MM Study on Mechanistic Photophysics of Alloxazine

Article pubs.acs.org/JPCA

QM/MM Study on Mechanistic Photophysics of Alloxazine Chromophore in Aqueous Solution Xue-Ping Chang, Xiao-Ying Xie, Shi-Yun Lin, and Ganglong Cui* Key Laboratory of Theoretical and Computational Photochemistry, Ministry of Education, College of Chemistry, Beijing Normal University, Beijing 100875, China S Supporting Information *

ABSTRACT: Compared with isoalloxazine, the core chromophore of biologically important flavins, alloxazine exhibits much lower fluorescence quantum yield and larger intersystem-crossing quantum yield. However, its efficient radiationless relaxation pathways are still elusive. In this work, we have used the QM(MS-CASPT2//CASSCF)/MM method to explore the mechanistic photophysics of alloxazine chromophore in aqueous solution. On the basis of the optimized minima, conical intersections, and crossing points in the lowest 1ππ*, 1 nπ*, 3ππ*, and 3nπ* states, we have proposed three energetically possible nonadiabatic relaxation pathways populating the lowest 3ππ* triplet state from the initially populated excited 1ππ* singlet state. The first is the direct 1ππ*→ 3 ππ* intersystem crossing via the 1ππ*/3ππ* crossing point. The second is an indirect 1ππ* → 3ππ* intersystem crossing relayed by the dark 1nπ* singlet state. In this route, the 1ππ* system first decays to the 1nπ* state via the 1ππ*/1nπ* conical intersection, followed by an 1nπ*→ 3ππ* intersystem crossing at the 1nπ*/3ππ* crossing point to arrive at the final 3ππ* state. The third is similar to the second one; but its intersystem crossing is relayed by the 3nπ* triplet state. The 1ππ* system first decays to the 3nπ* state via the 1ππ*/3nπ* crossing point; the generated 3nπ* state is then de-excited to the 3ππ* state through the 3nπ*→ 3ππ* internal conversion at the 3nπ*/3ππ* conical intersection. According to the classical El-Sayed rule, we suggest the second and third paths play a much more important role than the first one in the formation of the lowest 3ππ* state.

■

and polar protic solvents, and in cellulose matrix;19 in the same time, they have explored solvent effects on the spectroscopic properties of lumiflavins and lumichromes and found that the former fluorescence quantum yield is one order larger than the latter due to their lower nonradiative rate constants.20 Recently, they employed time-resolved spectral and photon counting kinetic results to observe excited-state double proton transfer catalyzed by a carboxylic acid molecule hydrogen-bonded with alloxazines.21 In addition to Sikorski group, Penzkofer and coworkers have very recently studied absorption and emission spectroscopic properties of alloxazine and lumichrome in aqueous solution at different pH values.22,23 Moyon and Mitra have explored fluorescence solvatochromism in lumichrome and excited-state tautomerization using steady-state and time-resolved fluorescence spectroscopy and found that the excited-state properties of lumichrome do not correlate with solvent polarity.24 Prukula et al. employed UV−vis spectroscopy and steady-state and time-resolved fluorescence techniques to explore the photophysical properties of lumichrome in four protonation/deprotonation states at different pH values.25 Douhal and co-workers reported photophysical studies of lumichrome in water at different pH values and in human

INTRODUCTION Alloxazines were studied in the last century when the proton transfer reactions in lumichrome (7,8-dimethyl alloxazine) were discovered.1−3 Experimental studies were mainly driven by the fact that they are structurally related to isoalloxazines and particularly to biologically relevant flavin chromophores as a result of excited-state proton transfer. They are focused on steady-state absorption and fluorescence, excited-state proton transfer in organic solvents (e.g., acetic acid, methanol−acetic acid mixtures, etc.), and methyl substitution effects on phototautomerism and photophysical and photochemical properties of alloxazines.4−13 Experimental studies of mechanistic photophysics and photochemistry of alloxazines were gradually intensified from the beginning of this century. In this regard, Sikorski’s group have contributed numerous works. In 2001, they explored the photophysics of lumichrome and its methyl-substituted variant in water solutions and measured fluorescence and triplet-state lifetimes, and quantum yield for the sensitized production of singlet oxygen.14 Later, they studied the spectroscopical and photophysical properties of methyllumichrome and dimethylalloxazines absorbed on cellulose and in water,15−17 and the acid−base properties of alloxazines and its methyl-substituted variants in their ground and first excited singlet states.18 In addition, they have explored singlet- and triplet-state properties of 1-methyllumichrome in a series of nonpolar, polar aprotic, © XXXX American Chemical Society

Received: March 15, 2016 Revised: July 15, 2016

A

DOI: 10.1021/acs.jpca.6b02669 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A serum albumin protein and β-cyclodextrin in neutral aqueous solutions.26 According to these experimental studies, it can be seen that in various solvents, the fluorescence quantum yield of alloxazines/lumichromes is much lower, e.g., 0.009/0.028 in acetonitrile and 0.048/0.055 in water for alloxazines; in comparison, the quantum yield of intersystem crossing is much higher, e.g., 0.36/0.73 in acetonitrile and 0.45/0.71 in water for alloxazines.7,27,28 In comparison with experimental studies, computational studies of alloxazines and lumichromes are rarely reported. Szymusiak et al. employed a semiempirical method to study the electronic structure of some monomethyl derivatives of alloxazines in the ground and excited singlet states (S0 and S1).12 Sikorski’s group exploited the TD-DFT and CIS methods to compute absorption and fluorescence spectra of a series of mono- and dimethyl substituted alloxazines in various solvents and to map the potential energy profiles of excited-state proton transfer in the S0 and S1 states in acetic acid.20,21,27,29,30 Recently, they have used the TD-DFT method to explore triplet states of some dimethylalloxazines and solvent and substitution effects on the ISC rate constants.31 Moyon and Mitra have studied the excited-state tautomerization process of lumichrome forming isoalloxazines using the DFT method.24 It is clear that these theoretical studies were dedicated to local spectroscopical properties and did not focus on the mechanistic photophysics of alloxazines and lumichromes until recently. Salzmann and Marian have employed the TD-DFT and combined density functional multireference configuration interaction (DFT/MRCI) methods to explore the excitedstate properties and deactivation pathways of alloxazine in vacuo and aqueous solution (COSMO model).28

Figure 1. Solvated QM/MM system used in this study. The alloxazine molecule is treated quantum mechanically, whereas the surrounding water molecules are treated by molecular mechanics. Also shown is the atomic labeling of the solute molecule.

is utilized to re-evaluate the energies of all optimized structures. In the MS-CASPT2 computations, a larger active space of 16 electrons in 12 orbitals with two additional lone-pair electrons is used (see Supporting Information); the Cholesky decomposition technique with unbiased auxiliary basis sets is used for accurate two-electron integral evaluation;39 the ionization potential-electron affinity (IPEA) shift is not applied;40 the imaginary shift technique (0.2 au) is employed to avoid intruder state issues.41 Vertical excitation energies at Franck− Condon points are also calculated using the MS-CASPT2 method. The 6-31G* and cc-pVTZ basis sets are employed for geometric optimizations and single-point energy refinements, respectively.42−44 This combined MS-CASPT2//CASSCF computational strategy is recently demonstrated to be able to give an accurate description for excited-state structures and energetics.45−49 All QM/MM calculations are carried out using the MOLCAS8.0 package that interfaces with the TINKER6.3.2 package.50−54

■

COMPUTATIONAL METHODS QM/MM Setup. Initial Cartesian coordinates of alloxazine are prepared manually. To explicitly consider solvation effects, alloxazine is solvated in a spherical water box with a radius of 30 Å. This solvated system is relaxed by sequentially carrying out constrained molecular mechanics (MM) minimizations (2000 steps) and molecular dynamics (2 ps, MD) runs in which water molecules more than 20 Å away from the center of mass of alloxazine chromophore are frozen. In the MM minimizations and MD dynamics runs, alloxazine is described using the CHARMM22 force field32 that is modified based on the isoalloxazine core of flavin chromophores; water molecules are modeled by the TIP3P model.33 All MM computations are carried out using the DL_POLY module of CHEMSHELL3.5 package.34−36 Then, the final MD snapshot that deletes frozen water molecules generates the starting structure for the following QM/MM computations (619 water molecules) in which only the atoms 12 Å away from the center of mass of alloxazine chromophore are frozen in geometry optimizations (see Figure 1). QM/MM Calculations. The state averaged complete active space self-consistent field (CASSCF) method is chosen as QM electronic structure method to optimize minima, conical intersections, and crossing points in the lowest 1ππ*, 1nπ*, 3 ππ*, 3nπ*, and S0 states. In the CASSCF computations, the active space consists of 14 electrons in 10 orbitals, which includes 12 π electrons in nine π and π* orbitals and two lonepair electrons in one n orbital (see Figure S1 in Supporting Information). Since the CASSCF method does not adequately capture dynamic correlation, the multistate complete active space second-order perturbation approach (MS-CASPT2)37,38

■

RESULTS AND DISCUSSION Local Spectroscopic Properties. The geometry of alloxazine chromophore itself in aqueous solution is nearly planar in the S0 state, as shown in Figure 2. At the QM(CASSCF)/MM level, the C2−O11 and C4−O12 bonds are calculated to be 1.191 and 1.185 Å, respectively, which are close to those in vacuo.28 The other bond lengths are shown in Figure 2. Table 1 compiles the vertical excitation energies to the lowest 1 ππ*, 1nπ*, 3ππ*, and 3nπ* electronically excited states at the Franck−Condon point (i.e., S0 minimum) computed by various high-level ab initio methods. Compared with experimental value (3.27 eV), TD-CAM-B3LYP/MM computation overestimates the vertical excitation energy to the lowest 1ππ* excited state more than 0.5 eV; TD-B3LYP/MM, MSCASPT2/6-31G*/MM, and MS-CASPT2/cc-pVTZ/MM computations are closer to the experimental value and previous DFT/MRCI/COSMO result (see Table 1). For the S0 → 1nπ* case, TD-B3LYP/MM and MS-CASPT2/cc-pVTZ/MM results are lower than those by TD-CAM-B3LYP/MM and MSCASPT2/6-31G*/MM methods but close to DFT/MRCI/ COSMO computation by Marian et al.28 For the S0 → 3ππ* vertical excitation, both MS-CASPT2 calculations give a little higher energies than those by TD-DFT and DFT/MRCI/ COSMO methods. However, for the S0 → 3nπ* vertical excitation, TD-DFT and MS-CASPT2/cc-pVTZ/MM computations supply similar data, which are slightly lower than those by MS-CASPT2/6-31G*/MM and DFT/MRCI/COSMO B


Article

The Journal of Physical Chemistry A

Figure 3. MS-CASSCF(16,12)/cc-pVTZ computed molecular orbitals (MOs) involved in the electronic transitions to the two lowest singlet states. For the S0 → 1ππ* vertical excitation, there are two main electronic configurations, i.e., πH → π*L and πH−1 → π*L (top two rows); while for the S0 → 1nπ* excitation only an electronic configuration, i.e., nN+N → π*L . See text for discussion.

Figure 2. CASSCF(14,10)/6-31G*/MM optimized minimum-energy structures in the lowest five electronic states (S0, 1nπ*, 1ππ*, 3nπ*, and 3 ππ*). Also shown are the selected bond lengths. See Supporting Information for their full Cartesian coordinates.

Table 2. MS-CASPT2(16,12)/cc-pVTZ/MM Computed Energies (in kcal/mol) of Four Excited-State Minima Relative to the S0 Minimum; See Figure2 for Schematic Structures

Table 1. Vertical Excitation Energies (in eV) to the Lowest 1 ππ*, 1nπ*, 3ππ*, and 3nπ* Electronic States at the Franck− Condon Point Computed by Different QM/MM Methodsa

ππ* nπ* 3 ππ* 3 nπ* 1 1

B3LYP

CAMB3LYP

MS-CASPT2 /6-31G*

3.41 3.48 2.48 2.88

3.81 3.85 2.45 2.93

3.57 3.90 2.99 3.24

MS-CASPT2 /cc-pVTZb 3.53 3.35 2.94 2.98

(81.4) (77.3) (67.8) (68.7)

ππ*

DFT/ MRCI 3.32 3.67 2.64 3.32

nπ*

1

1

67.7

68.5

ππ*

nπ*

3

3

58.0

61.2

the S0 counterparts is different, as shown in Figure 4. For the ππ* minimum, the largest deviation is from C6−C7 bond length, which is more than 0.1 Å at the CASSCF/6-31G*/MM level; the C8−C9 and C9−C9a bond lengths change more than 0.05 Å. For the 3ππ* minimum, the C6−C7 bond length still deviates the most, but there are more bond lengths with a deviation of more than 0.05 Å, e.g., C7−C8, C8−C9, C9a− N10, and C4a−C10a. Qualitatively different from the situation of 1ππ* and 3ππ*, the bond-length alternation of both 1nπ* and 3 nπ* states is nearly identical to each other, as shown in the right panel of Figure 4. For these two minima, the C4−C4a, C4a−N5, and C9a−N10 bond lengths change most significantly. On the energetic side, the 1ππ* and 3ππ* states remain the lowest excited singlet and triplet electronic states. Between them is the 3nπ* excited state. At the MS-CASPT2/cc-pVTZ/ MM level, the potential energies of the 1ππ*, 1nπ*, 3ππ*, and 3 nπ* minima relative to that of the S0 minimum are computed to be 67.7, 68.5, 58.0, and 61.2 kcal/mol, respectively. At the DFT/MRCI level, Salzmann and Marian estimated them to be 71.9, 78.9, 53.0, and 71.7 kcal/mol, respectively, by approximating solvent effects using spectral shifts computed at the Franck−Condon points.28 Conical Intersections. At the CASSCF/6-31G*/MM level, we have optimized two conical intersections between same spin states, i.e., 1ππ*/1nπ* and 3nπ*/3ππ*, and four crossing points between different spin states, i.e., 1nπ*/3ππ*, 1ππ*/3ππ*, 1 ππ*/3nπ*, and 3ππ*/S0. Figure 5 shows their schematic structures as well as selected bond lengths. Structurally, all intersection structures have an almost planar conformation, except the 3ππ*/S0 crossing point, in which the −C(6)H− group is significantly above the plane. In addition, 1

a

Only the QM methods used in the QM/MM computations are shown. DFT/MRCI values taken from the work of Salzmann and Marian.28. bEnergies in parentheses (kcal/mol).

methods. On the basis of the above comparison, we can find that MS-CASPT2/cc-pVTZ/MM method can provide a more balanced description for the involved four electronically excited states; thereby, MS-CASPT2/cc-pVTZ/MM method is chosen to refine the energies of all optimized structures and reaction paths to be presented below. Analysis of electronic structures shows that the π → π* singlet electronic transition stems from two main electronic configurations of πH → π*L and πH−1 → π*L . Their weights are comparable to each other and computed to be 0.43 and 0.13 at the MS-CASPT2/cc-pVTZ/MM level. Unlike the situation of the π → π* excitation, the n → π* singlet state is primarily from the nN+N → π*L electron configuration (weight: 0.73; see Figure 3). Minima. To explore the 1ππ* excited-state relaxation dynamics, we have employed the CASSCF/6-31G*/MM method to optimize the minimum-energy structures of aqueous alloxazine in the lowest 1ππ*, 1nπ*, 3ππ*, and 3nπ* states. Figure 2 schematically shows the optimized structures together with selected bond lengths; Table 2 collects their relative energies refined by the MS-CASPT2/cc-pVTZ/MM method. It can be easily found that the QM part, i.e., alloxazine chromophore, in these four QM/MM minima have similar planar conformations (see Figure 2). However, the bond-length alternation of the 1ππ*, 1nπ*, 3ππ*, and 3nπ* minima relative to C


Article


Figure 5. Optimized structures of six minimum-energy conical intersections and crossing points. Also shown are selected bond lengths (in Å) optimized at the CASSCF(14,10)/6-31G*/MM level. See Supporting Information for Cartesian coordinates.

populate the lowest triplet state from the initial populated singlet state because of the large intersystem-crossing quantum yield (0.45 in water) and the small fluorescence quantum yield (0.048 in water).7,27 In previous computational work, Salzmann and Marian mainly studied the mechanistic photophysics in vacuo using TD-B3LYP and DFT/MRCI methods.28 The photophysics in aqueous solution is explored using the microhydration scheme as well as the COSMO model. Solvent effects on excited-state alloxazine are simply considered by adding constant spectral shifts calculated at the Franck− Condon points. The shape of each excited state in aqueous solution is hence the same as that in vacuo (see Figures 8−9 in their work).28 In this situation, solvent effects are globally the same for each excited state. Our present calculations accurately consider solvent effects in a more realistic way by means of the QM/MM method. Based on our results in this study, we propose three energetically allowed relaxation pathways to the lowest 3ππ* triplet state from the initially populated 1ππ* electronic state. We have explored the feasibility of these three pathways using the linear interpolation internal coordinate (LIIC) paths (Figures 6−8). The LIIC paths have been extensively used by several groups to estimate the upper limit of the barrier of the related chemical reaction path.55−58 Figure 9 summarizes these three relaxation pathways. The first one is the direct 1ππ* → 3ππ* intersystem crossing. The energy gap of the 1ππ* and 3ππ* states at the Franck− Condon point is calculated to be about 13 kcal/mol at the MSCASPT2/cc-pVTZ/MM level (see Table 1), and it is further decreased to 1.1 kcal/mol as the system relaxes from the Franck−Condon point to its 1ππ* minimum. Figure 6 depicts the MS-CASPT2/cc-pVTZ/MM computed LIIC path connecting the 1ππ* minimum and the 1ππ*/ 3ππ* crossing point. It is easy to find that both 1ππ* and 3ππ* states are close to each other in energy along this path. At the 1ππ*/ 3ππ*

Figure 4. Bond-length difference of the 1ππ*, 1nπ*, 3ππ*, and 3nπ* minima compared with the counterparts of the S0 minimum. The positive and negative values mean they are longer and shorter than those of the S0 minimum, respectively.

we have found that the two CO bond lengths in all these intersection structures are not changed remarkably compared with the counterparts of the S0 minimum. This is consistent with the electronic structure character of the 1nπ* and 3nπ* electronic states because only the lone-pair electrons of the two N atoms are involved in these electronic transitions (Figure 3). All these intersection structures are approachable from energetics viewpoint. At MS-CASPT2/cc-pVTZ/MM level, the potential energies of the 1 ππ*/ 1 nπ*, 3 nπ*/ 3 ππ*, 1 ππ*/3ππ*, 1ππ*/3nπ*, 1nπ*/3ππ*, and 3ππ*/S0 relative to the S0 minimum are 69.1/70.5, 69.4/73.7, 67.6/68.2, 67.7/69.3, 61.7/63.4, and 79.4/76.4 kcal/mol, as shown in Table 3, all of which are lower than or close to the 1ππ* energy of 81.3 kcal/ mol at the Franck−Condon point (see Table 1). However, their importances for the mechanistic photophysics of alloxazine chromophore vary greatly (see below). These intersection structures are for the first time located computationally and provide important mechanistic information for understanding efficient radiationless processes of alloxazines and lumichromes. Paths to the 1ππ* State. Experimentally, it is believed that there should exist efficient excited-state relaxation pathways to D


Article


Table 3. MS-CASPT2(16,12)/cc-pVTZ/MM Computed Energies (in kcal/mol) of Six Intersection Structures Relative to the S0 Minimum; See Figure 5 for Schematic Structures 1

ππ*/1nπ*

1

nπ*/3ππ*

69.1/70.5

69.4/73.7

ππ*/3ππ*

1

ππ*/3nπ*

3

67.6/68.2

67.7/69.3

61.7/63.4

1

nπ*/3ππ*

3

ππ*/S0

79.4/76.4

the 1ππ* system can easily reach to the 1ππ*/1nπ* conical intersection, as demonstrated by the MS-CASPT2/cc-pVTZ/ MM computed LIIC path. As shown in the left panel of Figure 7, the energy of the 1nπ* state is reduced significantly along this path, while the energy of the 1ππ* state is merely increased to ca. 70 kcal/mol from 67.7 kcal/mol at the 1ππ* minimum; thus, the 1ππ*/1nπ* conical intersection is reachable along this path from energetics viewpoint. Once hopping to the dark 1nπ* state in the vicinity of the 1ππ*/1nπ* conical intersection, the system will relax to its 1nπ* minimum. Interestingly, near this region, both 1nπ* and 3ππ* are close to each other in energy, within 2 kcal/mol at the MS-CASPT2/cc-pVTZ/MM level (as evidenced by the right panel of Figure 7); therefore, at this crossing point, the lowest 3ππ* state is populated via the 1 nπ*/3ππ* crossing point. This nonadiabatic transition path is more efficient than the first path discussed above because the involved 1nπ* → 3ππ* intersystem crossing at the 1nπ*/3ππ* crossing point is allowed in terms of the El-Sayed rule (vide supra). The third one is similar to the second one and also an indirect 1ππ* → 3ππ* intersystem crossing; however, the dark 3 nπ* triplet state replaces the 1nπ* state as a relay excited state. The energy gap of the 1ππ* and 3nπ* states at the Franck− Condon point is calculated to be about 12 kcal/mol at the MSCASPT2/cc-pVTZ/MM level in Table 1; but it is decreased to 1.6 kcal/mol at the 1ππ* minimum. As indicated by the MSCASPT2/cc-pVTZ/MM computed LIIC path that connects the 1ππ* and 3nπ* minima, there turns out a 1ππ*/3nπ* crossing point along this relaxation path. Importantly, approaching this crossing point is nearly barrierless along this path (see the left panel of Figure 8). Upon hopping to the 3nπ* state, the system can further jump down to the lowest 3ππ* state via the subsequent 3nπ* → 3ππ* internal conversion via the 3nπ*/3ππ* conical intersection, which can be easily accessed structurally and energetically on the basis of the MS-CASPT2/cc-pVTZ/MM computed LIIC path that con-

Figure 6. MS-CASPT2(16,12)/cc-pVTZ/MM computed linear interpolation internal coordinate (LIIC) paths connecting the 1ππ* minimum and the 1ππ*/3ππ* crossing point. See text for discussion.

crossing point, the energy gap is finally reduced to 0.6 kcal/mol at the MS-CASPT2/cc-pVTZ/MM level. Even though this direct 1ππ*→ 3ππ* intersystem crossing is energetically and structurally allowed, this nonadiabatic path should play a minor role for the population of the lowest 3ππ* state concerning the classical El-Sayed rule, which says “The rate of intersystem crossing is relatively large if the radiationless transition involves a change of molecular orbital type, for example, a 1ππ* singlet could transition to a 3nπ* triplet state, but not to a 3ππ* triplet state and vice versa.”59,60 The second one is an indirect 1ππ* → 3ππ* intersystem crossing relayed by the dark 1nπ* singlet state. The energy gap of the 1ππ* and 1nπ* states at the Franck−Condon point is calculated to be about 4 kcal/mol at the MS-CASPT2/ccpVTZ/MM level in Table 1. After the system runs into its 1ππ* minimum, this gap is increased to ca. 12 kcal/mol. Fortunately,

Figure 7. MS-CASPT2(16,12)/cc-pVTZ/MM computed linear interpolation internal coordinate (LIIC) paths (left) connecting the 1ππ* minimum and the 1ππ*/1nπ* conical intersection and (right) connecting the 1nπ* minimum and the 3ππ* minimum. See text for discussion. E


Article


Figure 8. MS-CASPT2(16,12)/cc-pVTZ/MM computed linear interpolation internal coordinate (LIIC) paths (left) connecting the 1ππ* minimum and the 3nπ* minimum and (right) connecting the 3nπ* minimum and the 3nπ*/3ππ* conical intersection. See text for discussion.

are expected to be dominant) responsible for the formation of the lowest triplet state from the initially populated excited singlet state. The current work provides valuable mechanistic insights for understanding the photophysics of alloxazines, lumichromes, and similar variants.23,31,61,62

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.6b02669. Active orbitals in CASSCF computations, additional figures, and Cartesian coordinates of all optimized structures (PDF)

■

Figure 9. Suggested three nonadiabatic relaxation channels populating the lowest 3ππ* triplet state from the initially populated 1ππ* excited singlet state. See text for discussion.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

nects the 3 nπ* minimum and the 3 nπ*/ 3 ππ* conical intersection, as shown in the right panel of Figure 8. The third nonadiabatic transition path is as much efficient as the second one because the involved 1ππ*→ 3nπ* intersystem crossing at the 1ππ*/3nπ* crossing point is as well allowed in terms of the El-Sayed rule (vide supra). Finally, it should be noted that, although there exists a 3ππ*/ S0 crossing point capable of decaying the 3ππ* state to the S0 state, its energy relative to the 3ππ* minimum is too high so that the system cannot reach to this crossing point efficiently and will thus stay in the lowest 3ππ* state for a relatively long time (see Supporting Information). This viewpoint is consistent with experiments available and previous theoretical studies.23,28,31,61

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Natural Science Foundation of China (21522302 and 21520102005); G.C. is also grateful for financial support from the “Recruitment Program of Global Youth Experts” and “Fundamental Research Funds for Central Universities”.

■

REFERENCES

(1) Koziol, J. Absorption Spectra of Riboflavin Lumiflavin and Lumichrome in Organic Solvents. Experientia 1965, 21, 189−190. (2) Koziol, J. The Solvent Effect on the Fluorescence and Light Absorption of Riboflavin and Lumiflavin. Biochim. Biophys. Acta, Biophys. Incl. Photosynth. 1965, 102, 289−300. (3) Koziol, J. Studies on Flavins in Organic Solvents-I. Spectral Characteristics of Riboflavin, Riboflavin Tetrabutyrate and Lumichrome. Photochem. Photobiol. 1966, 5, 41−54. (4) Dekker, R. H.; Srinivas, B.-N.; Huber, J. R.; Weiss, K. Photochemistry of Flavins 0.1. Conventional and Laser FlashPhotolysis Study of Alloxazine. Photochem. Photobiol. 1973, 18, 457− 466. (5) Song, P. S.; Sun, M.; Koziolowa, A.; Koziol, J. Phototautomerism of Lumichromes and Alloxazines. J. Am. Chem. Soc. 1974, 96, 4319− 4323.

■

CONCLUSION We have employed the QM/MM method to explore for the first time the photophysics of alloxazine chromophore in aqueous solution. At the QM(CASSCF)/MM level, we have optimized minima, conical intersections, and crossing points in the lowest 1ππ*, 1nπ*, 3ππ*, and 3nπ* states, whose energies are further refined by the QM(MS-CASPT2)/MM method. Finally, based on the present computational results, we have proposed three nonadiabatic relaxation pathways (two of which F


Article

The Journal of Physical Chemistry A (6) Fugate, R. D.; Song, P. S. Lifetime Study of Phototautomerism of Alloxazine and Lumichromes. Photochem. Photobiol. 1976, 24, 479− 481. (7) Grodowski, M. S.; Veyret, B.; Weiss, K. Photochemistry of Flavins 0.2. Photophysical Properties of Alloxazines and Isoalloxazines. Photochem. Photobiol. 1977, 26, 341−352. (8) Koziolowa, A. Solvent and Methyl Substituent Effect on Phototautomerism and Ionization of Alloxazines. Photochem. Photobiol. 1979, 29, 459−471. (9) Choi, J. D.; Fugate, R. D.; Song, P. S. Nanosecond TimeResolved Fluorescence of Phototautomeric Lumichrome. J. Am. Chem. Soc. 1980, 102, 5293−5297. (10) Komasa, J.; Rychlewski, J.; Koziol, J. Electronic-Structure of Alloxazine and Its Methyl-Derivatives. J. Mol. Struct.: THEOCHEM 1988, 47, 205−212. (11) Koziolowa, A.; Visser, A. J. W. G.; Koziol, J. Phototautomerism of Lumichrome in Methanol Acetic-Acid Mixtures-a Steady-State and Time-Resolved Fluorescence Study. Photochem. Photobiol. 1988, 48, 7−11. (12) Szymusiak, H.; Konarski, J.; Koziol, J. An INDO/S MO Study of Alloxazine and Its Monomethyl Derivatives. J. Chem. Soc., Perkin Trans. 2 1990, 229−236. (13) Sikorska, E.; Koziolowa, A. Excited State Proton Transfer of Methyl- and Cyano-Substituted Alloxazines in the Presence of Acetic Acid. J. Photochem. Photobiol., A 1996, 95, 215−221. (14) Sikorski, M.; Sikorska, E.; Koziolowa, A.; Moreno, R. G.; Bourdelande, J. L.; Steer, R. P.; Wilkinson, F. Photophysical Properties of Lumichromes in Water. J. Photochem. Photobiol., B 2001, 60, 114− 119. (15) Sikorski, M.; Sikorska, E.; Moreno, R. G.; Bourdelande, J. L.; Worrall, D. R. Photophysics of Methyl Substituted Alloxazines in Water: Efficiency of Singlet Oxygen Generation. J. Photochem. Photobiol., A 2002, 149, 39−44. (16) Sikorski, M. Photophysics of Dimethyl-Alloxazines and 1Methyllumichrome on Cellulose. Phys. Chem. Chem. Phys. 2002, 4, 211−215. (17) Sikorski, M.; Sikorska, E.; Khmelinskii, I. V.; Gonzalez-Moreno, R.; Bourdelande, J. L.; Siemiarczuk, A. Photophysics of Alloxazines on Cellulose. Photoch. Photobio. Sci. 2002, 1, 715−720. (18) Sikorska, E.; Szymusiak, H.; Khmelinskii, I. V.; Koziolowa, A.; Spanget-Larsen, J.; Sikorski, M. Spectroscopy and Photophysics of Alloxazines Studied in Their Ground and First Excited Singlet States. J. Photochem. Photobiol., A 2003, 158, 45−53. (19) Sikorska, E.; Khemlinskii, I. V.; Worrall, D. R.; Williams, S. L.; Gonzalez-Moreno, R.; Bourdelande, J. L.; Koput, J.; Sikorski, M. Photophysics of 1-Methyllumichrome. J. Photochem. Photobiol., A 2004, 162, 193−201. (20) Sikorska, E.; Khmelinskii, I. V.; Prukala, W.; Williams, S. L.; Patel, M.; Worrall, D. R.; Bourdelande, J. L.; Koput, J.; Sikorski, M. Spectroscopy and Photophysics of Lumiflavins and Lumichromes. J. Phys. Chem. A 2004, 108, 1501−1508. (21) Sikorska, E.; Khmelinskii, I.; Hoffmann, M.; Machado, I. F.; Ferreira, L. F. V.; Dobek, K.; Karolczak, J.; Krawczyk, A.; Insinska-Rak, M.; Sikorski, M. Ground- and Excited-State Double Proton Transfer in Lumichrome/Acetic Acid System: Theoretical and Experimental Approach. J. Phys. Chem. A 2005, 109, 11707−11714. (22) Tyagi, A.; Penzkofer, A. Absorption and Emission Spectroscopic Characterization of Lumichrome in Aqueous Solutions. Photochem. Photobiol. 2011, 87, 524−533. (23) Penzkofer, A. Absorption and Emission Spectroscopic Investigation of Alloxazine in Aqueous Solutions and Comparison with Lumichrome. J. Photochem. Photobiol., A 2016, 314, 114−124. (24) Moyon, N. S.; Mitra, S. Fluorescence Solvatochromism in Lumichrome and Excited-State Tautomerization: A Combined Experimental and DFT Study. J. Phys. Chem. A 2011, 115, 2456−2464. (25) Prukala, D.; Sikorska, E.; Koput, J.; Khmelinskii, I.; Karolczak, J.; Gierszewski, M.; Sikorski, M. Acid-Base Equilibriums of Lumichrome and Its 1-Methyl, 3-Methyl, and 1,3-Dimethyl Derivatives. J. Phys. Chem. A 2012, 116, 7474−7490.

(26) Marchena, M.; Gil, M.; Martin, C.; Organero, J. A.; Sanchez, F.; Douhal, A. Stability and Photodynamics of Lumichrome Structures in Water at Different pHs and in Chemical and Biological Caging Media. J. Phys. Chem. B 2011, 115, 2424−2435. (27) Sikorska, E.; Khmelinskii, I. V.; Worrall, D. R.; Koput, J.; Sikorski, M. Spectroscopy and Photophysics of Iso- and Alloxazines: Experimental and Theoretical Study. J. Fluoresc. 2004, 14, 57−64. (28) Salzmann, S.; Marian, C. M. The Photophysics of Alloxazine: A Quantum Chemical Investigation in Vacuum and Solution. Photoch. Photobio. Sci. 2009, 8, 1655−1666. (29) Sikorska, E.; Khmelinskii, I. V.; Bourdelande, J. L.; Bednarek, A.; Williams, S. L.; Patel, M.; Worrall, D. R.; Koput, J.; Sikorski, M. Spectroscopy and Photophysics of Mono Methyl-Substituted Alloxazines. Chem. Phys. 2004, 301, 95−103. (30) Sikorska, E.; Khmelinskii, I. V.; Koput, J.; Bourdelande, J. L.; Sikorski, M. Electronic Structure of Isoalloxazines in Their Ground and Excited States. J. Mol. Struct. 2004, 697, 137−141. (31) Sikorski, M.; Prukala, D.; Insinska-Rak, M.; Khmelinskii, I.; Worrall, D. R.; Williams, S. L.; Hernando, J.; Bourdelande, J. L.; Koput, J.; Sikorska, E. Spectroscopy and Photophysics of DimethylSubstituted Alloxazines. J. Photochem. Photobiol., A 2008, 200, 148− 160. (32) MacKerell, A. D.; Bashford, D.; Bellott, M.; Dunbrack, R. L.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; et al. All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins. J. Phys. Chem. B 1998, 102, 3586−3616. (33) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926−935. (34) Smith, W.; Forester, T. R. DL−POLY−2.0: A General-Purpose Parallel Molecular Dynamics Simulation Package. J. Mol. Graphics 1996, 14, 136−141. (35) Metz, S.; Kastner, J.; Sokol, A. A.; Keal, T. W.; Sherwood, P. Chemshell-A Modular Software Package for QM/MM Simulations. Wires Comput. Mol. Sci. 2014, 4, 101−110. (36) ChemShell3.5, a Computational Chemistry Shell. www. chemshell.org (accessed July, 2016). (37) Andersson, K.; Malmqvist, P.-Å.; Roos, B. O.; Sadlej, A. J.; Wolinski, K. Second-Order Perturbation Theory with a CASSCF Reference Function. J. Phys. Chem. 1990, 94, 5483−5488. (38) Andersson, K.; Malmqvist, P.-Å.; Roos, B. O. Second-Order Perturbation Theory with a Complete Active Space Self-Consistent Field Reference Function. J. Chem. Phys. 1992, 96, 1218−1226. (39) Aquilante, F.; Lindh, R.; Pedersen, T. B. Unbiased Auxiliary Basis Sets for Accurate Two-Electron Integral Approximations. J. Chem. Phys. 2007, 127, 114107−114713. (40) Ghigo, G.; Roos, B. O.; Malmqvist, P.-Å. A Modified Definition of the Zeroth-Order Hamiltonian in Multiconfigurational Perturbation Theory (CASPT2). Chem. Phys. Lett. 2004, 396, 142−149. (41) Försberg, N.; Malmqvist, P.-Å. Multiconfiguration Perturbation Theory with Imaginary Level Shift. Chem. Phys. Lett. 1997, 274, 196− 204. (42) Ditchfield, R.; Hehre, W. J.; Pople, J. A. Self-Consistent Molecular-Orbital Methods. IX. An Extended Gaussian-Type Basis for Molecular-Orbital Studies of Organic Molecules. J. Chem. Phys. 1971, 54, 724−728. (43) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S.; DeFrees, D. J.; Pople, J. A. Self-Consistent Molecular Orbital Methods. XXIII. A Polarization-Type Basis Set for Second-Row Elements. J. Chem. Phys. 1982, 77, 3654−3665. (44) Dunning, T. H., Jr. Gaussian Basis Sets for Use in Correlated Molecular Calculations. I. The Atoms Boron through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007−1023. (45) Chang, X.-P.; Fang, Q.; Cui, G. L. Mechanistic Photodecarboxylation of Pyruvic Acid: Excited-State Proton Transfer and Three-State Intersection. J. Chem. Phys. 2014, 141, 154311−154318. (46) Cui, G. L.; Guan, P.-J.; Fang, W.-H. Photoinduced Proton Transfer and Isomerization in a Hydrogen-Bonded Aromatic Azo G


Article

The Journal of Physical Chemistry A Compound: A CASPT2//CASSCF Study. J. Phys. Chem. A 2014, 118, 4732−4739. (47) Xia, S.-H.; Liu, X.-Y.; Fang, Q.; Cui, G. L. Excited-State RingOpening Mechanism of Cyclic Ketones: A MS-CASPT2//CASSCF Study. J. Phys. Chem. A 2015, 119, 3569−3576. (48) Guan, P.-J.; Cui, G. L.; Fang, Q. Computational Photochemistry of the Azobenzene Scaffold of Sudan I and Orange II Dyes: ExcitedState Proton Transfer and Deactivation Via Conical Intersections. ChemPhysChem 2015, 16, 805−811. (49) Chang, X.-P.; Li, C.-X.; Xie, B.-B.; Cui, G. L. Photoprotection Mechanism of p-Methoxy Methylcinnamate: A CASPT2 Study. J. Phys. Chem. A 2015, 119, 11488−11497. (50) Ferré, N.; Olivucci, M. Probing the Rhodopsin Cavity with Reduced Retinal Models at the CASPT2//CASSCF/Amber Level of Theory. J. Am. Chem. Soc. 2003, 125, 6868−6869. (51) Ferré, N.; Cembran, A.; Garavelli, M.; Olivucci, M. CompleteActive-Space Self-Consistent-Field/Amber Parameterization of the Lys296-Retinal-Glu113 Rhodopsin Chromophore-Counterion System. Theor. Chem. Acc. 2004, 112, 335−341. (52) Ponder, J. W.; Richards, F. M. An Efficient Newton-Like Method for Molecular Mechanics Energy Minimization of Large Molecules. J. Comput. Chem. 1987, 8, 1016−1024. (53) Karlström, G.; Lindh, R.; Malmqvist, P.-Å.; Roos, B. O.; Ryde, U.; Veryazov, V.; Widmark, P. O.; Cossi, M.; Schimmelpfennig, B.; Neogrady, P.; et al. MOLCAS: A Program Package for Computational Chemistry. Comput. Mater. Sci. 2003, 28, 222−229. (54) Aquilante, F.; De Vico, L.; Ferré, N.; Ghigo, G.; Malmqvist, P.Å.; Neogrády, P.; Pedersen, T. B.; Pitoňaḱ , M.; Reiher, M.; Roos, B. O.; et al. MOLCAS 7: The Next Generation. J. Comput. Chem. 2010, 31, 224−247. (55) Boggio-Pasqua, M.; Groenhof, G. Controlling the Photoreactivity of the Photoactive Yellow Protein Chromophore by Substituting at the p-Coumaric Acid Group. J. Phys. Chem. B 2011, 115, 7021−7028. (56) Crespo-otero, R.; Mardykov, A.; Sanchezgarcia, E.; Sander, W.; Barbatti, M. Photo-Stability of Peptide-Bond Aggregates: N-methylformamide Dimers. Phys. Chem. Chem. Phys. 2014, 16, 18877−18887. (57) Plasser, F.; Crespo-Otero, R.; Pederzoli, M.; Pittner, J.; Lischka, H.; Barbatti, M. Surface Hopping Dynamics with Correlated SingleReference Methods: 9H-Adenine as a Case Study. J. Chem. Theory Comput. 2014, 10, 1395−1405. (58) Tuna, D.; Doslic, N.; Malis, M.; Sobolewski, A. L.; Domcke, W. Mechanisms of Photostability in Kynurenines: A Joint ElectronicStructure and Dynamics Study. J. Phys. Chem. B 2015, 119, 2112− 2124. (59) El-Sayed, M. A. Spin-Orbit Coupling and the Radiationless Processes in Nitrogen Heterocyclics. J. Chem. Phys. 1963, 38, 2834− 2838. (60) El-Sayed, M. A. Triplet State-Its Radiative and Nonradiative Properties. Acc. Chem. Res. 1968, 1, 8−161. (61) Prukala, D.; Khmelinskii, I.; Koput, J.; Gierszewski, M.; Pedzinski, T.; Sikorski, M. Photophysics, Excited-State Double-Proton Transfer and Hydrogen-Bonding Properties of 5-Deazaalloxazines. Photochem. Photobiol. 2014, 90, 972−988. (62) Prukala, D.; Gierszewski, M.; Kubicki, M.; Pedzinski, T.; Sikorska, E.; Sikorski, M. Spectroscopy and Photophysics of Trimethyl-Substituted Derivatives of 5-Deazaalloxazine. Experimental and Theoretical Approaches. J. Mol. Struct. 2015, 1079, 139−146.

H


MM Study on Mechanistic Photophysics of Alloxazine

Recommend Documents