Revealing Anharmonic Couplings and Energy Relaxation in DNA

11 Jun 2008 - Institut für Experimentalphysik, Freie Universität Berlin, Arnimallee 14, D-14195 Berlin, Germany, Max-Born Institut für Nichtlineare...
0 downloads 0 Views 1MB Size
J. Phys. Chem. B 2008, 112, 7909–7915

7909

Revealing Anharmonic Couplings and Energy Relaxation in DNA Oligomers by Ultrafast Infrared Spectroscopy K. Heyne,*,†,‡ G. M. Krishnan,§ and O. Ku¨hn*,§,| Institut fu¨r Experimentalphysik, Freie UniVersita¨t Berlin, Arnimallee 14, D-14195 Berlin, Germany, Max-Born Institut fu¨r Nichtlineare Optik and Kurzzeitspektroskopie, Max-Born Strasse 2A, D-12489 Berlin, Germany, and Institut fu¨r Chemie and Biochemie, Freie UniVersita¨t Berlin, Takustrasse 3, D-14195 Berlin, Germany ReceiVed: NoVember 28, 2007; ReVised Manuscript ReceiVed: March 31, 2008

The identification and characterization of NH2 hydrogen-bonded stretching vibrations [ν(NH2)] in DNA oligomers is usually hampered by the all-dominating absorption of the water stretching band in the spectral range of 3050-3600 cm-1. Here, we use the two-color IR pump-probe technique to overcome the limitations of linear absorption spectroscopy by exciting adenine-thymine (A-T) oligomer vibrations in the fingerprint region and analyzing induced transient spectral changes in the ν(NH2) spectral region. These transient changes are related to anharmonic couplings to the modes excited in the fingerprint region and to modes populated by intra- and intermolecular energy redistribution and relaxation. The combination of calculated anharmonic coupling parameters and experimental transient IR data allows the assignment of a transition at 3215 cm-1 to the ν(NH2) vibration of adenine in dA20-dT20 DNA oligomers. Introduction DNA is undoubtedly among the most prominent hydrogenbonded systems. Despite numerous experimental and theoretical investigations of vibrational spectra of nucleic acid bases,1–7 information on inter- and intramolecular interactions in base pairs and DNA oligomers is still relatively scarce.8–17 A recent example is the work on adenine-uracil (A-U) base pairs in the Watson-Crick geometry in solution, which showed strong enhancement of vibrational energy relaxation of the NH stretching vibration compared to the isolated uracil base.8 DNA oligomers adopt different types of conformations, in both the gaseous and condensed phases, such as the A, B, B′, C, D, and Z forms, depending on water and salt concentration, type of cations, pH, and base sequences.2,4,18–22 These forms might play specific roles in different biological processes such as building DNA-drug complexes.23 In the condensed phase, the conformations of DNA oligomers are stabilized by water molecules that form water networks, predominantly in the major and minor grooves and near the phosphate groups of the backbone.2 Among the different types of DNA base sequences, adeninethymine (A-T) oligomers are special because they do not undergo transitions from the B form to the A form upon reduction of the water content. Instead, A-T oligomers adopt the B′ form at low water concentrations, with four to six water molecules per base pair that can be hydrogen-bonded to the oligomer.2,24–26 In the B′ form of the A-T DNA oligomer, two hydrogen bonds are formed in the Watson-Crick configuration, as shown in the following diagram (the circle denotes the link to the phosphate backbone): * To whom correspondence should be addressed. E-mail: [email protected] (K.H.), [email protected] (O.K.). † Institut fu ¨ r Experimentalphysik, Freie Universita¨t Berlin. ‡ Max-Born Institut fu ¨ r Nichtlineare Optik and Kurzzeitspektroskopie. § Institut fu ¨ r Chemie and Biochemie, Freie Universita¨t Berlin. | New address: Institut fu ¨ r Physik, Universita¨t Rostock, Universita¨tsplatz 3, D-18051 Rostock, Germany.

The two hydrogen bonds between the base pairs are between the oxygen atom (O4) of the thymine and the NH2 group of the adenine (N6) and between the NH group of the thymine (N3) and the nitrogen atom (N1) of the adenine. Through the remainder of this article, vibrational stretching modes are indicated by the symbol ν, and vibrational bending modes by δ. The vibrational modes that are expected to be strongly influenced by the hydrogen bonding in the DNA helix are ν(C2dO2) at 1716 cm-1 and ν(C4dO4) and δ(NH2) both at 1665 cm-1.1,3,4,10,27–31 Note that, in contrast to H2O, for D2O, the δ(ND2) vibration of adenine and the carbonyl vibrations of thymine are decoupled, because of the frequency shift from δ(NH2) to δ(ND2).17,32 The δ(OH) vibration of water molecules in DNA samples typically absorbs in the same spectral region, i.e., around 1650 cm-1.3,27,33 A direct experimental assignment of ν(NH2) and ν(NH) in A-T DNA oligomers in the condensed phase is very difficult. Typically, symmetric and antisymmetric NH2 stretching vibrations absorb around 3300 cm-1.3 However, the spectral range from 3050 to 3600 cm-1 is dominated by the strong absorption of the water OH stretching vibration. Reducing the water content of the DNA oligomers does not solve this problem, because the DNA oligomers do not adopt a welldefined structure at extremely low water contents.

10.1021/jp711262y CCC: $40.75  2008 American Chemical Society Published on Web 06/11/2008

7910 J. Phys. Chem. B, Vol. 112, No. 26, 2008 In this contribution, we use ultrafast time-resolved infrared spectroscopy to address this issue. Ultrafast infrared spectroscopy is a powerful tool for investigating hydrogen-bond interactions and structure changes. For example, inter- and intramolecular couplings and energy relaxation dynamics have been studied extensively in various hydrogen-bonded systems.34–37 In this report, we investigate shifts in oligomer vibrational modes induced by excitation of the ν(C2dO2) or ν(C4dO4)/δ(NH2) oligomer fingerprint vibration. These shifts originate from interand intramolecular couplings among different vibrational modes of the DNA oligomer and depend on the strength of the couplings as well as the energy mismatch between different transitions. They are particularly pronounced if overtones or combinatorial modes match a fundamental vibrational transition (resonance enhancement). This affects the linear absorption band shape and the vibrational relaxation dynamics.38 A particular strength of ultrafast infrared pump-probe spectroscopy is the capability of uncovering vibrational spectral features that are not visible by linear spectroscopy because of excessive solvent absorption. This is demonstrated in the experiments presented here, where we excite oligomer vibrations between 1600 and 1760 cm-1 and probe for the oligomer ν(NH2) vibration in the region of 3050-3250 cm-1, which is dominated by water absorption. The experimental assignment of the adenine ν(NH2) vibration and the coupling pattern across the hydrogen bonds are supported by quantum chemical calculations of anharmonic couplings, which were used to obtain fundamental transition frequencies for six relevant modes of a microsolvated gas-phase A-T model. In fact, isolated and microsolvated base pairs have been studied theoretically, with a particular focus on the stability of different isomers; see, for example, the work by Hobza and co-workers,39 as well as that by Fonseca Guerra et al.40–42 Although a number of reports have been published on potential energy surfaces of base pairs in the harmonic approximation, there appear to be only a few calculations addressing anharmonicity in the context of issues such as proton transfer,5,41,42 coupling to the intermolecular hydrogen-bond vibration,13 or the assignment of different gas-phase isomers.43 Most notable in this respect are the recent studies of the anharmonic spectrum of a guanine-cytosine pair 9,44 and the development of a vibrational exciton model to describe nonlinear IR spectra involving DNA fingerprint modes.9,16,17,45 In principle, accurate theoretical modeling of the vibrational dynamics of DNA A-T base pairs requires that several effects be taken into account, including (i) the intermolecular double hydrogen bond between adenine and thymine; (ii) the interactions between different base pairs along the DNA strand; (iii) the charges and dynamics of the backbone; and (iv) the influence of water molecules, which might, for instance, make a hydrogen bond to the base pair. In this work, we are aiming at a semiquantitative model of the transient band shifts, whereby it is assumed that they are dominated by effect i, that is, the anharmonic coupling pattern due to the intermolecular hydrogen bond. In the calculations, we consider only a single isolated base pair. The effects of factors ii-iv can be of static nature (e.g., changes in the anharmonic frequencies and coupling constants) and also of dynamic nature (e.g., fluctuation of the energy levels). However, here, we focus only on the static influence of a well-defined environment determined by microsolvation of the A-T base pair by two water molecules. Experimental Section A-T DNA double-stranded oligomers with sodium counterions and a length of 20 base pairs were obtained from Biotherm

Heyne et al.

Figure 1. Normal-mode displacement vectors for the four modes included in the A-T model: (a) ν(NH2), (b) ν(C2dO2), (c) ν(C4dO4), and (d) δ(NH2).

and were dissolved in water and dried on a CaF2 window at 293 K in an atmosphere of 52% relative humidity (saturated solution of NaHSO4 · H2O at 20 °C46). This results in DNA samples with approximately four to six water molecules per base pair30 (sample thickness ≈ 6.5 µm). It has been reported that, under these conditions, A-T DNA oligomers adopt the B′ form.27 Femtosecond time-resolved infrared pump-probe experiments were performed with two independently tunable femtosecond pulses generated by parametric conversion processes pumped by a regenerative Ti:sapphire laser system (800 nm; repetition rate, 1 kHz; pulse duration, 100 fs).47 The central frequency of the pump pulse was varied from 1630 to 1760 cm-1, and the probe was centered around 1650 or 3200 cm-1. The cross-correlation between pump and probe-pulses typically had a temporal width of ∼130 fs (fwhm). With the pump pulse energy of 1 µJ, approximately 2% of the A-T base pairs in the sample volume were excited. After interaction with the sample, the probe-pulses were spectrally dispersed and detected with a HgCdTe detector array (resolution 5 cm-1). Theoretical Model In this section, we introduce a model that is tailored to the description of the anharmonic coupling between the CO stretching and NH2 bending vibrations around 1700 cm-1 and the NH2 stretching vibration around 3300 cm-1. The reported quantum chemistry calculations were performed on a single A-T DNA base pair in the Watson-Crick geometry (with exclusion of sugar moieties, phosphate groups, and associated counterions), using density functional theory with the B3LYP functional and the 6-31++G(d,p) basis set as implemented in Gaussian 03.48 Starting from the planar A-T Watson-Crick structure, the anharmonic potential energy surfaces were expanded into normal-mode coordinates up to fourth order. Third- and fourth-order anharmonic coupling constants were calculated using a combination of analytical second derivatives and finite differences.49 Terms that were off-resonant by more than ∼1000 cm-1 were neglected. Specifically, we included the four modes whose displacement vectors are shown in Figure 1, that is, the ν(NH2), ν(C4dO4), and ν(C2dO2) stretching vibrations and the δ(NH2) bending vibration. The other hydrogenbonded NH stretching modes were calculated to absorb at clearly different energies, with the asymmetric NH2 stretching mode above 3600 cm-1 and the N3-H thymine stretching mode below 3000 cm-1 for the gas-phase Watson-Crick pair.50

Study of DNA Oligomers by Ultrafast IR Spectroscopy

J. Phys. Chem. B, Vol. 112, No. 26, 2008 7911 TABLE 1: Fundamental Transition Frequencies (in cm-1) of the Modes Shown in Figures and 2a mode ν(NH2) ν(C2dO2) ν(C4dO4) δ(NH2)

A-T A-T A-T(H2O)2 A-T(H2O)2 harmonic 4D harmonic 6D 3410 1797 1728 1689

3330 1758 1719 1645

3401 1799 1713 1720

3280 1792 1702 1708

expt 3215b 17163,4,10 16653,4,10 16653,4,10

The anharmonic frequencies of the water modes are 3752 cm-1 [ν(H2O)] and 1588 cm-1 [δ(H2O)]. b This work. a

TABLE 2: Selected Third-Order Force Constants (in cm-1) for the Normal-Mode Displacements Shown in Figures and 2a ν(NH2) ν(NH2) ν(NH2) ν(NH2)

δ(NH2) ν(C4dO4) δ(NH2) δ(NH2)

δ(NH2) ν(C4dO4) ν(C4dO4) ν(C2dO2)

A-T

A-T(H2O)2

87 2 -17 8

96 11 -66 5

a Other couplings to DNA modes in the fingerprint region are negligible.

Figure 2. Optimized A-T(H2O)2 structure containing five hydrogen bonds. The presence of the water molecules lengthens the upper O4sN5 hydrogen bond (from 2.939 to 3.005 Å) while shortening the lower N3sN1 hydrogen bond (from 2.876 to 2.868 Å). Normal-mode displacement vectors are shown for the six modes included in the A-T(H2O)2 model (cf. Figure 1): (a) ν(H2O), (b) δ(H2O) (c) ν(NH2), (d) ν(C2dO2), (e) δ(NH2), and (f) ν(C4dO4).

In a second set of calculations [A-T(H2O)2 model], we included water molecules to model their effect on the anharmonic couplings. As discussed before, under the conditions of the experiment, there are four to six water molecules per A-T pair that can form different hydrogen bonds to the base pair. Because we are mainly interested in the influence on the NH2 and C4dO4 groups, we chose a minimum model of two water molecules as introduced previously in Fonseca Guerra et al.40 In the DNA, this corresponds to water situated in the major groove. Gas-phase geometry optimization yielded the structure shown in Figure 2. The modified target modes are shown in Figure 2c-f; they are supplemented by stretching and bending modes of the H2O molecule that is bonded to the adenine (Figure 2a,b). These modes were found to have the largest coupling to the ν(NH2) vibration. Overall, this gives a six-dimensional (6D) A-T(H2O)2 model. The vibrational transition frequencies were calculated from the Fourier-transform of the dipole-dipole autocorrelation function,51 assuming a linear dipole moment and negligible vibration-rotation coupling, for both the four- and sixdimensional models. The time propagation was performed for 5 ps using the multiconfiguration time-dependent Hartree (MCTDH) method52 as implemented in the Heidelberg program package.53 For all primitive bases, we used the Hermite discrete variable representation with 32 points in the interval [-0.8, 0.8] aB amu1/2. Furthermore, three single-particle functions per degree of freedom were sufficient to keep the natural orbital populations well below 1%. Results and Discussion Theoretical Results. In Table 1, we have compiled the fundamental transition frequencies obtained from both the harmonic and anharmonic analyses. Notice that, to highlight

the effects of anharmonicity, we did not attempt to scale the harmonic frequencies. The theoretical values for the isolated A-T and A-T(H2O)2 models are compared with experimental data on DNA A-T oligomers. Naturally, a single DNA base pair is a crude model for a double-stranded DNA oligomer. Therefore, our primary goal is to provide a semiquantitative guide for the interpretation of the experimental data. Comparing the two models, we notice that adding water improves the agreement with experiment for the selected target modes. Specifically, the ν(NH2) stretching transition is redshifted, and the frequencies of the ν(C4dO4) and δ(NH2) vibrations are brought almost into coincidence with those observed in the experimental spectrum. In both models, inclusion of anharmonicity causes a red shift of all fundamental transition frequencies as compared to the harmonic values. For instance, the diagonal anharmonicity of the ν(NH2) stretching coordinate in the hydrogen bond and the coupling to the other target modes reduce the transition frequency by ∼2.5% and ∼3.5%, for the A-T and A-T(H2O)2 models, respectively. The largest deviation from experiment was found for the ν(C2dO2) vibration, which, for the A-T(H2O)2 model, was overestimated by ∼4.5%. One possible reason within the present gas-phase model could be the neglect of the thymine N3-H bending vibration, which might be coupled to the ν(C2dO2) mode as suggested by the N3-H bending character in Figure 2d. Furthermore, inclusion of additional water molecules might lead to a hydrogen bond to the thymine O2, resulting in a downshift of this transition. The obvious conclusion that can be drawn is that the microenvironment affects the frequencies of fundamental transitions. More important is the extent to which the pattern of A-T anharmonic couplings is influenced. In Table 2, we report important third-order force constants that are responsible for the coupling between the fingerprint region of the CdO and δ(NH2) vibrations and the ν(NH2) stretching mode. We first focus on the ν(NH2) mode in the isolated A-T base pair. Here, the strongest coupling is a Fermi resonance interaction between the ν(NH2) stretching fundamental transition and the δ(NH2) bending overtone. A smaller coupling was found for the mixing of this fundamental transition with a combination of δ(NH2) and ν(C4dO4). In comparison, both the coupling via the ν(C4dO4) overtone and that through the ν(C2dO2) + δ(NH2) combination are considerably weaker than the coupling to the

7912 J. Phys. Chem. B, Vol. 112, No. 26, 2008

Heyne et al.

Figure 5. (a) Absorbance changes around 3200 cm-1 for various pump-probe delay times after excitation at 1630 cm-1 (fwhm 160 cm-1); data at 0-ps delay time were obtained by averaging from -200 to 200 fs to eliminate nonabsorbing signal contributions. (b) Same data after subtraction of the monoexponential 13-ps rise time of the hot water response.

Figure 3. (a) Scheme describing the pump-probe detection of the ν(NH2) fundamental transition in A-T base pairs. Because of the anharmonic coupling between the ν(C4dO4), δ(NH2), and ν(NH2) vibrations, the ν(NH2) ) 0 f 1 transition is bleached upon excitation of the ν(C4dO4) and δ(NH2) modes. Population of the ν(C4dO4) ) 1 and δ(NH2) ) 1 levels will locally heat the molecule, inducing a shift of the hot ground-state ν(NH2) transition. (b) Absorption spectra of A-T DNA oligomers prepared in 52% relative humidity (black solid line), neat water (gray solid line), and A-T DNA oligomers dried for two days in a N2 atmosphere (dashed line).

Figure 4. Absorption spectrum of A-T DNA oligomers around 3300 cm-1 (solid line) and absorbance difference spectra for various pump-probe delay times after excitation at 1740 cm-1 (fwhm 170 cm-1). Picosecond OH stretching response of water from 3600 to 3050 cm-1. Data at 0-ps delay time were obtained by averaging from -200 to 200 fs to eliminate nonabsorbing signal contributions.

δ(NH2) overtone. After inclusion of two water molecules, one finds that the Fermi resonance with the δ(NH2) overtone still dominates. In addition, there is a substantial increase of the coupling with the δ(NH2) + ν(C4dO4) combination transition. This is a consequence of the increased δ(NH2) character acquired by the ν(C4dO4) mode in the A-T(H2O)2 case (compare Figures 1c and 2f). In both cases, the ν(C2dO2) overtone vibration couples only marginally with the ν(NH2) fundamental (∼1 cm-1). To summarize, these calculations indicate that excitation of the δ(NH2) and ν(C4dO4) vibrations should result in significant frequency shifts of the ν(NH2) absorption band (see Figure 3a), whereas the effect of excitation of the ν(C2dO2) transition should be less pronounced.

Experimental Results. The absorption between 3050 and 3600 cm-1 (see Figure 3b, solid line) is more than 85% dominated by the broad O-H stretching absorption of water molecules. Therefore, it is not possible to determine the NH2 stretch absorption frequency of the DNA oligomer directly from this absorption spectrum. In the fingerprint region, the absorption of ν(C2dO2) is located at 1716 cm-1, and the combined absorption of δ(NH2) and ν(C4dO4) peaks at 1665 cm-1 (see Figure 3b, solid line).3,4,10,25,31,54 It has been reported that both the ν(C4dO4) and δ(NH2) vibrations absorb at 1665 cm-1, and therefore, they can not be excited separately in our experiments. Figure 4 shows results of femtosecond pump-probe experiments with excitation in the fingerprint region and probing between 3050 and 3600 cm-1. Excitation with a broad pump pulse at 1740 cm-1 (fwhm 170 cm-1) leads to an instantaneous spectrally narrow response around 3200 cm-1. Furthermore, a spectrally broad response over the entire range from 3050 to 3600 cm-1 is seen to increase on the picosecond time scale (see Figure 4). At 1740 cm-1, the pump pulse mainly excites the ν(C2dO2) stretching vibration. Given the photon energy and pulse intensity, excitation of vibrations around 3300 cm-1 by two-photon processes is negligible. As a consequence, both the instantaneous signal and the increasing broad negative signal must result from the excitation of vibrations in the fingerprint region. The broad negative signal, which becomes positive above 3530 cm-1, is known to correspond to the OH stretching vibration of hot bulk water. Excess energy in low-frequency vibrations of water (e.g., librations) weaken the hydrogen-bond strength, resulting in an increase of the OH stretching force constant and, therefore, a higher OH stretching frequency.37,55,56 The instantaneous narrow response around 3200 cm-1, after excitation at 1740 cm-1, decreases in time and should therefore correspond to a different process. Figure 5b shows absorbance changes in the range from 3050 to 3250 cm-1 upon excitation at 1630 cm-1 (fwhm 160 cm-1), after subtraction of the spectrally broad 13-ps component of the hot water formation, obtained from a global fit. At 1630 cm-1, the pump pulse mainly excites the δ(NH2) and ν(C4dO4) vibrations. An instantaneous bleach signal is observed with a maximum at 3215 cm-1 and a width of 50 cm-1, which decays on a subpicosecond time scale. The perturbed free induction decay of this band gives a total dephasing time, T2, of 0.5 ( 0.1 ps, which corresponds to a homogeneous line width of 21 ( 5 cm-1. This indicates that the observed 50 cm-1 width of the bleaching band has a different origin. The maximum of the

Study of DNA Oligomers by Ultrafast IR Spectroscopy

J. Phys. Chem. B, Vol. 112, No. 26, 2008 7913

TABLE 3: Time Constants of Transientsa

a

vibrational mode

pump (cm-1) (fwhm)

probe (cm-1)

time constants (ps)

ν(NH2) ν(OH)b ν(NH2) ν(C2dO2) ν(C2dO2) ν(C2dO2) ν(C4dO4)/δ(NH2) δ(OH) δ(OH) ν(C4dO4)/δ(NH2)

1630 (160) 1630 (160) 1730 (90) 1760 (100) 1760 (100) 1630 (130) 1630 (130) 1630 (130) 1630 (130) 1630 (130)

3215 3130 3215 1725 1685 1720 1665 1650 1640 1625

(d) 0.6 ( 0.2//(d) 3.0 ( 1.5//(r) 13 ( 2 (r) 0.4 ( 0.2//(d) 1.4 ( 0.4//(r) 13 ( 2 (d) 0.9 ( 0.4//(r) 4.0 ( 1.5//(r) 13.0 (d) 0.9 ( 0.1 (d) 0.7 ( 0.1 (d) 2.4 ( 0.2 (d) 0.4 ( 0.1//(d) 1.4 ( 0.4 (d) 0.2 ( 0.1//(d) 1.0 ( 0.2 (r) 0.6 ( 0.1 (d) 0.5 ( 0.1

r, rising signals; d, decaying signals. b Suggested assignment (see text).

Figure 6. Transients and fits to the dynamics probed at 3215 cm-1 after excitation at 1630 cm-1 (fwhm 160 cm-1; solid circles) and at 1730 cm-1 (fwhm 90 cm-1; open circles). The latter transient is multiplied by a factor of 7. Time constants are 0.6 ( 0.2, 3.0 ( 1.5, and 13 ( 2 ps after excitation at 1630 cm-1 (open circles) and 0.8 ( 0.4, 3.0 ( 1.5, and 13 ps after excitation at 1730 cm-1 (solid circles). Inset: Transients for longer pump-probe delay times.

instantaneous response at 3215 cm-1 decays with a 0.6 ( 0.2 ps time constant. The transients of the initial dynamics, together with the fits at 3215 cm-1, are presented in Figure 6. The time constants characterizing the kinetics for various pump-probe wavelength combinations are summarized in Table 3. To identify the origin of the instantaneous bleaching signal at 3215 cm-1, we compared transients at 3215 cm-1 after excitation at 1630 and 1730 cm-1. Excitation at 1730 cm-1 results in a signal that is ∼7 times weaker, and therefore, in Figure 6, these transients are scaled to each other at times >5 ps (see also the inset in Figure 6). Comparing the two transients, we observe that excitation at 1630 cm-1 leads to a threeexponential decay of the bleach with time constants of 0.6, 3, and 13 ps. A transient with similar time constants (0.9, 4, and 13 ps) but very different amplitudes is observed after pumping at 1730 cm-1, where mainly the C2dO2 stretching vibration is excited. Results of experiments in which both pumping and probing were performed in the fingerprint region are presented in Figure 7a,b. In Figure 7a, the A-T DNA oligomer sample was excited at 1760 cm-1 and probed between 1600 and 1760 cm-1. The pump-probe spectrum shows a negative band at 1725 cm-1 and a positive band at 1685 cm-1. The band at 1725 cm-1 decays monoexponentially with a time constant of 0.9 ( 0.1 ps, whereas the band at 1685 cm-1 decays with a time constant of 0.7 ( 0.1 ps. The positive signal can be assigned to the ν(C2dO2) 1 f 2 transition. A similar lifetime was obtained by Zanni et al. for measurements on G-C DNA oligomers with excitation in the same frequency range.9 The difference between the 0.7-ps excited-state lifetime and the 0.9-ps ground-state recovery time signals is possibly due to the initial conversion of the ν(C2dO2) energy into excitation of lower-frequency modes. As a consequence, the ground-state absorption frequency is shifted because of anharmonic coupling

to these lower-frequency modes and does not recover before these modes lose their excitation energy. Results for excitation of the A-T DNA oligomer at 1630 cm-1, presented in Figure 7b, show bleaching signals at the ground-state absorption positions of the ν(C2dO2) vibration (1716 cm-1), the ν(C4dO4) and δ(NH2) vibrations (both 1665 cm-1), and the water δ(OH) vibration (1650 cm-1).3,25,33,56 The ν(C2dO2) bleach recovers exponentially with a time constant of 2.4 ( 0.2 ps, whereas biexponential recoveries were observed for ν(C4dO4)/δ(NH2) with time constants of 0.4 ( 0.1 and 1.4 ( 0.4 ps and for the water δ(OH) with time constants of 0.2 ( 0.1 and 1.0 ( 0.2 ps. For the bending vibration δ(OH) of water molecules in bulk water, a lifetime of 170 ( 30 fs56 has been reported, which agrees with the fast component observed here for δ(OH). Increased absorption signals below 1640 cm-1 decay with a time constant of 0.5 ( 0.1 ps match the fast component of the ν(C4dO4)/δ(NH2) bleach recovery. In addition, a positive band appears around 1640 cm-1 that rises in 0.6 ( 0.1 ps. Signals at this spectral position have been assigned to the bending vibration of hot water molecules.56 The dynamics of experiments with different pump frequencies in the fingerprint region were compared to confirm that the bleach band at 3215 cm-1 originates from the NH2 stretching vibration of adenine and not from water. First, we point out that even the fastest recovery time of the bleach signal at 3215 cm-1 after excitation in the fingerprint region [where the water δ(OH) also absorbs] is 3 times slower than the reported lifetime of the water δ(OH) vibration.56 Second, the 50 cm-1 width of the bleaching band at 3215 cm-1 is considerably less than the 200 cm-1 estimate for the δ(OH) overtone band of bulk water.57 Having assigned the 3215 cm-1 bleaching signal to the A-T DNA oligomer ν(NH2) vibration, we now compare the dynamics at this frequency for excitation at 1630 and 1730 cm-1, which correspond to the absorption bands of both the ν(C4dO4) and δ(NH2) vibrations and the ν(C2dO2) vibration, respectively. For both excitation frequencies, the kinetics, shown in Figure 6, can be modeled by the same 13-ps time constant, corresponding to the rise of hot water signal due to energy transfer from the DNA oligomer to the water molecules. However, the initial picosecond dynamics are markedly different for these two excitation frequencies, further confirming that the two signals can not be due to the overtone excitation of the water δ(OH) vibration. Although the absorptions at the two frequencies are comparable, the signal strength after excitation at 1630 cm-1 is, in fact, 7 times stronger than that after excitation at 1730 cm-1, which indicates that the coupling to the ν(NH2) vibration is substantially weaker for the ν(C2dO2) vibration than for both the ν(C4dO4) and δ(NH2) vibrations. Comparison between Experiment and Theory. A model that accounts for both the theoretical and experimental results

7914 J. Phys. Chem. B, Vol. 112, No. 26, 2008

Heyne et al. ∼10 cm-1 for guanine-cytosine oligomers. This will lead to a delocalization of the carbonyl vibrations along the helix, accompanied by a downshift of the transition frequencies.17,32,45 Summary

Figure 7. Absorbance changes as a function of frequency for different pump-probe delay times. Data at 0-ps delay time were obtained by averaging from -200 to 200 fs to eliminate nonabsorbing signal contributions. (a) Excitation at 1760 cm-1 (fwhm 100 cm-1; dash-dotted line) and probing in the fingerprint region; mainly signals in the ν(C2dO2) region. (b) Excitation at 1630 cm-1 (fwhm 130 cm-1; dashdotted line); signals in the whole range. The solid line shows the relative absorption of the DNA sample (inverted).

is summarized in Figure 3a. The calculated couplings of the ν(C4dO4), δ(NH2), and ν(C2dO2) vibrations to the ν(NH2) vibration indicate that excitation of any of these modes should result in a bleaching signal due to the shifting of the ν(NH2) ) 0 f 1 transition. The bleach signals in Figure 5a,b around 3215 cm-1 agree with this theoretical result. Furthermore, the force constants in Table 2 predict that excitation of the ν(C4dO4) and δ(NH2) vibrations should result in a larger shift of the ν(NH2) vibration than of the ν(C2dO2) vibration. This is confirmed by the experimental data in Figure 6. Experimentally, one cannot distinguish between the contributions of the ν(C4dO4) and δ(NH2) modes. From the force constants in Table 2, however, we conclude that excitation of the δ(NH2) vibration is expected to have the most substantial effect on the shift of the ν(NH2) vibration. The agreement between theoretical predictions and IR pump-probe measurements allows us to assign the bleaching signal at 3215 cm-1 to the ν(NH2) vibration of adenine. This absorption band lies about 100 cm-1 lower in energy than the same mode in modified adenine-uracil Watson-Crick base pairs in solution.8 The lower frequency of the hydrogen-bonded ν(NH2) vibration in DNA films compared to single A-U base pairs in CDCl3 solution can be rationalized by significant interactions with neighboring base pairs and water molecules that decrease the force constant. The theoretical results further show that inclusion of water molecules leads to a hydrogen bond between a water molecule and the NH2 group of adenine, and therefore a coupling of the adenine ν(NH2) vibration and of the water bending vibration. As a general trend the interaction with water molecules (H2O) results in a downshift of the ν(C4dO4) vibration and an upshift of the δ(NH2) vibration. This is in accord with previous calculations of adenine-thymine Watson-Crick base pairs in D2O.17 These calculations also show a downshift of the ν(C2dO2) vibration upon interaction with heavy water. Coupling between stacked bases such as excitonic interactions of neighboring carbonyl vibrations can further influence the fundamental transitions. Because the carbonyl distances are smaller in the inner part of the helix, this effect should be more pronounced for the ν(C4dO4) vibrations. In fact the magnitude of interaction energy has been determined to be about

The symmetric NH2 stretching vibration ν(NH2) of adenine, hydrogen-bonded to thymine, is nearly impossible to identify for hydrated DNA, because the OH stretching absorption of water dominates the absorption between 3050 and 3600 cm-1. By combining femtosecond nonlinear IR pump-probe experiments on A-T DNA oligomers (prepared in 52% relative humidity) with calculations of vibrational spectra and anharmonic couplings on A-T and A-T(H2O)2 model systems, we conclude that the adenine ν(NH2) in A-T DNA oligomers absorbs at 3215 cm-1 and has pronounced anharmonic couplings to the ν(C4dO4) and δ(NH2) modes. This implies that the ν(NH2) vibration in A-T DNA oligomers lies about 100 cm-1 lower than in A-U base pairs in the Watson-Crick geometry in CDCl3 solution.8 The presented results demonstrate the capacity of combining IR pump-probe methods with calculations to reveal information on and identify hidden vibrational absorption bands. The kinetics of time-dependent frequency shifts of vibrational bands, which is a result of anharmonic coupling, provides possibilities to obtain insight into energy relaxation and redistribution pathways. Acknowledgment. We gratefully acknowledge financial support from the Deutsche Forschungsgemeinschaft (Sfb450). We thank Dr. Henk Fidder for valuable discussions on this work. References and Notes (1) Letellier, R.; Ghomi, M.; Taillandier, E. J. Biomol. Struct. Dyn. 1987, 4, 663. (2) Ouali, M.; Gousset, H.; Geinguenaud, F.; Liquier, J.; GabarroArpa, J.; LeBret, M.; Taillandier, E. Nucleic Acids Res. 1997, 25, 4816. (3) Tsuboi, M. Applied Spectroscopy ReViews 1969, 3, 45. (4) Clark, R. J. H.; Hester, R. E. AdVances in Infrared and Raman Spectroscopy; Wiley Heyden Ltd.: New York, 1985; Vol. 12. (5) Floria´n, J.; Hrouda, V.; Hobza, P. J. Am. Chem. Soc. 1994, 116, 1457. (6) Floria´n, J.; Leszczynski, J.; Johnson, B. G. J. Mol. Struct. 1995, 349, 421. (7) Shishkin, O. V.; Sponer, J.; Hobza, P. J. Mol. Struct. 1999, 477, 15. (8) Woutersen, S.; Cristalli, G. J. Chem. Phys. 2004, 121, 5381. (9) Krummel, A. T.; Mukherjee, P.; Zanni, M. T. J. Phys. Chem. B 2003, 107, 9165. (10) Howard, F. B.; Miles, H. T. J. Biol. Chem. 1965, 240, 801. (11) Nir, E.; Janzen, C.; Imhof, P.; Kleinermanns, K.; de Vries, M. S. Phys. Chem. Chem. Phys. 2002, 4, 740. (12) Brandl, M.; Lindauer, K.; Meyer, M.; Suhnel, J. Theor. Chem. Acc. 1999, 101, 103. (13) Spirko, V.; Sponer, J.; Hobza, P. J. Chem. Phys. 1997, 106, 1472. (14) Sponer, J.; Gabb, H. A.; Leszczynski, J.; Hobza, P. Biophys. J. 1997, 73, 76. (15) Sponer, J.; Leszczynski, J.; Hobza, P. J. Mol. Struct. (THEOCHEM) 2001, 573, 43. (16) Lee, C.; Park, K. H.; Kim, J. A.; Hahn, S.; Cho, M. J. Chem. Phys. 2006, 125, 114510. (17) Lee, C.; Cho, M. J. Chem. Phys. 2006, 125, 114509. (18) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New York, 1984. (19) Parvathy, V. R.; Bhaumik, S. R.; Chary, K. V. R.; Govil, G.; Liu, K. L.; Howard, F. B.; Miles, H. T. Nucleic Acids Res. 2002, 30, 1500. (20) Pattabiraman, N. Biopolymers 1986, 25, 1603. (21) Pichler, A.; Rudisser, S.; Winger, R. H.; Liedl, K. R.; Hallbrucker, A.; Mayer, E. Chem. Phys. 2000, 258, 391. (22) Lee, C.; Cho, M. J. Chem. Phys. 2007, 126, 145102. (23) Fritzsche, H. Nucleic Acids Res. 1994, 22, 787. (24) Taillandier, E.; Ridoux, J. P.; Liquier, J.; Leupin, W.; Denny, W. A.; Wang, Y.; Thomas, G. A.; Peticolas, W. L. Biochemistry 1987, 26, 3361. (25) Falk, M.; Hartman, K. A.; Lord, R. C. J. Am. Chem. Soc. 1963, 85, 387.

Study of DNA Oligomers by Ultrafast IR Spectroscopy (26) Vargason, J. M.; Henderson, K.; Ho, P. S. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 7265. (27) Adam, S.; Liquier, J.; Taboury, J. A.; Taillandier, E. Biochemistry 1986, 25, 3220. (28) Miles, H. T. Proc. Natl. Acad. Sci. U.S.A. 1964, 51, 1104. (29) Liquier, J.; Akhebat, A.; Taillandier, E.; Ceolin, F.; Dinh, T. H.; Igolen, J. Spectrochim. Acta A: Mol. Biomol. Spectrosc. 1991, 47, 177. (30) Sutherland, G. B. B. M.; Tsuboi, M. Proc. R. Soc. London A: Math. Phys. Sci. 1957, 239, 446. (31) Ghomi, M.; Letellier, R.; Liquier, J.; Taillandier, E. Int. J. Biochem. 1990, 22, 691. (32) Lee, C.; Park, K. H.; Cho, M. J. Chem. Phys. 2006, 125, 114508. (33) Benevides, J. M.; Thomas, G. J. Nucleic Acids Res. 1983, 11, 5747. (34) Heyne, K.; Nibbering, E. T. J.; Elsaesser, T.; Petkovic´, M.; Ku¨hn, O. J. Phys. Chem. A 2004, 108, 6083. (35) Cowan, M. L.; Bruner, B. D.; Huse, N.; Dwyer, J. R.; Chugh, B.; Nibbering, E. T. J.; Elsaesser, T.; Miller, R. J. D. Nature 2005, 434, 199. (36) Heyne, K.; Huse, N.; Dreyer, J.; Nibbering, E. T. J.; Elsaesser, T.; Mukamel, S. J. Chem. Phys. 2004, 121, 902. (37) Nibbering, E. T. J.; Elsaesser, T. Chem. ReV 2004, 104, 1887. (38) Giese, K.; Petkovic´, M.; Naundorf, H.; Ku¨hn, O. Phys. Rep. 2006, 430, 211. (39) Jurecka, P.; Sponer, J.; Cerny, J.; Hobza, P. Phys. Chem. Chem. Phys. 2006, 8, 1985. (40) Fonseca Guerra, C.; Bickelhaupt, F. M.; Snijders, J. G.; Baerends, E. J. J. Am. Chem. Soc. 2000, 122, 4117. (41) Villani, G. Chem. Phys. 2005, 316, 1. (42) Gorb, L.; Podolyan, Y.; Dziekonski, P.; Sokalski, W. A.; Leszczynski, J. J. Am. Chem. Soc. 2004, 126, 10119. (43) Krishnan, G. M.; Ku¨hn, O. Chem. Phys. Lett. 2007, 435, 132. (44) Brauer, B.; Gerber, R. B.; Kabelac, M.; Hobza, P.; Bakker, J. M.; Riziq, A. G. A.; de Vries, M. S. J. Phys. Chem. A 2005, 109, 6974. (45) Krummel, A. T.; Zanni, M. T. J. Phys. Chem. B 2006, 110, 13991. (46) Obrien, F. E. M. J. Sci. Instrum. Phys. Ind. 1948, 25, 73.

J. Phys. Chem. B, Vol. 112, No. 26, 2008 7915 (47) Kaindl, R. A.; Wurm, M.; Reimann, K.; Hamm, P.; Weiner, A. M.; Woerner, M. J. Opt. Soc. Am. B 2000, 17, 2086. (48) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; CuiA, Q.; Baboul, G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision B.04; Gaussian, Inc.: Pittsburgh PA, 2003. (49) Schneider, W.; Thiel, W. Chem. Phys. Lett. 1989, 157, 367. (50) Krishnan, G. M.; Kühn, O. Chem. Phys. Lett. 2007, 435, 132. (51) May, V.; Ku¨hn, O. Charge and Energy Transfer Dynamics in Molecular Systems, 2nd ed.; Wiley-VCH: Weinheim, Germany, 2004. (52) Meyer, H. D.; Worth, G. A. Theor. Chem. Acc. 2003, 109, 251. (53) Worth, G. A.; Beck, M. H.; Ja¨ckle, A.; Meyer, H.-D. The MCTDH Package, version 8.3; University of Heidelberg: Heidelberg, Germany, 2005. (54) Taillandier, E.; Liquier, J. Handbook of Vibrational Spectroscopy; John Wiley & Sons: Chichester, U.K., 2002; Vol. 5. (55) Nienhuys, H. K.; Woutersen, S.; van Santen, R. A.; Bakker, H. J. J. Chem. Phys. 1999, 111, 1494. (56) Huse, N.; Ashihara, S.; Nibbering, E. T. J.; Elsaesser, T. Chem. Phys. Lett. 2005, 404, 389. (57) Wang, Z. H.; Pakoulev, A.; Pang, Y.; Dlott, D. D. J. Phys. Chem. A 2004, 108, 9054.

JP711262Y