Photophysical Characters of Rationally Designed Hetero-Ring

Jul 31, 2008 - For a more comprehensive list of citations to this article, users are ... Absorption and Fluorescence Emission Spectroscopic Characters...
0 downloads 0 Views 2MB Size
Download PDF
J. Phys. Chem. B 2008, 112, 10723–10731

10723

Photophysical Characters of Rationally Designed Hetero-Ring-Expanded Guanine Analogues and Effect of Cytosine Pairing Laibin Zhang and Yuxiang Bu* The Center for Modeling & Simulation Chemistry, Institute of Theoretical Chemistry, Shandong UniVersity, Jinan, 250100, P. R. China ReceiVed: March 24, 2008; ReVised Manuscript ReceiVed: June 11, 2008

We present the results of the CIS and TDB3LYP calculations of the optical absorption and emission spectra of some newly designed guanine (G) analogues and their Watson-Crick base pairs. Compared with natural G, the onset absorption peaks of these newly designed analogues are red-shifted, while the fluorescence peaks are blue-shifted. In general, the first excited singlet states (ππ*) of these analogues are nonplanar for all bases considered here. But, the Stokes shifts for the designed G-analogues are much smaller than that of natural G, suggesting that they have stronger molecular rigidity and higher fluorescence quantum yields than those of natural G. The first excited states of these Watson-Crick base pairs essentially originate from those of their isolated purine moieties, as demonstrated from the S1 geometries of their Watson-Crick base pairs. For G and its analogues, A1 and A2 (they are ring-expanded with one-bond intercalation at the C5 site), the pairing with cytosine reduces the oscillator strengths of both the first absorption peak (by 27%-60%) and the fluorescent emission (by 19%-23%), while for the analogues A3, A4, and xG in which G is ring-expanded with a two-bond intercalation at the C5 site, the pairing, in contrast, increases the oscillator strengths of both the first absorption peak (by 11%-15%) and the fluorescent emission (by 3%-20%). These observations indicate that the pairing with cytosine can quench the fluorescence for G, A1, and A2 but enhance the fluorescence quantum yields for A3, A4, and xG. The significant shifts induced by ring-expansion in the ring-expanded G with a two-bond intercalation at the C5 site reveal a possibility for their fluorescent detections. 1. Introduction Fluorescence spectroscopic techniques have been widely used in nucleic acid research to study structures and dynamics as well as the kinetics of interactions between DNA/RNA and other molecules. Fluorescent probes enable researchers to detect particular components of complex biomolecular assemblies, including live cells, with exquisite sensitivity and selectivity. As we all know, natural nucleobases display extremely low fluorescent quantum yields and ultrashort excited-state lifetimes in both solution and gas phases. The ability of DNA and RNA to absorb ultraviolet light without genetic damage and fluorescence is a vital property for life exposed to sunlight. Thus, in order to more easily probe DNA strand conformational dynamics with spectroscopic techniques, the creation of fluorescent DNA base analogues is very important. These nucleobase analogues must have at least two features: (a) their structures and functionalities must be similar to the natural nucleobases so that they can mimic many of their properties; (b) they must have sufficiently high fluorescence quantum yields when incorporated into DNA/RNA. Environment-sensitive fluorescent analogues of the natural nucleobases have been used extensively to study conformational changes of both RNA and DNA.1 A number of nucleobase analogue probes are now available, some even commercially, for incorporation into oligonucleotides for biophysical and biochemical studies. For example, the adenine analogue 2-aminopurine (2AP) has been used as a site-specific probe of nucleic acid structure and dynamics2–5 because it can form base pairs with thymine in Watson-Crick (WC) geometry2 or with cytosine in a wobble configuration.5,6 The cytosine * Corresponding author. E-mail: [email protected].

analogue, pyrrolocytosine, whose fluorescence yield is sensitive to the local structure of the biomolecule when incorporated into a nucleic acid, is used as a fluorescent probe in experimental studies of nucleic acid dynamics.7 Hawkins and colleagues have described the synthesis of several pteridine base analogues,8 some of which are highly fluorescent with ΦF ranging from 0.03 to 0.88. Clearly, these investigations have provided some useful information for the creation and design of unnatural nucleobases. Recently, creating unnatural nucleobases has gained increasing attention, primarily motivated by their potential applications in biotechnology and medicinal chemistry, as well as in material science. For example, Kool’s group9 synthesized a new class of size-expanded analogues of adenine (A), guanine (G), thymine (T), and cytosine (C) (named as xA, xG, xT, and xC, respectively) in which the fusion of a benzo ring increases their size by 2.4 Å. These x-bases were found to be able to form stable DNA-like structures and to have smaller energy gaps between the highest occupied molecular orbitals (HOMO) and the lowest unoccupied molecular orbitals (LUMO) than the natural ones, suggesting that they could enlarge the genetic alphabet and also have potential application in biotechnologies.10,11 Inspired by Kool’s benzo-expansion scheme, our group has designed a series of new size-expanded G-analogues that are hetero-(5/6)-membered and ring-expanded.12 We have demonstrated that all these newly designed G-analogues can also form the stable DNA-like structures and that the HOMO-LUMO energy gaps for most of them and their WC base pairs are considerably lower than those of the corresponding natural base and base pairs, suggesting that they may be considered as DNA genetic motifs and that they may serve as building blocks for

10.1021/jp802556a CCC: $40.75  2008 American Chemical Society Published on Web 07/31/2008

10724 J. Phys. Chem. B, Vol. 112, No. 34, 2008

Zhang and Bu

Figure 1. Optimized geometries of guanine and its five analogues at the MP2/6-311G(d,p) level with Cs symmetry. The dipole moments (µ, Debye) and vertical ionization potentials (Ip, eV) are also given. The midpoints of the C4-C5 axis are used as references for all angles for dipole moments, and the angles of the dipoles relative to C4-C5 line are given. The ionization potentials have been reported before at the B3LYP/ 6-311+G(d) level.12

potential biological applications and the development of molecular electronic devices. In addition to their modified electronic properties, the increased sizes of the expanded nucleobases give them greater π-conjugation, potentially yielding useful optical properties. Experimental studies have revealed that x-bases are brightly fluorescent compared with the natural DNA nucleobases.9 Consequentially, a question we must face is whether these newly designed G-analogues can be fluorescent-active as those x-bases. Clearly, the fluorescent-activity of these base analogues presents a potential basis for designing fluorescent probes in nucleic acid research. In this work, to understand the electronic spectroscopic properties of these newly designed G analogues, their photophysical characters are explored using ab initio calculations, and the pairing effect with their complementary base, cytosine, on both the absorption and emission processes were examined as well. 2. Computational Details Accurate ab initio calculations on electronically excited states still present a major challenge to theoreticians even for relatively small molecules. To date, two ab initio methods are available for accurately modeling the excited states: the multireference configuration interaction method (MRCI) and the complete active space self-consistent-field method (CASSCF) with second order perturbative correction (CASPT2). But these two methods are computationally rather expensive, especially for larger molecules, and they may therefore be used only for a limited number of calculations. In the present work, because of the large number of monomers and base pairs to be considered, the relatively cheaper methods, the configuration interaction singles (CIS13) and the time-depended density functional theory (TDDFT) methods, were used for modeling excited states. These two methods have been successfully applied to studies of excited-state properties of a variety of molecules (e.g., CIS method,14–17 and TDDFT method18–21). All electronic structure calculations were performed with the help of the Gaussian 03 suite of programs.22 The ground-state geometries of G and its five analogues (A1-A4 and xG, Figure

1) were optimized using both the HF and MP2 methods under Cs symmetry. Although vibrational analysis revealed one negative frequency associated with the pyramidalization of the external -NH2 group in the six monomers, the Cs symmetry restrictions were not removed because the planar structures would be expected in the helical environments. It should be mentioned that several previous theoretical studies on G also assumed the geometry to be planar.23 The ground-state structures of the WC hydrogen-bonded dimers between the six bases and their natural counterpart C were optimized using the HF method without symmetry constraints. It should be noted that the B3LYP and MP2 methods are much more suitable with respect to both H-bond interactions and the corresponding bond lengths.24 In this article, the HF method was used to locate the ground-state structures of the WC base pairs because the S1 state structures were optimized at the CIS level, which is at the same level as the HF method. Then comparisons can be made between the ground- and excited-state structures. This strategy has been demonstrated to be suitable for these kinds of systems by several other groups.15–17 The geometries of the isolated monomers and the corresponding WC base pairs in the first singlet ππ* state (lowest-energy) were determined employing the CIS method without symmetry restraints imposed. The 6-311G(d, p) basis set was used in all calculations. The frequency analyses were performed on these stationary points to confirm that the optimized structures were true minima, characterized by positive vibrational frequencies in all cases. Vertical transition energies to the low-lying singlet states were computed for all of the six bases using both the CIS and TDDFT (at the B3LYP25 level) methods, while the vertical transition energy from the optimized ππ* excited-state (S1) to the ground state corresponding to the fluorescence emission was determined using the TDB3LYP method. To determine the influence of the pairing with the corresponding complementary base C, the vertical transition energies associated with both the absorption and emission processes of the WC hydrogen-bonded base pairs were calculated using both the CIS and TDB3LYP methods.

Hetero-Ring-Expanded Guanine Analogues 3. Results and Discussion 3.1. Structural Characters of the Ground State. The groundstate (S0) structures of G and its five analogues (A1-A4, xG) optimized at the MP2/6-311G(d,p) level of theory with Cs symmetry imposed are shown in Figure 1, in which the dipole moments (µ, Debye) and vertical ionization potentials (Ip, eV) are also given. The optimized ground-state geometry of the GC base pair was found to be planar at the HF/6-311G(d, p) level, in line with the previous results obtained by several authors using the RHF method with various basis sets.26 Similarly, the A2C and A4C base pairs also have planar ground-state geometries, while the ground-state geometries of the other base pairs (A1C, A3C, and xGC) optimized under Cs symmetry represent saddle points on the ground-state potential energy surfaces. Relaxation of the Cs symmetry constraint leads to, to some extent, propeller-twisted complexes. The twisting of the base pairs results from the inherent pyramidalization of the amino groups involved in the hydrogen bonding. It should be mentioned that a similar structure of the GC base pair had been previously reported at the MP2/6-31G(d) level by Sobolewski et al.27 The deformation energies defined as the energy differences between the optimized isolated bases (planar symmetry) and those within the optimized framework of the base pairs were calculated to be smaller than 0.80 kcal/mol at the MP2/ 6-311G(d,p) level, suggesting that the geometrical deformation is small in going from individual bases to the base pair complexes. The ground-state electronic properties of these newly designed G-analogues and their corresponding WC base pairs were studied exhaustively in our recent paper.12 An important finding in our previous work is that their HOMO-LUMO energy gaps and ionization potentials (Ip) are considerably lower than that of natural G. Since their HOMO-LUMO energy gaps are smaller compared with that of natural G, their absorption peaks in the low energy region are expected to be red-shifted, which will be shown below. 3.2. Absorption Spectra of Isolated Bases. G has keto and enol tautomers28 in addition to the 7H and 9H tautomers as found in adenine. The excited-state properties of the 7H-keto and 9H-keto tautomers have been extensively investigated because these two forms are predicted to be the lowest in energy in solution.15,23b,29–31 Vertical transition energies, oscillator strengths, and state assignments to the low-lying excited singlet states for the six bases calculated using both the CIS and TDB3LYP methods with the 6-311G(d,p) basis set are given in Table S1 (Supporting Information). Transitions up to the first nπ* state have been included since the nπ* state is thought to be vitally important for ultrafast internal conversion.31 It is well known that the energies of the ππ* states predicted by the CIS method are too large; therefore, all energies obtained using the CIS method in this work have been scaled by a factor of 0.72, as recommended by Broo and Holme´n.14 Both the CIS and TDB3LYP methods predict the first excited state (S1) of these base monomers to be a ππ* state. This electronic excitation is dominated by the configuration HOMO (H) f LUMO (L). As mentioned above, all the newly designed analogues have smaller HOMO-LUMO gaps;12 therfore, the absorption peak of the analogue corresponding to this transition is red-shifted compared with that of natural G. As shown in Figure S1 (Supporting Information), with the Kohn-Sham HOMO-LUMO gaps decreasing from G to xG, the excitation energy corresponding to the first excited state is reduced from 5.02 to 3.95 eV, indicating that the absorption wavelength is red-shifted.

J. Phys. Chem. B, Vol. 112, No. 34, 2008 10725 For G, the CIS method predicts that the three lowest excited singlet states (S1, S2, and S3) are ππ*, nπ*, and ππ*, respectively, while Shukla et al. once reported that, using the same method, the three lowest excited singlet states of G with planar symmetry in the gas phase are of ππ* (S1), nπ* (S2), and nπ* (S3) characters.26a The disagreement between them may be attributed to the use of different methods and basis sets for locating the ground-state geometry because the photophysical properties depend sensitively on the ground-state geometry used in the calculations. However, the TDB3LYP method predicts that there are two ππ* states (S1 and S3) before the first nπ* state (S4) with the second one being more intense and that the vertical transition energies are quite close to the previous TDDFT data.30 Similarly, results obtained at a higher level of theory (CASPT2) also predicted that there are two ππ* transitions before the first nπ* transition.23b,31a Thus, it seems that the TDDFT method can give more accurate results. In the case of xG, the transition energy of the first excited state is predicted to be 3.95 eV at the TDB3LYP/6-311G(d,p) level (corresponding to 314 nm), in good agreement with the experimental value of free dxG in methanol solution (320 nm).9d In addition, this peak is 67 nm red-shifted compared with that of G (247 nm). Especially, an experiment revealed that there are two stronger absorption peaks following the onset absorption peak and that the third peak is the strongest among them for dxG in methanol solution.9d Here, it can be seen from Table S1 (Supporting Information) that the results obtained by the CIS method reproduced well this behavior for xG, while the TDB3LYP method revealed an intermediate absorption peak (S2, 4.54 eV) between the onset peak and the most intense peak that develops at higher energy (S4, 5.16 eV), which is in agreement with the recent TDDFT investigation by Varsano et al.21 In order to distinguish these base monomers more clearly by using their spectroscopic fingerprints, Figure 2 gives their absorption spectra in the ultraviolet region (200 to 400 nm) with the TDB3LYP results. The UV-vis spectra were calculated with the help of the SWizard program, revision 4.4,32 using the Gaussian model. The half-bandwidths were taken to be equal to 500 cm-1. It should be mentioned that the absorption peaks in Figure 2 are prevalent due to the ππ* transitions (bright states) since the nπ* states and the Rydberg-like states (dark states) have very small oscillator strengths and therefore cannot be detected experimentally. However, the location of the first nπ* state is lined out on the absorption spectrum. As shown in Figure 2, there are two ππ* transitions before the nOπ* state (where nO denotes the O-centered lone pair orbital) for all the bases with an exception of A1, for which there are three ππ* transitions before the nOπ* state, but the third ππ* state and the nOπ* state are nearly degenerated, suggesting that the populated third ππ* state could transfer to the nOπ* state by internal conversion if the coupling of these two states is strong enough. Upon inspection of Figure 2, one can find that these base monomers exhibit different optical behaviors in this ultraviolet region. First, the onset peak is gradually red-shifted from G to xG (the corresponding absorption wavelengths were calculated to be 247, 279, 289, 290, 313, and 314 nm, respectively.), as mentioned above. In addition, A1 is predicted to have the strongest onset absorption peak since it has the largest oscillator strength (0.277) corresponding to this transition. Second, they have different numbers of absorption peaks in this ultraviolet region. In detail, only two absorption peaks are observed for G, but more absorption peaks are observed for A1 (5 peaks), A2 (4 peaks), A3 and A4 (6 peaks), and xG (7

10726 J. Phys. Chem. B, Vol. 112, No. 34, 2008

Figure 2. Absorption spectra of isolated bases in the ultraviolet region between 200 and 400 nm determined from the TDB3LYP results. Locations of the first nOπ* dark states are indicated by a dashed olive line.

peaks), which has the largest number of absorption peaks among these bases. Note that A1 and A3 have the same spacer ring and that the only difference in structures between them is the direction of the 5-membered spacer ring. Clearly, these observations have presented a possibility to distinguish these two pyrrole-expanded G analogues by inspecting their absorption spectra: A3 has more bright absorption peaks than A1 in ultraviolet region 200-400 nm. Third, the strongest absorption peak of G is the second peak (in the direction of increasing the transition energy) with a wavelength of 234 nm. In the cases of A1 or A2, the first absorption peak is also the strongest one with the absorption wavelength of 279 or 289 nm, respectively. As for A3 and xG, the strongest peak is the third one for each, and their corresponding absorption wavelengths were calculated to be 227 and 240 nm, respectively. In the case of A4, the strongest absorption peak is calculated to occur around 208 nm, corresponding to the fifth peak in Figure 2. Finally, the energy gaps between the first two absorption peaks of the newly designed analogues are larger than that of G. In addition, for G and A3, the second peak is more intense than the first one, while for the others, the onset peak is more intense than the second one. The above analyses reveal that insertion of a conjugated spacer ring to G not only red-shifts the first absorption peak but also results in more bright absorption peaks in the ultraviolet region of 200-400 nm. Considering that all the newly designed G-analogues have greater π-conjugation than G, these results are not unexpected. Quantum mechanics teaches us that electrons confined to bigger boxes have more closely spaced energy levels than those in smaller boxes. Thus, when we

Zhang and Bu compare, for example, the ππ* transitions of a series of conjugated molecules, we find that the maximum absorption wavelength increases (red shift) and more absorption peaks appear with the size of the π system. For the bases considered in this work, clearly, the excitation maxima of these newly designed analogues are red-shifted, and their absorption peaks are more than those in natural G. Furthermore, wavelengths of the excitation maxima of the bases expanded with 6-membered spacer ring (A4 and xG) are longer than those expanded with 5-membered spacer ring. Compared with A1, A2 could be viewed as the substitution product of N-H with oxygen in the 5-membered spacer ring. The oxygen substitution changes the molecular orbital energy levels, which in turn results in the shifts of the absorption peaks (Figure 2). For instance, HOMO and LUMO levels of A2 are both decreased compared with those of A1 but LUMO was decreased with a larger magnitude than HOMO; thus, the HOMO-LUMO gap becomes smaller after substitution, which in turn leads to the red shift of the onset absorption peak. But as a whole, there is one-to-one correspondence of the absorption peaks between A1 and A2. Similarly, substitution of nitrogen with C-H in the 6-membered spacer ring of A4 also results in the shifts of the absorption peaks. In summary, the increased sizes of the expanded nucleobases give them greater π-conjugation, which in turn results in some new characters of the electronic absorption spectra that have been discussed above. 3.3. Structural Characters and Emission Spectra of the S1 Excited State of Base Monomers. The equilibrium geometries of the lowest energy singlet ππ* state (S1) of six base monomers optimized at the CIS/6-311G(d,p) level were shown in Figure 3, together with the corresponding side-viewed structures. The fully optimized structures, characterized by all positive frequencies, have different characters. Some selected bond angles and dihedral angles in the excited states, which reflect the nonplanarity of the S1 excited states, are shown in Table S2 (Supporting Information). Geometry of G in the S1 state is highly nonplanar, and such nonplanarity is mainly localized at the six-membered ring zone. The most notable feature of this structure is the significant out-of-plane distortion of the C2 atom and the pyramidalization at the N1 and C5 atoms in the pyrimidine ring. Furthermore, the amino group pyramidalization is also increased in the excited state. The difference between 360° and the summation of bond angles at an atom is a measure of the pyramidalization of that atom (termed as pyramidalization degree). It can be found that for G the pyramidalization degree at the N1 and C5 atoms is 25.7° and 12.0°, respectively. These results are in line with several previous CIS studies.15 As shown in Figure 4, A1, A2 and A3 also have largely distorted S1 geometries, but the pyramidalization of the amino group and that at N1 atom are reduced compared with that of G (Table S2, Supporting Information). Furthermore, for A1 and A3, the S1 excited singlet states also exhibit the pyramidalization at the nitrogen atoms of the fivemembered spacer rings with the pyramidalization degrees of 26.5° and 11.3°, respectively. Different from that of G, the S1 geometries of A4 and xG are nearly planar, and the amino group pyramidalization degree is small, especially for A4, whose amino group is almost planar. Essentially, these different characters for the S1 geometries of A4 and xG from the others can be interpreted by their transition configurations. As mentioned above, the lowest singlet ππ* state is mainly dominated by the H f L configuration at the TDB3LYP/6-311G(d,p) level. The orbitals involved in this transition are shown in Figure 4. It is evident that HOMOs are similar for all the isolated base

Hetero-Ring-Expanded Guanine Analogues

J. Phys. Chem. B, Vol. 112, No. 34, 2008 10727

Figure 3. Lowest singlet ππ* excited-state structures of G and its five analogues obtained at the CIS/6-311G(d,p) level. Bond lengths are given in Å.

Figure 4. Orbitals involved in the lowest transitions of the isolated base monomers obtained at the TDB3LYP/6-311G(d,p) level.

TABLE 1: Absorption Wavelengths, Fluorescent Wavelengths, and Stokes Shifts (S.S) for the Six Base Monomers Calculated at the TDB3LYP/6-311G(d,p) Level, and the Ground- and Excited-State Energy Gaps Calculated at the MP2/6-311G(d,p) Levela structure

absorption/nm

emission/nm

S.S./nm

S0-S1 gap/eV

G A1 A2 A3 A4 xG

239 (0.151) 266 (0.290) 271 (0.275) 271 (0.116) 297 (0.159) 296 (0.077)

452 (0.072) 361 (0.197) 363 (0.215) 319 (0.130) 331 (0.153) 337 (0.080)

213 95 92 48 34 41

1.51 0.46 0.41 0.22 0.11 0.16

a For the absorption process, the ground-state structures (S0) used here are non-planar ones optimized at the HF/6-311G(d,p) level; the oscillator strengths are given in parentheses.

monomers and are the π-type orbital, while LUMOs are of π*type and can be classified into two groups. That is, LUMOs of A4 and xG are similar, but are different from LUMOs of G, A1, A2, and A3, which are similar to each other. The difference in LUMOs appears responsible for the different behaviors in the S1 geometries discussed above. On the basis of the optimized S1 state geometries, the S1 f S0 transition energies corresponding to the fluorescence emission were calculated using the TDB3LYP method, and the results are compiled in Table 1. It is interesting to note that the fluorescent wavelengths of the newly designed G-analogues are blue-shifted compared with that of G and that A3 is expected

to have the strongest emission intensity as it has the largest oscillator strength (0.215). For G, the fluorescence emission would be expected to occur around 452 nm, and the associated oscillator strength is 0.072. It has been reported that the normal fluorescence of G in aqueous solution has a peak near 332 nm and a clear shoulder near 450 nm.33 Shukla et al. reported that the normal fluorescence spectrum of G with a peak at 332 nm would originate from the lowest excited singlet state of 7Hketo G, while the fluorescence of oxygen-rich aqueous G solution near 450 nm would originate mainly from the lowest excited singlet state of the 9H-keto form.26a Their calculations predicted that the fluorescence from the 9H-keto form of G in aqueous solution was expected to peak at 449 nm, which is in agreement with the value predicted here (452 nm) for the fluorescence in vacuum. The fluorescence from A1 to A4 would be expected to occur around 361, 363, 319, and 331 nm, respectively. As for xG, the fluorescence emission would be expected to occur around 337 nm. Fluorescence from free nucleotide dxG in methanol solution was reported to peak at 413 nm,9d which is close to the value predicted here for the fluorescence of xG in vacuum. Note that the experimental value was obtained in solution and that the neglect of the solvent effect in our calculations may account for the observed difference as suggested by previous theoretical studies.34 On the contrary, the neglect of the sugar-phosphate backbone should not be relevant at low energy regions since the sugar and the phosphate contributions to the absorption spectra only start to be important at higher energies. As similarly demonstrated by Hardman and Thompson,16 the sugar and the phosphate groups only have a very slight effect on the absorption and fluorescence properties of 2-aminopurine. Table 1 further shows that the Stokes Shift of G is much larger than those of the newly designed analogues. The Stokes Shift, defined as the difference in wavelength between absorbed and emitted quanta, is a measure of the rigidity of a molecule. When the geometry difference is large between ground and excited states, the Stokes Shift is large. On the contrary, the Stokes Shift is small if the geometry difference is small. Clearly, the smaller Stokes Shifts of these size-expanded G-analogues

10728 J. Phys. Chem. B, Vol. 112, No. 34, 2008 imply the smaller geometric deviations between ground and excited states compared with that of natural G. These observations indicate that all the newly designed analogues have stronger molecular rigidity than G, and thus, the rate constant for the internal conversion (kIC) between S1 and S0 of these analogues are expected to be smaller than that of G. This conclusion can be also drawn by comparing the energy gaps (∆E) between the ground and excited states. As shown in Table 1, the S1-S0 gaps of the newly designed analogues are much smaller compared with that of G. Considering that these analogues have greater π-conjugation and stronger molecular rigidity than G, their fluorescence quantum yields are expected to be higher than that of G. In fact, xG has been experimentally reported to be brightly fluorescent (ΦF ) 0.40) compared with G,9d for which the strong molecular rigidity must be one of the factors. Recently, Chen and Li studied the nonradiative deactivation pathways of excited 9H-guanine and demonstrated that the lowest ππ* excited-state is most likely to decay directly to the ground state through an S1/S0 conical intersection without an nπ* intermediate state.31a As xG is more fluorescence-active than natural G, it can be concluded that if a similar conical intersection links the ππ* state of xG to the ground state, the barrier between the (ππ*)min and the conical intersection must be considerably larger than the 1.7 kcal/mol recommended by Chen and Li for G.31a 3.4. Effect of Cytosine Pairing on Both Absorption and Emission Processes. It is well known that photophysical properties of nucleic acids are very complicated and influenced by a variety of factors. In such systems, base stacking, hydrogen bonding of the WC base pairs, and surrounding factors play a very important role. However, computational resources do not at present allow for a complete quantum mechanical calculation of such complex systems. Therefore, it is a reasonable approach to start from individual bases and move to base pair complexes. In this section, influences of hydrogen bonding due to the cytosine pairing on both the absorption and emission processes are discussed. As we all know, although the TDB3LYP method can be used to accurately predict energies and transition dipole moments for valence states, it grossly under-predicts the energies of charge-transfer states.35 As reported by Thompson et al., when TDB3LYP is used to study the fluorescence of 2-aminopurine and pyrrolocytosine in nucleic acids, it inevitably leads to meaningless low-lying charge-transfer states.16,17 The reasons for this have been fully discussed in a review by Dreuw and Head-Gordon35 and the references therein but are briefly introduced here. It is known that charge-transfer states should be characterized by a long-range Coulombic potential (the electrostatic attraction between positive and negative charges). As this stabilizing Coulombic interaction falls off with 1/R (where R is the distance coordinate between the separated charges of the charge-transfer state), the energy of the chargetransfer state should increase with 1/R, i.e., the potential energy curve of the charge-transfer state should exhibit the correct 1/R asymptote, but the curve calculated using the TDB3LYP method exhibits an asymptotic behavior of 0.2/R because the local exchange-correlation functional does not exhibit the correct long-range behavior.35b Thus, prediction of the low-lying chargetransfer states by the TDB3LYP method should be treated with extreme caution. We also used TDB3LYP to model the excitedstates of the WC base pairs and found that it predicted the first excited-state to be a meaningless charge-transfer state; therefore, we only present the CIS results here. It was found in early studies that the computed and scaled electronic transitions at

Zhang and Bu TABLE 2: Vertical Transition Energies, Oscillator Strengths, and Assignments to Low-Lying Singlet Excited-States of the WC Base Pairs Computed at the CIS/6-311G(d,p) Level WC base pairs states E/eV G

A1

A2

A3

A4

xG

S1 S4 S11 S1 S7 S9 S1 S6 S9 S1 S5 S11 S1 S4 S16 S1 S5 S14

4.63 5.28 5.88 4.33 5.36 5.41 4.23 5.34 5.54 4.08 5.36 5.68 3.83 4.86 5.93 3.82 5.08 5.75

f

assignments

0.113 0.001 0.001 0.336 0.001 0.005 0.306 0.001 0.009 0.458 0.003 0.004 0.446 0.001 0.004 0.175 0.001 0.001

πGπG*b nGbπG*b π G πC * πA1πA1* nA1bπA1* πA1πC* πA2πA2* nA2bπA2* πA2πC* πA3πA3* nA3bπA3* πA3πC* πA4πA4*b nA4bπA4*b πA4πC*c πxGπxG* nxGπxG* πxGπC*

Isolated base monomersa E/eV

f

assignments

4.59(4.63) 0.281(0.279) 4.97(5.04) 0.001(0.001)

ππ* nπ*

4.26(4.30) 0.461(0.458) 5.03(4.93) 0.001(0.001)

ππ* nπ*

4.15(4.19) 0.420(0.415) 5.00(5.07) 0.001(0.004)

ππ* nπ*

4.08(4.09) 0.429(0.446) 5.04(5.10) 0.001(0.001)

ππ* nπ*

3.90(3.91) 0.389(0.378) 4.79(4.82) 0.001(0.001)

ππ* nπ*

3.87(3.84) 0.125(0.123) 4.80(4.86) 0.000(0.000)

ππ* nπ*

a Represent the structures of purine moieties within the optimized geometrical framework of the WC base pairs; values in parentheses correspond to the base monomers optimized at the HF/6-311G** level under planar symmetry restriction. b This orbital has a small distribution on cytosine. c This orbital has a small distribution on the purine moiety; all CIS energies are scaled by a factor of 0.72.

the CIS level are in good agreement with the corresponding experimental data.14,16,17,26a To predict the extent of effect originating from the base pairing on the excitations of individual bases, vertical excitation energies of the purine constituents within the optimized geometrical framework of the respective base pairs were also determined, as listed in Table 2. The relevant low-lying transitions up to the charge-transfer state (excited from HOMO of purine to LUMO of cytosine) of the WC base pairs are given in Supporting Information. Here, we only analyze the effect of the hydrogen bonding on the first ππ* and nπ* states of the purine moieties since these two types of transitions are very important in nucleic acids, as focused on in a lot of theoretical and experimental works. It should be mentioned that the nth transition of the WC base pairs may not necessarily correspond to the nth state of the isolated monomers. As shown in Table 2, the first excited state for all the hydrogen bonded complexes is of ππ* character and is essentially identical to the first transition of the isolated base monomers, and therefore can be classified as the local excitation of the purine moiety. Varsano et al. also reported that, for xGC base pair, the first excited state is of ππ* nature and is clearly a local excitation of xG.21 An inspection of Table 2 reveals that pairing with C does not yield significant effect on the transition energies of the first ππ* state with the changes within 0.08 eV. In detail, for G, A1, and A2, the C pairing slightly blue-shifts the first ππ* states (by 0.04-0.08 eV) relative to their corresponding isolated bases within the optimized geometrical framework of the respective base pairs. However, for the other bases, the energies of the first ππ* states are slightly red-shifted (by 0.0-0.07 eV) after pairing. There is good consensus for all the bases that the nπ* transition energies are considerably increased upon base pairing. It is interesting to note that the C pairing yields different effects on the oscillator strengths corresponding to the first ππ* states of these base monomers. For G, the oscillator strength of the first ππ* state is reduced by 60% upon pairing. Similarly, the C pairing can

Hetero-Ring-Expanded Guanine Analogues

J. Phys. Chem. B, Vol. 112, No. 34, 2008 10729

Figure 5. Lowest singlet ππ* (S1) excited-state structures of the WC base pairs optimized at the CIS/6-311G(d,p) level. Bond lengths are given in Å.

reduce the oscillator strengths corresponding to the first ππ* states for A1 and A2 by 53% and 27%, respectively, but increase the oscillator strengths of the first ππ* states by 15% and 11% for A3 and A4, respectively, compared with those of the isolated base monomers. In particular, as for xG, the pairing with C can red-shift the first ππ* state by 0.05 eV and increase the corresponding oscillator strength by 11%, in agreement with the recent theoretical results reported by Varsano et al.21 The charge-transfer state excited from HOMO of purine to LUMO of C is also included in Table 2 as this type of transition plays an important role in the radiationless process for the WC DNA base pairs. Ab initio calculations27,36–39 and experiments40,41 have revealed that for the WC base pairs excited-state coupled electron-proton transfer can take place over a shallow potential barrier. In the vicinity of the charge-transfer state, a conical intersection with the ground state could be located. For the GC base pair, this mechanism arises via a G f C ππ* chargetransfer state (excitation from HOMO of G to LUMO of C). After photoexcitation and charge transfer, a proton is spontaneously transferred along the central hydrogen bond from G to C, driven by charge compensation. After crossing the conical intersection of this charge transfer state with the ground state, the GC WC base pair recovers to its original structure by electron-proton back transfers.27,36 This mechanism was proposed to account for, in part, the photostability of the WC base pairs. As shown in Table 2, the charge transfer state is determined to be 1.25 eV above the local excited ππ* state for the GC base pair. In the case of A1C and A2C base pairs, the corresponding energy gaps are 1.08 and 1.31 eV, respectively, close to that of the GC base pair, while those for A3C, A4C, and xGC base pairs are much larger than that of the GC base pair (1.60, 2.1, and 1.93 eV versus 1.25 eV). The lifetime of the local excited ππ* state should sensitively depend on this energy gap,37,40 and thus, the lifetimes of the local excited ππ* states of A1C and A2C are expected to be close to that of the GC base pair, while the lifetimes of A3C, A4C, and xGC are

expected to be longer than that of GC since the energy gaps between the charge transfer state and the first local excited ππ* state are much larger than that of GC base pair. On the basis of the strength changes and the energy gaps caused by the C pairing, the six bases considered here can be classified into two groups. The first group includes G, A1, and A2, for which the oscillator strengths of the first transitions are reduced by 27%-60% after the C pairing, and the yielded energy gaps between the first local excited ππ* state and the charge transfer state are close to each other. The second group includes the remainder (A3, A4, xG), for which the C pairing can cause the oscillator strengths of the first transitions to increase by 11%-15%, and the resulted energy gaps between two corresponding states are much larger than that of G. Next, we turn to analyze the effect of the WC hydrogen bonds on the fluorescence of these base monomers by considering the vertical transition from the relaxed geometry of the ππ* state to the ground state. The first (lowest-energy) excited ππ* states of the WC base pairs were fully optimized at the CIS/ 6-311G(d,p) level as shown in Figure 5 and Supporting Information (full geometric details). Figure 5 reveals that the geometric changes induced by electronic excitation are localized mostly on the purine moiety but not on the C moiety, thus supporting the assignment of the first transition to the local excitation of the purine moiety. It should be noted that the amino groups involved in the hydrogen bonding are almost planar and that the pyramidalization degree at the nitrogen atoms of the amino groups are smaller than 1°. For the GC base pair, the nonplanarity originates mainly from the pyramidalization at the N1 and C4 atoms of the G moiety with the pyramidalization degrees of 26.5° and 11.3° at these two atoms, in line with the previous theoretical results reported by Sobolewski and Domcke.27 Similarly, the local electronic transitions on the purine moieties of the A1C, A2C, and A3C base pairs lead to the pyramidalization of the N1 and C4 atoms on the purine fragments. As found in the S1 geometries of the A1 and A3

10730 J. Phys. Chem. B, Vol. 112, No. 34, 2008

Zhang and Bu

TABLE 3: Vertical Transition Energies, Oscillator Strengths, and Assignments Corresponding to the Fluorescence Emission of the WC Base Pairs Calculated at the CIS/6-311G(d,p) Level isolated base monomersa

WC base pairs structures E/eV G A1 A2 A3 A4 xG

3.18 3.02 2.87 3.47 3.45 3.37

f 0.176 0.173 0.248 0.385 0.445 0.227

assignments E/eV πGπG* πA1πA1* πA2πA2*b πA3πA3*b πA4πA4*b πxGπxG*b

3.18 2.97 2.83 3.46 3.52 3.45

f

assignments

0.220 0.225 0.305 0.373 0.396 0.189

ππ* ππ* ππ* ππ* ππ* ππ*

a Represent the structures of the purine constituents within the optimized S1 geometrical framework of the base pairs. b Slightly contaminated by excitation to the virtual orbital of the cytosine unit; all CIS energies are scaled by a factor of 0.72.

base monomers, the locally excited states (S1) of their corresponding WC base pairs are also pyramidal at the nitrogen atoms of the five-membered spacer rings in the purine moieties. The pyramidalization degrees at these two atoms are found to be 25.2° and 11.7°, respectively. The S1 geometries of the A4C and xGC base pairs are almost planar. On the basis of the optimized S1 geometries, the S1 f S0 transition energies corresponding to the fluorescence emission were determined using the CIS method, as compiled in Table 3, together with the vertical excitation energies of the purine constituents within the optimized S1 geometrical frameworks of the base pairs. For G, as occurred in the absorption spectrum, the pairing with C results in a 20% reduction in the oscillator strength for the fluorescent transition. Similarly, for A1 and A2, the oscillator strengths of fluorescence transitions are reduced by 23% and 19%, respectively, and the fluorescence peaks are slightly blue-shifted for these two bases after pairing with C. While the oscillator strengths corresponding to the fluorescence emission of A3, A4, and xG are increased by 3%, 12%, and 20%, respectively, after pairing. As known, the rate constant for the fluorescent process (kF) is proportional to the oscillator strength (f), and the fluorescence quantum yield is defined as

φF )

kF kF + kNR

(1)

where kNR stands for the rate constant of the relevant nonradiative processes. Clearly, if kNR unchanged, the quantum yield φF is determined by kF. Thus, this relationship implies that a reduction of 19%-23% in the oscillator strength corresponding to the fluorescent transition of G, A1, and A2 induced by pairing with C may result in a reduction in their fluorescence quantum yield. But for A3, A4, and xG, the pairing with C increases the oscillator strengths of the fluorescent transitions by 3%-20%, implying an enhancement in the fluorescence quantum yields. It should be mentioned that all the WC base pairs formed by the newly designed analogues with C may exhibit similar photostability to GC base pair according to the proposed excitedstate coupled electron-proton transfer mechanism by Sobolewski and Domcke.27,36 However, their photophysical characters may be different, and a detailed exploration is still underway. 4. Conclusions In this work, both the CIS and TDDFT methods were used to explore the photophysical properties of some newly designed guanine analogues and their corresponding WC base pairs. Some interesting phenomena and characters have been observed. For

the isolated bases, the onset absorption peaks of these newly designed analogues are red-shifted compared with that of natural G, while the fluorescence wavelengths are blue-shifted. The calculated excitation energies are in good qualitative agreement with the measured data, where experimental results are available. In general, the S1 singlet excited states (ππ* states) are nonplanar for these newly designed base analogues, and the corresponding Stokes Shifts are much smaller than that for G, suggesting that these size-expanded bases have stronger molecular rigidity than G. Therefore, their fluorescence quantum yields are expected to be higher than that of natural G. When the effect of the pairing with C is taken into account, these bases can be divided into two groups. The first group includes G, A1, and A2 (one-bond-intercalated at the C5 site), and their parings with C could reduce the oscillator strengths of both the first ππ* transitions by 27%-60% and the fluorescence emissions by 19%-23%. In addition, the energy gaps between the first local excited ππ* state and the charge transfer state are close to each other for their corresponding base pairs. The other three bases (A3, A4, and xG) considered here make up the second group, which are two-bondsintercalated at the C5 sites. The pairing with C increases the oscillator strengths of both the first ππ* transitions by 11%-15% and the fluorescence emissions by 3%-20%, and the corresponding energy gaps are much larger than that of the GC pair. In general, if the pairing with C does not affect the nonradiative process, it can reduce the fluorescence quantum yield for G, A1, and A2, but enhance it for A3, A4, and xG. The interesting photophysical properties of these newly designed G-analogues make them potential candidates for exploring the conformations, functions, and dynamics of RNA and DNA, as well as other biomacromolecules. Acknowledgment. This work was supported by NSFC (20633060, 20573063), NCET, and Virt Laboratory for Comput Chem & SCC of CNIC-CAS, MCBILIN at MSU, and HPCC at SDU. Supporting Information Available: Tables of vertical transition energies, oscillator strengths and state assignments to the low-lying singlet transitions corresponding to the isolated monomers and their WC base pairs; a table of some selected bond angles and dihedral angles in the excited-states reflecting the nonplanarity of the S1 geometries of isolated monomers; Cartesian coordinates of the ground- and excited-state (S1) structures of the monomers and the S1 state structures of WC base pairs; and a figure depicting the HOMO-LUMO gaps and the vertical transition energies of the first transition states for all monomers. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) For more information, see reviews: (a) Manuela, J. R.; Marino, J. P. Curr. Org. Chem. 2002, 6, 775. (b) Hawkins, M. E. Top. Fluoresc. Spectrosc. 2003, 7, 151. (2) (a) Ward, D. C.; Reich, E.; Stryer, L. J. Biol. Chem. 1969, 244, 1228. (b) Nordlund, T. M.; Andersson, S.; Nilsson, L.; Rigler, R.; Gra¨slund, A.; McLaughlin, L. W. Biochemistry 1989, 28, 9095. (c) Nordlund, T. M.; Xu, D.; Evans, K. O. Biochemistry 1993, 32, 12090. (3) (a) Menger, M.; Tuschl, T.; Eckstein, F.; Porschke, D. Biochemistry 1996, 35, 14710. (b) Allan, B. W.; Reich, N. O.; Beechem, J. M. Biochemistry 1999, 38, 5308. (4) (a) Guest, C. R.; Hochstrasser, R. A.; Sowers, L. C.; Millar, D. P. Biochemistry 1991, 30, 3271. (b) Bandwar, R. P.; Patel, S. S. J. Biol. Chem. ¨ jva´ri, A.; Martin, C. T. Biochemistry 1996, 35, 2001, 276, 14075. (c) UA 14574. (d) Jia, Y.; Kumar, A.; Patel, S. S. J. Biol. Chem. 1996, 271, 30451.

Hetero-Ring-Expanded Guanine Analogues (5) (a) Stivers, J. T. Nucleic Acids Res. 1998, 26, 3837. (b) Rachofsky, E. L.; Seibert, E.; Stivers, J. T.; Osman, R.; Ross, J. B. A. Biochemistry 2001, 40, 957. (6) Beechem, J. M.; Otto, M. R.; Bloom, L. B.; Eritja, R.; Reha-Krantz, L. J.; Goodman, M. F. Biochemistry 1998, 37, 10144. (7) (a) Liu, C.; Martin, C. T. J. Mol. Biol. 2001, 308, 465. (b) Liu, C.; Martin, C. T. J. Biol. Chem. 2002, 277, 2725. (8) (a) Hawkins, M. E. Cell Biochem. Biophys. 2001, 34, 257. (b) Hawkins, M. E.; Pfleiderer, W.; Balis, F. M.; Porter, D.; Knutson, J. R. Anal. Biochem. 1997, 244, 86. (c) Hawkins, E.; Pfleiderer, W.; Jungmann, O.; Balis, F. M. Anal. Biochem. 2001, 298, 231. (9) (a) Liu, H.; Gao, J.; Lynch, S. R.; Saito, Y. D.; Maynard, L.; Kool, E. T. Science 2003, 302, 868. (b) Gao, J.; Liu, H.; Kool, E. T. J. Am. Chem. Soc. 2004, 126, 11826. (c) Liu, H.; Gao, J.; Kool, E. T. J. Am. Chem. Soc. 2005, 127, 1396. (d) Liu, H.; Gao, J.; Kool, E. T. J. Org. Chem. 2005, 70, 639. (e) Krueger, A. T.; Kool, E. T. J. Am. Chem. Soc. 2008, 130, 3989. (10) Leconte, A. M.; Romesberg, F. E. Nature 2006, 444, 553. (11) (a) Fuentes-Cabrera, M.; Sumpter, B. G.; Wells, J. C. J. Phys. Chem. B 2005, 109, 21135. (b) Fuentes-Cabrera, M.; Sumpter, B. G.; Lipkowski, P.; Wells, J. C. J. Phys. Chem. B 2006, 110, 6379. (12) Zhang, J. M.; Cukier, R. I.; Bu, Y. X. J. Phys. Chem. B 2007, 111, 8335. (13) Foresman, J. B.; Head-Gordon, M.; Pople, J. A.; Frisch, M. J. J. Phys. Chem. 1992, 96, 135. (14) (a) Broo, A.; Holme´n, A. J. Phys. Chem. A 1997, 101, 3589. (b) Holme´n, A.; Broo, A.; Norden, B. J. Am. Chem. Soc. 1997, 119, 12240. (15) (a) Shukla, M. K.; Leszczynski, J. J. Phys. Chem. A 2002, 106, 4709. (b) Shukla, M. K.; Leszczynski, J. J. Phys. Chem. B 2005, 109, 17333. (16) Hardman, S. J. O.; Thompson, K. C. Biochemistry 2006, 45 (30), 9145. (17) Thompson, K. C.; Miyake, N. J. Phys. Chem. B 2005, 109, 6012. (18) (a) Jodicke, C. J.; Luthi, H.-P. J. Chem. Phys. 2002, 117, 4146. (b) Jodicke, C. J.; Luthi, H.-P. J. Chem. Phys. 2003, 119, 12852. (c) Jodicke, C. J.; Luthi, H.-P. J. Am. Chem. Soc. 2003, 125, 252. (19) (a) Zhao, G. J.; Han, K. L. J. Phys. Chem. A 2007, 111, 2469. (b) Zhao, G. J.; Liu, J. Y.; Zhou, L. C.; Han, K. L. J. Phys. Chem. B 2007, 111, 8940. (c) Zhao, G. J.; Han, K. L. J. Phys. Chem. A 2007, 111, 9218. (20) Kumar, A.; Sevilla, M. D. J. Am. Chem. Soc. 2008, 130 (7), 2130. (21) Varsano, D.; Garbesi, A.; Felice, R. D. J. Phys. Chem. B 2007, 111, 14012. (22) 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.; 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.; Ortz, J. V.; Cui, Q.; Baboul, A. F.;

J. Phys. Chem. B, Vol. 112, No. 34, 2008 10731 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.; Nanyakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision B.03; Gaussian, Inc.: Pittsburgh, PA, 2003. (23) (a) Petke, J. D.; Maggiora, G. M.; Christoffersen, R. F. J. Am. Chem. Soc. 1990, 112, 5452. (b) Fu¨lscher, M. P.; Serrano-Andre´s, L.; Roos, B. O. J. Am. Chem. Soc. 1997, 119, 6168. (24) (a) Fonseca Guerra, C.; Bickelhaupt, F. M.; Baerends, E. J. Cryst. Growth Des. 2002, 2, 239. (b) Sˇponer, J.; Jurecka, P.; Hobza, P. J. Am. Chem. Soc. 2004, 126, 10142. (25) (a) Beck, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B: Condens. Matter. 1988, 37, 785. (26) (a) Shukla, M. K.; Mishra, S. K.; Kumar, A.; Mishra, P. C. J. Comput. Chem. 2000, 21, 826. (b) Florian, J.; Leszczynski, J.; Johnon, B. G. J. Mol. Struct. 1995, 349, 421. (c) Shishkin, O. V.; Sˇponer, J.; Hobza, P. J. Mol. Struct. 1999, 477, 15. (27) Sobolewski, A. L.; Domcke, W. Phys. Chem. Chem. Phys. 2004, 6, 2763. (28) (a) Sabio, M.; Topiol, S.; Lumma, W. C., Jr. J. Phys. Chem. 1990, 94, 1366. (b) Petke, J. D.; Maggiora, G. M.; Christoffersen, R. E. J. Am. Chem. Soc. 1990, 112, 5452. (c) Gorb, L.; Leszczynski, J. J. Am. Chem. Soc. 1998, 120, 50249. (29) Mennucci, B.; Toniolo, A.; Tomasi, J. J. Phys. Chem. A 2001, 105, 7126. (30) Crespo-Herna´ndez; C. E. ; Marai; C. ; Kohler, B. unpublished results. (31) (a) Chen, H.; Li, S. H. J. Chem. Phys. 2006, 124, 154315. (b) Serrano-Andre´s, L.; Mercha´n, M.; Borin, A. C. J. Am. Chem. Soc. 2008, 130, 2473. (32) Gorelsky, S. I. SWizard Program; CCRI, University of Ottawa: Ottawa, Canada, 2008. http://www.sg-chem.net/. (33) Santhosh, C.; Mishra, P. C. J. Mol. Struct. 1989, 198, 327. (34) (a) Mennucci, B.; Toniolo, A.; Tomasi, J. J. Phys. Chem. A 2001, 105, 7126. (b) Mishra, S. K.; Shukla, M. K.; Mishra, P. C. Spectrochim. Acta, Part A 2000, 56, 1355. (35) (a) Dreuw, A.; Head-Gordon, M. Chem. ReV. 2005, 105, 4009. (b) Dreuw, A.; Weisman, J. L.; Head-Gordon, M. J. Chem. Phys. 2003, 119, 2943. (36) Sobolewski, A. L.; Domcke, W. Chem. Phys. 2004, 294, 73. (37) Sobolewski, A. L.; Domcke, W.; Ha¨ttig, C. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 17903. (38) Perun, S.; Sobolewski, A. L.; Domcke, W. J. Phys. Chem. A 2006, 110, 9031. (39) Markwick, P. R. L.; Doltsinis, N. L. J. Chem. Phys. 2007, 126, 175102. (40) Schultz, T.; Samoylova, E.; Tadloff, W.; Hertel, I. V.; Sobolewski, A. L.; Domcke, W. Science 2004, 306, 175. (41) Schwalb, N. K.; Temps, F. J. Am. Chem. Soc. 2007, 129, 30.

JP802556A