Letter pubs.acs.org/JPCL
Intermolecular Coulombic Decay in Biology: The Initial Electron Detachment from FADH− in DNA Photolyases Philipp H. P. Harbach, Matthias Schneider, Shirin Faraji, and Andreas Dreuw* Interdisciplinary Center for Scientific Computing, Ruprecht-Karls University, Im Neuenheimer Feld 368, 69120 Heidelberg, Germany S Supporting Information *
ABSTRACT: Intermolecular coulombic decay (ICD) is an efficient mechanism of low-energy electron generation in condensed phases and is discussed as their potential source in living cells, tissues, and materials. The first example of ICD as an operating mechanism in real biological systems, that is, in the DNA repair enzymes photolyases, is presented. Photolyase function involves lightinduced electron detachment from a reduced flavin adenine dinucleotide (FADH−), followed by its transfer to the DNA-lesion triggering repair of covalently bound nucleobase dimers. Modern quantum chemical methods are employed to demonstrate that the transferred electron is efficiently generated via a resonant ICD process between the antenna pigment and the FADH− cofactors. SECTION: Spectroscopy, Photochemistry, and Excited States
I
the original, undamaged bases.14−16 Until today, the structures of several photolyases from various organisms have been determined (see, for example, refs 17−20) that share common features. Most notably, all photolyases bind a reduced flavine adenine dinucleotide (FADH−) cofactor as electron donor as well as an augmenting antenna pigment, which can be 8hydroxy-5-deazaflavin (HDF), 5-methyltetrahydrofolate (MTHF), flavinmononucleotide (FMN), or a similar chromophore.21 The initial step of the repair mechanism involves absorption of a photon and generation of a reactive electron, which is transferred to the DNA photolesion (Figure 2). Despite the rather long distance between FADH− and the antenna pigment of ∼16 Å, the quantum yield of this process has been shown to be near unity.22 Previous investigations of this photoinduced electron transfer have pointed out the role of the specific environment in the directionality of the process.23−27 For example, electron tunneling pathways have been identified,24,27 and the time scales of the transfer have been calculated using nonadiabatic simulations.26 In this work, these aspects of the electron transfer step play only a minor role and are not questioned, in contrast to the mechanism of the initial detachment of the transferred electron from FADH−. The latter we show corresponds to an ICD process rather than to a classical Fö rster EET. A clear distinction between these related processes is, however, crucially important for its proper theoretical description. Whereas Förster EET occurs between
ntermolecular coulombic decay (ICD) is a recently discovered general mechanism of very fast and efficient low-energy electron generation by ionized or excited atoms or molecules in condensed phases.1,2 It is speculated that ICD is relevant in biological systems, cells, and tissues as a source for low-energy electrons, which cause severe damage to biological material, for example, DNA strand breaks.3 However, until today, it has not been proven to be indeed operative in a real biological system.4 In general, ICD can occur after absorption of a photon by an atom or molecule when its ionization potential (IP) is higher than the one of a neighboring molecule of the environment (Figure 1). Then, electron emission occurs from the neighbor instead of from the initially excited molecule, provided that the excitation energy suffices.5 ICD has been predicted by theory6,7 and was later experimentally confirmed.8,9 Most theoretical and experimental studies on ICD reported so far focused on diatomic and small molecular clusters to demonstrate its efficiency and to understand the underlying physics of the process.10,11 In general, ICD is much faster and more efficient than regular excitation energy transfer (EET) because the acceptor state is not a single-bound electronic state but a state embedded in a continuum. This always guarantees resonance conditions because the emitted free electron can take up an arbitrary amount of kinetic energy. Here we demonstrate that ICD is in fact the physical mechanism of electron generation in photolyases, eventually triggering DNA damage repair. Photolyases are capable of restoring major UV-light-induced DNA damages, for example, cyclobutane pyrimidine dimers (CPDs) or pyrimidine-pyrimidone (6−4) lesions12,13 utilizing themselves near-UV light. The repair process involves the splitting of the covalently bound lesions and the restoration of © 2013 American Chemical Society
Received: January 17, 2013 Accepted: March 5, 2013 Published: March 5, 2013 943
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
The Journal of Physical Chemistry Letters
Letter
Figure 1. General mechanism of resonant ICD. Upon initial excitation of one molecule (here: 8-HDF), a neighboring molecule of the environment is ionized (here: FADH−).
of the deazariboflavin and flavin rings of HDF and FADH−, respectively, which suffice for the description of the initial photon absorption, excitation energy, and electron transfer processes because the involved molecular orbitals are all localized on these rings. Moreover, as we will see in the following, all relevant energetic quantities are consistently shifted upward by 0.3 to 0.4 eV by going from the full systems to the model systems, and because the relative energies are not affected, the operating physical mechanism can be conclusively identified using the reduced molecular models. The EDE of FADH− and its reduced molecular model have been calculated as the difference of the total energies of the neutral and anionic systems at the DFT/B3LYP as well as SOSMP238 levels of theory (Table S1 in the Supporting Information).35 A dielectric continuum has been employed within a self-consistent reaction field approach39 with dielectric constants varying from weakly to strongly polar media (ε = 2, 5, 10, 20, 80) to simulate the influence of the protein as a polar environment. Over this large range of dielectric constants, the EDE of FADH− changes from 2.2 (ε = 2) to 2.45 eV (ε = 80) and the EDE of the model changes from 2.61 to 2.85 at the level of DFT/B3LYP, respectively. Although the true value of the dielectric constant of a protein is still a matter of ongoing debate, a value between 5 and 10 seems reasonable.40 Hence, we can conclude that the EDE of FAHD− and the molecular model range are in the area of 2.4 and 2.8 eV at the theoretical level of DZP/B3LYP (Table S1 in the Supporting Information), respectively, and about 0.1 eV lower at the theoretical level of SOS-MP2. The energetically lowest excited states of HDF have been computed with TDDFT, and they show the typical blue shift of all states with increasing HF amount (Table S4 in the Supporting Information); however, the order of the energetically lowest and thus relevant states remains the same. Hence, a deteriorating charge-transfer problem of the TDDFT results can be excluded.41,42 For further corroboration of the quality of the TDDFT results and the small influence of the protein environment, the excitation energies were also computed for the reduced molecular model of HDF (Figure S1 in the Supporting Information) using high-level wave function-based CIS(D), ADC(2)-s, and CC2 methods. In fact, the excitation energies obtained for the HDF model at TDDFT/B3LYP level agree with the ones obtained at the higher levels of theory (Table S2 in the Supporting Information). Subsequently, the protein was simulated via a dielectric continuum with increasing dielectric constant from ε = 2 to 80 at the level of TDDFT/
Figure 2. Sketch of the initial absorption of a photon and the excitation energy and electron-transfer processes between HDF (orange) and FADH− (green) in the deoxyribodipyrimidine photolyase of Thermus thermophilus.
two electronically bound and thus stable states, resonant ICD occurs between one bound excited state and an unbound hence unstable excited state, the latter of which requires a different and more elaborate theoretical treatment, as will be outlined below. Many further experimental and theoretical investigations aim at explaining the repair mechanism occurring after the electron transfer.20,28−34 In this context, it has recently been demonstrated theoretically and experimentally that the final repair of DNA photolesions involves a proton transfer from a nearby protonated histidine residue coupled to the electron transfer, which steers the repair toward the restoration of the bases.35−37 For resonant ICD to be the operative mechanism of electron generation in photolyases, the initial excitation energy of the antenna pigment has to lie above the electron detachment energy (EDE) of FADH− in the protein. Hence, in a first step the vertical excitation energies of the antenna pigment and FADH− as well as the EDE of the latter have been computed in the protein environment using the crystal structure of deoxyribodipyrimidine photolyase of Thermus thermophilus as the starting point.21 The details of the calculations are compiled in the Supporting Information. Most importantly, reduced molecular models of HDF and FADH− (Figure S1 in the Supporting Information) have been employed for testing the accuracy of the chosen theoretical methods, for the calculation of the electronically unbound excited states and for the calculation of the ICD rates (see below). These models consist 944
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
The Journal of Physical Chemistry Letters
Letter
Figure 3. Graphical representation of the results of the stabilization method for the excited states of FADH−. With augmentation of the standard ccpVDZ basis set, continuum states (crosses) drop in energy, whereas only local resonance states (orange dots) stabilize at their resonance energy.
vertical excitation energies of FADH− and its molecular model lie ∼0.4 eV above the corresponding electron detachment threshold. Hence, the lowest excited state is undoubtedly electronically unstable and can decay via emission of an electron. In contrast with HDF, the ground-state equilibrium geometry of FADH−, however, adopts a bent conformation. Reoptimization of the crystal-structure geometries of the molecule pairs FADH·/FADH− and HDF/HDF·− at the B3LYP/cc-pVDZ level of theory demonstrates that the former is bent only when the additional electron is added, whereas the latter is always planar. This has already been previously observed.45 The electronically unbound nature of the excited states of FADH−, which are energetically above the EDE, implies difficulties in their theoretical description because their wave functions possess free electron-like parts describing the outgoing electron. In principle, methods originating from scattering theory are required for the description of such decaying states, and standard quantum chemical methods exploiting square-integrable basis functions are generally not applicable. However, theoretically rigorous approaches exist, most notably the so-called complex scaling technique46−48 or the use of complex absorbing potentials,49,50 which allow for a theoretically sound treatment of resonance states with standard quantum chemical procedures.51 In these procedures, the energies become complex numbers, of which the real part describes the resonance position and the imaginary part describes the width or equivalent lifetime of the state. Unfortunately, these approaches are computationally very demanding and currently not applicable to molecules as large as FADH−. Still, a viable approach to estimate the resonance energy of an electronically unbound state is provided by the socalled stabilization method,52,53 which is to be employed in the following to study the excitation energies of the lowest excited states of FADH− The gist of the stabilization method is to smoothly increase the size of the basis set by adding more diffuse basis functions to it. Following this procedure, the continuum states as well as the resonances are better described and become lower in
B3LYP. It turned out that the excitation energies are practically not influenced by the presence of a dielectric continuum (Table S5 in the Supporting Information), that is, by a polar environment. The lowest bright S1 state of HDF exhibits an excitation energy of 3.11 eV at the TDDFT/B3LYP level, which is in good agreement with the SOS-CIS(D) value of 3.19 eV. This excitation energy corresponds to an absorption wavelength of ∼390 nm, which lies in the near-UV region and agrees well with the experimentally known absorption of HDF around 400 nm.43,44 This state corresponds to a typical ππ* excited state, which is best represented in a molecular orbital picture as an electron transition from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO). The optimization of the geometry of HDF on the S1 potential energy surface does not lead to significant structural changes but only to minor changes in the geometrical parameters. The fluorescence wavelength thus exhibits a calculated Stoke’s shift of 70 nm and occurs at 460 nm at the theoretical level of TDDFT/B3LYP. This is again in good agreement with the observed experimental value of 450 nm.43,44 Thus the S1 state can be assigned to the emitting state of HDF, that is, the initial state of the EET process between HDF and FADH−. Turning to the energy acceptor, FADH−, its experimental absorption spectrum is dominated by a weak absorption around 400 nm, which corresponds to an excitation energy of ∼3.1 eV. Comparison of this excitation energy with the calculated EDE value reveals that the energy of the excited ππ* state does in fact lie above the EDE. This is further corroborated by our calculations using TDDFT/B3LYP with a standard cc-pVDZ basis set, which has been previously shown to be generally capable of describing the vertical excited states of FADH− with good accuracy.35 At the applied level of theory, the energetically lowest excited ππ* excited state with an oscillator strength of 0.1 (S5) exhibits a calculated vertical excitation energy of 2.81 eV in the protein (Table S5 in the Supporting Information),35 which can thus be assigned to the experimental lowest experimental absorption band. At this level of theory, the 945
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
The Journal of Physical Chemistry Letters
Letter
ICD rates, which can then be approximated as the sum of the VPE rates between the single bound donor state of HDF (DS1) and the energetically accessible electron detached acceptor states of FADH− (ASn).
energy until they converge toward the EDE, that is, the electron detached ground state and a free electron with zero kinetic energy. However, in the course of the increase in the basis set, resonances are known to stabilize at the resonance energy for a broad range of basis set size. This energy is to be identified as the vertical excitation energy of the resonance state. This approach is expected to work better the more closely the state of interest resembles a bound electronic state and the less it lies above the EDE. These requirements are fulfilled for the lowest excited states of FADH−, which indeed lie only 0.4 eV above the EDE and that can thus be expected to represent a rather long-lived resonance state. The result of the stabilization method applied to the 30 lowest excited states of FADH− is graphically represented in Figure 3. In this Figure, the vertical excited states are given as a function of the diffuseness of the basis set. It is apparent that with increasing the number of diffuse functions an increasing number of continuum states are generated. These continuum states are characterized by a molecular orbital structure, which resembles the one of the neutral radical FADH· and a very diffuse excited additional electron, representing the free, unbound electron. These states drop continuously in energy with increasing diffuseness of the basis. As a consequence, the 30 lowest excited states cover an excitation energy up to 6.5 eV when the standard cc-pVDZ basis set is used, whereas for an augmentation with 4 diffuse s- and p-type functions, that is, with the cc-pVDZ(+4s,+4p) basis, the same number of states covers only a range up to 3.1 eV. Local excited resonance states behave differently though. The excitation energies of typical ππ* excited states, however, decrease only slightly. These excited states are easily identified by their molecular orbital structure and oscillator strengths. These states are marked with a thick orange dot in Figure 3. Using the nondiffuse cc-pVDZ basis set, only one single isolated excited ππ* state with oscillator strength is identified as S2 state with an excitation energy of 3.11 eV. With increasing diffuseness and basis set size, or in other words, the better the basis set can describe the free unbound electron, the excitation energies of the allowed ππ* excited states stabilize at 3 eV before it drops further to 2.7 eV for the largest cc-pVDZ(+4s,+4p) basis set. Hence it is the best estimate of the true excitation energy of the lowest allowed, although the electronically unstable excited state of FADH− is 3 eV, corresponding to an absorption wavelength of ∼413 nm. Calculation of an ICD rate in general requires, in principle, consideration of all possible electron detached states that represent possible final states of the ICD process and that are infinitely many. In our calculations, these states correspond to all FADH− excited states that lie in the range of the energy provided by the HDF antenna pigment, whose excitation energy is eventually transferred. Because the pigments HDF and FADH− have a large intermolecular distance in photolyases of 16 Å, the energy exchange mechanism of the ICD process can be described as virtual photon exchange (VPE), which is dominated by the Coulomb interaction of the corresponding transition densities of HDF and FADH−. In this case, the ICD rate decays with intermolecular distance R as R−6.6 At short distances, such a description is certainly not appropriate because orbital interactions and electron correlation effects come into play and contribute significantly. VPE is also the physical mechanism of classical long-range EET processes between two bound electronic excited states of the interacting chromophores. Hence, the developed methodologies for the calculation of VPE transfer rates can be adapted to estimate
kICD =
∑ k VPE(DS
1
→ A Sn)
(1)
n
For the large intermolecular separation of HDF and FADH− in photolyases, it is certainly the most appropriate for the computation of the VPE rates to consider only the electrostatic long-range interactions. Then, the rates for VPE are given as k VPE(DS1 → A Sn) =
2π f (DS1)f (A Sn) × |V long|2 × δ(ω DS − ω A S ) 1 n ℏ (2)
where f denotes the probability of electronic transition of the donor and acceptor and the δ function tests the resonance condition between the excitation energies of donor and acceptor ωDS1 and ωASn, respectively. Because we are calculating the VPE rate between a bound and unbound electronic states, the outgoing electron can always take up the appropriate amount of kinetic energy to ensure the resonance condition. Hence the δ function can generally be set to one in this case. Utilizing the transition densities ρTA and ρTD on the acceptor and donor, respectively, the long-range coupling element is then simply given as V
long
=
∫ dr ∫ dr ′
ρDT (r )ρAT (r′) S1
Sn
|r − r′|
(3)
The integral over the transition density is evaluated using the so-called transition density cube (TDC) method,54 within which the continuous transition density is integrated numerically on a 3D grid of finite-sized cubic volume elements. For this task, we implemented a TDC approach into a development version of Q-Chem 4.0 because here the long-range coupling requires the integration of the transition densities over an enormous volume with a fine grid due to the high diffuseness of the excited states. (See the Supporting Information.) Using this approach to evaluate the ICD rate according to eq 1, two more aspects require attention. One aspect is the lifetime of the donor state, that is, the S1 state of HDF prior to VPE. If this state is long-lived and the rate of VPE is low, then the excited HDF molecules have sufficient time to structurally relax into their equilibrium structures on the S1 surface. The wavelength of the virtually exchanged photon to be used in eq 1 should then correspond to the fluorescence wavelength of HDF. If the ICD rate is very high and the VPE mechanism is ultrafast, then HDF has practically no time to relax structurally and the wavelength of the exchanged photon should correspond to the absorption wavelength of HDF. Of course, these are the two extreme cases and most ICD processes are intermediate cases in reality. Therefore, the ICD rate of the HDF·FADH− system has been computed for both (Table 1). The second aspect concerns the number of excited FADH− states to be included within the summation occurring in eq 1. Here, in principle, all accessible states need to be included, and hence the summation can be limited to all states with an excitation energy below 4 eV because energetically higher lying states are not accessible. 946
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
The Journal of Physical Chemistry Letters
Letter
Table 1. Calculated ICD Rates between HDF and FADH− in 109 s−1 Depending on the Diffuseness of the Basis Set and the Choice of Absorption (A) or Fluorescence (F) Wavelength of HDF basis set
A
F
cc-pVDZ +1s,+1p +2s,+2p +3s,+3p +4s,+4p
2.32 1.81 1.81 1.77 1.68
1.66 1.45 1.41 1.39 2.63
low kinetic energies because the energy difference between the initial excitation energy and the EDE is small. One may argue that in photolyases and in biological systems in general not a true free electron is created because FADH− is embedded into a protein environment and the electron is eventually funneled toward the acceptor, the DNA photolesion, triggering its repair. From a supermolecular perspective, the final state hence corresponds to an electronically stable electron transfer state. However, the initial excitation of FADH− is a local event. Then, for an understanding of the electronic structure and the underlying physical processes, the global EDE is not relevant, but moreover the local EDE of FADH− is relevant. Here we have clearly demonstrated that the initially excited state of FADH− lies clearly above its EDE, and thus the physical mechanism generating the free or, in the protein, “solvated” electron corresponds to an intermolecular Coulombic decay process.
The calculated values for the ICD rate are given in Table 1. In general, the computed rates using the fluorescence wavelength of HDF are generally smaller than the rates obtained when the absorption wavelength is employed in eq 1. Only the most diffuse basis set is an exception; however, this can be due to numerical limitations in the evaluation of the long-range coupling element (eq 3) between the HDF S1 and the very diffuse FADH− states via the TDC method and thus may not have physical meaning. Neglecting this outlying value, the calculated ICD rates converge to values between 1.7 × 109 and 1.4 × 109 s−1 depending on whether the absorption or fluorescence wavelengths and transition dipole moments of HDF are employed. Considering the level of approximation in the underlying theoretical methodology, this is in remarkable agreement with reported experimental values for the energy transfer rate in photolyases. Using time-resolved fluorescence spectroscopy, rates of 1.9 × 1010 and 4.6 × 109 s−1 could be determined for the energy transfer from HDF to FADH2 in Anacystis nidulans photolyase and from 10-methenyltetrahydrofolate (MTHF) to FADH2 in Escherichia coli photolyase.55,56 The experimental numbers most likely represent upper bounds to the rates because the reduction of free fluorescence of HDF in the presence of FADH2 has been assumed to be exclusively due to VPE neglecting other possible decay channels.55,56 On the contrary, the calculated ICD rates most likely correspond to lower bounds because it is assumed that excited HDF transfers a virtual photon exclusively from S1. Our findings are also in excellent agreement with previous theoretical analyses of VPE between HDF and FADH−.45 There, the accepting state of FADH− is treated as the bound electronic state and VPE is treated as Förster EET in the dipole approximation using the semiempirical quantum chemical ZINDO/S method and including spectral line-broadening via thermal fluctuations.45 Already in this simplified approach an efficiency of VPE of 95% has been found, in good agreement with the experimental estimate of 98%.55 Here it has been demonstrated that the electron catalyzing the repair of DNA lesions in photolyases is generated via an ICD process between HDF and FADH− and not, as it has been assumed until now, via a “classical” Förster-type EET followed by electron detachment. The key difference between these seemingly equivalent processes is the electronically unbound nature of the accepting state in the ICD process, which makes the exchange of energy especially efficient because the outgoing electron can always take up the right amount of kinetic energy to ensure perfect resonance conditions. The unbound nature of the accepting state requires special theoretical attention because this resonance is embedded into the continuum of states of the free electron (with arbitrary energy) and the FADH radical. It is worth noting that the emitted electron will generally have very
■
ASSOCIATED CONTENT
S Supporting Information *
Computational details, electron detachment energies of FADH− and its molecular model as well as the ionization potential of HDF, ten lowest excited singlet states of the reduced HDF model and FADH−, five energetically lowest excited singlet states of the reduced FADH− model and HDF, and molecular structure of HDF and FADH−. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work has been performed within the Research Unit “Interatomic and Intermolecular Intercoulombic Decay” (FOR 1789) of the Deutsche Forschungsgemeinschaft.
■
REFERENCES
(1) Cederbaum, L. S.; Zobeley, J.; Tarantelli, F. Giant Intermolecular Decay and Fragmentation of Clusters. Phys. Rev. Lett. 1997, 79, 4778− 4781. (2) Zobeley, J.; Cederbaum, L. S.; Tarantelli, F. Highly Excited Electronic States of Molecular Clusters and Their Decay. J. Chem. Phys. 1998, 108, 9737−9751. (3) Boudaiffa, B.; Cloutier, P.; Hunting, D.; M. A. Huels, L. S. Resonant Formation of DNA Strand Breaks by Low-Energy (3 to 20 eV) Electrons. Science 2000, 287, 1658−1660. (4) Mucke, M.; Braune, M.; Barth, S.; Förstel, M.; Lischke, T.; Ulrich, V.; Arion, T.; Becker, U.; Bradshawand, A.; Hergenhahn, U. A Hitherto Unrecognized Source of Low Energy Electrons in Water. Nat. Phys. 2012, 6, 143−146. (5) Zobeley, J.; Tarantelli, F.; Cederbaum, L. S. Intermolecular Coulombic Decay of Molecular Clusters: Identification of the Decay Mechanism Using a New Hole-Population Analysis. J. Phys. Chem. A 1999, 103, 11145−11160. (6) Santra, R.; Cederbaum, L. S. Non-Hermitian Electronic Theory and Applications to Clusters. Phys. Rep. 2002, 368, 1−117. (7) Averbukh, V.; Demekhin, P.; Kolorenc, P.; Scheit, S.; Stoychev, S.; Kuleff, A.; Chiang, Y.-C.; Gokhberg, K.; Kopelke, S.; Sisourat, N.; Cederbaum, L. Interatomic Electronic Decay Processes in Singly and Multiply Ionized Clusters. J. Electron Spectrosc. Relat. Phenom. 2011, 183, 36−47.
947
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
The Journal of Physical Chemistry Letters
Letter
(8) Marburger, S.; Kugeler, O.; Hergenhahn, U.; Mö ller, T. Experimental Evidence for Interatomic Coulombic Decay in Ne Clusters. Phys. Rev. Lett. 2003, 90, 203401−203404. (9) Jahnke, T.; Czasch, A.; Schöffler, M.; Schössler, S.; Knapp, A.; Käsz, M.; Titze, J.; Wimmer, C.; Kreidi, K.; Grisenti, R.; Staudte, A.; Jagutzki, O.; Hergenhahn, U.; Schmidt-Böcking, H.; Dörner, R. Experimental Observation of Interatomic Coulombic Decay in Neon Dimers. Phys. Rev. Lett. 2004, 93, 163401−163404. (10) Stoychev, S. D.; Kuleff, A. I.; Cederbaum, L. S. Intermolecular Coulombic Decay in Small Biochemically Relevant Hydrogen-Bonded Systems. J. Am. Chem. Soc. 2011, 133, 6817−6824. (11) Hergenhahn, U. Interatomic and Intermolecular Coulombic Decay: The Early Years. J. Electron Spectrosc. Relat. Phenom. 2011, 184, 78−90. (12) Carrier, W. L.; Snyder, R. D.; Regan, J. D. The Science of Photomedicine; Plenum: New York, 1982. (13) Taylor, J.-S. DNA, Sunlight and Skin Cancer. Pure Appl. Chem. 1995, 67, 183−190. (14) Matsumura, Y.; Ananthaswamy, H. N. Front. Biosci. 2002, 7, D765−783. (15) Reardon, J. T.; Sancar, A. Recognition and Repair of the Cyclobutane Thymine Dimer, a Major Cause of Skin Cancers, by the Human Excision Nuclease. Genes Dev. 2003, 17, 2539−2551. (16) Sancar, A. Structure and Function of DNA Photolyase and Cryptochrome Blue-Light Photoreceptors. Chem. Rev. 2003, 103, 2203−2237. (17) Lin, C.; Todo, T. The Cryptochromes. Genome Biol. 2005, 6, 220−228. (18) Todo, T.; Kim, S. T.; Hitomi, K.; Otoshi, E.; Inui, T.; Morioka, H.; Kobayashi, H.; Ohtsuka, E.; Toh, H.; Ikenaga, M. Flavin Adenine Dinucleotide as a Chromophore of the Xenopus (6−4) Photolyase. Nucleic Adic Res. 1997, 25, 764−768. (19) Yi, Y.; Yi, C.; Qian, L.; Min, L.; Long, C.; Linhan, B.; Zhirong, Y.; Dairong, Q. Cloning and Sequence Analysis of the Gene Encoding (6−4) Photolyase from Dunaliella Salina. Biotechnol. Lett. 2006, 28, 309−314. (20) Glas, A. F.; Schneider, S.; Maul, M. J.; Hennecke, U.; Carell, T. Crystal Structure of the T(6−4)C Lesion in Complex with a (6−4) DNA Photolyase and Repair of UV-Induced (6−4) and Dewar Photolesions. Chemistry 2009, 15, 10387−10396. (21) Klar, T.; Kaiser, G.; Hennecke, U.; Carell, T.; Batschauer, A.; Essen, L.-O. Natural and Non-Natural Antenna Chromophores in the DNA Photolyase from Thermus Thermophilus. Chem. Biol. Chem. 2006, 7, 1798−1806. (22) Malhotra, K.; Kim, S. T.; Walsh, C.; Sancar, A. Roles of FAD and 8-Hydroxy-5-deazaflavin Chromophores in Photoreactivation byAnacystis Nidulans DNA Photolyase. J. Biol. Chem. 1992, 267, 15406−15411. (23) Heelis, P. F.; Hartman, R. F.; Rose, S. D. Energy and Electron Transfer Processes in Flavoprotein-Mediated DNA Repair. J. Photochem. Photobiol., A 1996, 95, 89−98. (24) Cheung, M. S.; Daizadeh, I.; Stuchebrukhov, A. A.; Heelis, P. F. Pathways of Electron Transfer in Escherichia coli DNA Photolyase: Trp306 to FADH. Biophys. J. 1999, 76, 1241−1249. (25) Ai, Y.-J.; Zhang, F.; Chen, S.-F.; Luo, Y.; Fang, W.-H. Importance ofthe Intramolecular Hydrogen Bond on the Photochemistry of Anionic Hydroquinone (FADH−) in DNA Photolyase. J. Phys. Chem. Lett. 2010, 1, 743−747. (26) Woiczikowski, P. B.; Steinbrecher, T.; Kumar, T.; Elstner, M. Nonadiabatic QM/MM Simulations of Fast Charge Transfer in Escherichia coli DNA Photolyase. J. Phys. Chem. B 2011, 115, 9846− 9863. (27) Liu, Z.; Guo, X.; Tan, C.; Li, J.; Kao, Y.-T.; Wang, L.; Sancar, A.; Zhong, D. Electron Tunneling Pathways and Role of Adenine in Repair of Cyclobutane Pyrimidine Dimer by DNA Photolyase. J. Am. Chem. Soc. 2012, 134, 8104−8114. (28) Sancar, A. Structure and Function of Photolyase and In Vivo Enzymology: 50th Anniversary. J. Biol. Chem. 2008, 283, 32153− 32157.
(29) Carell, T.; Burgdorf, L. T.; Kundu, L. M.; Cichon, M. The Mechanism of Action of Photolyase. Curr. Opin. Chem. Biol. 2001, 5, 491−498. (30) Essen, L.-O.; Klar, T. Light Driven DNA Repair by Photolyases. Cell. Mol. Life Sci. 2006, 63, 1266−1277. (31) Yamamoto, J.; Hitomi, K.; Hayashi, R.; Getzoff, E. D.; Iwai, S. Role of the Carbonyl Group of the (6−4) Photoproduct in the (6−4) Photolyase Reaction. Biochemistry 2009, 48, 9306−9312. (32) Borg, O. A.; Eriksson, L. A.; Durbeej, B. Electron-Transfer Induced Repair of 6−4 Photoproducts in DNA: A Computational Study. J. Phys. Chem. A 2007, 111, 2351−2361. (33) Domratcheva, T.; Schlichting, I. Electronic Structure of (6−4) DNA Photoproduct Repair Involving a Non-Oxetane Pathway. J. Am. Chem. Soc. 2009, 131, 17793−17799. (34) Sadeghian, K.; Bocola, M.; Merz, T.; Schütz, M. Theoretical Study on the Repair Mechanism of the (6−4) Photolesion by the (6− 4) Photolyase. J. Am. Chem. Soc. 2012, 132, 16285−16295. (35) Harbach, P. H. P.; Borowka, J.; Bohnwagner, M.-V.; Dreuw, A. Photolesion Repair Occurs in the Electronic Ground State of the TT Dinucleotide Dimer Radical Anion. J. Phys. Chem. Lett. 2010, 1, 2556− 2560. (36) Li, J.; Liu, Z.; Guo, X.; Wang, L.; Sancar, A.; Zhong, D. Dynamics and Mechanism of Repair of Ultraviolet-Induced (6−4) Photoproduct by Photolyase. Nature 2010, 466, 887−890. (37) Faraji, S.; Dreuw, A. Proton-Transfer-Steered Mechanism of Photolesion Repair by (6−4)-Photolyases. J. Phys. Chem. Lett. 2012, 3, 227−230. (38) Jung, Y.; Lochan, R. C.; Dutoi, A. D.; Head-Gordon, M. Scaled Opposite-Spin Second Order Møller-Plesset Correlation Energy: an Economical Electronic Structure Method. J. Chem. Phys. 2004, 121, 9793−9803. (39) Tapia, O.; Goscinski, O. Self-Consistent Reaction Field Theory of Solvent Effects. Mol. Phys. 1975, 29, 1653−1661. (40) Schutz, C. N.; Warshel, A. What Are the Dielectric Constants of Proteins and How to Validate Electrostatic Models? Proteins 2001, 44, 400−417. (41) Dreuw, A.; Head-Gordon, M. Single-Reference Ab Initio Methods for the Calculation of Excited States of Large Molecules. Chem. Rev. 2005, 105, 4009−4037. (42) Harbach, P. H. P.; Dreuw, A. In Molecular Modeling; WileyVCH: Weinheim, Germany, 2011; p 29. (43) Rokita, S. E.; Walsh, C. T. Flavin and 5-Deazaflavin Photosensitized Cleavage of Thymine Dimer: A Model of in Vivo Light-Requiring DNA Repair. J. Am. Chem. Soc. 1984, 106, 4589− 4595. (44) Todo, T.; Takemori, H.; Ryo, H.; Ihara, M.; Matsunaga, T.; Nikaido, O.; Sato, K.; Nomura, T. A New Photoreactivating Enzyme That Specifically Repairs Ultraviolet Light-Induced (6−4) Photoproducts. Nature 1993, 361, 371−374. (45) Zheng, X.; Garcia, J.; Stuchebrukhov, A. A. Theoretical Study of Excitation Energy Transfer in DNA Photolyase. J. Phys. Chem. B 2008, 112, 8724−8729. (46) Nuttall, J.; Cohen, H. L. Method of Complex Coordinates for Three-Body Calculations above the Breakup Threshold. Phys. Rev. 1969, 188, 1542−1544. (47) Balslev, E.; Combes, J. M. Spectral Properties of Many-Body Schrödinger Operators with Dilatation-Analytic Interactions. Commun. Math. Phys. 1971, 22, 280−294. (48) Simon, B. Quadratic Form Techniques and the Balslev-Combes Theorem. Commun. Math. Phys. 1972, 27, 1−9. (49) Riss, U. V.; Meyer, H.-D. Calculation of Resonance Energies and Widths Using the Complex Absorbing Potential Method. J. Phys. B 1993, 26, 4503−4513. (50) Riss, U. V.; Meyer, H.-D. Reflection-Free Complex Absorbing Potentials. J. Phys. B 1995, 28, 1475−1488. (51) Santra, R.; Cederbaum, L. S.; Meyer, H.-D. Electronic Decay of Molecular Clusters: Non-Stationary States Computed by Standard Quantum Chemistry Methods. Chem. Phys. Lett. 1999, 303, 413−419. 948
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
The Journal of Physical Chemistry Letters
Letter
(52) Eliezer, I.; Taylor, H. S.; William, J. K. Resonant States of H2−. J. Chem. Phys. 1967, 47, 2165−2178. (53) Volobuev, Y. L.; Truhlar, D. G. An MIMD Strategy for Quantum Mechanical Reactive Scattering. Comput. Phys. Commun. 2000, 128, 516−526. (54) Krueger, B. P.; Scholes, G. D.; Fleming, G. R. Calculation of Couplings and Energy Transfer Pathways Between the Pigments of LH2 by the Ab Initio Transition Density Cube Method. J. Phys. Chem. B 1998, 102, 5378−5386. (55) Kim, S. T.; Heelis, P. F.; Sancar, A. Energy Transfer (Deazaflavin → FADH2) and Electron Transfer (FADH2 → TT) Kinetics in Anacystic nidulans Photolyase. Biochemistry 1992, 31, 11244−11248. (56) Kim, S. T.; Heelis, P. F.; Okamura, T.; Hirata, Y.; Mataga, N.; Sancar, A. Determination of Rates and Yields of Interchromophore (Folate → Flavin) Energy Transfer and Intermolecular (Flavin → DNA) Electron Transfer in Escherichia coli Photolyase by TimeResolved Fluorescence and Absorption Spectroscopy. Biochemistry 1991, 30, 11262−11270.
949
dx.doi.org/10.1021/jz400104h | J. Phys. Chem. Lett. 2013, 4, 943−949
