ARTICLE pubs.acs.org/JPCA
Specific Photodynamics in Thymine Clusters: The Role of Hydrogen Bonding Yevgeniy Nosenko,† Maksim Kunitski, and Bernhard Brutschy* Institut f€ur Physikalische und Theoretische Chemie, Goethe-Universit€at, Max-von-Laue-str. 7, 60438 Frankfurt/Main, Germany
bS Supporting Information ABSTRACT: A photoionization detected IR study of thymine and 1-methylthymine monohydrates and of their homodimers was carried out to shed some light on the structure of the thymine clusters whose complex photodynamics has recently been the subject of great interest. Under supersonic jet conditions, thymine forms doubly H-bonded cyclic clusters with water or another base preferentially via its N1H group and the adjacent carbonyl group. This hydrate is of no biological relevance since the N1H group is the sugar binding site in thymidine. On the other hand, 1-methylthymine forms the donor H-bonds only via the N3H group. Hence, properties of the N1H and the N3H bound clusters of thymine can be studied using thymine and 1-methylthymine molecules, respectively. No biologically relevant conformations of the dimers and hydrates of thymine, contrary to those of 1-methylthymine, are observed under supersonic jet conditions. Thymine homodimer, which extensively fragments upon UV ionization by formation of a protonated monomer, exhibits two N1H 3 3 3 OdC2 hydrogen bonds. The photodynamics of hydrated thymines is found to be extremely sensitive to the hydration site: ranging from an ultrafast relaxation in less than 100 fs up to formation of a dark state with the lifetime on the microsecond time scale.
1. INTRODUCTION Most photolesions in DNA concern the pyrimidine bases thymine and cytosine.1 This can be related to the fact that isolated pyrimidine nucleobases dissipate electronic energy after optical excitation slower than the purines, having thus higher probability for photoreactions.2 Understanding of the photophysics and photochemistry of the pyrimidine bases is therefore crucial for understanding the nature of UV-induced carcino- and mutagenesis. The photodynamics of thymine (T) has been extensively studied both in solution3 and in the gas phase.2,411 In the latter, intrinsic properties may be distinguished at the molecular level from effects imposed by the environment. Of particular interest are properties of the isolated microhydrates due to the fundamental role of water in biology. Isolated T reveals a broad vibrationally featureless UV absorption spectrum at λ < 280 nm, explained in terms of either the large geometry change upon electronic excitation or an ultrafast relaxation of the optically accessible state.4 Indeed, electronically excited T (at 267 nm) decays via an ∼100 fs and a 59 ps transient, as measured in femtosecond pumpprobe ionization experiments.2,7,10,11 Ullrich et al. have found transients with even shorter lifetimes, that is, 97%) and 1-methylthymine (white, polycrystalline) samples were purchased from Sigma-Aldrich and used as delivered. Isolated T (mT), its dimers, and its hydrates were produced in a pulsed supersonic expansion of helium (3 bar), seeded with the vapor of T (mT) and water, utilizing a commercially available Even-Lavie valve.28 The nozzle temperature was maintained at 200 and 160 °C for the experiments with T and mT, respectively. The water concentration was controlled by a bubbler type gas mixing station. In some of the measurements, water contamination of the gas line/carrier gas was sufficient for observing hydrates. Because the fragmentation of larger molecular clusters can significantly affect the double-resonance vibrational spectra of smaller species,29 the cluster size distributions were optimized to be as narrow as possible by minimizing the signal in the 1:2 hydrate channels to a trace level. Additionally, the IR spectra were recorded also in the ion channels of the heavier clusters, where available, to disentangle an eventual contribution to the vibrational fingerprints of the investigated species. The IR/fsMPI measurements of jet-cooled thymines and their complexes were performed as described previously25 using a home-built tunable seeded narrowband IR-OPO as an IR pump and the third harmonics (267 nm, autocorrelation of 240 fs fwhm) of the chirp amplified Ti:Sapphire system for the onecolor UV-ionization probe. Here, the double-resonance effect is caused by different photoionization rates for cold and vibrationally excited species originating, in turn, from different Franck Condon overlaps and fragmentation rates. The pumpprobe time delay was typically 80 ns. Calibration of the IR frequency was accomplished by measuring the red and green wavelengths used for the difference frequency generation of the seed radiation by means of a wave-meter (ATOS LM-007). Thus, the final accuracy was (1 cm1 in the tuning range 28003800 cm1. The ultrafast excited state dynamics of the mT monohydrate was measured by the femtosecond pumpprobe technique using 267 nm for excitation (pump) and 800 nm for ionization (probe). The time delay between the pump and the probe pulses was controlled by an optical delay stage. Pump pulses with an energy of 18 μJ were collimated with a long focal-length lens (4 m), whereas the probe pulses with an energy of 280 μJ were focused with a 50 cm lens into the ionization region of a linear time-offlight mass spectrometer. Both beams were spatially overlapped by means of a dichroic beam combiner and introduced into the vacuum system perpendicularly to the molecular beam. The cross-correlation of the pump and probe pulses (240 fs) was estimated from a time-resolved ionization spectrum of indole, which was added at a small concentration to the gas mixture. 9430
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A
ARTICLE
Table 1. ZPVE and BSSE Corrected B3LYP/6-311þG** and PW91/6-311þþG** Stabilization Energies (in kcal/mol) of Doubly H-Bonded Thymine and 1-Methylthymine Monohydrates and Homodimers in Comparison with Literature Data
a
B3LYP/ 6-31þþG
MP2/
MP2/
MP2/ 6-31G*
(d,p)a
DZPib
TZVPPc
(0.25)d
PW91
B3LYP
mT34:W
8.3
6.5
mT32:W
7.5
6.0
T12:W
7.7
10.2
9.9
9.5
8.0
T34:W
6.2
8.6
8.1
8.2
6.4
T32:W mT34:mT43
5.9
8.5
8.2
7.6 11.4
6.1 9.4
mT34:mT23
10.6
10.7
9.0
mT32:mT23
10.5
10.1
8.7
T12:T21
15.9
17.1
15.2
T12:T43
13.0
14.1
12.1
T12:T23
12.8
13.3
11.6
T34:T43
10.6
11.5
9.5
T34:T23 T32:T23
10.6
10.8 10.3
9.2 8.9
Ref 21. b Ref 22. c Ref 23. d Ref 24.
Ground state geometries, binding energies, and vibrations of thymine hydrates were calculated at the B3LYP30,31/6-311þG** level of theory, suitable for vibrational analysis of small molecular hydrates.32 We have also tested the PW9133 functional in combination with the 6-311þþG** basis set, as recently recommended for base pairs.34 All optimized structures were checked for the absence of imaginary vibrational frequencies ensuring that real energy minima were found. Basis set superposition error (BSSE) of the binding energies was accounted for by the counterpoise correction.35 The harmonic B3LYP vibrational frequencies, which are higher and lower than 2000 cm1, were scaled by factors of 0.95832 and 0.985, respectively. The latter was optimized for the NH in plane bending and carbonyl stretch vibrations of gas phase thymine.36 In addition, vertical electronic spectra were calculated for selected clusters at their ground state geometries by means of the TDDFT/B3LYP/6-311þG** method. The calculations were performed using the Gaussian 03 program package.37
3. RESULTS 3.1. Calculations. We limit our consideration to cyclic doubly H-bonded species, because no evidence of other types of clusters was found in the spectroscopic data. In particular, we observed no indication of π-stacked structures, as will be discussed later. This choice is also justified by similar molecular systems reported in literature.9,19,20,38,39 A thymine molecule possesses two carbonyl and two NH groups whose topology provides three possible hydration sites for a water molecule forming a doubly H-bonding bridge between the adjacent CdO and NH groups of the base. As the mT molecule lacks the N1H group, only two such monohydrate structures are possible (Figure 1). Similarly, six and three isomers of the homodimers of T and mT, respectively, can be formed by means of two antiparallel H-bonds. We encode the isomeric species by numbers, which indicate the functional groups of the bases involved in the H-bonding. The T34:T43, T34:T23, and T32:T23 isomers of T dimer directly
Figure 2. Ion signal (open circles) of 1-methylthymine (bottom) and its monohydrate (top) as a function of the delay time between the pump (267 nm) and probe (800 nm) pulses. The best fit and residuals are depicted with lines. Note the different time scales.
reflect the mT dimer forms. The binding energies of all these structures are listed in Table 1. The PW91 functional yielded interaction energies closer to the MP2 values than the B3LYP functional. The stability order was found to be independent of the functional used. Furthermore, neither the BSSE (e1 kcal/mol) nor the ZPVE (12 kcal/mol) corrections did alter the stability order of the clusters. In the vibrational analysis, we employed the B3LYP/6-311þG** results because these provided better overall agreement of the measured vibrational frequencies. In agreement with previously reported calculations, the N1H bound isomers are more stable than the N3H bound ones. For example, the T12:T21 dimer is by 40% more stable than the T32: T23 isomer. So, the formation of differently bound species is expected in the experiments with T and mT. Although the N3H/C4dO bound complexes are systematically more stable than the N3H/C2dO bound ones, the energy difference is not big enough for ruling out the less stable clusters. At the MP2 level of theory, these structures tend to be degenerate. The calculated vibrational spectra will be discussed aside the corresponding experimental ones. 3.2. Time-Resolved Femtosecond PumpProbe Ionization Spectra. The ion signals of mT and its monohydrate as a function of time delay between the pump (267 nm) and the probe (800 nm) are shown in Figure 2. The 3-fold decay of the mT monomer transient ionization signal is similar to that of thymine and is discussed in a separate paper.15 The time-dependent ion signal of the mT monohydrate, in contrast to the monomer, showed only a single exponential decay with a constant of 100 fs or less. A similar dynamics is observed for the T monohydrate by Schultz and co-workers.11 However, in addition to the intense 100 fs decay component they also detected a very weak 7 ps decay. 3.3. IR Spectra of 1-Methylthymine. 3.3.1. Monomer. The IR/fsMPI vibrational spectra of mT, its monohydrate, and its dimer are compared in Figure 3 with the corresponding simulated IR spectra (B3LYP/6-311þG**). The spectrum of the mT monomer (Figure 3a) is characterized by the NH stretch (νN3H) at 3443 cm1 and by a group of bands in the region from 2900 to 3000 cm1 assigned to the CH3 stretches. The 9431
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A
ARTICLE
Table 2. Experimental NH Stretch Frequencies (in cm1) of Isolated 1-Methylthymine (mT) and Thymine (T) Monomers νN1H mT
T
Figure 3. IR/fsMPI spectra of (a) 1-methylthymine, (b) its monohydrate, and (c) its dimer in comparison with the corresponding spectra calculated at the B3LYP/6-311þG** level. Calculated positions of the βOH2, νC5dC6, νC4dO, and νC2dO overtones for hydrates and the βuN3H, νuC5dC6, νuC4dO, and νuC2dO overtones for dimers in energy ascending order are plotted with sticks.
νC6H mode is seen as a weak band at 3051 cm1. In addition, broader bands appear at 3170 and 3070 cm1, which we assign to fragmentation bands of the mT dimer, as it is evident from a comparison with the dimer spectrum (Figure 3c). The νN3H mode (3443 cm1) was found shifted by 9 cm1 to the blue, as compared with the previously reported value of 3434 cm1 determined with IR/nsR2PI.9 Because vibrations of the enol tautomers of mT are expected at significantly higher frequencies,
νN3H
environment/reference
3443
jet/this work
3434 3434
jet/9 Ar matrix/40
3417
N2 matrix/40
3435
gas cell/41
3481
3435
gas cell/41
3484
3437
gas cell/36
3484.9
3439.4
jet/19
3494.2
3444.5
He droplet/42
at around 3600 cm1, the discrepancy must be attributed to different calibration procedures. The comparison of the frequency of this mode measured in a supersonic jet with those from FTIR spectra recorded in cold matrices40 and gas cell36,41 is limited due to matrix and temperature effects, respectively. Both are capable of shifting the frequency and broadening the band. Because methylation at the N1 position does not affect the νN3H frequency41 we refer for comparison to thymine. Casaes et al. have reported a value of 3439.4 cm1 for the νN3H mode observed for jet-cooled T by IR cavity ringdown spectroscopy.19 Measurements in helium nanodroplets, with presumably minimum matrix-induced vibrational perturbations give 3444.5 cm1.20,42 These two values, obtained for T, are in good agreement with the presently reported frequency of 3443 cm1, measured for mT (Table 2). 3.3.2. 1-Methylthymine Monohydrate. The IR/fsMPI spectrum of the mT monohydrate (Figure 3b) agrees well with the calculated IR spectra of both doubly H-bonded monohydrate isomers with the mT34:W isomer fitting slightly better. Thus, the bands at 3724, 3442, and 3288 cm1 are readily assigned to the free OH (νOHfree), H-bonded OH (νOHhb), and H-bonded NH (νN3Hhb) stretching vibrations, respectively. The shoulder peak at 3393 cm1 can be explained in three ways: either in terms of a fragmenting dihydrate, another isomer, or a Fermi resonance. The first option has been invoked by Busker et al.9 based on the MP2 vibrational analysis. The IR/fsMPI experiment was performed under conditions, where the dihydrate ion signal was too weak for measuring its IR spectrum. So, we neither prove nor reject this possibility. The contribution of an isomeric hydrate is supported by the fact that for the mT32:W and mT34:W forms the calculated νOHhb frequency differ by 37 cm1, in good agreement with the observed splitting of 49 cm1. However, the intensities within the 3393/3442 cm1 doublet would suggest in this case that the mT32:W isomer is the global minimum, in contrast to the calculated relative stabilities of the hydrates (Table 1). Therefore, we prefer the third interpretation, where the 3393/3442 cm1 doublet is rationalized by a Fermi resonance of the νOHhb fundamental and an overtone/ combination mode. The unperturbed frequencies of the Fermi resonant modes can be estimated assuming a zero intensity for the unperturbed overtone (or combination) and the interaction of only two vibrational states. The observed intensity ratio of the overtone and fundamental (Io/If) can then be expressed in terms of the observed spectral separation (δ) and an unperturbed one (δ0) as follows:43 Io =If ¼ R ¼ ðδ δ0 Þ=ðδ þ δ0 Þ w δ0 ¼ δð1 RÞ=ð1 þ RÞ 9432
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A
ARTICLE
Table 3. Experimental Vibrational Frequencies/H-Bonding Induced Spectral Shiftsa (in cm1) of Thymine (T), 1-Methylthymine (mT), Uracil (U), and 2-Pyridone (2PY) Monohydrates νOHfree
b
νNHhb
mT:Wb
3724
3442/265
3288/155
mT:Wc
3712
3455/252
3284/150
T:W
3729
3456/251
3328/166
U12:Wd
3727.1
3467.5/240
3317.3/176.6
U34:Wd
3722.3
3467.5/240
3271/172.7
d
U32:W
3727.8
3500.7/206
3256/187.7
2PY:We
3723
3347/360
3289/159
b
a
νOHhb
Relative frequencies with respect to the corresponding νOH This work. c Ref 9. d Ref 20. e Ref 39.
free
Table 4. Experimental Absolute (ν) and Relative (ννOHfree) Vibrational Frequencies (in cm1) of the 1-Methylthymine Monohydrate and the Corresponding Theoretical Values for Its Doubly H-Bonded Isomers Calculated at the B3LYP/6-311þG** Level of Theory theory mT32:W
experiment mT34:w
IR/R2PIa
IR/fsMPIb
ν ννOHfree ν ννOHfree ν ννOHfree ν ννOHfree νOHfree 3732 νOHhb 3494 νNHhb 3273 a
0 238 459
3731 3457 3264
0 274 467
3712 3455 3284
0 257 428
3724 3442 3288
0 282 436
Ref 9. b Present work.
values.
Due to the spectral saturation and the irregular shape of the bands we can only roughly estimate an intensity ratio R = 17/21.5 by integration of the doublet components. Given the observed splitting δ of 49 cm1, the above equation yields a separation of the unperturbed vibrational states of 6 cm1. The corresponding unperturbed frequencies of the νOHhb fundamental and the overtone/combination modes are then obtained by a symmetric narrowing the δ gap at 3421 and 3415 cm1, respectively. The vibrations of thymine measured in gas phase36 suggest the overtone of the νC4dO mode (1725 cm1) as the best candidate for the Fermi resonance component at 3393 cm1. Applying the anharmonicity parameter of the CO molecule (ωeχe = 13 cm1) the νC4dO overtone frequency can be estimated at 3424 cm1. The H-bonding to a water molecule (in the mT34:W isomer) should further downshift the frequency. For example, a red shift of 14 cm1 has been observed for the νCdO mode in hydrated 2-pyridone by Mikami and coworkers.38 Such a frequency shift would result in a value of 3396 cm1 for the νC4dO overtone in mT, in good agreement with both the estimated unperturbed frequency of 3415 cm1 and the calculated DFT value of 3387 cm1. Another possible candidate is the combination of both carbonyl stretches (3440 cm1) whose calculated overtone frequencies are displayed in Figure 3 by sticks at 3387 and 3493 cm1. All in all, the observed IR/fsMPI spectrum can be satisfactorily explained by the lowest energy monohydrate isomer (mT34:W). Despite a general similarity of the IR/fsMPI and the IR/nsR2PI9 spectra there were some quantitative differences, not explainable by different calibration procedures. Whereas slightly different intensity ratios, in particular, a stronger relative depletion at 3288 cm1 in the IR/fsMPI spectrum, can be explained by different photoionization schemes, providing in general specific Franck Condon factors and ionization probabilities, the mismatch of the vibrational frequencies summarized in Table 3 requires more attention. Both sets of frequencies in this table are compared with those of the structurally related monohydrates of uracil and 2-pyridone. Comparison of the νOHfree mode of different monohydrates (Table 3) supports the correctness of our calibration, whereas the IR/nsR2PI data are shifted to lower energies by ≈10 cm1. Noteworthy, the frequency difference is similar to that for the νN3H vibration of the monomer (3443 vs 3434 cm1, Table 2). Remarkably, the discrepancy between the frequencies measured in the current work and those from the IR/nsR2PI study is not systematic. While the νOHfree and νNHhb bands are by 12 and 4 cm1 higher in the present work,
the νOHhb mode is lower by 14 cm1. Spectral contamination of the monohydrate ion signals by fragmenting larger hydrates to a different and undefined extent, as already mentioned above, may partially explain the discrepancies. Another possible reason is that different hydrate isomers are detected by the femtosecond (2 267 nm) and the delayed nanosecond (273.3 nm þ 213 nm) ionization schemes. Assuming this let us compare the relevant experimental and calculated vibrational frequencies listed in Table 4. Because the calculated νOHfree mode is within 1 cm1, the same for both the mT34:W and mT32:W isomer contrary to the observed splitting of 12 cm1, it is convenient to consider the νOHhb and νNHhb frequencies relative to the free OH ones. Thus, the relative frequency values of the calculated νOHhb and νNHhb modes of the mT34:W (mT32:W) isomer are 274 (238) and 467 (459) cm1 (Table 4), with the calculated isomeric splittings of 36 and 8 cm1, respectively. The differences of the relative νOHhb and νNHhb frequencies in the IR/fsMPI and IR/nsR2PI studies are 25 and 8 cm1, respectively, with the IR/fsMPI data revealing lower relative frequencies. Because these splitting values are in very good agreement with the calculated isomeric ones of 36 and 8 cm1, respectively, we conclude that the spectrum of Busker et al. represents the mT32:W form, while we analyzed the mT34:W species. The reason why the two photoionization schemes select different isomers of the mT monohydrate will be discussed later. 3.3.3. 1-Methylthymine Homodimer. The dimer of mT revealed IR absorption bands in the frequency interval between 2800 and 3200 cm1 (Figure 3c). Polycrystalline mT40 as well as the structurally related 1-cyclohexyluracil clusters (in CDCl3 solution)44 reveal similar IR spectra, albeit with broader spectral features due to environment effects. Noteworthy, the solid state mT sample is composed of mT34:mT43 dimers according to the X-ray data of Hoogsteen.45 The jet-cooled dimer of 2-pyridone, considered as a mimic of T, is characterized by a broad (∼230 cm1) band at ∼2800 cm1 indicating stronger H-bonding than in the dimer of mT.38 From a comparison with the monomer spectrum (Figure 3a), we assign the narrow bands at 2942, 2969, and 2985 cm1 to the νCH3 modes. More importantly, the lack of the νN3Hfree vibration of the monomer (3443 cm1) confirms a cyclic doubly H-bonded structure of the dimer and excludes stacked isomers. Three such geometries with similar binding energies are possible (Table 1). The calculated IR spectra of these three isomers are also similar in the spectral region under study (Figure 3) and characterized by one strong vibrational mode representing 9433
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A
Figure 4. IR/fsMPI spectra recorded in parallel for the (a) thymine dihydrate, (b) thymine monohydrate, and (c) thymine monomer in comparison to (df) the calculated IR spectra of the thymine monohydrate isomers. Calculated positions of the βN1H, βOH2, νC5dC6, νC4dO, and νC2dO overtones in energy ascending order are indicated with sticks.
a cooperative antiphase stretch of two H-bonded N3H groups (νaN3Hhb). The IR/fsMPI spectrum is, however, not that simple. Two intense absorption features may alternatively be associated with the νaN3Hhb mode: either the band at 3071 cm1 or the irregularly shaped band centered at 3170 cm1 and possessing a peak at 3193 cm1 and a low-energy shoulder. We assign the strongest depletion band at 3071 cm1 to the νaN3Hhb mode of the mT34:mT43 isomer. Although the calculated frequency, 3158 cm1, is closer to the observed band at 3170 cm1, it is likely overestimated because the value was obtained by scaling of the harmonic frequencies with a uniform scaling factor, ideal for the free NH and OH stretches. Strong H-bonding, however, implies larger anharmonicity coefficients for the bound stretch modes and, thus, smaller scaling factors.
ARTICLE
According to Biemann et al., who studied the 1-cyclohexyluracil clusters (in CDCl3 solution),44 the bands at 3071 and 3170 cm1 might correspond to C4dO and C2dO bound N3H stretches, respectively, implying the presence not only of the mT34:mT43 dimer, but also of the mT34:mT23 species. Though this isomeric assignment is not unreasonable, it does not fully explain the magnitude of the IR depletion observed. In its maximum, this reached up to 80%, which is too intense for a mixture of different isomers. Also, the calculated splitting of these C4dO and C2dO bound N3H stretches (1 μs.9 The question arises: why is the mT34:W isomer obviously not trapped in the dark state? Busker et al. reports that, in the vertical excitation spectrum, calculated at the CC2/aug-cc-pVTZ level of theory, the bright 1 ππ* state of this hydrate is lower than the 1nπ* state in contrast to the mT monomer where the situation is just opposite.9 Etinski and Marian predict by ab initio calculations the same effect for the microhydrated and solvated thymine derivatives and point out that such excited states reversal might decrease the population of both the 1nπ* and the triplet states thus facilitating the photostability of the nucleobase.48 We agree with these
ARTICLE
arguments and point out in addition that, given the hydration induced stabilization of the optically accessible 1ππ* state, the hydrate acquires at the same excitation wavelength (267 nm) a higher energy excess than the monomer, enabling a faster passage of a barrier on the way to the 1ππ*/S0 conical intersection. This explains perfectly the ultrafast (e100 fs), barrierless decay of the photoexcited mT34:W isomer (Figure 2). The excited states of the mT32:W isomer, however, which is not the lowest energy structure, have not been studied until now in quantum chemistry literature to the best of our knowledge. To fill this gap we calculated the vertical energies of the 1nπ* and 1 ππ* states of the mT monomer and its monohydrates using the TD/B3LYP/6-311þG** method. The results are depicted in Figure 6. There is perfect agreement to within 0.05 eV of these energy values for the mT monomer and the mT34:W hydrate with the CC2/aug-cc-pVTZ//CC2/cc-pVDZ data.9 This justifies the use of the low-cost TD-DFT method also for other isomers. Thus the mT32:W hydrate reveals the same order of the 1 nπ* and 1ππ* states like the mT monomer with an even larger energy separation between these states. Consequently, we expect a stronger coupling of the bright 1ππ* state to the 1nπ* state, which either itself slows down electronic relaxation or mediates a transition into the long-lived triplet state.14 The observation of the mT32:W species by the pumpprobe ionization with the ns to μs time delays9 corroborates this hypothesis. It should be pointed out that both isomers whose binding energies differ by 0.5 kcal/mol only or less (Table 1) are formed in a supersonic expansion in amounts enough for detection. However, the mT32:W isomer is not detected by fsMPI. There are a few explanations for this observation. First, the mT32:W hydrate, being not the most stable, is less abundant than the mT34:W one. Second, the fsMPI is often not sensitive enough to slow transients because the corresponding states are populated with small quantum yields, as discussed in the literature.11 Finally, the long-lived transients of the A and T clusters are prone to fragmentation under multiphoton ionization.8 Assuming these, we conclude that a possible ultrafast relaxation of the mT32:W isomer contributes to the fsMPI signal of the mT monohydrate in a negligible way compared to the mT34:W isomer. Otherwise, its vibrational fingerprint, though partially overlapping with that of the mT34 hydrate, should be visible in the IR/fsMPI spectrum in the region of the νOHhb mode (Table 3). The vibrational analysis of the IR/fsMPI spectra leaves no doubts that hydration of thymine in a supersonic jet involves predominantly the H-bonding to the N1H group. So, the T12: W isomer exhibits the biexponential decay of the femtosecond pumpprobe ionization signal with an intense component of e100 fs and a weak 7 ps component.11 On the longer time scale, the monohydrated thymine is reported to decay slightly faster than the monomer, that is, in 12 versus 22 ns.5 For stability reasons (Table 1), we believe that the same T12:W isomer is also responsible for the observed nanosecond dynamics. The less stable isomers (T34:W and T32:W) should be significantly less abundant than the mT32:W isomer in the case of the mT monohydrates (Table 1). Hence, we assign all three transients with lifetimes in the fs, ps, and ns ranges to the T12:W isomer. The vertical electronic spectra calculated by Schultz and coworkers at the CC2/aug-cc-pVTZ//MP2/cc-pVDZ level of theory are practically identical for the T monomer and the T12:W hydrate with the 1nπ* state being lower than the 1ππ* state. This qualitatively explains the similarity of the electronic 9436
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A relaxation of thymine and its T12:W hydrate, both exhibiting decay components e100 fs, 7 ps, and a few tens of nanoseconds.5,11 As we already mentioned, under hydration, the nanosecond relaxation component of T becomes faster while that of mT gets slower.5,9 Busker et al. have discussed this difference in terms of different hydration conditions in the corresponding experiments. According to the present structural findings, we propose that the different nanosecond dynamics of the T and mT hydrates is caused by the different hydration sites. Similarly, we recently observed an isomer-specific photodynamics of the adenine monohydrate.25 Here, the fragmentation pattern of the adenine monohydrate upon fsMPI with 267 nm pulses strongly depends on the water binding position. We have explained this difference by postulating two specific relaxation paths: a fast internal conversion and a direct excited state fragmentation for the amino- and the N9H bound hydrates, respectively. In the case of the hydrated thymines, the fragmentation was found to vary from no effect for mT34:W to an IR induced fragmentation of the T12:W species into the monomer ion channel. However, the difference was not that distinct as for the adenine hydrates. This is reasonable because, the fsMPI detection is generally not sensitive to the dynamics on the nanosecond time scale, where the isomer specific electronic relaxation of the thymine hydrates occurs. Hence, this problem should be investigated at nanosecond resolution using a twocolor two-photon ionization probe,5,9 which is presently not available in our laboratory. Nonetheless, the hydration site dependence seems to be the most plausible explanation for the observed diversity in the photodynamics of hydrated thymines. We found that the relative vertical energies of the 1nπ* and 1 ππ* excited states can qualitatively explain the dependence of photodynamics of the T and mT hydrates on the H-bonding site, as observed experimentally. For a quantitative analysis of the observed transient lifetimes, knowledge is required on the conical intersections availability, energy barriers, spinorbit couplings, and so on, depending on the hydration site. This subject deserves a special theoretical investigation. Given that electronic relaxation mechanism of thymine is still under debate, the presently reported structural data must facilitate the theoretical treatment of the problem in particular for the monohydrates. 4.2. Dimers. By vibrational analysis of the IR/fsMPI spectra, we could give evidence that the T dimer under supersonic jet conditions exists as its lowest energy isomer T12:T21. Its vibrational spectrum in the region of the H-bonded NH stretches is comparably complex as that of the mT homodimer. This complexity is not attributed to overlapping spectra of different isomers. Hence, the latter can indeed be assigned to one isomer, mT34:mT43. For the ultimate elucidation of a possible contribution of other isomers, however, the ionization detected double resonance IR-IR spectroscopy might be useful. The different H-bonding topologies of the mT and T homodimers are also reflected in their fragmentation patterns. While the mT dimer reveals some fragmentation to the monomer ion channel, the dimer of T distinctly fragments into the protonated monomer channel. The same is concluded by Samoylova et al. based on the time-resolved femtosecond pumpprobe ionization measurements.10 Hence, we attribute the photodynamics of the T dimer to the particular structure: T12:T21. The T12:T21 species, as well as that of the adeninethymine dimer, reveals in its excited state dynamics a 40 ps decay component, first discovered by Samoylova et al.10 and discussed in
ARTICLE
terms of an excited state electron-driven proton transfer process, as proposed by Perun et al.49 In the latter theoretical study the authors have shown that for the WatsonCrick isomer a charge transfer state connects the optically bright 1ππ* state with the ground state in a barrierless manner by a sequence of conical intersections along the proton transfer coordinate. This relaxation channel has been found inactive for the energetically lowest isomer of the A:T base pair, which has most probably been studied by Samoylova et al. The lowest energy A:T isomer is formed by the (A)N9H 3 3 3 OdC2(T) and (A)N3 3 3 3 HN1(T) H-bonds.24 Under supersonic jet conditions, both the T:T and A:T base pairs reveal the 40 ps excited state decay component. Therefore, T should be H-bonded in the jet-cooled A:T complex by its N1H/C2dO site, like in the T homodimer. The A:T base pair observed by Samoylova et al. is not the WatsonCrick but rather the lowest energy isomer. Consequently, we conclude that either the 40 ps transient has nothing to do with the WatsonCrick specific electron-driven proton transfer relaxation pathway, or theory, which excludes this pathway for the most stable A:T isomer,49 is yet not accurate enough. Anyway, a rigorous theoretical investigation of the excited state relaxation of the T dimer, whose structure is now known from the present work, should help to understand the observed photodynamics not only in the homodimer of T but also in the jet-cooled A:T base pair.
5. CONCLUSIONS The ground state IR spectra of the homodimers and monohydrates of 1-methylthymine and thymine were recorded by double-resonance ion depletion spectroscopy using femtosecond multiphoton ionization (fsMPI) detection. This method was mandatory due to the ultrashort lifetimes of the investigated species in their electronically excited states. The analysis of the vibrational spectra of the clusters by means of DFT calculations and by comparison with the available data of the monomers reveals that in T and mT molecules different sites are involved in H-bonding under supersonic jet conditions. A water or another T molecule form a double H-bonding bridge preferentially between the N1H and the C2dO groups of thymine in contrast to that of mT, where this binding is blocked by methylation. Because N1 is the sugar binding site, the jet-cooled T complexes are biologically not relevant. This is in contrast to mT clusters, where only N3H bound cyclic doubly H-bonded structures were detected and no indication of π-stacked structures could be found. In fact, every of the reported four vibrational spectra can be satisfactorily interpreted in terms of the corresponding lowest energy hydrate or dimer. To accomplish these assignments, we had to consider Fermi resonances that seem to be active in the investigated spectral region for all the studied thymine clusters. Such vibrational couplings might facilitate vibrational cooling of thymine derivatives in biological systems. In this work, we provide structural evidence that the 40 ps excited state transient in the thymine homodimer and the adeninethymine base pair10 occurs in geometries with thymine H-bonded at its sugar binding site. Hence, the corresponding electronic relaxation process is biologically not relevant. By the vibrational analysis for the detected species we get the structural background for the diverse photodynamics recently reported for the isolated thymine clusters.5,811 Our results strongly suggest a unique photodynamics for all of the three isomeric monohydrates of thymine. While the mT34:W form 9437
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A exhibits only an ultrafast excited state decay of e100 fs, a significant fraction of the mT32:W hydrates remains electronically excited during more than 1 μs after photoexcitation.9 The T12:W species is characterized by dynamics with both the short-lived e100 fs/7 ps11 and the nanosecond transient states.5 Although there may be a long way from hydrates to base pairs in comparing their dynamics after electronic excitation, we notice that hydration at the C4dO/N3H site of 1-methylthymine, involved also in the WatsonCrick base pairing of adeninethymine, leads to the fastest electronic relaxation of the thymine nucleobase.
’ ASSOCIATED CONTENT
bS
Supporting Information. Tables of the calculated IR frequencies and intensities. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Fax: þ49 69 79829560. Present Addresses †
Fachbereich Chemie, Technische Universit€at Kaiserslautern, Erwin-Schr€odinger-Str. 52, Kaiserslautern, 67663, Germany.
’ ACKNOWLEDGMENT We thank German Science Foundation for financial support by the DFG Project BR848/7-1. ’ REFERENCES (1) Cadet, J.; Sage, E.; Douki, T. Mutat. Res., Fundam. Mol. Mech. Mutagen. 2005, 571, 3. (2) Kang, H.; Lee, K. T.; Jung, B.; Ko, Y. J.; Kim, S. K. J. Am. Chem. Soc. 2002, 124, 12958. (3) Crespo-Hernandez, C. E.; Cohen, B.; Hare, P. M.; Kohler, B. Chem. Rev. 2004, 104, 1977. (4) Brady, B. B.; Peteanu, L. A.; Levy, D. H. Chem. Phys. Lett. 1988, 147, 538. (5) He, Y. G.; Wu, C. Y.; Kong, W. J. Phys. Chem. A 2004, 108, 943. (6) Ullrich, S.; Schultz, T.; Zgierski, M. Z.; Stolow, A. Phys. Chem. Chem. Phys. 2004, 6, 2796. (7) Canuel, C.; Mons, M.; Piuzzi, F.; Tardivel, B.; Dimicoli, I.; Elhanine, M. J. Chem. Phys. 2005, 122. (8) Gador, N.; Samoylova, E.; Smith, V. R.; Stolow, A.; Rayner, D. M.; Radloff, W. G.; Hertel, I. V.; Schultz, T. J. Phys. Chem. A 2007, 111, 11743. (9) Busker, M.; Nispel, M.; Haber, T.; Kleinermanns, K.; Etinski, M.; Fleig, T. ChemPhysChem 2008, 9, 1570. (10) Samoylova, E.; Schultz, T.; Hertel, I. V.; Radloff, W. Chem. Phys. 2008, 347, 376. (11) Gonzalez-Vazquez, J.; Gonzalez, L.; Samoylova, E.; Schultz, T. Phys. Chem. Chem. Phys. 2009, 11, 3927. (12) Hudock, H. R.; Levine, B. G.; Thompson, A. L.; Satzger, H.; Townsend, D.; Gador, N.; Ullrich, S.; Stolow, A.; Martinez, T. J. J. Phys. Chem. A 2007, 111, 8500. (13) Serrano-Perez, J. J.; Gonzalez-Luque, R.; Merchan, M.; SerranoAndres, L. J. Phys. Chem. B 2007, 111, 11880. (14) Etinski, M.; Fleig, T.; Marian, C. A. J. Phys. Chem. A 2009, 113, 11809. (15) Kunitski, M.; Nosenko, Y.; Brutschy, B. ChemPhysChem 2011, DOI: 10.1002/cphc.201000985. (16) Page, R. H.; Shen, Y. R.; Lee, Y. T. J. Chem. Phys. 1988, 88, 4621.
ARTICLE
(17) Riehn, C.; Lahmann, C.; Wassermann, B.; Brutschy, B. Chem. Phys. Lett. 1992, 197, 443. (18) Nosenko, Y.; Kunitski, M.; Thummel, R. P.; Kyrychenko, A.; Herbich, J.; Waluk, J.; Riehn, C.; Brutschy, B. J. Am. Chem. Soc. 2006, 128, 10000. (19) Casaes, R. N.; Paul, J. B.; McLaughlin, R. P.; Saykally, R. J.; van Mourik, T. J. Phys. Chem. A 2004, 108, 10989. (20) Choi, M. Y.; Miller, R. E. Phys. Chem. Chem. Phys. 2005, 7, 3565. (21) Chandra, A. K.; Nguyen, M. T.; Zeegers-Huyskens, T. J. Phys. Chem. A 1998, 102, 6010. (22) van Mourik, T.; Benoit, D. M.; Price, S. L.; Clary, D. C. Phys. Chem. Chem. Phys. 2000, 2, 1281. (23) Rejnek, J.; Hanus, M.; Labelac, M.; Ryjacek, F.; Hobza, P. Phys. Chem. Chem. Phys. 2005, 7, 2006. (24) Kabelac, M.; Hobza, P. J. Phys. Chem. B 2001, 105, 5804. (25) Nosenko, Y.; Kunitski, M.; Riehn, C.; Harbach, P. H. P.; Dreuw, A.; Brutschy, B. Phys. Chem. Chem. Phys. 2010, 12, 863. (26) Chin, W.; Mons, M.; Piuzzi, F.; Tardivel, B.; Dimicoli, I.; Gorb, L.; Leszczynski, J. J. Phys. Chem. A 2004, 108, 8237. (27) Crews, B.; Abo-Riziq, A.; Grace, L.; Callahan, M.; Kabelac, M.; Hobza, P.; de Vries, M. S. Phys. Chem. Chem. Phys. 2005, 7, 3015. (28) Even, U.; Jortner, J.; Noy, D.; Lavie, N.; Cossart-Magos, C. J. Chem. Phys. 2000, 112, 8068. (29) Nosenko, Y.; Kunitski, M.; Riehn, C.; Thummel, R. P.; Kyrychenko, A.; Herbich, J.; Waluk, J.; Brutschy, B. J. Phys. Chem. A 2008, 112, 1150. (30) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (31) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (32) Wiosna-Salyga, G.; Nosenko, Y.; Kijak, M.; Thummel, R. P.; Brutschy, B.; Waluk, J. J. Phys. Chem. A 2010, 114, 3270. (33) Perdew, J. P.; Wang, Y. Phys. Rev. B 1992, 45, 13244. (34) Frey, J. A.; Muller, A.; Frey, H. M.; Leutwyler, S. J. Chem. Phys. 2004, 121, 8237. (35) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553. (36) Colarusso, P.; Zhang, K. Q.; Guo, B. J.; Bernath, P. F. Chem. Phys. Lett. 1997, 269, 39. (37) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. J. A.; 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.; Bakken, V.; 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.; Cui, Q.; Baboul, A. 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 C.02; Gaussian, Inc.: Wallingford, CT, 2004. (38) Matsuda, Y.; Ebata, T.; Mikami, N. J. Chem. Phys. 1999, 110, 8397. (39) Florio, G. M.; Gruenloh, C. J.; Quimpo, R. C.; Zwier, T. S. J. Chem. Phys. 2000, 113, 11143. (40) Morzyk-Ociepa, B.; Nowak, M. J.; Michalska, D. Spectrochim. Acta, Part A 2004, 60, 2113. (41) Plutzer, C.; Hunig, I.; Kleinermanns, K.; Nir, E.; de Vries, M. S. ChemPhysChem 2003, 4, 838. (42) Choi, M. Y.; Miller, R. E. J. Phys. Chem. A 2007, 111, 2475. (43) McKean, D. C. Spectrochim. Acta, Part A 1973, A 29, 1559. (44) Biemann, L.; Haber, T.; Kleinermanns, K. J. Chem. Phys. 2009, 130, 5. (45) Hoogsteen, K. Acta Crystallogr. 1963, 16, 28. 9438
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
The Journal of Physical Chemistry A
ARTICLE
(46) Lewis, T. P.; Miles, H. T.; Becker, E. D. J. Phys. Chem. 1984, 88, 3253. (47) Brutschy, B. Chem. Rev. 2000, 100, 3891. (48) Etinski, M.; Marian, C. M. Phys. Chem. Chem. Phys. 2010, 12, 4915. (49) Perun, S.; Sobolewski, A. L.; Domcke, W. J. Phys. Chem. A 2006, 110, 9031.
9439
dx.doi.org/10.1021/jp111373t |J. Phys. Chem. A 2011, 115, 9429–9439
