovació, Universitats i Empresa of the Generalitat de Catalunya (grant no. 2008BPB00108). A.A.V. is grateful to the Spanish Ministry of Education and Science for financial support (project CTQ2009- 12346). Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/anie.201004732. Zuschriften 1860  2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. 2011, 123, 1860 –1862