PCCP View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2015, 17, 19797
View Journal | View Issue
Excess electron interaction with radiosensitive 5-bromopyrimidine in aqueous solution: a combined ab initio molecular dynamics and time-dependent wave-packet study† Changzhe Zhang and Yuxiang Bu* Radiation-generated secondary electrons can induce resonance processes in a target molecule and fragment it via different pathways. Although the associating electronic resonant states at equilibrium geometry have been well studied for many target molecules in the gas phase, vibrational resonance contributions and the solvent effect are still poorly understood for relevant processes in solution. Taking a radiosensitive drug, 5-bromopyrimidine (5-BrPy), as an example, we here present a combined ab initio molecular dynamics simulation and time-dependent wave packet study with an emphasis on vibrational resonance and solvation effects on excess electron interaction with 5-BrPy in solution. The gaseous results reveal two primary channels for the electron induced C–Br bond on vertical potential cleavage: the highest vibrational resonance tunneling
energy curve via a tunneling mechanism e þ 5-BrPy ! 5-BrPy ! Br þ Py , and auto-dissociation relaxation along repulsive relaxed potential energy curve e þ 5-BrPy ! 5-BrPy ! Br þ Py , which account for the two peaks at 0.2 and 0 eV observed in Modelli’s experiment. However, a strong solvation effect modifies the mechanism and dynamics of the dissociation of the electron 5-BrPy system. On one hand, the spontaneous dissociation becomes unfavorable due to a barrier on the relaxed free energy surface created
by the coupling between the p* and s* states. Seven vibrational resonances (v = 0–6) are identified for the solution process and only the high-level v = 5, 6 with non-negligible quantum tunneling coefficient tunneling localization can cause the dissociation e þ 5-BrPy ! 5-BrPy ! Brd Pyd ! Br þ Py . On the
Received 10th May 2015, Accepted 23rd June 2015 DOI: 10.1039/c5cp02693h
other hand, protonation is also observed at the N sites of the hydrated 5-BrPy anion localization relaxation e þ 5-BrPy ! 5-BrPy ! Prt-5-BrPy , and this inhibits the dissociation along the C–Br bond, suggesting a competing pathway against C–Br bond cleavage. Clearly, this work provides a combination strategy using an ab initio molecular dynamics technique and time-dependent wave packet method to explore the effects of vibrational resonances and solvation on the interaction of radio-
www.rsc.org/pccp
generated excess electrons with target biological molecules in complicated solution surroundings.
1 Introduction Radiotherapy is widely utilized against human cancer in the clinic. However, the exposure of living organisms to ionizing radiation results in lethal lesions not only towards cancer but also to normal cells. Generally, DNA molecules are the main targets in cell death by ionizing radiation and above 80% of the damage is induced by secondary species generated from an initial ionization.1–3 Low-energy (o20 eV) electrons (LEEs)
School of Chemistry and Chemical Engineering, Institute of Theoretical Chemistry, Shandong University, Jinan, 250100, People’s Republic of China. E-mail: [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/c5cp02693h
This journal is © the Owner Societies 2015
have been proven to be one of the most abundant secondary species,4–6 which gradually lose energy through inelastic collisions and eventually become solvated within 1 ps in biological media.7 Before getting solvated, LEEs can attach to a resonance of a molecular target, leading to the secondary damage. On one hand, LEEs can directly ionize components of DNA when the kinetic energy reaches the ionization thresholds of target molecules (ca. 7–8 eV), generating more reactive fragments. On the other hand, the electrons having kinetic energies below 7–8 eV and even as low as 0.1 eV which can no longer produce secondary ionizations can also induce DNA strand rupture due to dissociative electron attachment (DEA).8–15 However, the above findings raise new questions as to how to restrict patient side-effects caused by dissociative secondary LEE attachment
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19797
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
Paper
to DNA. Along this line, a set of uracil analogues have been proposed to substitute thymidine into cancer cell DNA as a radiosensitizing agent with potential clinical applications.16–24 Halogenated pyrimidines represent an important class of molecules due to the fact that they are elementary building blocks of the drugs mentioned above employed in cancer treatment. The DEA to these gas phase constituents has been the concern of several recent publications explorating the involved radiosensitization mechanism which is believed to be operating through different resonant dissociation channels. For example, Modelli and coworkers reported the electronic resonance states of halopyrimidines experimentally by electron transmission spectroscopy and theoretically by scaling the virtual orbital energies,25 and also measured the yield of negative ions with DEA spectroscopy. They attributed the observed two signals (at 0 and 0.2 eV, respectively) to electron attachments in the first two p* resonances. Thereafter, Barbosa and coworkers calculated the integral cross section for collisions of low-energy electrons with halopyrimidines by the Schwinger multichannel method.26 The computed cross sections revealed three p* and one s* resonances which is consistent with Modelli’s results. However, the p*2 resonance observed by Barbosa and coworkers lies above 0.5 eV, and is too high to account for the peak located at 0.2 eV. This is presumably due to the fact that their calculations are only performed at the equilibrium geometry of the target without considering the nuclear motions. Clearly, the close spacing between the two peaks observed in the DEA spectroscopy by Modelli and coworkers is evocative of taking the vibrational contribution into account which is inevitable for a comprehensive description of the possible damage or benefit of the DEA process. In addressing this, the time-dependent wave packet (TDWP) method should be a more appropriate tool because it has long been applied to investigate vibrational resonance structures in electron molecule interactions like electron N2, electron H2, electron CO scattering27–29 and is also becoming quite popular in the field of electron-induced biomolecule damage due to the advantages of numerical efficiency and conceptual simplicity.30–33 Through the TDWP calculations along the corresponding potential energy curves (PECs), vibrational resonance distribution as well as its contributions to the C–Br bond cleavage can be revealed in an effort to complement potential DEA pathways and provide new insights into the resonance phenomena, e.g. those in Modelli’s experiment. Furthermore, as is well known, the secondary electroninduced damage mostly occurs in the biological environment that contains 70–80% water. Thus, it is necessary to take into account the solvent effect. It should be emphasized that the DEA pathway is quite sensitive to the PEC character. Clearly, the primary challenge, compared with the gaseous case, is to properly mimic the solvent environment and construct reliable PECs for structure changes in solution. Although the general polarizable continuum model (PCM) works well for many important processes in solution, it meets large difficulties when the reactants interact strongly with solvent molecules. Thus, the ab initio molecular dynamics (AIMD) simulation technique
19798 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
PCCP
can be a tool for reasonably taking into account the solvent effect in constructing reliable PECs in solution. Now, the AIMD simulation technique has been extensively applied to the studies of solution systems because it is not only able to model water molecules at the same level as the solute and allows for protonation of solutes,34 but it can also treat the enthalpic and entropic effects. Accordingly, it should be useful to construct an accurate free energy surface (FES) along the reaction coordinate and then to investigate the vibrational resonance structures and reaction mechanism of the electron interacting systems in solution (e.g. the electron hydrated 5-BrPy here) in conjunction with the TDWP method. In this work, we reveal that, in the gas phase, quantum tunneling through high-level vibrational resonance (v = 3) on the vertical PECs and spontaneous dissociation along the relaxed PECs are the likely responsible mechanisms for the two peaks at 0.2 and 0 eV observed in Modelli’s experiment, respectively.25 However, in solution, an increased energy barrier due to the coupling between the p* and s* states exists on the FES which induces stabilization of the p*-type anion instead of cleavage of the C–Br bond. We identify 7 vibrational bound states above the C–Br dissociation limit for the fully hydrated 5-BrPy and only v = 5, 6 can contribute to the bond cleavage via the quantum tunneling path, while the rest is proven to be further stabilized by proton transfer from a surrounding water molecule to the anion. The remainder of this paper is organized as follows. Section 2 illustrates details of the used methodology, and discussion of prominent results is given thereafter. A brief summary of the investigation concludes this paper.
2 Methodology 1. Calculations and AIMD simulations of potential energy surfaces The first ingredient in our calculations is the PECs of the target. However, a full-dimensional description of the reaction is out of reach in a full quantum model. Thus, we just consider the essential C–Br dissociation coordinate and all other degrees of freedom are taken into account via full geometry relaxations. The gaseous PECs were computed using Gaussian 03 code at the DFT-B3LYP level35 with a 6-311++G(d,p) basis set for C, H, N atoms and a LANL2DZ basis set36,37 for the Br atom. However, for the solution case, the corresponding energy is free energy which can be determined by the constrained AIMD simulation and thermodynamic integration of the constraint force using the CP2K/Quickstep package.38 We construct the system including 100 H2O molecules around a 5-BrPy molecule in a cubic box with r = 1.0 g cm3 and box size = 14.82 Å. 20 sampling windows are chosen for each fixed C–Br distance which varies from 3.1 to 5.8 a0. Each window undergoes 2 ps equilibration and 2 ps production time with a typical timestep of 0.5 fs. The tempera´–Hoover chain thermostat. ture was kept at 300 K by a Nose Energies and forces were evaluated by the Becke, Lee, Yang, and Parr (BLYP) exchange–correlation function39,40 and the triple-z
This journal is © the Owner Societies 2015
View Article Online
PCCP
Paper
basis set augmented with two polarization functions (TZV2P). The constraint mean force [ f (r)] was achieved through the Shake algorithm41 by averaging the last 2 ps trajectories and was then integrated to produce the free energy profile Ðb DAa!b ¼ a h f ðrÞidr .
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
2. TDWP method for state density spectrum On the basis of the PECs or FESs, the TDWP approach was subsequently performed. This method has been described in detail previously,42–46 and here we only give an outline of it. After excess electron attachment to 5-BrPy, the nuclear motion, evolved under the influence of the anionic Hamiltonian (HA), can be described as cðR; tÞ ¼ eiHA ðRÞt=h cðR; t ¼ 0Þ
(1)
h2 2 r þ EA ðRÞ and EA(R) denoting anionic 2mA potential energies. c(R,t = 0) is the initial nuclear wavefunction in the form of " 1=4 # 2 R R0 2 cðR; t ¼ 0Þ ¼ (2) exp ikR pd0 d0 with HA ðRÞ ¼
where d0 denotes the initial width, and R0 is the initial position of the wave packet. Subsequently, the wavefunction at an arbitrary time of the system C(R,t) can be obtained by using the split-operator propagation scheme. To avoid unphysical reflection of the wave packet, we performed complex absorbing potentials at the edge of the grid. Finally, the state-density spectrum was extracted from Fourier transformation of the autocorrelation function: ð1 s¼ expðiEt=hÞhcðR; t ¼ 0ÞjcðR; tÞidt (3) 1
helpful to first know relevant information about it in the gas phase. Thus, in the following analyses, the gaseous mechanism of DEA to 5-BrPy is first discussed in detail to obtain a basic understanding of the relevant processes. Then, a complicated solution case is further discussed to clarify how the solvent effect can modify the DEA mechanism and dynamics using a combined AIMD simulation and time-dependent wave packet method. (a) Dissociative electron attachment in the gas phase AIMD simulation was first carried out to reveal the dynamics of DEA to 5-BrPy in the gas phase. An excess electron was vertically attached to 5-BrPy in its stable configuration optimized at the DFT-B3LYP level. As a result, two separate stages are distinguished. Initially, the electron attaches to the vacant p* orbital and the C–Br bond length fluctuates around RC–Br E 3.75 a0 when t o 50 fs. After a short duration of the formed p*-type transient anion (5-BrPy*), the electron transfers to a s* orbital gradually and promotes the C–Br bond dissociation into Br plus the Py radical (Fig. S1 and S2, ESI†). This phenomenon can be described by the nuclear motions along the corresponding PECs. We suggest that, in the early period (t o 50 fs), the system evolves on the vertical PEC. However, after the configuration rearrangement to adapt the electron attachment, the relaxed PEC should be responsible for the consequent evolution (t > 50 fs). As shown in Fig. 1, the relaxed PEC presents a repulsive character and the probability amplitude on it can spread to the region featuring a large C–Br bond length (RC–Br) with increasing time (Fig. S4, ESI†). In contrast, the probability amplitude in the vertical anionic state is localized at RC–Br E 3.75 a0 due to the presence of a barrier at RC–Br E 4.0 a0 with height E0.2 eV (Fig. S3, ESI†). To explore the origin of the barrier, we first examined a set of low-lying unoccupied
3. Tunneling coefficient To evaluate the vibrational state contributions to the dissociation, the transmission coefficient (T) was determined. Assuming a potential barrier with the form UðxÞ ¼
U0 cosh 2 ðaxÞ
(4)
where U0 is the height of the potential barrier, a is the width parameter, and x is the reaction coordinate, T can be obtained by using the following relation47 T¼
sinh 2 ðpk=aÞ
1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sinh 2 ðpk=aÞ þ cosh 2 p 8mU0 =h2 a2 1 2
(5)
ð2mEÞ1=2 , m is the reduced mass and E is the energy h of the bound vibrational state.
where k ¼
3 Results and discussion Although the main objective of this work is to explore the interaction of an excess electron with 5-BrPy in solution, it is
This journal is © the Owner Societies 2015
Fig. 1 Virtual orbital energies and potential energy curves as a function of C–Br distance. (upper panel) Corresponding virtual orbital energies of the optimized neutral molecules. (lower panel) The vertical and relaxed anionic potential energy curves for 5-BrPy.
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19799
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
Paper
PCCP
Fig. 2 Energy spectrum of state-density for anionic 5-Brpy extracted from the wave packet propagation calculations: (A) in the gas phase; (B) in the aqueous phase.
molecular orbitals (LUMOs) of neutral 5-BrPy around RC–Br = 4.0 a0 (Tables S1 and S2, ESI†), and observed that the state with a shorter C–Br bond before the barrier [RC—Br o 4.0 a0 (2.12 Å)] shows a p* character, while the state in the region after the barrier, i.e., a larger C–Br bond (RC–Br > 4.0 a0) becomes of the repulsive C–Br s* character. In addition, we illustrated a series of virtual molecular orbital energies, which include two empty p* MOs and one s* MO, of neutral 5-BrPy varying with RC–Br (Fig. 1, upper panel). Two avoided crossing points are detected: one is between the s* MO and the p*2 MO at RC–Br E 3.8 a0, and the other is between the s* MO and the p*1 MO at RC–Br E 4.0 a0. Clearly, the above results provide evidence for the fact that the coupling of the p*1 and s* states around RC–Br E 4.0 a0 is responsible for a kinetic barrier.48 It is apparent that transfer of an excess electron from the p*1 to s* state and destabilization of the C–Br bond occur only if these two states are equal in energy. To reach the point on the corresponding PEC, the C–Br bond has to elongate, which suggests the importance of taking into account the vibrational resonance contribution. Based on the PEC, we carried out the TDWP calculations, generating a vibrational state-density spectrum for anionic 5-BrPy as shown in Fig. 2(A). As we can see, despite the somewhat quasi-continuous peaks just above the barrier, four relatively sharp spectral peaks are identified in the energy range below the dissociation barrier at RC–Br E 4.0 a0 which represent the corresponding bound vibrational resonances located on
Table 1
the vertical PEC. To further understand their contributions to the dissociation, we computed relevant transmission coefficients (T) for the vibrational states (Table 1) using eqn (5). In general, the barrier width a from the PECs is difficult to ensure, which stimulated us to use a model potential barrier with the form of eqn (4) for finding out the same. Different combinations of U0 and a were tested, and the best choice (U0 = 6.6 103 a.u., a = 14.5 a01) was compared with the anionic PEC in Fig. S5 (ESI†). It can be seen from Table 1 that the T values for v = 0–2 are negligibly small, while for v = 3 it is significant. This seems reasonable because the reduced mass of the C–Br bond is too large to allow the probability density from states (v = 0–2) to flow towards the dissociative region. However, the higher resonance [v = 3, E = 6.1 103 a.u. (0.17 eV)] corresponds to high vibrational energies and a low barrier width which weakens the effect of the reduced mass of the C–Br bond and enhances the tunneling probability. The non-negligible T value of v = 3 indicates a potential path for the C–Br dissociation along the vertical anionic PEC. Compared with Modelli’s experiment, we have demonstrated that, in the gas phase, an excess electron immediately occupies the p*1 orbital of 5-BrPy to form 5-BrPy* and this induces the molecule to dissociate through two pathways. On one hand, the electron induces vibrational excitation on the vertical PEC allowing C–Br bond cleavage through the quantum tunneling pathway. The calculated resonance position (0.17 eV) is quite close to the experimental value (0.2 eV). On the
Transmission coefficient (T) for anionic vibrational states calculated using eqn (5)
Gas phase Vibrational state
Energy (a.u.)
0 1 2 3 4 5 6
8.6 2.4 4.2 6.1 — — —
104 103 103 103
Aqueous phase Tunneling coefficient (T) 2.64 6.91 8.72 4.63 — — —
19800 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
104 103 102 101
Energy (a.u.) 7.0 2.1 3.4 4.7 6.1 7.4 8.7
104 103 103 103 103 103 103
Tunneling coefficient (T) 1.07 2.04 3.94 5.18 7.10 7.95 4.00
107 106 105 104 103 102 101
This journal is © the Owner Societies 2015
View Article Online
PCCP
other hand, geometry relaxation of the system invokes a spontaneous electron state conversion from the p*1 to s* orbital, leading to a dissociation along the relaxed PEC. Clearly, this mechanism can account for the 0 eV signal.
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
(b)
Electron-impacted dissociation in solution
It is well known, however, that the secondary electron induced damage mainly occurs in the biological microsurrounding or aqueous solution. Therefore, it is necessary to include the solvent effect in the extended calculations to clarify how the solvent effect modifies DEA to 5-BrPy. To model the process in aqueous solution, we constructed a system containing 100 H2O molecules around a target 5-BrPy. After a 3 ns equilibration with a classical molecular dynamics simulation, a representative configuration was chosen for another 3 ps AIMD simulation to obtain an equilibrated solution system. Further, AIMD simulation was conducted for the negatively charged system by considering periodic boundary condition for 1 ps with the initial conformation extracted from the neutral AIMD equilibrium trajectory mentioned above. Fig. 3 and 4(A) illustrate the time evolution of
Fig. 3 Time evolutions of the C–Br bond length and the spin density on 5-BrPy after vertically adding an excess electron to the system.
Paper
spin density on 5-BrPy and related snapshot configurations, respectively. Different from that in the gas phase, a vertically injected excess electron distributes over the surrounding water molecules and the target molecule, exhibiting a delocalized state. During about 40 fs of relaxation,22 it is gradually concentrated in the p* orbital of 5-BrPy with an elongation of the C–Br bond from 3.65 a0 (the bond length of neutral 5-BrPy) to 3.75 a0 (that of the corresponding anion, Fig. 3). This observation indicates that (a) localization of an excess electron on the fully hydrated 5-BrPy occurs simultaneously with structure relaxation of the system, i.e. reorientation of the solvent molecules and structural fluctuation of the pyrimidine base; (b) the stable p*-type anion state is preferred after electron attachment rather than auto-dissociation as observed in the gas phase. These conclusions were also confirmed by additional AIMD simulations starting with different initial configurations, exchange–correlation functions, and basis sets (Fig. S6–S9, ESI†). Clearly, the observed difference in the preliminary products due to excess electron attachment to 5-BrPy in the gas phase and aqueous solution should be attributed to their different PECs regarding the electron-impacted C–Br bond variations. In particular, in solution, the relevant thermodynamic energy curve should be the FES which takes energy exchange with the environment into account. Thus, we constructed the FES by using the thermodynamic integration method. A series of AIMD simulations with fixed C–Br bond lengths were performed. The timeaveraged constraint force [ f (r)] (Fig. S17, ESI†) was employed as Ðb the mean force for the integration DAa!b ¼ a h f ðrÞidr .49 Fig. 5 displays the constrained mean forces (white circle) and their integration along the C–Br stretching coordinate (black circle). It is found that there is an energy barrier with a height of B0.25 eV on the relaxed FES at RC–Br E 4.2 a0. This observation is different from that (no barrier) in the gas phase. In fact, as shown earlier, the barrier comes from the coupling between the p* and s* states during electron transfer. To confirm the validity of this picture in solution, we examined the relevant molecular orbitals and energies both in PCM (Fig. S10, S11 and Table S3, ESI†) and
Fig. 4 Snapshots at different times extracted from: (A) trajectory after an excess electron is vertically added to the solvated 5-Brpy; (B) trajectory about the proton transfer process. The green and yellow opaque shades are for positive and negative parts of the wave function (isovalue = 0.02), respectively.
This journal is © the Owner Societies 2015
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19801
View Article Online
Paper
PCCP
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
that in the gas phase. This observation should be attributed to the fact that the solvent effect can considerably stabilize the anionic complex, leading to a high and wide barrier which weakens the probability density from p* to s* states. Overall, we suggest that the electron attachment to aqueous 5-BrPy can be divided into two cases: attachment of an electron with energy E > 0.2 eV can excite the system to high-level vibrational states (v = 5, 6) and induces the C–Br bond cleavage via a quantum tunneling pathway tunneling localization e þ 5-BrPy ! 5-BrPy ! Brd Pydd ! Br þ Py ; while attachment of an electron with E o 0.2 eV can produce a stable p*-type anion but not the further dissociated products (the s*-type anionic complex or separate hydrated Br and Py ). (c) Fig. 5 The constrained mean forces and their integration with respect to the C–Br bond length of 5-BrPy.
explicit water environment (Fig. S12–S16, ESI†) which present similar behaviors to those in the gas phase. In addition, it should be noted that another barrier occurs when stretching the C–Br bond beyond RC–Br = B5.3 a0. Clearly, this barrier corresponds to that of dissociation of the solvent-stabilized s*-type complex into Br plus the Py radical in solution. To further prove the existence of such a barrier, we performed an AIMD simulation on the system with the C–Br bond length being RC–Br > 5.0 a0 (Fig. S18 and S19, ESI†), and found that the excess electron steadily resides in the s* orbital and the C–Br bond fluctuates around RC–Br = B5.3 a0, neither being considerably elongated towards the dissociated state nor being further shortened to go back to the stable anionic complex structure. This observation indicates that a barrier separates the s*-type anionic complex (Brd Pyd) and the dissociated products (Br + Py ). This s*-type anion complex is a stable ionic pair connected by a 3-electron bond with a bond length of about 5.2 a0 (2.8 Å) and the excess electron is shared by two moieties (Br and Py). This is reasonable because the Py radical has considerably large electron affinity that is further enhanced by solvation (Table S4, ESI†), so that the attached excess electron only partially transfers to Br to form the (Brd Pyd) state when the C–Br bond elongates. Clearly, the above mentioned reliable FES provides us with the basis to explore the resonance distribution of the fully solvated 5-BrPy anion. Fig. 2(B) represents the calculated state-density spectrum as a function of the energy produced by Fourier-transforming the time autocorrelation function. Compared with the gas phase results, we identify seven separate sharp resonance peaks located below the dissociation threshold. That is, there are seven vibrational resonance states (v = 0–6) for solvated anionic 5-BrPy in solution. Similar to the calculations in the gas phase, we introduce a model as eqn (4) to mimic the barrier and the best choice (U0 = 9 103 a.u., a = 2.9 a01) is utilized (Fig. S20, ESI†). The calculated T values for the states are collected in Table 1, indicating that they are negligible for the v = 0–4 states, non-negligible for the v = 5 state, and significant for the v = 6 state. It should be noted that the transmission coefficients in solution are generally smaller than
19802 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
Proton transfer effect
In fact, the generated intermediate, p*-type anion, is metastable because two possible processes can develop. One is the further dissociation through a tunneling pathway, as mentioned above, while another is protonation of the negatively charged 5-BrPy at its N site through a proton transfer mechanism from an adjacent solvent H2O molecule. The associated dynamics with this proton transfer reaction can be monitored by time evolution of the Mulliken charge distribution and the Nb–Hw distance as shown in Fig. 6. Analysis of the dynamics trajectory suggests a threestep process. (a) The electron localization stage. At the beginning (t o 100 fs), the injected excess electron gradually localizes towards 5-BrPy and at the same time the Nb–Hw distance continuously decreases due to attraction of the Nb site with the gradually increased extra negative charge. (b) The formation and lasting stage of the H-bond. The H-bond between the Nb of 5-BrPy and Hw of an adjacent H2O molecule is gradually formed at 100–520 fs and then fluctuates steadily at B1.6 Å. In this process, the H-bond network of the surrounding water molecules constantly reorients. (c) The proton transfer stage. Spontaneous proton transfer is triggered upon an adequate hydration structure that is established and finished rapidly within about 30 fs (in 520–550 fs) with protonation at the Nb site of 5-BrPy. This process can be described by the radial distribution
Fig. 6 Time evolution of the Nb–Hw distance (RNb–Hw) and the charge distributions on the corresponding parts. The subscript b denotes the N atom of 5-BrPy and the subscript w denotes the water molecule releasing a proton.
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
PCCP
Paper
sensitizer, 5-BrPy, using a combined AIMD simulation and time-dependent wave packet method in this work as well as potential reaction pathways. In particular, an explicit water environment is considered to emphasize the importance of the solvent effect in scaling the electron interaction with solvent H2O molecules and the target 5-BrPy. In the gas phase, an excess electron directly attaches to the 5-BrPy target molecule and occupies its p* orbital to form a transient anion, and induces the C–Br bond dissociation via two primary channels: (i) a quantum tunneling mechanism through the highest vibrational resonance (v = 3) on a vertical PEC tunneling e þ 5-BrPy ! 5-BrPy ! Br þ Py , and (ii) dissociation relaxation along a relaxed PEC e þ 5-BrPy ! 5-BrPy ! Br þ Py . Fig. 7 Time evolution of radial distribution functions (g(r)) between Ow of the H2O releasing a proton and H of other surrounding water molecules. At 520–550 fs, the peak shifts towards short Ow H distance, suggesting that proton transfer occurs.
functions (RDFs)50,51 between the Or of the H2O releasing a proton and its surrounding water hydrogen atoms. Fig. 7 provides the time evolution of RDFs extracted from the proton transfer trajectory. It can be seen that during the formation of the H-bond (t o 520 fs), the water H atoms around the Or mainly distribute with ROr–H > 1.75 Å. However, the RDF peak moves to the region at B1.6 Å (ROr–H) when the proton transfer occurs (t = 520–550 fs), meaning that strong H-bonds are formed between the Or and its surrounding water molecules as well as sharing the Or attraction with the Hr and the promoted Hr tends to the Nb site of 5-BrPy. The protonation was also observed at another N site in a parallel AIMD simulation which yields a similar conclusion (Fig. S21–S24, ESI†). Encouraged by the above analyses, we further evaluated the proton transfer influence on DEA. Examination reveals that, after protonation, the excess electron still occupies the p* orbital, but the center of electron spin density moves towards the Nb (from 1.3 to 1.2 Å) and leaves the Br (from 3.2 to 3.3 Å) (Fig. S25, ESI†). Clearly, on one hand, the shift of the spin density center weakens the electron Br interaction. In fact, shortening of the average C–Br bond length is observed from 1.98 Å to 1.93 Å [the C–Br bond length (3.65 a0) of neutral 5-BrPy in solution] after protonation (Fig. S26, ESI†). On the other hand, the shift of the spin density center stabilizes the anion with an enhancement of the barrier (Fig. S27, ESI†) to conversion from the p* to s* state (5-BrPy* - [Brd Pyd]). However, it has little effect on further conversion from the s*-type anion to the completely dissociated Br plus Py radical ([Brd Pyd] - Br + Py ) because the protonated 5-BrPy is deprotonated when the C–Br bond (RC–Br) is stretched beyond 4.2 a0 (Fig. S28 and S29, ESI†). This observation about deprotonation is attributed to the excess electron transfer to the s* orbital located at the Br–C bond zone and the excess electron no longer distributes at the Nr.
4 Conclusion In summary, we explored the vibrational resonance contributions to the mechanism of DEA of the selected tumor-specific
This journal is © the Owner Societies 2015
We suggest that these two channels are responsible for the two peaks at 0.2 and 0 eV observed experimentally by Modelli and coworkers. In contrast, in solution, a vertically injected excess electron is found to distribute on the target 5-BrPy and solvent H2O molecules in a delocalized state and then localizes on 5-BrPy within 40 fs along with a solution structure relaxation. The formed metastable hydrated anion does not spontaneously dissociate due to a barrier on the relaxed FES originating from the coupling between the p* and s* states. Seven resonant states (v = 0–6) are identified on the FES and only the high-level v = 5, 6 with non-negligible quantum tunneling probability can lead to the dissociation tunneling localization eþ5-BrPy ! 5-BrPy ! Brd Pydd !Br þPy ; while the other low-level resonant states (v = 0–4) lead to localization protonation at a N site of 5-BrPy eþ5-BrPy ! relaxation
5-BrPy ! Prt-5-BrPyÞ, inhibiting the C–Br bond cleavage. The protonation undergoes a three-step process and is sensitive to H-bond network rearrangement among the surrounding water molecules. It should be noted that proton transfer from a surrounding H2O molecule to a N site of 5-BrPy is actually an electron localization coupled nonsynchronous process and also presents a competition with the C–Br bond cleavage by neutralizing the anion core. This work provides a deep understanding of the radio-sensitivity of the 5-BrPy drug in solution and reveals the important impact of the low energy electron triggered vibrational excitations and also the importance of the solvent effect.
Acknowledgements This work was supported by NSFC (21373123, 20633060, and 20973101), NSF (ZR2013BM027) of Shandong Province. A part of the calculations were carried out at National Supercomputer Center in Jinan, Shanghai Supercomputer Center, and HighPerformance Supercomputer Center at SDU-Chem.
References 1 P. P. Connell, S. J. Kron and R. R. Weichselbaum, Relevance and irrelevance of DNA damage response to radiotherapy, DNA Repair, 2004, 3, 1245–1251.
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19803
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
Paper
2 L. Sanche, Low energy electron-driven damage in biomolecules, Eur. Phys. J. D, 2005, 35, 367–390. 3 E. Alizadeh and L. Sanche, Precursors of Solvated Electrons in Radiobiological Physics and Chemistry, Chem. Rev., 2012, 112, 5578–5602. 4 M. Toulemonde, E. Surdutovich and A. V. Solov’yov, Temperature and pressure spikes in ion-beam cancer therapy, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2009, 80, 031913. 5 E. Scifoni, E. Surdutovich and A. V. Solov’yov, Spectra of secondary electrons generated in water by energetic ions, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2010, 81, 021903. 6 S. M. Pimblott and J. A. LaVerne, Production of low-energy electrons by ionizing radiation, Radiat. Phys. Chem., 2007, 76, 1244–1247. 7 G. Hanel, B. Gstir, S. Denifl, P. Scheier, M. Probst, ¨rk, B. Farizon, M. Farizon, E. Illenberger and T. D. Ma Electron Attachment to Uracil: Effective Destruction at Subexcitation Energies, Phys. Rev. Lett., 2003, 90, 188104. 8 J. Gu, J. Wang and J. Leszczynski, Electron AttachmentInduced DNA Single Strand Breaks: C3 0 –O3 0 s-Bond Breaking of Pyrimidine Nucleotides Predominates, J. Am. Chem. Soc., 2006, 128, 9322–9323. 9 Y. Zheng, P. Cloutier, D. J. Hunting, J. R. Wagner and L. Sanche, Glycosidic Bond Cleavage of Thymidine by LowEnergy Electrons, J. Am. Chem. Soc., 2004, 126, 1002–1003. 10 X. Li, M. D. Sevilla and L. Sanche, Density Functional Theory Studies of Electron Interaction with DNA: Can Zero eV Electrons Induce Strand Breaks?, J. Am. Chem. Soc., 2003, 125, 13668–13669. 11 I. Anusiewicz, J. Berdys, M. Sobczyk, P. Skurski and J. Simons, Effects of Base p-Stacking on Damage to DNA by Low-Energy Electrons, J. Phys. Chem. A, 2004, 108, 11381–11387. 12 I. Anusiewicz, M. Sobczyk, J. Berdys-Kochanska, P. Skurski and J. Simons, A Theoretical Model for Indirect Dissociative Electron Attachment, J. Phys. Chem. A, 2004, 109, 484–492. 13 J. Berdys, P. Skurski and J. Simons, Damage to Model DNA Fragments by 0.25–1.0 eV Electrons Attached to a Thymine p* Orbital, J. Phys. Chem. B, 2004, 108, 5800–5805. 14 C.-R. Wang, J. Nguyen and Q.-B. Lu, Bond Breaks of Nucleotides by Dissociative Electron Transfer of Nonequilibrium Prehydrated Electrons: A New Molecular Mechanism for Reductive DNA Damage, J. Am. Chem. Soc., 2009, 131, 11320–11322. 15 B. Boudaı¨ffa, P. Cloutier, D. Hunting, M. A. Huels and L. Sanche, Resonant Formation of DNA Strand Breaks by Low-Energy (3 to 20 eV) Electrons, Science, 2000, 287, 1658–1660. 16 C.-R. Wang and Q.-B. Lu, Real-Time Observation of a Molecular Reaction Mechanism of Aqueous 5-Halo-2 0 deoxyuridines under UV/Ionizing Radiation, Angew. Chem., Int. Ed., 2007, 46, 6316–6320. 17 S. Cecchini, S. Girouard, M. A. Huels, L. Sanche and D. J. Hunting, Interstrand Cross-Links: A New Type of g-Ray Damage in Bromodeoxyuridine-Substituted DNA, Biochemistry, 2005, 44, 1932–1940.
19804 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
PCCP
18 T. Chen, G. P. Cook, A. T. Koppisch and M. M. Greenberg, Investigation of the Origin of the Sequence Selectivity for the 5-Halo-2 0 -deoxyuridine Sensitization of DNA to Damage by UV-Irradiation, J. Am. Chem. Soc., 2000, 122, 3861–3866. 19 X. Li, L. Sanche and M. D. Sevilla, Dehalogenation of 5-Halouracils after Low Energy Electron Attachment: A Density Functional Theory Investigation, J. Phys. Chem. A, 2002, 106, 11248–11253. 20 X. Li, M. D. Sevilla and L. Sanche, DFT Investigation of Dehalogenation of Adenine–Halouracil Base Pairs upon Low-Energy Electron Attachment, J. Am. Chem. Soc., 2003, 125, 8916–8920. 21 L. Chomicz, M. Zdrowowicz, F. Kasprzykowski, J. Rak, A. Buonaugurio, Y. Wang and K. H. Bowen, How to Find Out Whether a 5-Substituted Uracil Could Be a Potential DNA Radiosensitizer, J. Phys. Chem. Lett., 2013, 4, 2853–2857. ´r, P. Wityk, J. Czub, L. Chomicz and J. Rak, A first22 M. Wieczo principles study of electron attachment to the fully hydrated bromonucleobases, Chem. Phys. Lett., 2014, 595–596, 133–137. 23 H. Abdoul-Carime, M. A. Huels, E. Illenberger and L. Sanche, Sensitizing DNA to Secondary Electron Damage: Resonant Formation of Oxidative Radicals from 5-Halouracils, J. Am. Chem. Soc., 2001, 123, 5354–5355. 24 S. D. Wetmore, R. J. Boyd and L. A. Eriksson, A theoretical study of 5-halouracils: electron affinities, ionization potentials and dissociation of the related anions, Chem. Phys. Lett., 2001, 343, 151–158. 25 A. Modelli, P. Bolognesi and L. Avaldi, Temporary Anion States of Pyrimidine and Halopyrimidines, J. Phys. Chem. A, 2011, 115, 10775–10782. 26 A. S. Barbosa and M. H. F. Bettega, Shape resonances in low-energy-electron collisions with halopyrimidines, J. Chem. Phys., 2013, 139, 214301. 27 M. Sarma, S. Adhikari and M. K. Mishra, Simple systematization of vibrational excitation cross-section calculations for resonant electron-molecule scattering in the boomerang and impulse models, J. Chem. Phys., 2007, 126, 044309. 28 M. Sarma, S. Adhikari and M. K. Mishra, Mechanistic investigation of vibrational fine structure in e-H2 scattering using local complex potential-based time dependent wave packet approach, Int. J. Quantum Chem., 2008, 108, 1044–1051. 29 R. Singh, M. Sarma, A. Jain, S. Adhikari and M. Mishra, Calculation of vibrational excitation cross-sections in resonant electron-molecule scattering using the time-dependent wave packet (TDWP) approach with application to the 2P CO shape resonance, J. Chem. Sci., 2007, 119, 385–389. 30 T. Takayanagi, T. Asakura and H. Motegi, Theoretical Study on the Mechanism of Low-Energy Dissociative Electron Attachment for Uracil, J. Phys. Chem. A, 2009, 113, 4795–4801. 31 R. Bhaskaran and M. Sarma, Effect of quantum tunneling on single strand breaks in a modeled gas phase cytidine nucleotide induced by low energy electron: a theoretical approach, J. Chem. Phys., 2013, 139, 045103. 32 S. Bhowmick, R. Bhaskaran, M. K. Mishra and M. Sarma, Investigation of dissociative electron attachment to 2 0 -deoxycytidine-3 0 -monophosphate using DFT method and
This journal is © the Owner Societies 2015
View Article Online
PCCP
33
34
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
35
36
37
38
39
40
41
time dependent wave packet approach, J. Chem. Phys., 2012, 137, 064310. R. Bhaskaran, S. Bhowmick, M. K. Mishra and M. Sarma, Low-Energy Electron-Induced Single Strand Breaks in 2 0 -Deoxycytidine-3 0 -monophosphate Using the Local Complex Potential Based Time-Dependent Wave Packet Approach, J. Phys. Chem. A, 2011, 115, 13753–13758. M. Smyth and J. Kohanoff, Excess Electron Interactions with Solvated DNA Nucleotides: Strand Breaks Possible at Room Temperature, J. Am. Chem. Soc., 2012, 134, 9122–9125. M. Frisch, G. Trucks, H. Schlegel, G. Scuseria, M. Robb, J. Cheeseman, J. Montgomery Jr, T. Vreven, K. Kudin and J. Burant, Gaussian 03, Revision D. 01, Gaussian. Inc., Wallingford, CT, 2004. W. R. Wadt and P. J. Hay, Ab initio effective core potentials for molecular calculations. Potentials for main group elements Na to Bi, J. Chem. Phys., 1985, 82, 284–298. S. P. Ananthavel and M. Manoharan, A theoretical study on electron donor–acceptor complexes of Et2O, Et2S and Me3N with interhalogens, I–X (X = Cl and Br), Chem. Phys., 2001, 269, 49–57. J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing and J. Hutter, Quickstep: fast and accurate density functional calculations using a mixed Gaussian and plane waves approach, Comput. Phys. Commun., 2005, 167, 103–128. A. D. Becke, Density-functional exchange-energy approximation with correct asymptotic behavior, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098–3100. C. Lee, W. Yang and R. G. Parr, Development of the ColleSalvetti correlation-energy formula into a functional of the electron density, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789. J.-P. Ryckaert, G. Ciccotti and H. J. C. Berendsen, Numerical integration of the cartesian equations of motion of a system
This journal is © the Owner Societies 2015
Paper
42
43
44
45
46
47
48 49
50 51
with constraints: molecular dynamics of n-alkanes, J. Comput. Phys., 1977, 23, 327–341. T. Takayanagi and A. Wada, Theoretical prediction of the lifetime of the metastable helium compound: HHeF, Chem. Phys. Lett., 2002, 352, 91–98. T. Takayanagi, H. Motegi, Y. Taketsugu and T. Taketsugu, Accurate ab initio electronic structure calculations of the stable helium complex: HeBeO, Chem. Phys. Lett., 2008, 454, 1–6. Q.-T. Meng, G.-H. Yang and K.-L. Han, Time-dependent wave packet approach to Rabi oscillation in strong laser field, Int. J. Quantum Chem., 2003, 95, 30–36. Q.-T. Meng, G.-H. Yang, H.-L. Sun, K.-L. Han and N.-Q. Lou, Theoretical study of the femtosecond-resolved photoelectron spectrum of the NO molecule, Phys. Rev. A: At., Mol., Opt. Phys., 2003, 67, 063202. R.-F. Lu, P.-Y. Zhang and K.-L. Han, Attosecond-resolution quantum dynamics calculations for atoms and molecules in strong laser fields, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2008, 77, 066701. R. Marom, C. Levi, T. Weiss, S. Rosenwaks, Y. Zeiri, R. Kosloff and I. Bar, Quantum Tunneling of Hydrogen Atom in Dissociation of Photoexcited Methylamine, J. Phys. Chem. A, 2010, 114, 9623–9627. J. Simons, How Do Low-Energy (0.1–2 eV) Electrons Cause DNA-Strand Breaks?, Acc. Chem. Res., 2006, 39, 772–779. Y. Mei, D. M. Sherman, W. Liu and J. Brugger, Ab initio molecular dynamics simulation and free energy exploration of copper(I) complexation by chloride and bisulfide in hydrothermal fluids, Geochim. Cosmochim. Acta, 2013, 102, 45–64. J. Cohn, Theory of the radial distribution function, J. Phys. Chem., 1968, 72, 608–616. B. L. Nguyen and B. M. Pettitt, Effects of acids, bases and heteroatoms on proximal radial distribution functions for proteins, J. Chem. Theory Comput., 2015, 11, 1399–1409.
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19805
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2015, 17, 19797
View Journal | View Issue
Excess electron interaction with radiosensitive 5-bromopyrimidine in aqueous solution: a combined ab initio molecular dynamics and time-dependent wave-packet study† Changzhe Zhang and Yuxiang Bu* Radiation-generated secondary electrons can induce resonance processes in a target molecule and fragment it via different pathways. Although the associating electronic resonant states at equilibrium geometry have been well studied for many target molecules in the gas phase, vibrational resonance contributions and the solvent effect are still poorly understood for relevant processes in solution. Taking a radiosensitive drug, 5-bromopyrimidine (5-BrPy), as an example, we here present a combined ab initio molecular dynamics simulation and time-dependent wave packet study with an emphasis on vibrational resonance and solvation effects on excess electron interaction with 5-BrPy in solution. The gaseous results reveal two primary channels for the electron induced C–Br bond on vertical potential cleavage: the highest vibrational resonance tunneling
energy curve via a tunneling mechanism e þ 5-BrPy ! 5-BrPy ! Br þ Py , and auto-dissociation relaxation along repulsive relaxed potential energy curve e þ 5-BrPy ! 5-BrPy ! Br þ Py , which account for the two peaks at 0.2 and 0 eV observed in Modelli’s experiment. However, a strong solvation effect modifies the mechanism and dynamics of the dissociation of the electron 5-BrPy system. On one hand, the spontaneous dissociation becomes unfavorable due to a barrier on the relaxed free energy surface created
by the coupling between the p* and s* states. Seven vibrational resonances (v = 0–6) are identified for the solution process and only the high-level v = 5, 6 with non-negligible quantum tunneling coefficient tunneling localization can cause the dissociation e þ 5-BrPy ! 5-BrPy ! Brd Pyd ! Br þ Py . On the
Received 10th May 2015, Accepted 23rd June 2015 DOI: 10.1039/c5cp02693h
other hand, protonation is also observed at the N sites of the hydrated 5-BrPy anion localization relaxation e þ 5-BrPy ! 5-BrPy ! Prt-5-BrPy , and this inhibits the dissociation along the C–Br bond, suggesting a competing pathway against C–Br bond cleavage. Clearly, this work provides a combination strategy using an ab initio molecular dynamics technique and time-dependent wave packet method to explore the effects of vibrational resonances and solvation on the interaction of radio-
www.rsc.org/pccp
generated excess electrons with target biological molecules in complicated solution surroundings.
1 Introduction Radiotherapy is widely utilized against human cancer in the clinic. However, the exposure of living organisms to ionizing radiation results in lethal lesions not only towards cancer but also to normal cells. Generally, DNA molecules are the main targets in cell death by ionizing radiation and above 80% of the damage is induced by secondary species generated from an initial ionization.1–3 Low-energy (o20 eV) electrons (LEEs)
School of Chemistry and Chemical Engineering, Institute of Theoretical Chemistry, Shandong University, Jinan, 250100, People’s Republic of China. E-mail: [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/c5cp02693h
This journal is © the Owner Societies 2015
have been proven to be one of the most abundant secondary species,4–6 which gradually lose energy through inelastic collisions and eventually become solvated within 1 ps in biological media.7 Before getting solvated, LEEs can attach to a resonance of a molecular target, leading to the secondary damage. On one hand, LEEs can directly ionize components of DNA when the kinetic energy reaches the ionization thresholds of target molecules (ca. 7–8 eV), generating more reactive fragments. On the other hand, the electrons having kinetic energies below 7–8 eV and even as low as 0.1 eV which can no longer produce secondary ionizations can also induce DNA strand rupture due to dissociative electron attachment (DEA).8–15 However, the above findings raise new questions as to how to restrict patient side-effects caused by dissociative secondary LEE attachment
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19797
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
Paper
to DNA. Along this line, a set of uracil analogues have been proposed to substitute thymidine into cancer cell DNA as a radiosensitizing agent with potential clinical applications.16–24 Halogenated pyrimidines represent an important class of molecules due to the fact that they are elementary building blocks of the drugs mentioned above employed in cancer treatment. The DEA to these gas phase constituents has been the concern of several recent publications explorating the involved radiosensitization mechanism which is believed to be operating through different resonant dissociation channels. For example, Modelli and coworkers reported the electronic resonance states of halopyrimidines experimentally by electron transmission spectroscopy and theoretically by scaling the virtual orbital energies,25 and also measured the yield of negative ions with DEA spectroscopy. They attributed the observed two signals (at 0 and 0.2 eV, respectively) to electron attachments in the first two p* resonances. Thereafter, Barbosa and coworkers calculated the integral cross section for collisions of low-energy electrons with halopyrimidines by the Schwinger multichannel method.26 The computed cross sections revealed three p* and one s* resonances which is consistent with Modelli’s results. However, the p*2 resonance observed by Barbosa and coworkers lies above 0.5 eV, and is too high to account for the peak located at 0.2 eV. This is presumably due to the fact that their calculations are only performed at the equilibrium geometry of the target without considering the nuclear motions. Clearly, the close spacing between the two peaks observed in the DEA spectroscopy by Modelli and coworkers is evocative of taking the vibrational contribution into account which is inevitable for a comprehensive description of the possible damage or benefit of the DEA process. In addressing this, the time-dependent wave packet (TDWP) method should be a more appropriate tool because it has long been applied to investigate vibrational resonance structures in electron molecule interactions like electron N2, electron H2, electron CO scattering27–29 and is also becoming quite popular in the field of electron-induced biomolecule damage due to the advantages of numerical efficiency and conceptual simplicity.30–33 Through the TDWP calculations along the corresponding potential energy curves (PECs), vibrational resonance distribution as well as its contributions to the C–Br bond cleavage can be revealed in an effort to complement potential DEA pathways and provide new insights into the resonance phenomena, e.g. those in Modelli’s experiment. Furthermore, as is well known, the secondary electroninduced damage mostly occurs in the biological environment that contains 70–80% water. Thus, it is necessary to take into account the solvent effect. It should be emphasized that the DEA pathway is quite sensitive to the PEC character. Clearly, the primary challenge, compared with the gaseous case, is to properly mimic the solvent environment and construct reliable PECs for structure changes in solution. Although the general polarizable continuum model (PCM) works well for many important processes in solution, it meets large difficulties when the reactants interact strongly with solvent molecules. Thus, the ab initio molecular dynamics (AIMD) simulation technique
19798 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
PCCP
can be a tool for reasonably taking into account the solvent effect in constructing reliable PECs in solution. Now, the AIMD simulation technique has been extensively applied to the studies of solution systems because it is not only able to model water molecules at the same level as the solute and allows for protonation of solutes,34 but it can also treat the enthalpic and entropic effects. Accordingly, it should be useful to construct an accurate free energy surface (FES) along the reaction coordinate and then to investigate the vibrational resonance structures and reaction mechanism of the electron interacting systems in solution (e.g. the electron hydrated 5-BrPy here) in conjunction with the TDWP method. In this work, we reveal that, in the gas phase, quantum tunneling through high-level vibrational resonance (v = 3) on the vertical PECs and spontaneous dissociation along the relaxed PECs are the likely responsible mechanisms for the two peaks at 0.2 and 0 eV observed in Modelli’s experiment, respectively.25 However, in solution, an increased energy barrier due to the coupling between the p* and s* states exists on the FES which induces stabilization of the p*-type anion instead of cleavage of the C–Br bond. We identify 7 vibrational bound states above the C–Br dissociation limit for the fully hydrated 5-BrPy and only v = 5, 6 can contribute to the bond cleavage via the quantum tunneling path, while the rest is proven to be further stabilized by proton transfer from a surrounding water molecule to the anion. The remainder of this paper is organized as follows. Section 2 illustrates details of the used methodology, and discussion of prominent results is given thereafter. A brief summary of the investigation concludes this paper.
2 Methodology 1. Calculations and AIMD simulations of potential energy surfaces The first ingredient in our calculations is the PECs of the target. However, a full-dimensional description of the reaction is out of reach in a full quantum model. Thus, we just consider the essential C–Br dissociation coordinate and all other degrees of freedom are taken into account via full geometry relaxations. The gaseous PECs were computed using Gaussian 03 code at the DFT-B3LYP level35 with a 6-311++G(d,p) basis set for C, H, N atoms and a LANL2DZ basis set36,37 for the Br atom. However, for the solution case, the corresponding energy is free energy which can be determined by the constrained AIMD simulation and thermodynamic integration of the constraint force using the CP2K/Quickstep package.38 We construct the system including 100 H2O molecules around a 5-BrPy molecule in a cubic box with r = 1.0 g cm3 and box size = 14.82 Å. 20 sampling windows are chosen for each fixed C–Br distance which varies from 3.1 to 5.8 a0. Each window undergoes 2 ps equilibration and 2 ps production time with a typical timestep of 0.5 fs. The tempera´–Hoover chain thermostat. ture was kept at 300 K by a Nose Energies and forces were evaluated by the Becke, Lee, Yang, and Parr (BLYP) exchange–correlation function39,40 and the triple-z
This journal is © the Owner Societies 2015
View Article Online
PCCP
Paper
basis set augmented with two polarization functions (TZV2P). The constraint mean force [ f (r)] was achieved through the Shake algorithm41 by averaging the last 2 ps trajectories and was then integrated to produce the free energy profile Ðb DAa!b ¼ a h f ðrÞidr .
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
2. TDWP method for state density spectrum On the basis of the PECs or FESs, the TDWP approach was subsequently performed. This method has been described in detail previously,42–46 and here we only give an outline of it. After excess electron attachment to 5-BrPy, the nuclear motion, evolved under the influence of the anionic Hamiltonian (HA), can be described as cðR; tÞ ¼ eiHA ðRÞt=h cðR; t ¼ 0Þ
(1)
h2 2 r þ EA ðRÞ and EA(R) denoting anionic 2mA potential energies. c(R,t = 0) is the initial nuclear wavefunction in the form of " 1=4 # 2 R R0 2 cðR; t ¼ 0Þ ¼ (2) exp ikR pd0 d0 with HA ðRÞ ¼
where d0 denotes the initial width, and R0 is the initial position of the wave packet. Subsequently, the wavefunction at an arbitrary time of the system C(R,t) can be obtained by using the split-operator propagation scheme. To avoid unphysical reflection of the wave packet, we performed complex absorbing potentials at the edge of the grid. Finally, the state-density spectrum was extracted from Fourier transformation of the autocorrelation function: ð1 s¼ expðiEt=hÞhcðR; t ¼ 0ÞjcðR; tÞidt (3) 1
helpful to first know relevant information about it in the gas phase. Thus, in the following analyses, the gaseous mechanism of DEA to 5-BrPy is first discussed in detail to obtain a basic understanding of the relevant processes. Then, a complicated solution case is further discussed to clarify how the solvent effect can modify the DEA mechanism and dynamics using a combined AIMD simulation and time-dependent wave packet method. (a) Dissociative electron attachment in the gas phase AIMD simulation was first carried out to reveal the dynamics of DEA to 5-BrPy in the gas phase. An excess electron was vertically attached to 5-BrPy in its stable configuration optimized at the DFT-B3LYP level. As a result, two separate stages are distinguished. Initially, the electron attaches to the vacant p* orbital and the C–Br bond length fluctuates around RC–Br E 3.75 a0 when t o 50 fs. After a short duration of the formed p*-type transient anion (5-BrPy*), the electron transfers to a s* orbital gradually and promotes the C–Br bond dissociation into Br plus the Py radical (Fig. S1 and S2, ESI†). This phenomenon can be described by the nuclear motions along the corresponding PECs. We suggest that, in the early period (t o 50 fs), the system evolves on the vertical PEC. However, after the configuration rearrangement to adapt the electron attachment, the relaxed PEC should be responsible for the consequent evolution (t > 50 fs). As shown in Fig. 1, the relaxed PEC presents a repulsive character and the probability amplitude on it can spread to the region featuring a large C–Br bond length (RC–Br) with increasing time (Fig. S4, ESI†). In contrast, the probability amplitude in the vertical anionic state is localized at RC–Br E 3.75 a0 due to the presence of a barrier at RC–Br E 4.0 a0 with height E0.2 eV (Fig. S3, ESI†). To explore the origin of the barrier, we first examined a set of low-lying unoccupied
3. Tunneling coefficient To evaluate the vibrational state contributions to the dissociation, the transmission coefficient (T) was determined. Assuming a potential barrier with the form UðxÞ ¼
U0 cosh 2 ðaxÞ
(4)
where U0 is the height of the potential barrier, a is the width parameter, and x is the reaction coordinate, T can be obtained by using the following relation47 T¼
sinh 2 ðpk=aÞ
1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sinh 2 ðpk=aÞ þ cosh 2 p 8mU0 =h2 a2 1 2
(5)
ð2mEÞ1=2 , m is the reduced mass and E is the energy h of the bound vibrational state.
where k ¼
3 Results and discussion Although the main objective of this work is to explore the interaction of an excess electron with 5-BrPy in solution, it is
This journal is © the Owner Societies 2015
Fig. 1 Virtual orbital energies and potential energy curves as a function of C–Br distance. (upper panel) Corresponding virtual orbital energies of the optimized neutral molecules. (lower panel) The vertical and relaxed anionic potential energy curves for 5-BrPy.
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19799
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
Paper
PCCP
Fig. 2 Energy spectrum of state-density for anionic 5-Brpy extracted from the wave packet propagation calculations: (A) in the gas phase; (B) in the aqueous phase.
molecular orbitals (LUMOs) of neutral 5-BrPy around RC–Br = 4.0 a0 (Tables S1 and S2, ESI†), and observed that the state with a shorter C–Br bond before the barrier [RC—Br o 4.0 a0 (2.12 Å)] shows a p* character, while the state in the region after the barrier, i.e., a larger C–Br bond (RC–Br > 4.0 a0) becomes of the repulsive C–Br s* character. In addition, we illustrated a series of virtual molecular orbital energies, which include two empty p* MOs and one s* MO, of neutral 5-BrPy varying with RC–Br (Fig. 1, upper panel). Two avoided crossing points are detected: one is between the s* MO and the p*2 MO at RC–Br E 3.8 a0, and the other is between the s* MO and the p*1 MO at RC–Br E 4.0 a0. Clearly, the above results provide evidence for the fact that the coupling of the p*1 and s* states around RC–Br E 4.0 a0 is responsible for a kinetic barrier.48 It is apparent that transfer of an excess electron from the p*1 to s* state and destabilization of the C–Br bond occur only if these two states are equal in energy. To reach the point on the corresponding PEC, the C–Br bond has to elongate, which suggests the importance of taking into account the vibrational resonance contribution. Based on the PEC, we carried out the TDWP calculations, generating a vibrational state-density spectrum for anionic 5-BrPy as shown in Fig. 2(A). As we can see, despite the somewhat quasi-continuous peaks just above the barrier, four relatively sharp spectral peaks are identified in the energy range below the dissociation barrier at RC–Br E 4.0 a0 which represent the corresponding bound vibrational resonances located on
Table 1
the vertical PEC. To further understand their contributions to the dissociation, we computed relevant transmission coefficients (T) for the vibrational states (Table 1) using eqn (5). In general, the barrier width a from the PECs is difficult to ensure, which stimulated us to use a model potential barrier with the form of eqn (4) for finding out the same. Different combinations of U0 and a were tested, and the best choice (U0 = 6.6 103 a.u., a = 14.5 a01) was compared with the anionic PEC in Fig. S5 (ESI†). It can be seen from Table 1 that the T values for v = 0–2 are negligibly small, while for v = 3 it is significant. This seems reasonable because the reduced mass of the C–Br bond is too large to allow the probability density from states (v = 0–2) to flow towards the dissociative region. However, the higher resonance [v = 3, E = 6.1 103 a.u. (0.17 eV)] corresponds to high vibrational energies and a low barrier width which weakens the effect of the reduced mass of the C–Br bond and enhances the tunneling probability. The non-negligible T value of v = 3 indicates a potential path for the C–Br dissociation along the vertical anionic PEC. Compared with Modelli’s experiment, we have demonstrated that, in the gas phase, an excess electron immediately occupies the p*1 orbital of 5-BrPy to form 5-BrPy* and this induces the molecule to dissociate through two pathways. On one hand, the electron induces vibrational excitation on the vertical PEC allowing C–Br bond cleavage through the quantum tunneling pathway. The calculated resonance position (0.17 eV) is quite close to the experimental value (0.2 eV). On the
Transmission coefficient (T) for anionic vibrational states calculated using eqn (5)
Gas phase Vibrational state
Energy (a.u.)
0 1 2 3 4 5 6
8.6 2.4 4.2 6.1 — — —
104 103 103 103
Aqueous phase Tunneling coefficient (T) 2.64 6.91 8.72 4.63 — — —
19800 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
104 103 102 101
Energy (a.u.) 7.0 2.1 3.4 4.7 6.1 7.4 8.7
104 103 103 103 103 103 103
Tunneling coefficient (T) 1.07 2.04 3.94 5.18 7.10 7.95 4.00
107 106 105 104 103 102 101
This journal is © the Owner Societies 2015
View Article Online
PCCP
other hand, geometry relaxation of the system invokes a spontaneous electron state conversion from the p*1 to s* orbital, leading to a dissociation along the relaxed PEC. Clearly, this mechanism can account for the 0 eV signal.
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
(b)
Electron-impacted dissociation in solution
It is well known, however, that the secondary electron induced damage mainly occurs in the biological microsurrounding or aqueous solution. Therefore, it is necessary to include the solvent effect in the extended calculations to clarify how the solvent effect modifies DEA to 5-BrPy. To model the process in aqueous solution, we constructed a system containing 100 H2O molecules around a target 5-BrPy. After a 3 ns equilibration with a classical molecular dynamics simulation, a representative configuration was chosen for another 3 ps AIMD simulation to obtain an equilibrated solution system. Further, AIMD simulation was conducted for the negatively charged system by considering periodic boundary condition for 1 ps with the initial conformation extracted from the neutral AIMD equilibrium trajectory mentioned above. Fig. 3 and 4(A) illustrate the time evolution of
Fig. 3 Time evolutions of the C–Br bond length and the spin density on 5-BrPy after vertically adding an excess electron to the system.
Paper
spin density on 5-BrPy and related snapshot configurations, respectively. Different from that in the gas phase, a vertically injected excess electron distributes over the surrounding water molecules and the target molecule, exhibiting a delocalized state. During about 40 fs of relaxation,22 it is gradually concentrated in the p* orbital of 5-BrPy with an elongation of the C–Br bond from 3.65 a0 (the bond length of neutral 5-BrPy) to 3.75 a0 (that of the corresponding anion, Fig. 3). This observation indicates that (a) localization of an excess electron on the fully hydrated 5-BrPy occurs simultaneously with structure relaxation of the system, i.e. reorientation of the solvent molecules and structural fluctuation of the pyrimidine base; (b) the stable p*-type anion state is preferred after electron attachment rather than auto-dissociation as observed in the gas phase. These conclusions were also confirmed by additional AIMD simulations starting with different initial configurations, exchange–correlation functions, and basis sets (Fig. S6–S9, ESI†). Clearly, the observed difference in the preliminary products due to excess electron attachment to 5-BrPy in the gas phase and aqueous solution should be attributed to their different PECs regarding the electron-impacted C–Br bond variations. In particular, in solution, the relevant thermodynamic energy curve should be the FES which takes energy exchange with the environment into account. Thus, we constructed the FES by using the thermodynamic integration method. A series of AIMD simulations with fixed C–Br bond lengths were performed. The timeaveraged constraint force [ f (r)] (Fig. S17, ESI†) was employed as Ðb the mean force for the integration DAa!b ¼ a h f ðrÞidr .49 Fig. 5 displays the constrained mean forces (white circle) and their integration along the C–Br stretching coordinate (black circle). It is found that there is an energy barrier with a height of B0.25 eV on the relaxed FES at RC–Br E 4.2 a0. This observation is different from that (no barrier) in the gas phase. In fact, as shown earlier, the barrier comes from the coupling between the p* and s* states during electron transfer. To confirm the validity of this picture in solution, we examined the relevant molecular orbitals and energies both in PCM (Fig. S10, S11 and Table S3, ESI†) and
Fig. 4 Snapshots at different times extracted from: (A) trajectory after an excess electron is vertically added to the solvated 5-Brpy; (B) trajectory about the proton transfer process. The green and yellow opaque shades are for positive and negative parts of the wave function (isovalue = 0.02), respectively.
This journal is © the Owner Societies 2015
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19801
View Article Online
Paper
PCCP
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
that in the gas phase. This observation should be attributed to the fact that the solvent effect can considerably stabilize the anionic complex, leading to a high and wide barrier which weakens the probability density from p* to s* states. Overall, we suggest that the electron attachment to aqueous 5-BrPy can be divided into two cases: attachment of an electron with energy E > 0.2 eV can excite the system to high-level vibrational states (v = 5, 6) and induces the C–Br bond cleavage via a quantum tunneling pathway tunneling localization e þ 5-BrPy ! 5-BrPy ! Brd Pydd ! Br þ Py ; while attachment of an electron with E o 0.2 eV can produce a stable p*-type anion but not the further dissociated products (the s*-type anionic complex or separate hydrated Br and Py ). (c) Fig. 5 The constrained mean forces and their integration with respect to the C–Br bond length of 5-BrPy.
explicit water environment (Fig. S12–S16, ESI†) which present similar behaviors to those in the gas phase. In addition, it should be noted that another barrier occurs when stretching the C–Br bond beyond RC–Br = B5.3 a0. Clearly, this barrier corresponds to that of dissociation of the solvent-stabilized s*-type complex into Br plus the Py radical in solution. To further prove the existence of such a barrier, we performed an AIMD simulation on the system with the C–Br bond length being RC–Br > 5.0 a0 (Fig. S18 and S19, ESI†), and found that the excess electron steadily resides in the s* orbital and the C–Br bond fluctuates around RC–Br = B5.3 a0, neither being considerably elongated towards the dissociated state nor being further shortened to go back to the stable anionic complex structure. This observation indicates that a barrier separates the s*-type anionic complex (Brd Pyd) and the dissociated products (Br + Py ). This s*-type anion complex is a stable ionic pair connected by a 3-electron bond with a bond length of about 5.2 a0 (2.8 Å) and the excess electron is shared by two moieties (Br and Py). This is reasonable because the Py radical has considerably large electron affinity that is further enhanced by solvation (Table S4, ESI†), so that the attached excess electron only partially transfers to Br to form the (Brd Pyd) state when the C–Br bond elongates. Clearly, the above mentioned reliable FES provides us with the basis to explore the resonance distribution of the fully solvated 5-BrPy anion. Fig. 2(B) represents the calculated state-density spectrum as a function of the energy produced by Fourier-transforming the time autocorrelation function. Compared with the gas phase results, we identify seven separate sharp resonance peaks located below the dissociation threshold. That is, there are seven vibrational resonance states (v = 0–6) for solvated anionic 5-BrPy in solution. Similar to the calculations in the gas phase, we introduce a model as eqn (4) to mimic the barrier and the best choice (U0 = 9 103 a.u., a = 2.9 a01) is utilized (Fig. S20, ESI†). The calculated T values for the states are collected in Table 1, indicating that they are negligible for the v = 0–4 states, non-negligible for the v = 5 state, and significant for the v = 6 state. It should be noted that the transmission coefficients in solution are generally smaller than
19802 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
Proton transfer effect
In fact, the generated intermediate, p*-type anion, is metastable because two possible processes can develop. One is the further dissociation through a tunneling pathway, as mentioned above, while another is protonation of the negatively charged 5-BrPy at its N site through a proton transfer mechanism from an adjacent solvent H2O molecule. The associated dynamics with this proton transfer reaction can be monitored by time evolution of the Mulliken charge distribution and the Nb–Hw distance as shown in Fig. 6. Analysis of the dynamics trajectory suggests a threestep process. (a) The electron localization stage. At the beginning (t o 100 fs), the injected excess electron gradually localizes towards 5-BrPy and at the same time the Nb–Hw distance continuously decreases due to attraction of the Nb site with the gradually increased extra negative charge. (b) The formation and lasting stage of the H-bond. The H-bond between the Nb of 5-BrPy and Hw of an adjacent H2O molecule is gradually formed at 100–520 fs and then fluctuates steadily at B1.6 Å. In this process, the H-bond network of the surrounding water molecules constantly reorients. (c) The proton transfer stage. Spontaneous proton transfer is triggered upon an adequate hydration structure that is established and finished rapidly within about 30 fs (in 520–550 fs) with protonation at the Nb site of 5-BrPy. This process can be described by the radial distribution
Fig. 6 Time evolution of the Nb–Hw distance (RNb–Hw) and the charge distributions on the corresponding parts. The subscript b denotes the N atom of 5-BrPy and the subscript w denotes the water molecule releasing a proton.
This journal is © the Owner Societies 2015
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
PCCP
Paper
sensitizer, 5-BrPy, using a combined AIMD simulation and time-dependent wave packet method in this work as well as potential reaction pathways. In particular, an explicit water environment is considered to emphasize the importance of the solvent effect in scaling the electron interaction with solvent H2O molecules and the target 5-BrPy. In the gas phase, an excess electron directly attaches to the 5-BrPy target molecule and occupies its p* orbital to form a transient anion, and induces the C–Br bond dissociation via two primary channels: (i) a quantum tunneling mechanism through the highest vibrational resonance (v = 3) on a vertical PEC tunneling e þ 5-BrPy ! 5-BrPy ! Br þ Py , and (ii) dissociation relaxation along a relaxed PEC e þ 5-BrPy ! 5-BrPy ! Br þ Py . Fig. 7 Time evolution of radial distribution functions (g(r)) between Ow of the H2O releasing a proton and H of other surrounding water molecules. At 520–550 fs, the peak shifts towards short Ow H distance, suggesting that proton transfer occurs.
functions (RDFs)50,51 between the Or of the H2O releasing a proton and its surrounding water hydrogen atoms. Fig. 7 provides the time evolution of RDFs extracted from the proton transfer trajectory. It can be seen that during the formation of the H-bond (t o 520 fs), the water H atoms around the Or mainly distribute with ROr–H > 1.75 Å. However, the RDF peak moves to the region at B1.6 Å (ROr–H) when the proton transfer occurs (t = 520–550 fs), meaning that strong H-bonds are formed between the Or and its surrounding water molecules as well as sharing the Or attraction with the Hr and the promoted Hr tends to the Nb site of 5-BrPy. The protonation was also observed at another N site in a parallel AIMD simulation which yields a similar conclusion (Fig. S21–S24, ESI†). Encouraged by the above analyses, we further evaluated the proton transfer influence on DEA. Examination reveals that, after protonation, the excess electron still occupies the p* orbital, but the center of electron spin density moves towards the Nb (from 1.3 to 1.2 Å) and leaves the Br (from 3.2 to 3.3 Å) (Fig. S25, ESI†). Clearly, on one hand, the shift of the spin density center weakens the electron Br interaction. In fact, shortening of the average C–Br bond length is observed from 1.98 Å to 1.93 Å [the C–Br bond length (3.65 a0) of neutral 5-BrPy in solution] after protonation (Fig. S26, ESI†). On the other hand, the shift of the spin density center stabilizes the anion with an enhancement of the barrier (Fig. S27, ESI†) to conversion from the p* to s* state (5-BrPy* - [Brd Pyd]). However, it has little effect on further conversion from the s*-type anion to the completely dissociated Br plus Py radical ([Brd Pyd] - Br + Py ) because the protonated 5-BrPy is deprotonated when the C–Br bond (RC–Br) is stretched beyond 4.2 a0 (Fig. S28 and S29, ESI†). This observation about deprotonation is attributed to the excess electron transfer to the s* orbital located at the Br–C bond zone and the excess electron no longer distributes at the Nr.
4 Conclusion In summary, we explored the vibrational resonance contributions to the mechanism of DEA of the selected tumor-specific
This journal is © the Owner Societies 2015
We suggest that these two channels are responsible for the two peaks at 0.2 and 0 eV observed experimentally by Modelli and coworkers. In contrast, in solution, a vertically injected excess electron is found to distribute on the target 5-BrPy and solvent H2O molecules in a delocalized state and then localizes on 5-BrPy within 40 fs along with a solution structure relaxation. The formed metastable hydrated anion does not spontaneously dissociate due to a barrier on the relaxed FES originating from the coupling between the p* and s* states. Seven resonant states (v = 0–6) are identified on the FES and only the high-level v = 5, 6 with non-negligible quantum tunneling probability can lead to the dissociation tunneling localization eþ5-BrPy ! 5-BrPy ! Brd Pydd !Br þPy ; while the other low-level resonant states (v = 0–4) lead to localization protonation at a N site of 5-BrPy eþ5-BrPy ! relaxation
5-BrPy ! Prt-5-BrPyÞ, inhibiting the C–Br bond cleavage. The protonation undergoes a three-step process and is sensitive to H-bond network rearrangement among the surrounding water molecules. It should be noted that proton transfer from a surrounding H2O molecule to a N site of 5-BrPy is actually an electron localization coupled nonsynchronous process and also presents a competition with the C–Br bond cleavage by neutralizing the anion core. This work provides a deep understanding of the radio-sensitivity of the 5-BrPy drug in solution and reveals the important impact of the low energy electron triggered vibrational excitations and also the importance of the solvent effect.
Acknowledgements This work was supported by NSFC (21373123, 20633060, and 20973101), NSF (ZR2013BM027) of Shandong Province. A part of the calculations were carried out at National Supercomputer Center in Jinan, Shanghai Supercomputer Center, and HighPerformance Supercomputer Center at SDU-Chem.
References 1 P. P. Connell, S. J. Kron and R. R. Weichselbaum, Relevance and irrelevance of DNA damage response to radiotherapy, DNA Repair, 2004, 3, 1245–1251.
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19803
View Article Online
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
Paper
2 L. Sanche, Low energy electron-driven damage in biomolecules, Eur. Phys. J. D, 2005, 35, 367–390. 3 E. Alizadeh and L. Sanche, Precursors of Solvated Electrons in Radiobiological Physics and Chemistry, Chem. Rev., 2012, 112, 5578–5602. 4 M. Toulemonde, E. Surdutovich and A. V. Solov’yov, Temperature and pressure spikes in ion-beam cancer therapy, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2009, 80, 031913. 5 E. Scifoni, E. Surdutovich and A. V. Solov’yov, Spectra of secondary electrons generated in water by energetic ions, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2010, 81, 021903. 6 S. M. Pimblott and J. A. LaVerne, Production of low-energy electrons by ionizing radiation, Radiat. Phys. Chem., 2007, 76, 1244–1247. 7 G. Hanel, B. Gstir, S. Denifl, P. Scheier, M. Probst, ¨rk, B. Farizon, M. Farizon, E. Illenberger and T. D. Ma Electron Attachment to Uracil: Effective Destruction at Subexcitation Energies, Phys. Rev. Lett., 2003, 90, 188104. 8 J. Gu, J. Wang and J. Leszczynski, Electron AttachmentInduced DNA Single Strand Breaks: C3 0 –O3 0 s-Bond Breaking of Pyrimidine Nucleotides Predominates, J. Am. Chem. Soc., 2006, 128, 9322–9323. 9 Y. Zheng, P. Cloutier, D. J. Hunting, J. R. Wagner and L. Sanche, Glycosidic Bond Cleavage of Thymidine by LowEnergy Electrons, J. Am. Chem. Soc., 2004, 126, 1002–1003. 10 X. Li, M. D. Sevilla and L. Sanche, Density Functional Theory Studies of Electron Interaction with DNA: Can Zero eV Electrons Induce Strand Breaks?, J. Am. Chem. Soc., 2003, 125, 13668–13669. 11 I. Anusiewicz, J. Berdys, M. Sobczyk, P. Skurski and J. Simons, Effects of Base p-Stacking on Damage to DNA by Low-Energy Electrons, J. Phys. Chem. A, 2004, 108, 11381–11387. 12 I. Anusiewicz, M. Sobczyk, J. Berdys-Kochanska, P. Skurski and J. Simons, A Theoretical Model for Indirect Dissociative Electron Attachment, J. Phys. Chem. A, 2004, 109, 484–492. 13 J. Berdys, P. Skurski and J. Simons, Damage to Model DNA Fragments by 0.25–1.0 eV Electrons Attached to a Thymine p* Orbital, J. Phys. Chem. B, 2004, 108, 5800–5805. 14 C.-R. Wang, J. Nguyen and Q.-B. Lu, Bond Breaks of Nucleotides by Dissociative Electron Transfer of Nonequilibrium Prehydrated Electrons: A New Molecular Mechanism for Reductive DNA Damage, J. Am. Chem. Soc., 2009, 131, 11320–11322. 15 B. Boudaı¨ffa, P. Cloutier, D. Hunting, M. A. Huels and L. Sanche, Resonant Formation of DNA Strand Breaks by Low-Energy (3 to 20 eV) Electrons, Science, 2000, 287, 1658–1660. 16 C.-R. Wang and Q.-B. Lu, Real-Time Observation of a Molecular Reaction Mechanism of Aqueous 5-Halo-2 0 deoxyuridines under UV/Ionizing Radiation, Angew. Chem., Int. Ed., 2007, 46, 6316–6320. 17 S. Cecchini, S. Girouard, M. A. Huels, L. Sanche and D. J. Hunting, Interstrand Cross-Links: A New Type of g-Ray Damage in Bromodeoxyuridine-Substituted DNA, Biochemistry, 2005, 44, 1932–1940.
19804 | Phys. Chem. Chem. Phys., 2015, 17, 19797--19805
PCCP
18 T. Chen, G. P. Cook, A. T. Koppisch and M. M. Greenberg, Investigation of the Origin of the Sequence Selectivity for the 5-Halo-2 0 -deoxyuridine Sensitization of DNA to Damage by UV-Irradiation, J. Am. Chem. Soc., 2000, 122, 3861–3866. 19 X. Li, L. Sanche and M. D. Sevilla, Dehalogenation of 5-Halouracils after Low Energy Electron Attachment: A Density Functional Theory Investigation, J. Phys. Chem. A, 2002, 106, 11248–11253. 20 X. Li, M. D. Sevilla and L. Sanche, DFT Investigation of Dehalogenation of Adenine–Halouracil Base Pairs upon Low-Energy Electron Attachment, J. Am. Chem. Soc., 2003, 125, 8916–8920. 21 L. Chomicz, M. Zdrowowicz, F. Kasprzykowski, J. Rak, A. Buonaugurio, Y. Wang and K. H. Bowen, How to Find Out Whether a 5-Substituted Uracil Could Be a Potential DNA Radiosensitizer, J. Phys. Chem. Lett., 2013, 4, 2853–2857. ´r, P. Wityk, J. Czub, L. Chomicz and J. Rak, A first22 M. Wieczo principles study of electron attachment to the fully hydrated bromonucleobases, Chem. Phys. Lett., 2014, 595–596, 133–137. 23 H. Abdoul-Carime, M. A. Huels, E. Illenberger and L. Sanche, Sensitizing DNA to Secondary Electron Damage: Resonant Formation of Oxidative Radicals from 5-Halouracils, J. Am. Chem. Soc., 2001, 123, 5354–5355. 24 S. D. Wetmore, R. J. Boyd and L. A. Eriksson, A theoretical study of 5-halouracils: electron affinities, ionization potentials and dissociation of the related anions, Chem. Phys. Lett., 2001, 343, 151–158. 25 A. Modelli, P. Bolognesi and L. Avaldi, Temporary Anion States of Pyrimidine and Halopyrimidines, J. Phys. Chem. A, 2011, 115, 10775–10782. 26 A. S. Barbosa and M. H. F. Bettega, Shape resonances in low-energy-electron collisions with halopyrimidines, J. Chem. Phys., 2013, 139, 214301. 27 M. Sarma, S. Adhikari and M. K. Mishra, Simple systematization of vibrational excitation cross-section calculations for resonant electron-molecule scattering in the boomerang and impulse models, J. Chem. Phys., 2007, 126, 044309. 28 M. Sarma, S. Adhikari and M. K. Mishra, Mechanistic investigation of vibrational fine structure in e-H2 scattering using local complex potential-based time dependent wave packet approach, Int. J. Quantum Chem., 2008, 108, 1044–1051. 29 R. Singh, M. Sarma, A. Jain, S. Adhikari and M. Mishra, Calculation of vibrational excitation cross-sections in resonant electron-molecule scattering using the time-dependent wave packet (TDWP) approach with application to the 2P CO shape resonance, J. Chem. Sci., 2007, 119, 385–389. 30 T. Takayanagi, T. Asakura and H. Motegi, Theoretical Study on the Mechanism of Low-Energy Dissociative Electron Attachment for Uracil, J. Phys. Chem. A, 2009, 113, 4795–4801. 31 R. Bhaskaran and M. Sarma, Effect of quantum tunneling on single strand breaks in a modeled gas phase cytidine nucleotide induced by low energy electron: a theoretical approach, J. Chem. Phys., 2013, 139, 045103. 32 S. Bhowmick, R. Bhaskaran, M. K. Mishra and M. Sarma, Investigation of dissociative electron attachment to 2 0 -deoxycytidine-3 0 -monophosphate using DFT method and
This journal is © the Owner Societies 2015
View Article Online
PCCP
33
34
Published on 24 June 2015. Downloaded by Fudan University on 19/10/2017 11:51:44.
35
36
37
38
39
40
41
time dependent wave packet approach, J. Chem. Phys., 2012, 137, 064310. R. Bhaskaran, S. Bhowmick, M. K. Mishra and M. Sarma, Low-Energy Electron-Induced Single Strand Breaks in 2 0 -Deoxycytidine-3 0 -monophosphate Using the Local Complex Potential Based Time-Dependent Wave Packet Approach, J. Phys. Chem. A, 2011, 115, 13753–13758. M. Smyth and J. Kohanoff, Excess Electron Interactions with Solvated DNA Nucleotides: Strand Breaks Possible at Room Temperature, J. Am. Chem. Soc., 2012, 134, 9122–9125. M. Frisch, G. Trucks, H. Schlegel, G. Scuseria, M. Robb, J. Cheeseman, J. Montgomery Jr, T. Vreven, K. Kudin and J. Burant, Gaussian 03, Revision D. 01, Gaussian. Inc., Wallingford, CT, 2004. W. R. Wadt and P. J. Hay, Ab initio effective core potentials for molecular calculations. Potentials for main group elements Na to Bi, J. Chem. Phys., 1985, 82, 284–298. S. P. Ananthavel and M. Manoharan, A theoretical study on electron donor–acceptor complexes of Et2O, Et2S and Me3N with interhalogens, I–X (X = Cl and Br), Chem. Phys., 2001, 269, 49–57. J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing and J. Hutter, Quickstep: fast and accurate density functional calculations using a mixed Gaussian and plane waves approach, Comput. Phys. Commun., 2005, 167, 103–128. A. D. Becke, Density-functional exchange-energy approximation with correct asymptotic behavior, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098–3100. C. Lee, W. Yang and R. G. Parr, Development of the ColleSalvetti correlation-energy formula into a functional of the electron density, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789. J.-P. Ryckaert, G. Ciccotti and H. J. C. Berendsen, Numerical integration of the cartesian equations of motion of a system
This journal is © the Owner Societies 2015
Paper
42
43
44
45
46
47
48 49
50 51
with constraints: molecular dynamics of n-alkanes, J. Comput. Phys., 1977, 23, 327–341. T. Takayanagi and A. Wada, Theoretical prediction of the lifetime of the metastable helium compound: HHeF, Chem. Phys. Lett., 2002, 352, 91–98. T. Takayanagi, H. Motegi, Y. Taketsugu and T. Taketsugu, Accurate ab initio electronic structure calculations of the stable helium complex: HeBeO, Chem. Phys. Lett., 2008, 454, 1–6. Q.-T. Meng, G.-H. Yang and K.-L. Han, Time-dependent wave packet approach to Rabi oscillation in strong laser field, Int. J. Quantum Chem., 2003, 95, 30–36. Q.-T. Meng, G.-H. Yang, H.-L. Sun, K.-L. Han and N.-Q. Lou, Theoretical study of the femtosecond-resolved photoelectron spectrum of the NO molecule, Phys. Rev. A: At., Mol., Opt. Phys., 2003, 67, 063202. R.-F. Lu, P.-Y. Zhang and K.-L. Han, Attosecond-resolution quantum dynamics calculations for atoms and molecules in strong laser fields, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2008, 77, 066701. R. Marom, C. Levi, T. Weiss, S. Rosenwaks, Y. Zeiri, R. Kosloff and I. Bar, Quantum Tunneling of Hydrogen Atom in Dissociation of Photoexcited Methylamine, J. Phys. Chem. A, 2010, 114, 9623–9627. J. Simons, How Do Low-Energy (0.1–2 eV) Electrons Cause DNA-Strand Breaks?, Acc. Chem. Res., 2006, 39, 772–779. Y. Mei, D. M. Sherman, W. Liu and J. Brugger, Ab initio molecular dynamics simulation and free energy exploration of copper(I) complexation by chloride and bisulfide in hydrothermal fluids, Geochim. Cosmochim. Acta, 2013, 102, 45–64. J. Cohn, Theory of the radial distribution function, J. Phys. Chem., 1968, 72, 608–616. B. L. Nguyen and B. M. Pettitt, Effects of acids, bases and heteroatoms on proximal radial distribution functions for proteins, J. Chem. Theory Comput., 2015, 11, 1399–1409.
Phys. Chem. Chem. Phys., 2015, 17, 19797--19805 | 19805
