Intrinsic Lifetimes of the Excited State of DNA and RNA Bases Hyuk Kang, Kang Taek Lee, Boyong Jung, Yeon Jae Ko, and Seong Keun Kim* School of Chemistry, Seoul National UniVersity, Seoul 151-747, Korea Received July 10, 2002 The four bases of DNA are the chromophores of DNA. When a DNA is irradiated by ultraviolet light, the bases are electronically excited and should, in principle, become prone to photochemical reactions that lead to mutagenesis and carcinogenesis. 1-4 Such lethal consequences are largely avoided, however, because the excited bases are believed to have a lifetime so short that they relax to their ground state before a photochemical reaction takes place. It has been suggested that the ultrashort lifetime of the photoexcited DNA bases is what led Nature to adopt the four nucleobases as the durable carriers of genetic information. 5 Consider, for example, the isomers, 2-aminopurine and adenine (6-aminopurine). It appears that thermodynamics of base pairing played little role in the selection of the latter as a nucleobase because, for instance, these two bases have comparable association constants for pairing with uracil. 6 It is the lifetime of the excited state of these bases that is widely different between them: tens of ns for 2-aminopurine versus 1 ps for adenine. 5 The excited-state lifetimes of DNA bases have been measured to be generally very short (of the order of 1 ps or even less) in aqueous solution, 7-14 but it has never been tested whether these short lifetimes are an intrinsic property of these molecules or a fortuitous result of their interaction with solvents. Since the solution has often proved to be a vastly different medium from the structurally confined biological environment devoid of strong dielectric effect, a short lifetime in a labile solution may not mean the same in an in vivo organism unless it is intrinsic to the molecule. Here we show explicitly for the first time, by measuring the lifetime for an isolated molecule, that the ultrashort lifetime of all nucleobases is their genuine molecular property, resulting from an extremely facile internal conversion. The lifetime should therefore be largely independent of the medium, that is whether in vacuo, in solution, or in vivo. Details of our experimental setup for femtosecond pump-probe ionization have been published elsewhere. 15 In short, we used a third harmonic pulse of a Ti:sapphire laser to pump the molecule to its electronically excited state, and then after a given time delay, we probed the population of the excited state by multiphoton ionization using the fundamental light. An effusive beam condi- tion was employed throughout this experiment because dissociative ionization of clusters generated by supersonic expansion often yields misleading conclusions about the genuine property of a molecule. 16 The powder sample of DNA bases was heated in a metal or Teflon oven to 170 °C (250 °C for guanine) to attain a vapor pressure of a few mTorr and effused into the interaction region. The sample was purchased from Aldrich Chemical Co. and used without further purification. Figure 1 shows the pump-probe transient of adenine at the pump wavelength of 267 nm. The experiment was performed at a high level of probe power where one of the transitions was saturated to achieve a sufficiently large signal-to-noise ratio. Although three photons of probe pulse were needed to ionize the excited state of adenine, the dependence of the transient ion signal on the probe power was quadratic, indicating the saturation. Since the sum of the energy of one pump photon and one probe photon matches the energy of a highly excited state with an absorption band near 200 nm, 17 this resonant state may be saturated first in the ionization step. The dependence of the transient ion signal on the pump power remained linear throughout the experiment. The transient has a large spike around the time zero, followed by a decay that drops completely to the background level, in contrast to the decay curve of an earlier picosecond study. 18 It could be well fitted into the sum of a Gaussian function (dashed line in Figure 1) and a single exponential decay (dotted line) convoluted by the cross-correlation of the lasers. The time constant of the exponential decay was 1.0 ps. The ratio of the Gaussian component to the exponential component was increased as the probe power was increased, but the decay time of the exponential component remained constant. The most likely explanation for the Gaussian component is coherent absorption of the pump and probe photons, especially when their intensities are high. For example, when the pump and the probe pulses coincide at zero delay time, excitation to the aforementioned resonant state at 200 nm is possible by the (1 + 1′) excitation with one pump and one probe photons, in addition to the excitation to the first excited state by a single pump photon. Both kinds of excitation contribute to the ion signal. When the pump and probe pulses do not overlap in time, however, the former type of excitation can no longer occur. The ion signal in this case is due only to those molecules that have been excited to the first excited state by the pump pulse alone. Therefore, the Gaussian component is strongly centered around time zero. The * Corresponding author. E-mail: seongkim@plaza.snu.ac.kr. Figure 1. Pump-probe transient ionization signal of adenine in the gas phase at the pump wavelength centered at 267 nm. Hollow circles: experimental data; solid line: theoretical fit to the data; dashed line: Gaussian component of the fit; dotted line: exponential component of the fit. The Gaussian component results from coherent absorption of the pump and probe pulses. The averaged exponential decay time is 1.0 ps, which is the intrinsic lifetime of the excited state of adenine. Published on Web 10/15/2002 12958 9 J. AM. CHEM. SOC. 2002, 124, 12958-12959 10.1021/ja027627x CCC: $22.00 © 2002 American Chemical Society