Intersystem Crossing Enables 4-Thiothymidine to Act as a

Jul 22, 2014 - Intersystem Crossing Enables 4-Thiothymidine to Act as a ... structure of their S1 states, and (3) environmental effects play an import...
15 downloads 0 Views 2MB Size
Letter pubs.acs.org/JPCL

Intersystem Crossing Enables 4‑Thiothymidine to Act as a Photosensitizer in Photodynamic Therapy: An Ab Initio QM/MM Study Ganglong Cui*,† and Walter Thiel* Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 Mülheim an der Ruhr, Germany S Supporting Information *

ABSTRACT: Motivated by its potential use as a photosensitizer in photodynamic therapy, we report the first ab initio quantum mechanics/molecular mechanics (QM/MM) study of 4-thiothymidine in aqueous solution. The core chromophore 4-thiothymine was described using the multiconfigurational CASSCF and CASPT2 QM methods, while the ribose and the solvent water molecules were treated at the MM level (CHARMM and TIP3P, respectively). The minima of the five lowest electronic states (S0, S1, S2, T1, and T2) and six minimum-energy intersections were fully optimized at the QM(CASSCF)/MM level, and their energies were further refined by single-point QM(CASPT2)/MM and CASPT2 calculations. The relevant spin−orbit couplings were also computed. We find that (1) there are three efficient photophysical pathways that account for the experimentally observed ultrafast formation of the lowest triplet state with a quantum yield of nearly unity, (2) the striking qualitative differences in the photophysical behavior of 4thiothymine and thymine originate from the different electronic structure of their S1 states, and (3) environmental effects play an important role. The present QM/MM calculations provide mechanistic insight that may guide the design of improved photosensitizers for photodynamic therapy. SECTION: Spectroscopy, Photochemistry, and Excited States

P

hotodynamic therapy (PDT)1−5 has attracted much experimental interest in view of potential medical applications in the treatment of cancers, infections, and heart diseases.6,7 In PDT, photosensitizers are first incorporated into the diseased cells. After one- or two-photon excitation to a spectroscopically bright singlet state, the lowest triplet state of the photosensitizer is populated via intersystem crossing (ISC) with a high quantum yield. This triplet state then transfers excitation energy to a ground-state oxygen molecule, generating cytotoxic singlet oxygen, which is exceptionally reactive in a biological medium and causes oxidative degradation of harmful biomolecules in the target cells. Therefore, in an ideal PDT photosensitizer, internal conversion (IC) to the ground state should be minimized, and simultaneously, ISC to the lowest triplet state should be maximized. Moreover, radiative processes and competing reactive channels generating other photoproducts should be suppressed both in the singlet and triplet states. One prerequisite to design better photosensitizers for a specific PDT process is to understand the underlying photophysical and photochemical mechanisms at the atomistic level. In this regard, reliable high-level electronic structure calculations can play an essential role because they allow us to identify the relevant excited-state topologies and deactivation channels. For a realistic modeling of PDT processes, the effects of the cellular environment should also be considered (at least by accounting for surrounding water molecules). © 2014 American Chemical Society

The photophysics and photochemistry of azo- and thiosubstituted nucleobases have recently been studied extensively because of their potential as drugs for the treatment of several diseases such as cancer.8−21 In these nucleobase derivatives, the lowest triplet state is formed with a high quantum yield (close to unity). This is qualitatively different from the behavior of natural nucleobases in which IC to the ground state is the predominant deactivation channel that protects them from harmful photodamage.22−37 A simple C-to-N or O-to-S substitution in these natural nucleobases largely suppresses IC to the ground state and instead allows for efficient ISC to the lowest triplet state. The photophysical mechanisms of IC and ISC have not yet been thoroughly explored in these nucleobase variants. For example, how does ISC compete with IC? How does the system avoid the efficient IC to the ground state that is observed in all five natural nucleobases? How does the topology of the excited-state potential energy surfaces inside and outside of the Franck−Condon region influence the photophysical processes that are relevant for the design of better photosensitizers? To shed light on these questions, we have chosen the potential 4-thiothymidine PDT photosensitizer as our target, Received: June 7, 2014 Accepted: July 22, 2014 Published: July 22, 2014 2682

dx.doi.org/10.1021/jz501159j | J. Phys. Chem. Lett. 2014, 5, 2682−2687

The Journal of Physical Chemistry Letters

Letter

CASSCF) treatment of the core chromophore with a realistic MM description of the ribose/water environment. Our goal is to elucidate the working mechanism of aqueous 4-thiothymidine and to gain detailed insight into the photoinduced processes by locating and characterizing all relevant excitedstate minima, conical intersections (for IC), and minimumenergy crossing points (for ISC). Using the ChemShell implementation of the QM(CASSCF)/MM method,38 we have optimized the minimumenergy structures of our system in the S0, S1, S2, T1, and T2 electronic states (referred to as S0-MIN, S1-MIN, S2-MIN, T1MIN, and T2-MIN) as well as six relevant conical intersections and crossing points (S2S1-MIN, S2T2-MIN, T2T1-MIN, S1T1-MIN, S1S0-MIN, and T1S0-MIN). Table 1 collects selected bond lengths and relative energies obtained from single-point QM(CASPT2)/MM calculations and from CASPT2 calculations of the unrelaxed QM region (last column). At the Franck−Condon point (S0-MIN), the S1 state is mainly of nsπ* character; the lone-pair orbital is localized at the sulfur atom, while the π* orbital is delocalized over the whole molecule. The S2, T1, and T2 states arise from ππ* electronic excitations; in the case of T1, there is some nsπ* admixture. These assignments agree with previous results.12,13 At the QM(CASPT2)/MM [CASPT2] level, the vertical excitation energies to S1 and S2 are computed to be 3.0 [2.7] and 4.1 [4.0] eV, respectively. They are much higher than the previous values of 2.3 and 3.4 eV from TD-B3LYP calculations in vacuo12 but closer to those computed at the TD-PBE0 level, 2.99 and 4.13 eV in vacuo, 3.26 and 4.02 eV in solution (PCM model), and 3.22 and 4.0 eV in solution (two explicit water molecules plus the PCM model).13 The S2 state is spectroscopically bright, while S1 is dark. The excited-state minima have distinct geometries that reflect the differences in their electronic structure. Comparing the S1 and T1 minima (S1-MIN and T1-MIN) with the S0 minimum (S0-MIN), the most notable structural change concerns the C4−S4′ bond. It is a partial CS double bond in S0-MIN (1.68 Å) and becomes a typical C−S single bond in S1-MIN (1.81 Å) and T1-MIN (1.78 Å). In the S1 and T1 states, there is no π bonding in the C4−S4′ moiety because the bonding π and antibonding π* molecular orbital are each occupied by one electron, leading to a vanishing π bond order. In addition, because the π* orbital is delocalized, the C5−C6 bond length also increases slightly. T2-MIN is structurally similar to S1-

motivated by recent femtosecond pump−probe experiments in aqueous solution12,13 and in ionic liquids.14 All of these experiments show an ultrafast formation of the lowest triplet state with a high quantum yield; however, mechanistic details are under debate.12,13 In these experimental studies, DFT and TD-DFT electronic structure calculations were performed in vacuo and in solution (polarizable continuum model, PCM) to support the assignment of the observed spectra.12,13 These calculations on 4-thiothymidine provided the optimized ground-state structure, the lowest vertical excitation energies, and a characterization of the corresponding electronic states,12,13 the relevant molecular orbitals, and spin−orbit matrix elements,12 as well as an optimized triplet structure along with the associated vertical phosphorescence emission energy.13 These previous calculations did not explore the excited-state potential energy surfaces beyond the Franck− Condon region nor did they consider environmental effects explicitly. We address these issues in the present multiconfigurational QM(CASPT2//CASSCF)/MM study (Figure 1) by combining an accurate ab initio QM(CASPT2//

Figure 1. QM/MM computational protocol. 4-Thiothymidine was solvated in a spherical water ball of 20 Å. All atoms within 10 Å from any atom of 4-thiothymidine were allowed to move during geometry optimizations (1133 atoms), and the remaining 2196 atoms were frozen at the positions adopted after the initial equilibrium dynamics simulations. The QM subsystem (4-thiothymine chromophore) was described using the CASPT2//CASSCF method, while the MM subsystem (all other atoms) was represented by the CHARMM force field (ribose) and the TIP3P model (water). A hydrogen link atom was used in combination with the charge-shift scheme at the QM/MM boundary (black line). Color code: sulfur in yellow, nitrogen in blue, oxygen in red, carbon in gray, and hydrogen in white. Also shown is the atomic numbering. See the Supporting Information for a more detailed specification of the computational protocol.

Table 1. Selected Bond Lengths (Å) of Stationary Points and Intersection Structures Optimized at the QM(CASSCF)/MM Level and Single-Point Relative Energies (kcal/mol) Obtained at the QM(CASPT2//CASSCF)/MM Level (see the Supporting Information for CASPT2 Relative Energies of the Unrelaxed QM Region)

S0-MIN T1-MIN S1-MIN T2-MIN S2-MIN S1S0-MIN T1S0-MIN S1T1-MIN S2S1-MIN S2T2-MIN T2T1-MIN

N1−C2

C2−O2′

C2−N3

N3−C4

C4−S4′

C4−C5

C5−C6

QM/MM energy

1.36 1.37 1.37 1.37 1.40 1.38 1.37 1.38 1.40 1.40 1.37

1.23 1.24 1.24 1.24 1.23 1.24 1.24 1.24 1.23 1.23 1.24

1.34 1.33 1.33 1.33 1.33 1.33 1.32 1.32 1.33 1.33 1.33

1.37 1.43 1.40 1.41 1.39 1.42 1.41 1.39 1.39 1.39 1.41

1.68 1.78 1.81 1.80 1.74 1.98 2.34 1.88 1.74 1.75 1.80

1.45 1.38 1.38 1.37 1.37 1.30 1.38 1.37 1.37 1.37 1.39

1.34 1.39 1.38 1.40 1.42 1.52 1.40 1.39 1.42 1.42 1.38

0.00 52.8 58.6 55.8 71.8 91.7(S0)/94.6(S1) 91.4(S0)/92.6(T1) 60.0(T1)/59.7(S1) 72.5(S1)/74.3(S2) 71.7(T2)/73.0(S2) 60.2(T1)/60.5(T2)

2683

dx.doi.org/10.1021/jz501159j | J. Phys. Chem. Lett. 2014, 5, 2682−2687

The Journal of Physical Chemistry Letters

Letter

MIN and T1-MIN, whereas S2-MIN has shorter C4−S4′ and C6−N1 bonds and a longer N1−C2 bond (see Table 1). The three lowest excited states (S1, T1, and T2) are energetically close to each other but much below the S2 state (see Table 1). At the QM(CASPT2)/MM level, the S1, T1, and T2 energies relative to S0-MIN (adiabatic excitation energies) are computed to be 58.6, 52.8, and 55.8 kcal/mol, respectively, compared with 71.8 kcal/mol for the S2 state. Hence, the S2 state lies 16.0 kcal/mol above the T2 state, which is derived from the same ππ* excitation, while the energy gap between the S1 and T1 states is much smaller (5.8 kcal/mol). The energetic proximity of the three lowest excited states makes ISC to the triplet very efficient. Conical intersections and crossing points among potential energy surfaces are at the heart of nonadiabatic photophysics and photochemistry. Here, we have employed a QM/MMbased penalty function approach (see the Supporting Information for details) to optimize six conical intersections and crossing points. Of these structures, S2S1-MIN, S2T2MIN, T2T1-MIN, and S1T1-MIN can be easily accessed from S2-MIN, T2-MIN, and S1-MIN, respectively, because of small energy gaps, each of which is less than 5.0 kcal/mol at the QM(CASPT2)/MM level (see Table 1). By contrast, the nonadiabatic channels to the S0 state via S1S0-MIN (IC) and T1S0-MIN (ISC) are inefficient due to much larger energy gaps (more than 30 kcal/mol). The relative energies of S1S0MIN and T1S0-MIN are close to the available total energy of 95.1 kcal/mol, that is, the S2 energy at the Franck−Condon point, so that they can hardly be reached dynamically in the presence of competing downhill processes. Hence, 4thiothymidine can serve as a PDT photosensitizer because both IC and ISC to the ground state are effectively suppressed.12,13 On the basis of the optimized structures and energies of the lowest five electronic states as well as the relevant spin−orbit couplings (SOCs), three main nonadiabatic pathways can be proposed to account for the ultrafast formation and high quantum yield of the lowest triplet state T1 from the initially populated bright excited singlet state S2.12,13 In the first pathway, referred to as P-I, the S2 state efficiently decays to S1 via an easily accessible S2/S1 conical intersection S2S1-MIN (∼2 kcal/mol higher than the S2 minimum). ISC from the S1 to the T1 state can then rapidly occur around the S1/T1 crossing point, which is located only ∼1 kcal/mol above S1-MIN. In the second pathway, P-II, the S2 system first switches at the S2/T2 crossing point S2T2-MIN to a relay T2 state, from which the lowest T1 state is subsequently accessed by a T2 → T1 IC at T2T1-MIN. The third pathway, P-III, is a direct S2 → T1 ISC mediated by vibronic coupling and strong SOC interactions (see QM/MM SOC values in Table 2; a Z component of 93.2 cm−1). It bypasses both the intermediate T2 and S1 states. Figure 2 illustrates these three main decay channels. Kasha’s rule39 states that in a given spin manifold, IC from higher excited states to the lowest one will be faster than competing processes; accordingly, in the singlet manifold, the deexcitation Sn → S1 (n > 1) should be ultrafast, and fluorescence emission, chemical reactions, and triplet-state population should start from the lowest singlet excited state (S1). Among the three proposed pathways, only P-I complies with this rule: an S2 → S1 IC at S2S1-MIN populates the S1 state followed by ISC to the T1 state. The other two pathways bypass the intermediate S1 state on two different routes. In P-II, the T2 state acts as a relay as a result of S2 → T2 ISC at S2T2-

Table 2. Absolute Values of QM(CASSCF)/MM Calculated Components of SOC Matrix Elements (Ĥ ′: SOC operator; unit: cm−1) at Selected Pointsa structure S2T2-MIN S1T1-MIN S2-MIN S2-MIN S1-MIN

|⟨S2|Ĥ ′|T2⟩| |⟨S1|Ĥ ′|T1>| |⟨S2|Ĥ ′|T1⟩| |⟨S2|Ĥ ′|T2⟩| |⟨S1|Ĥ ′|T1⟩|

X, Y

Z

3.0 1.8 0.8 2.8 0.7

3.1 10.2 93.2 3.0 60.6

a

See the Supporting Information for further discussion and for values computed without the MM point charges.

MIN; in P-III, the lowest T1 state is directly populated via vibronic coupling because of very strong S2/T1 SOC at the S2 minimum S2-MIN (see Table 2). A comprehensive mechanistic scenario thus requires consideration not only of the lowest singlet and triplet excited states but also of energetically accessible higher states. The photophysics of 4-thiothymine (the core chromophore) is completely different from that of the natural nucleobase thymine. In the latter, the S1/S0 conical intersection lies much below the Franck−Condon point of the initially populated S1 state (by 0.9 eV40); it is energetically easily accessible and funnels the excited S1 state to the ground state so as to quickly dissipate the excess energy from sunlight, thus avoiding photodamage.22−30,32,36,37,40 This viewpoint has been seconded by ab initio and semiempirical trajectory-based surface-hopping dynamics simulations.31,33−35,41 The simple single-atom substitution O4 → S4′ (see Figure 1) makes a big difference to the photophysics. In aqueous 4-thiothymidine, the S1/S0 conical intersection is found to be very high in energy, ∼93 kcal/mol at S1S0-MIN, compared to the S2 energy of 95.1 kcal/mol at the Franck−Condon point, and therefore, the S1 → S0 IC becomes inefficient and is effectively inhibited. As an alternative, different ISC pathways to the triplet states open up (see above) because the heavy-atom effect of sulfur enhances SOC and thus facilitates population of the T1 state. However, the T1/S0 crossing point (T1S0-MIN) also occurs at very high energy, at around 92 kcal/mol at the QM(CASPT2)/MM level and, thus, only 3 kcal/mol below the Franck−Condon point of the initially populated S2 state, so that the T1 → S0 ISC process is also impeded. Hence, one may safely expect that 4thiothymidine will stay in the T1 state for a relatively long time, which will allow for the generation of reactive singlet oxygen species that can “kill” pathological cells in PDT. We now consider the computed structures of the intersection points, all of which have been determined through optimizations starting from the corresponding excited-state minimum (see the Supporting Information for details). As already discussed, four of these structures (S2S1-MIN, S1T1MIN, S2T2-MIN, and T2T1-MIN) are energetically very close to the starting points (S2-MIN, S1-MIN, and T2-MIN), and it is thus not surprising that they are also geometrically close (see Table 1). By contrast, the other two structures (S1S0-MIN and T1S0-MIN) are much higher in energy and are therefore much more distorted. Taking S2S1-MIN and S1S0-MIN as examples (Figure 3), the CH moiety at the C6 atom is fully extruded out of the molecular plane in S1S0-MIN (right panel, C4−C5− C6−H6′: 95.7°), whereas it remains almost in the molecular plane in S2S1-MIN (left panel, C4−C5−C6−H6′: 171.3°). The qualitative differences in the mechanistic photophysics of 4-thiothymine and thymine (see above) are reflected in the 2684

dx.doi.org/10.1021/jz501159j | J. Phys. Chem. Lett. 2014, 5, 2682−2687

The Journal of Physical Chemistry Letters

Letter

Figure 2. Proposed possible pathways for the formation of the lowest triple state (T1) of aqueous 4-thiothymidine on the basis of the present QM/ MM calculations. The P-I (S2 → S1 → T1) pathways comply with Kasha’s rule; the P-II (S2 → T2 → T1) and P-III (S2 → T1) pathways bypass the lowest singlet excited state S1 and thus do not follow Kasha’s rule. The ISC processes in P-I and P-II proceed via minimum-energy crossing points, while those in P-III involve vibronic coupling and SOC. The three schemes are only meant to visualize the possible pathways, without reference to any particular reaction coordinate. See the text.

Figure 3. QM(CASSCF)/MM optimized S2/S1 (S2S1-MIN, left) and S1/S0 (S1S0-MIN, right) conical intersection structures with two selected dihedral angles (see Figure 1 for the atomic numbering and Table 1 for specific geometric parameters). Also shown in the middle panel is their spatial overlap. See the text.

for aqueous 4-thiothymidine. Concerning energies (see Table 3 of the Supporting Information), single-point QM calculations of the QM region yield relative energies for the three lowest excited states (S1, T1, and T2) that are close to the QM/MM values and also close to each other, whereas they overestimate the S2 energy by 12.1 kcal/mol (compared with QM/MM). The influence of the MM environment is more pronounced in the case of the conical intersections and crossing points; in the single-point QM calculations, the two relevant states remain close to each other only for S1T1-MIN and T2T1-MIN, but they considerably differ in energy in the other cases (up to 23.3 kcal/mol for S2T2-MIN). This demonstrates the importance of environmental effects on the geometries and energies of crossing points. The MM environment also has a significant influence on the computed SOCs; the surrounding water molecules and the ribose exert state-specific, nonadditive polarization effects such that the SOC values depend rather sensitively on the inclusion of the MM point charges (see Table 4 of the Supporting Information). To summarize, we have performed the first QM/MM electronic structure investigation of the mechanistic photophysics of the potential PDT photosensitizer 4-thiothymidine in aqueous solution.12,13 We identified three efficient photophysical pathways to account for the experimentally observed ultrafast formation of the lowest triplet state with high quantum yield; one of these conforms to the classical Kasha rule, and the other two do not. The differences in the electronic structures of

computed structures of their S1/S0 conical intersections (S1S0MIN), which are of different character and shape (see Table 2 of the Supporting Information). The X4′−C4−C5−C5′ dihedral angle is 1.2° in 4-thiothymine (X = S) and ∼95° in thymine (X = O), indicating that the methyl substituent is coplanar with the ring in the former and essentially perpendicular in the latter. The H6 atom is distorted out of the ring plane more strongly in 4-thiothymine than that in thymine, as can be seen from the C4−C5−C6−H6′ dihedral angles of 95.7 and ∼130°, respectively. Moreover, S1S0-MIN has a very elongated C4−S4′ bond in 4-thiothymine (1.98 Å, larger than the typical C−S single bond length of ∼1.8 Å), while the C4−O4′ bond remains short (∼1.2 Å, in the range of a typical CO double bond). In both cases, the C5C6 double bond in S0-MIN is broken and becomes a single bond in S1S0-MIN, albeit with a more pronounced elongation in 4thiothymine (1.52 Å) than that in thymine (1.47 Å).42 The S1 state is of ππ* character in thymine but of mixed ππ* and nsπ* character in 4-thiothymine. As a consequence, the S1/S0 conical intersection has an ethylene-like structure in thymine but adopts another shape in 4-thiothymine, with a much larger C4−S4 distance and a different kind of out-of-plane distortion of the ring. The different photophysics of 4-thiothymine and thymine is thus ultimately rooted in the distinct electronic structure of their excited states. Finally, we address the effect of the MM environment (ribose and solvent water molecules) on the present QM/MM results 2685

dx.doi.org/10.1021/jz501159j | J. Phys. Chem. Lett. 2014, 5, 2682−2687

The Journal of Physical Chemistry Letters

Letter

efficient intersystem crossing and singlet oxygen photosensitization. J. Phys. Chem. B 2010, 114, 8782. (11) Etinski, M.; Marian, C. M. Overruling the energy gap law: fast triplet formation in 6-azauracil. Phys. Chem. Chem. Phys. 2010, 12, 15665. (12) Harada, Y.; Okabe, C.; Kobayashi, T.; Suzuki, T.; Ichimura, T.; Nishi, N.; Xu, Y. Ultrafast intersystem crossing of 4-thiothymidine in aqueous solution. J. Phys. Chem. Lett. 2010, 1, 480−484. (13) Reichardt, C.; Crespo-Hernández, C. Room-temperature phosphorescence of the DNA monomer analogue 4-thiothymidine in aqueous solutions after UVA excitation. J. Phys. Chem. Lett. 2010, 1, 2239−2243. (14) Reichardt, C.; Crespo-Hernández, C. Ultrafast spin crossover in 4-thiothymidine in an ionic liquid. Chem. Commun. 2010, 46, 5963− 5965. (15) Reichardt, C.; Guo, C.; Crespo-Hernández, C. Excited-state dynamics in 6-thioguanosine from the femtosecond to microsecond time scale. J. Phys. Chem. B 2011, 115, 3263−3270. (16) Martínez-Fernández, L.; González, L.; Corral, I. An ab initio mechanism for efficient population of triplet states in cytotoxic sulfur substituted DNA bases: the case of 6-thioguanine. Chem. Commun. 2012, 48, 2134−2136. (17) Gobbo, J. P.; Borin, A. C.; Serrano-Andrés, L. On the relaxation mechanisms of 6-azauracil. J. Phys. Chem. B 2011, 115, 6243. (18) Gobbo, J.; Borin, A. On the mechanisms of triplet excited state population in 8-azaadenine. J. Phys. Chem. B 2012, 116, 14000−14007. (19) Cui, G.; Fang, W. State-specific heavy-atom effect on intersystem crossing processes in 2-thiothymine: a potential photodynamic therapy photosensitizer. J. Chem. Phys. 2013, 138, 044315. (20) Gobbo, J.; Borin, A. On the population of triplet excited states of 6-aza-2-thiothymine. J. Phys. Chem. A 2013, 117, 5589−5596. (21) Pollum, M.; Crespo-Hernández, C. Communication: The dark singlet state as a doorway state in the ultrafast and efficient intersystem crossing dynamics in 2-thiothymine and 2-thiouracil. J. Chem. Phys. 2014, 140, 071101. (22) Pecourt, J.; Peon, J.; Kohler, B. DNA excited-state dynamics: ultrafast internal conversion and vibrational cooling in a series of nucleosides. J. Am. Chem. Soc. 2001, 123, 10370. (23) Ismail, N.; Blancafort, L.; Olivucci, M.; Kohler, B.; Robb, M. Ultrafast decay of electronically excited singlet cytosine via a π,π* to nO,π* state switch. J. Am. Chem. Soc. 2002, 124, 6818−6819. (24) Crespo-Hernández, C.; Cohen, B.; Hare, P.; Kohler, B. Ultrafast excited-state dynamics in nucleic acids. Chem. Rev. 2004, 104, 1977. (25) Matsika, S. Radiationless decay of excited states of uracil through conical intersections. J. Phys. Chem. A 2004, 108, 7584−7590. (26) Perun, S.; Sobolewski, A.; Domcke, W. Ab initio studies on the radiationless decay mechanisms of the lowest excited singlet states of 9H-adenine. J. Am. Chem. Soc. 2005, 127, 6257−6265. (27) Marian, C. A new pathway for the rapid decay of electronically excited adenine. J. Chem. Phys. 2005, 122, 104314. (28) Perun, S.; Sobolewski, A.; Domcke, W. Conical intersections in thymine. J. Phys. Chem. A 2006, 110, 13238. (29) Marian, C. The guanine tautomer puzzle: quantum chemical investigation of ground and excited states. J. Phys. Chem. A 2007, 111, 1545−1553. (30) Yamazaki, S.; Domcke, W.; Sobolewski, A. Nonradiative decay mechanisms of the biologically relevant tautomer of guanine. J. Phys. Chem. A 2008, 112, 11965−11968. (31) Barbatti, M.; Lischka, H. Nonadiabatic deactivation of 9 Hadenine: a comprehensive picture based on mixed quantum-classical dynamics. J. Am. Chem. Soc. 2008, 130, 6831−6839. (32) Asturiol, D.; Lasorne, B.; Robb, M.; Blancafort, L. Photophysics of the thymine: MS-CASPT2 minimum-energy paths and CASSCF on-the-fly dynamics. J. Phys. Chem. A 2009, 113, 10211. (33) Szymczak, J.; Barbatti, M.; Soo Hoo, J.; Adkins, J.; Windus, T.; Nachtigallova, D.; Lischka, H. Photodynamics simulations of thymine: relaxation into the first excited singlet state? J. Phys. Chem. A 2009, 113, 12686.

the excited states of 4-thiothymidine and natural thymine are responsible for their distinct mechanistic photophysics. Environmental effects are found to be significant and need to be included to arrive at realistic mechanisms for photosensitizers in PDT; after all, the corresponding medical treatment is carried out in a cellular environment, not in vacuum.1−4,6,7 The current ab initio QM/MM study offers insight into the mechanisms by which aqueous 4-thiothymidine may act as a PDT photosensitizer but does not establish the relative importance of the three proposed pathways. For this purpose, excited-state dynamics simulations are required, which will be the subject of future work. We hope that these issues will also be investigated experimentally, for example, by ultrafast time-resolved spectroscopy.21



ASSOCIATED CONTENT

S Supporting Information *

System setup, computational details, further energy and geometry data, spin−orbit couplings, local hydrogen-bonding networks, additional references, force field parameters, and Cartesian coordinates of all optimized structures. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (G.C.). *E-mail: [email protected] (W.T.). Present Address †

G.C.: Key Laboratory of Theoretical and Computational Photochemistry, Ministry of Education, College of Chemistry, Beijing Normal University, Beijing 100875, China. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by an Alexander von Humboldt Fellowship (G.C.). REFERENCES

(1) Henderson, B.; Dougherty, T. How does photodynamic therapy work? Photochem. Photobiol. 1992, 55, 145−157. (2) Dougherty, T.; Gomer, C.; Henderson, B.; Jori, G.; Kessel, D.; Korbelik, M.; Moan, J.; Peng, Q. Photodynamic therapy. J. Natl. Cancer Inst. 1998, 90, 889−905. (3) Macdonald, I.; Dougherty, T. Basic principles of photodynamic therapy. J. Porphyrins Phthalocyanines 2001, 5, 105−129. (4) Dolmans, D.; Fukumura, D.; Jain, R. Photodynamic therapy for cancer. Nat. Rev. Cancer 2003, 3, 380−387. (5) Kilin, D.; Tsemekhman, K.; Prezhdo, O.; Zenkevich, E.; von Borczyskowski, C. Ab initio study of exciton transfer dynamics from a core−shell semiconductor quantum dot to a porphyrin-sensitizer. J. Photochem. Photobiol., A 2007, 190, 342−351. (6) Bonnett, R. Photosensitizers of the porphyrin and phthalocyanine series for photodynamic therapy. Chem. Soc. Rev. 1995, 24, 19−33. (7) Bonnett, R. Chemical aspects of photodynamic therapy; CRC Press, Boca Raton, FL, 2000. (8) Kobayashi, T.; Harada, Y.; Suzuki, T.; Ichimura, T. Excited state characteristics of 6-azauracil in acetonitrile: drastically different relaxation mechanism from uracil. J. Phys. Chem. A 2008, 112, 13308. (9) Kobayashi, T.; Kuramochi, H.; Harada, Y.; Suzuki, T.; Ichimura, T. Intersystem crossing to excited triplet state of aza analogues of nucleic acid bases in acetonitrile. J. Phys. Chem. A 2009, 113, 12088. (10) Kuramochi, H.; Kobayashi, T.; Suzuki, T.; Ichimura, T. Excitedstate dynamics of 6-aza-2-thiothymine and 2-thiothymine: highly 2686

dx.doi.org/10.1021/jz501159j | J. Phys. Chem. Lett. 2014, 5, 2682−2687

The Journal of Physical Chemistry Letters

Letter

(34) Barbatti, M.; Aquino, A.; Szymczak, J.; Nachtigallova, D.; Hobza, P.; Lischka, H. Relaxation mechanisms of UV-photoexcited DNA and RNA nucleobases. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 21453. (35) Lan, Z.; Fabiano, E.; Thiel, W. Photoinduced nonadiabatic dynamics of pyrimidine nucleobases: on-the-fly surface-hopping study with semiempirical methods. J. Phys. Chem. B 2009, 113, 3548−3555. (36) Middleton, C.; de La Harpe, K.; Su, C.; Law, Y.; CrespoHernández, C.; Kohler, B. DNA excited-state dynamics: from single bases to the double helix. Annu. Rev. Phys. Chem. 2009, 60, 217. (37) Kohler, B. Nonradiative decay mechanisms in DNA model systems. J. Phys. Chem. Lett. 2010, 1, 2047. (38) ChemShell3.5, a Computational Chemistry Shell. www. chemshell.org (2014). (39) Kasha, M. Characterization of electronic transitions in complex molecules. Discuss. Faraday Soc. 1950, 9, 14−19. (40) Merchán, M.; González-Luque, R.; Climent, T.; Serrano-Andrés, L.; Rodríguez, E.; Reguero, M.; Peláez, D. Unified model for the ultrafast decay of pyrimidine nucleobases. J. Phys. Chem. B 2006, 110, 26471−26476. (41) Hudock, H.; Levine, B.; Thompson, A.; Satzger, H.; Townsend, D.; Gador, N.; Ullrich, S.; Stolow, A.; Martı ́nez, T. Ab initio molecular dynamics and time-resolved photoelectron spectroscopy of electronically excited uracil and thymine. J. Phys. Chem. A 2007, 111, 8500− 8508. (42) Zechmann, G.; Barbatti, M. Photophysics and deactivation pathways of thymine. J. Phys. Chem. A 2008, 112, 8273−8279.

2687

dx.doi.org/10.1021/jz501159j | J. Phys. Chem. Lett. 2014, 5, 2682−2687