Elucidation of the Mechanistic Pathways of the Hydroxyl Radical

Aug 2, 2012 - genesis in a hairless mice model.11. Recently, it has been demonstrated that ..... An interesting revelation is the spin density distrib...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCA

Elucidation of the Mechanistic Pathways of the Hydroxyl Radical Scavenging Reaction by Daidzein Using Hybrid QM/MM Dynamics Sandipan Chakraborty and P. K. Biswas* Laboratory of Computational Biophysics & Bioengineering Department of Physics, Tougaloo College, Tougaloo Mississippi 39174, United States

ABSTRACT: Employing a hybrid QM/MM simulation we explored the reaction dynamics of the hydroxyl radical scavenging activity of daidzein, a soy isoflavone. Our simulations illustrate that the highly reactive hydroxyl radical can participate in hydrogen abstraction reaction with both OH functional groups of daidzein and can form stable daidzein radicals. We found that the reaction involving the 4′-OH site of daidzein is energetically favorable over the other reaction pathway involving the 7-OH site of daidzein by ∼29 kcal/mol. The high enthalpic stabilization involved in daidzein radical formation at the 4′-OH site can be partly attributed to better solvation through hydrogen-bonding interactions with water and higher electron density delocalization of radical over the adjacent aromatic ring. As evident from the QM/MM dynamics, both HAT pathways led to formation of ketones at the 7-OH and 4′-OH sites of daidzein, respectively, and the adjacent aromatic rings appear in a p-quinonoid form, a highly stable resonating structure. The suitability of the QM/MM methodology to study the reaction mechanism, identification of intermediate states, and pathways of flavonoid radical stabilization reported here opens up a new possibility to study a similar reaction mechanism in other systems.

1. INTRODUCTION

along with remarkable efficacy further extends their application as alternative therapeutic drugs as well as functional foods. Daidzein belongs to a subclass of flavonoids known as isoflavonoids and is a biologically active compound found in food sources such as soybeans and in a number of plants and herbs like Pueraria mirif ica and Pueraria lobata.6 Daidzein exhibits a wide range of physiological and pharmacological effects relevant to human health.7,8 It has been documented

Bioflavonoids are a class of polyphenolic compounds that is widely distributed in the plant kingdom as secondary metabolites.1 They are the most important plant pigments that play a crucial role in flower coloration to attract pollinator animals. In higher plants, flavonoids are involved in UV filtration, symbiotic nitrogen fixation, and floral pigmentation.2 Recent research on various bioactive flavonoids reveals that they are highly effective against a wide range of free-radicalmediated and other diseases (e.g., atherosclerosis, ischemia, neuronal degeneration, cancers, allergies, cardiac problems, inflammation, AIDS, etc.)2−5 in human. Low systemic toxicity © XXXX American Chemical Society

Received: April 12, 2012 Revised: July 4, 2012

A

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

that a low incidence rates of breast cancer was observed among Asian women consuming high soy food intake.9 It is highly effective as an antioxidant and plays a crucial role in reducing the radical oxygen species involved in various cellular damage and exhibited high antioxidant potency in the liposome system.10 It has been demonstrated that topical treatment using soybean isoflavone inhibits UVB-induced skin tumor genesis in a hairless mice model.11 Recently, it has been demonstrated that daidzein possesses strong ability to scavenge hydroxyl radicals which are known to cause oxidative damage to DNA and lead to diseases like genetic aberration, carcinogenesis, and aging.12,13 However, the detailed mechanism of the radical scavenging and consequent stabilization pathways of the flavonoid radical are unknown. The antioxidant potency of polyphenols has been widely interpreted in terms of calculated ionization potential (IP) and bond dissociation enthalpy (BDE) using semiempirical14−16 and ab initio calculations including DFT.17−21 However, the dynamics of the reaction pathways and consequent stabilization processes in the presence of explicit water have been never explored. Simulating reaction dynamics is beyond the scope of classical molecular mechanics (MM). Quantum mechanics (QM) is highly suitable to understand the electronic rearrangements during bond breaking and/or formation process. However, simulating such a reaction in the presence of explicit water (to mimic physiological condition) using QM is computationally highly expensive and thus has never been attempted. Employment of a combined quantum-mechanical molecular-mechanical (QM/MM) method provides a practical approach to treat such reaction phenomenon.22 In a QM/MM method, the total system is divided into QM and MM subsystems with the QM subsystem containing the active site and the MM subsystem containing the rest. During molecular dynamics, the QM subsystem wave function or the electron density for a DFT calculation is optimized at every step in the presence of the external field of MM atoms. We, for the first time, studied the dynamical aspects of the chemical reaction between •OH radical and daidzein using the GROMACSCPMD hybrid QM/MM methodology.22 The daidzein and • OH radical have been treated with the QM method, while the explicit solvents are treated with a classical molecularmechanical (MM) force field.

Scheme 1. Structure of Daidzein Shown in Stick Representationa

a

Three rings of daidzein (A, B, and C) are shown in the center of those rings. Atom labels with the numbering scheme of the daidzein structure considered in this study are also shown.

of daidzein molecule was assigned the proper atom type definition as per the OPLS-AA parameter set. The van der Waals and torsional parameters and partial charges for the ligand were obtained by group analogy in the OPLS-AA set.24 The daidzein molecule was first energy minimized using MM and then solvated in a water box containing SPC water molecules, such that the daidzein atoms were at a distance equal to or greater than 1 nm from the box edges. To eliminate any bad contacts with water, a short classical energy minimization followed by position-restrained dynamics of 5 ps was performed on the system using GROMACS,25,26 and this system is considered to prepare all systems studied here. From this solvated system, three different systems were prepared: (i) Solvated daidzein with 4′-OH is targeted by a radical •OH, (ii) solvated daidzein with 7-OH is targeted by a radical •OH, and (iii) solvated daidzein with both 4′-OH and 7OH sites are targeted with two •OH radicals. To prepare a • OH radical near the active site of daidzein, we target a water molecule closest to the active site, delete a hydrogen from it, change the residue type from water to OH, and change the atom types and partial charges suitable for a OH radical. This is to emphasize here that the selected atom types and partial charges will have no bearing or influence in the QM/MM simulation performed here as the OH is being treated within the QM. Now, in order to prepare a solvated daidzein−•OH system that will allow us to monitor the HAT reactions this is to be noted that the probability to reach a transition structure (TS) for the daidzein−OH system will be too small in the available simulation time if one allows the •OH to diffuse through water from any arbitrary position. Therefore, in order to be close to a transition state, we manually dock the •OH radical to a hydrogen-bonding distance (∼2 Å) with the daidzein active site. This will allow us to increase the probability that the reaction might happen and elucidate the reaction mechanism and molecular pathways of radical stabilizations. After manual docking of the radical •OH on the daidzein active site the solvated system was given a short steepestdescent energy minimization followed by a 10 ps position restrained dynamics at 300 K using GROMACS.25,26 This allows proper equilibration of the system required to start a QM/MM dynamics. After equilibration, an appropriate frame has been selected from the position-restrained trajectories where the hydroxyl radical is ∼2 Å from the diadzein active site. It is mentioned here that during the classical positionrestrained dynamics the Lennard−Jones parameters assigned to

2. METHODOLOGY One of the known pathways by which polyphenolic compounds show antioxidant activity is the hydrogen-atom transfer (HAT) reaction.23 In this pathway there is direct H transfer from polyphenol (Ph−OH) to the radical (•OH), as summarized in eq 1. Ph−OH +• OH → Ph−O• + HOH

(1)

Daidzein possesses two hydroxyl groups capable of participating in a HAT reaction with a hydroxyl radical. Thus, we consider the HAT reaction from both the 7-OH and the 4′-OH sites separately. We also studied the HAT reactions by two • OH targeted on the two hydroxyl groups of daidzein (7-OH and 4′-OH) simultaneously to understand the chronology of the events. A schematic representation of the structure of daidzein considered in the study is shown in Scheme 1. 2.1. System Preparation. The 3-D coordinates of daidzein were obtained from the PubChem database (ID 5281708) and converted into PDB format. Initial daidzein parameters were generated according to the OPLS-AA force field.24 Each atom B

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 1. (A) Stable geometry of daidzein in water obtained from QM/MM dynamics. Explicit water molecules are not shown due to clarity. (B) Electron density plot of stable daidzein.

phase, daidzein structure remains stable throughout the simulation period. The stable geometry of the daidzein in water as obtained from the QM/MM dynamics is shown in Figure 1. Daidzein in water exhibits a nonplanar structure with its B ring making an angle to the chromone moiety formed by the A and C rings. In our QM/MM dynamics of solvated daidzein, the dihedral angle between the B and the C rings is found to be hovering around 46°. Jiaheng et al., in their calculations on daidzein, genistein, formononetin, and biochanin A, found all isoflavones to be nonplanar with the dihedral angle between the B ring and the chromone moiety in the range of 41−43°. For daidzein, they found the dihedral angle to be 42.2°.32 Nevertheless, it is to be noted that the dihedral angle varies broadly for different classes of flavonoids, and substituent groups also have a profound influence on the observed dihedral angle. Moreover, in different calculations,33−35 quercetin, a flavonol, was reported to have planar as well as nonplanar geometry with a dihedral angle of 27.3°33 between the B and the C rings. Serge Antonczak addressed the apparent discrepancy by performing a dihedral angle scan using the B3LYP/6-31(+)G* level of theory. He reported that the rotation of the B ring ±30° around the minima is associated to an energy increase of less than 0.3 kcal/mol.36 These calculations further support that the flavonoids can accommodate both planar and nonplanar structures. It is to be noted that all theoretical calculations on flavonoids were performed either in the gas phase or in the continuum water model where hydrogen-bonding possibilities are completely ignored. We performed a QM/MM dynamics study where solvent effects are explicitly considered. The presence of explicit water can certainly influence the planarity between the B ring and the chromone moiety through hydrogen bonding with water molecules. The bond length distribution (Figure 1A) and electron density distribution analysis (Figure 1B) indicate that both the A and the B rings are aromatic in nature with significant delocalization of the π-electron cloud throughout both rings, which results in an equal bond distance distribution of 1.40 Å. The C4−O4 bond is 1.23 Å and double bonded in nature. On the other hand, C4−C3 and C4−C10 bond distances are 1.49 Å. Thus, a significant single-bonded character is predicted for these two bonds. This is further supported from electron density analysis (Figure 1B), which shows a reduced π-electron density on those two bonds. On the contrary, C2−C3 and C3− C1′ bond lengths are 1.40 Å, which is equal to the bond lengths of the B ring. Thus, there is extensive delocalization predicted

the hydrogen atoms of the OH groups of daidzein in OPLS-AA are found to generate strong repulsion from the 1/r12 term and present an artifact as also noted by Stephanie et al.28 Thus, while preparing the systems using the classical OPLS-AA force field, we kept the Lennard−Jones interactions between the hydrogen of the daidzein and the •OH to zero as has also been done by Stephanie et al.28 As mentioned earlier in this section, none of the daidzein and •OH classical parameters will have any impact on the OH−daidzein HAT reaction studied here as both the •OH and daidzein will be treated with QM. 2.2. GROMACS-CPMD QM/MM Dynamics Protocol. To understand the HAT mechanism, employing the GROMACSCPMD22 QM/MM simulation protocol, the daidzein and •OH radical are being treated with the QM code CPMD27 which uses a plane-wave-based DFT model to optimize the wave function of the QM subsystem (here daidzein and OH• radical) in the presence of the MM atom charges (here all the water molecules) treated with the classical code GROMACS. In the GROMACS-CPMD QM/MM program, GROMACS controls the molecular dynamics and CPMD provides the energy and forces on QM atoms. CPMD also provides QM/MM forces on the MM subsystem atoms as arises due to the QM subsystem. As GROMACS controls the simulation, the QM electron density will follow a Born−Oppenheimer molecular dynamics in the GROMACS-CPMD QM/MM protocol. Accordingly, we found a time step of 1.0 fs is applicable for the QM/MM dynamics. We performed constant pressure and temperature (NPT) MD simulation with a reference temperature of 300 K using an external bath with a coupling constant of 0.1 ps. Pressure was kept constant (1 bar) using the time constant for pressure coupling set to 0.5 ps. For the QM system, wave function optimization was performed in a QM simulation cell of size (12.6 Å × 12.6 Å × 7 Å) together with Poisson solver of Martyna and Tuckerman29 for wave function minimization in CPMD. For wave function optimization we used the normconserving ultrasoft Vanderbilt pseudopotentials with an energy cutoff of 25 Ry (Ry), local spin density approximation (LSDA), and BLYP exchange correlation.30,31

3. RESULTS AND DISCUSSION 3.1. Daidzein Structure in Water from QM/MM Dynamics. Initially, to understand the structure of daidzein in water, we performed a QM/MM dynamics of daidzein in water without the •OH radical. Daidzein electron charge densities undergo a redistribution according to its MM environment (water here) which leads to some structural reorientation of the molecule. After this initial equilibration C

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 2. (A) Time sequence of the HAT for a 1000 fs QM/MM dynamics targeting the 7-OH site of daidzein. Daidzein and the radical are shown in stick representation. Oxygen, hydrogen, and carbon are colored as red, white, and cyan. Water molecules are not shown for clarity. (B) Distance dynamics between H (7-OH site of daidzein) and O (OH• radical). (C) Evolution of the angle between H7‑OH and OH radical in solution.

between the B ring and the C2−C3 bond through the C3−C1′, bond which is also evident from electron density distribution analysis (Figure 1B). There are two hydroxyl groups, 7-OH and 4′-OH, of daidzein that can participate in HAT reaction with •OH. We provide a systematic study of HAT reactions from these two sites separately and also simultaneously using QM/MM dynamics. 3.2. Reaction of •OH with Daidzein Through the 7-OH Site in the A Ring. To study the HAT reaction dynamics from the 7-OH site of daidzein by radical •OH, we treat the daidzein and radical •OH with QM and perform a 1000 fs QM/MM dynamics on the system with the •OH targeted on 7-OH through hydrogen bonding as described in the system preparation in section 2.1. The results of this simulation are summarized in Figure 2. As is evident from Figure 2A, daidzein can scavenge a hydroxyl radical by HAT reaction through its 7-OH site. The time sequence of the QM/MM dynamics displayed in Figure 2A reveals that from a hydrogen-bonding configuration the radical approaches toward the diadzein around ∼67 fs and involves in a bonding interaction with the hydrogen at the 7OH group to form the activated complex. At around 120 fs, HAT reaction is complete with formation of a water molecule. It is to be noted that these time scales do not signify any

reaction rate as our initial frames are targeted to study the dynamics of HAT reaction and stabilization pathways of daidzein radical. Figure 2B shows the distance dynamics between the •OH radical and the hydrogen atom of 7-OH. It has been observed that the transfer of hydrogen occurs between t = ∼90 and ∼120 fs. The new bond formed due to the transfer of hydrogen from the 7-OH site of daidzein to the radical oxygen becomes stable to form water and never dissociates again. An interesting observation is the angular dependency of hydrogen-atom transfer, shown in Figure 2C. During new bond formation, the •OH radical orients itself such that the H−O•− H7 angle is close to that in water, which is ∼105°. When the angular criterion is satisfied, we found that there is an immediate hydrogen-atom transfer. After hydrogen-atom transfer, the angle also fluctuates ∼105°, which is typically the angle of water. This angular correlation in HAT reaction has also been noticed by Ramin et al.37 while studying hydrogen abstraction from guanine base by radical •OH. We further analyzed the electron and spin density distribution along with changes in bond lengths to interpret the daidzein radical stabilization pathways, and results are shown in Figure 3. It is to be noted that the A ring of the daidzein molecule is highly aromatic in character with a symmetric distribution of the π-electron cloud throughout the ring, as is evident from the D

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 3. (A) Changes in electron (left) and spin density (right) during the HAT reaction from the 7-OH site of daidzein. (B) Structure of daidzein radical formed after hydrogen abstraction from the 7-OH site. Bond length distributions are displayed for the A and C rings of daidzein radical obtained after 1000 fs QM/MM dynamics. (C) Possible mechanism of daidzein radical electron delocalization and most stable resonating structure (middle).

Figure 4. (A) Time sequence of the HAT of 1000 fs QM/MM dynamics targeting the 4′-OH site of daidzein in the B ring. Daidzein and the radical are shown in stick representation. Oxygen, hydrogen, and carbon are colored red, white, and cyan. (B) Distance dynamics between H (4′-OH site of daidzein) and O (•OH radical). (C) Evolution of the angle between H4′‑OH and •OH radical in solution. E

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 5. (A) Changes in electron (left) and spin density (right) during the HAT reaction from the 4′-OH site of daidzein. (B) Structure of daidzein radical formed after HAT from the 4′-OH site. Bond length distributions are displayed for the B ring of daidzein radical obtained after 1000 fs QM/ MM dynamics. (C) Possible mechanism of daidzein radical electron delocalization (left) and most stable resonating structure (middle).

electron density distribution analysis shown in Figure 1B. Around 67 fs of this QM/MM simulation trajectory, when the • OH radical first approached the 7-OH group of daidzein to form the initial activated complex, most of the spin densities were on the oxygen atom of the •OH radical. In this configuration, the A ring of diadzein is also aromatic in character with a nearly symmetric π-electron distribution over the entire ring. This activated complex undergoes significant alteration in terms of electronic and spin density distribution with the angular reorientation of the •OH radical. As a result of this reorientation, we observe (around 91 fs) a spin transfer from the •OH radical to the oxygen atoms of the 7-OH group and O4 of daidzein. This spin redistribution is concomitant with the changes in electron density delocalization in the adjoining A ring of daidzein, which result in a loss of its aromatic character. Electron density increases on both the C8− C9 and the C5−C6 bonds, while it reduces significantly on the C7−C8 and C7−C6 bonds. These changes are followed by an immediate hydrogen transfer from the 7-OH group to the radical •OH to form water and daidzein radical. With time, the daidzein radical electron density undergoes further redistribution to form a stable radical. The radical spin density seems to be delocalized and distributed over the 7-OH group and O4 oxygens and C4 and C10 carbons of the daidzein. The C7−O7 bond distance is 1.36 Å (Figure 1), and after HAT it decreases to 1.27 Å (Figure 3A), signifying ketone formation. A closer look into other A-ring bond-distance dynamics reveals that both the C8−C9 and the C5−C6 bond distances decrease with an increase in π-electron densities on those two bonds, which signifies an increase in the double-bond character in those two bonds. On the contrary, C7−C8 and

C7−C6 bond distances increase to 1.48 Å with a significant reduction in π-electron densities on those two bonds, indicating shifting toward the single-bonded nature of those two bonds. Likewise, the C9−C10 and C5−C10 bonds are also stabilized with an increase in the single-bond nature after HAT, as evident from the observed increase in bond length to 1.44 Å. On the basis of these observed changes in the bond lengths and the electron and spin density distribution obtained from QM/MM dynamics after HAT, a probable mechanism of radical stabilization is shown in Figure 3C. Radical delocalization in the A ring produces numerous alternate possible resonating structures (shown in Figure 3C, left), and the most stable form of daidzein radical is shown in Figure 3C (middle structure). As evident from Figure 3, after HAT, the 7-OH daidzein radical undergoes extensive delocalization of electrons and its electron density appeared to be shifting from O7 to C10 with formation of the para-quinonoid structure, a highly stable resonating structure. It is to be mentioned that the A ring of daidzein is in conjugation with the C ring. Thus, the presence of radical electron density on C10 due to electron delocalization induces additional electron sharing between C10 and C4, resulting in a decrease in the C10−C4 bond (from 1.49 to 1.48 Å) length and an increase in the C4−O4 bond (from 1.23 to 1.25 Å). It is to be noted that during spin transfer from the OH• radical to daidzein there is a significant spin density on O4 atom, and with further electronic redistribution in the A ring of the daidzein radical with simulation time the spin density on the O4 atom increases further. This gives rise to another less probable alternating resonating structure, as shown in Figure 3C (right). F

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 6. (A) Time sequence of the hydrogen-atom transfer of 1000 fs QM/MM dynamics targeting both the 7-OH and the 4′-OH sites of daidzein. Daidzein and the radical are shown in stick representation. Oxygen, hydrogen, and carbon are colored red, white, and cyan. (B) Distance dynamics between H (of 4′-OH and 7-OH sites of daidzein, respectively) and O (of •OH radical) with simulation time. (C) Structure of daidzein diradical after HATs from both 7-OH and 4′-OH sites.

3.3. HAT Reaction of •OH with Daidzein Through the 4′-OH Group in the B Ring. The 4′-OH site of daidzein can also participate in the HAT reaction with •OH to form a stable water as evident from 1000 fs QM/MM dynamics. Results are summarized in Figure 4. Figure 4 shows the HAT pathway involving the 4′-OH site of daidzein. The time sequence of the QM/MM dynamics reveals that the radical approaches toward the B ring of diadzein, and at ∼11 fs, it attacks the hydrogen at the 4′-OH group to form the activated complex. At ∼38 fs the HAT reaction is complete with formation of a water molecule. Figure 4B shows the distance dynamics between the •OH radical and the hydrogen atom of the 4′-OH site of daidzein. Again, a similar angular dependency of HAT has been observed, shown in Figure 4C. HAT takes place when the H−O•−H4′ angle is ∼105°, the typical angle of water. After HAT there is formation of daidzein radical, and the pathways of radical stabilization are shown in Figure 5. The B ring of diadzein is highly aromatic in character with a highly delocalized π-electron cloud over the entire ring (Figure 1B). This gives rise a symmetric bond-length distribution of 1.40 Å (Figure 1A) throughout the B ring. After HAT from the 4′-OH site, there is radical formation at this site which undergoes extensive delocalization with the adjoining B ring. Interesting to note is the spin distribution in the B ring during the HAT reaction. Around 11 fs of this QM/MM simulation trajectory, when the •OH radical first approached the 4′-OH site to form the initial activated complex, most of the spin density is on the oxygen atom of the •OH radical. In this structure the B ring is still highly aromatic in nature with a symmetric π-electron delocalization. This complex undergoes significant spin and electronic redistribution with angular reorientation of the •OH radical in the complex. Around 32 fs, there is a spin transfer from the •OH radical to the 4′-OH oxygen of daidzein, which is concomitant with the π-electron redistribution in the B ring. There is an increase in the π-electron densiy on both the C2′−

C3′ and the C5′−C6′ bonds, while it decreases significantly for the C3′−C4′, C4′−C5′, C1′−C6′, and C1′−C2′ bonds. It is immediately followed by dissociation of the activation complex to form water and daidzein radical. In the remaining simulation time, the daidzein radical electron density undergoes further delocalization to form a stable radical. In stable 4′-OH daidzein radical there is ketone formation between C4′−O4′, as evident from a decrease in bond length from 1.38 to 1.28 Å. Subsequent to ketone formation, the C2′− C3′ and C5′−C6′ bond lengths decrease from 1.40 to 1.38 Å along with a significant increase in π-electron density, which indicates an increase in the double-bond character of these two resonating bonds. On the other hand, after HAT reaction, the C3′−C4′ and C4′−C5′ bond distances increase from 1.40 to 1.45 Å and the C1′−C6′ and C1′−C2′ bond distances also increase from 1.40 to 1.44 Å. It is interesting to note that there is considerable reduction in the π-electron density in all those bonds. This indicates a shift toward the single-bond character of these bonds. This bond distance dynamics with electron and spin distribution dynamics reveals a probable mechanism of radical stabilization, as shown in Figure 5C. The daidzein radical is stabilized to a stable resonating structure as shown in Figure 5C (middle) by extensive delocalization of electrons over the entire B ring, as evidenced from the changes in the bond lengths and electron distributions. In the most stable resonating structure, the B ring appears to be in para-quinonoid form, which is further supported by the fact that there is a significant spin density on the C1′ atom. In the daidzein molecule, there is extensive conjugation of the electron cloud between the B and the C rings through the C3−C1′ bond, which has significant double-bond character (Figure 1). It is evident that after hydrogen-atom transfer from the O4′ site the bond distance of the C3−C1′ bond increases from 1.40 to 1.46 Å. Also, the π-electron density has been significantly reduced. This indicates that the bond loses its double-bond nature. However, still there is π-electron delocalization between the B G

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 7. Changes in electron density distribution (A) and spin density distribution (B) during the HAT reaction from both 4′-OH and 7-OH sites of daidzein.

and the C rings through the C3−C1′ bond. This gives rise to a second less probable resonating structure shown in Figure 5C (left). This observation is further supported by the fact that in a stable 4′-OH daidzein radical there is significant spin density on the C2 atom of the C ring. 3.4. HAT Reaction of Two OH• Radicals with Daidzein. We further studied the simultaneous exposure of both OH groups of daidzein with two hydroxyl radicals. Daidzein can scavenge two •OH, and the HAT reaction pathways and radical stabilization dynamics have been analyzed using QM/MM simulation. Results are summarized in Figure 6. Changes in electron and spin densities during the HAT reaction with two • OH radicals are also analyzed and shown in Figure 7. As evident from Figure 6, there is hydrogen-atom transfer from both hydroxyl sites of daidzein. As per the initial frame chosen here, the HAT reaction from the 4′-OH site is completed within 42 fs of the simulation when the other hydroxyl radical on the 7-OH site is not able to react with the C7−OH group of daidzein. Simulation timeline reveals that the HAT reaction from the B ring does not have any influence on the HAT reaction from the A ring of daidzein, as evident from the comparable timeline of HAT observed when only a single hydroxyl radical is present at the 7-OH site of daidzein in the A ring. This may be due to the fact that after HAT from the B ring the electron delocalization between the B and the C rings reduces significantly. Thus, alteration of the electronic cloud due to HAT in the B ring does not have any influence on the electronic distribution of the A ring of daidzein. Figure 6C shows the structure of daidzein diradical with the obtained bond-length distribution. Bond-length distribution analysis reveals that HATs lead to formation of diketone on both the C7−O7 and the C4′−O4′ sites of daidzein. Electron density distribution analysis reveals (Figure 7A) that both the A and the B rings appear as the para-quinonoid form. The

electronic delocalization of the C ring remains unaltered after diradical formation. QM/MM dynamics performed here reveals that daidzein has the ability to scavenge two hydroxyl radicals, and the daidzein diradical stabilizes with formation of diketone and para-quinonoid structure. An interesting revelation is the spin density distribution during formation of daidzein monoand diradicals. During formation of daidzein monoradical due to the HAT reaction from the 4′-OH site the radical spin density was first localized on the 4′O oxygen but then the spin density diffuses over the B ring of daidzein. The spin density is found to be mostly concentrated on C1′, C5′, and 4′-O atoms. During the second HAT reaction from the 7-OH site we see spin-density transfers from the oxidation sites to the central C ring. After the second HAT, the daidzein diradical ground state will be in a zero spin state. In the immediate aftermath of the second HAT we see localized spin densities on the C ring and at the 4′-O and 7-O oxygens though the total integrated spin density of the system appears zero. However, as the system relaxes over time, the localized spin densities mostly disperse, leaving some residual spin densities on 7-O and 4′-O. 3.5. Energetics of HAT Reactions. A reaction coordinate analysis has been performed to identify the possible reaction intermediates, activated complex, and minimum energy structure of the product. The donor−acceptor distances of H4′···O (of radical •OH) and H7···O (of second •OH radical) are considered as reaction coordinates for the HAT reactions. Plotting the energy with the reaction coordinate reveals a possible activated complex (TS) and product minima, shown in Figure 8. For the energy of the active site we considered summation of the energy of the QM system (EQM) and the interaction energy between the QM and the MM systems (EQM/MM). Initially the hydroxyl radical is 1.75 Å away from the reaction center of the diadzein. As it approaches close to the daidzein, energy H

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Interestingly, the calculated activation energies for the forward reactions involving both sites of daidzein are quite similar with a difference of ∼1.5 kcal/mol. Since, the hydroxyl radical is highly reactive, hydrogen-atom abstraction from both the sites is highly feasible at room temperature. There is an apparent paradox that HAT from the 4′-OH site happens much quicker over the other HAT reaction, although the activation energy required for the former reaction is ∼1.5 kcal/mol higher over the other reaction pathway. This may be due to apparent consideration of the initial structure for HAT reaction. It is to be noted that HAT is highly dependent on the initial position of the radical and the angular dependency of the HAT reaction coordinate. HAT reaction from the 4′-OH site leads to formation of a more stable product than the other reaction pathway. The calculated difference in ΔH for the HAT reaction involving the two reaction pathways is 29.49 kcal/mol. Thus, HAT reaction involving the 4′-OH site of daidzein is enthalpically more favorable. This observation is in accordance with the current hypothesis that the B ring of polyphenolic compounds is the more preferred site to scavenge a radical.14 Our result is also in accordance with the calculated bond dissociation enthalpy for daidzein obtained using B3LYP/631G(p) basis for daidzein in vacuo, which shows that the BDE is ∼7 kcal/mol less for the 4′-OH bond compared to the 7-OH bond.38 It is to be noted that explicit consideration of water molecules provides further enthalpic stabilization of the HAT reaction pathway involving 4′-OH. The activation energy for the HAT reaction involving 7-OH being ∼1.5 kcal/mol less than the HAT reaction involving the other site of daidzein is attributed to stabilization of the transition state (TS) due to solvation as evident from Figure 10. The TS of the HAT pathway involving the 7-OH site stabilizes more by hydrogen-bonding interactions with nearby water molecules compared to the other TS involving the 4′-OH site. Contrasting to this observation, the product daidzein radical formed after HAT reaction involving the 4′-OH site is involved in more hydrogen-bonding interactions with nearby water compared to the daidzein radical formed after HAT reaction involving the 7-OH site. This stabilization obtained through the hydrogen-bonding network can be partly interpreted by the high enthalpic stabilization obtained in HAT reaction involving the 4′-OH site.

Figure 8. Reaction coordinate diagram for the HAT reaction involving both sites of daidzein.

increases and reaches a maximum when the distance is ∼1.6− 1.5 Å. This energy maximum can be interpreted as a transition state or activated complex energy, where the radical makes a loose bonding interaction with the hydrogen atom of the hydroxyl group of daidzein. This is followed by energy downhill to the product minima with complete transfer of the hydrogen atom from daidzein to the hydroxyl radical. During the rest of the simulation time the product remains stable and its energy fluctuates around the product minima energy funnel. From the product energy funnel, energy minima have been identified and an energy profile diagram has been computed (Figure 9).

4. CONCLUSIONS We studied the possible mechanism of hydroxyl radical scavenging pathways by daidzein isoflavone in the presence of explicit water by employing the combined QM/MM molecular dynamics as implemented in GROMACS-CPMD. The QM/ MM dynamics allows us to study the electron density and electron spin density redistribution on the fly during hydrogenatom transfer (HAT) reactions from the 4′-OH and 7-OH sites of daidzein and formation and stabilization of daidzein monoand diradicals. We found that daidzein can scavenge hydroxyl radicals using both of its OH functional groups. The low reaction energy barrier indicates the high feasibility of the HAT reaction on both sites, particularly considering the high reactivity of the hydroxyl radical. HAT reaction involving the 4′-OH site is enthalpically highly favorable over the other reaction pathway involving the 7-OH site by ∼29 kcal/mol. Thus, it can be concluded that the radical scavenging reaction involving the B ring of daidzein is the most preferred site, and this is in accordance with the bond dissociation enthalpy (BDE) calculations.38 The high stabilization in terms of

Figure 9. Reaction energy diagram for the HAT reaction involving both sites of daidzein.

It is to be noted that the QM/MM dynamics is highly CPU expensive, and it is not possible to consider diffusion of the radical to the reaction site. We started the simulation by considering the hydroxyl radical within the hydrogen-bonded distances. Thus, our starting geometry is very close to the transition state geometry of the complex. To compute the initial reactant structure we started with the optimized daidzein structure and calculated the energy of the system by placing a hydroxyl radical far away from the reaction site >10 Å, such that the interactions between them is negligible. Energy diagram analysis reveals that the HAT reaction is highly exothermic. I

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

Figure 10. Solvated structure of the activated complex (TS, left) and product (right) for (A) HAT reaction involving the 7-OH group and (B) HAT reaction involving the 4′-OH group.

■

ACKNOWLEDGMENTS Authors acknowledge financial support from MS-INBRE (USM-GR04015-05-9; NIH/NCRR P20RR016476) and EPSCoR (EPS-0903787; Sub-contract: 190200-362492-10).

reaction enthalpy can be attributed partly to the effect of explicit solvent and consequent hydrogen-bonding interactions with the neighboring water molecules. The HAT reaction is highly exothermic in nature. After HAT, there is ketone formation at the corresponding reaction site of daidzein. Subsequent electron delocalization further stabilizes the newly formed daidzein radical with the adjacent aromatic ring appear in a para-quinonoid structure, a highly stable resonating form. Our result can explain the observed high antioxidant potency of daidzein and is in accordance with the experimental observations. We report for the first time the applicability of the QM/MM methodology to provide mechanistic insight into the radical scavenging activity of flavonoids. The QM/MM dynamics allows hydrogen bonding between the QM and the MM subsystems and explicit polarization of the QM subsystem by the MM subsystem atoms, thus allowing us to account for the effect of environment (here water) on the reaction site (here daidzein 4′-OH and 7-OH sites together with radical • OH). The applicability of QM/MM methodology, where the reaction center requires ab initio level quantum description but warrants the effect of the environment to be taken into account, opens up new opportunities to study the reaction mechanism of biomolecular systems.

■

■

REFERENCES

(1) Tapas, A. R.; Sakarkar, D. M.; Kakde, R. B. Trop. J. Pharm. Res. 2008, 7, 1089−1099. (2) Woo, H.-H.; Jeong, B. R.; Hawes, M. C. Biotechnol. Lett. 2005, 27, 365−374. (3) Meijer, D. K. F.; Van der sluijs, P. Pharm. Res. 1989, 6, 105−118. (4) Lamson, D. W.; Brignall, M. S. Altern. Med. Rev. 2000, 5, 196− 208. (5) Chakraborty, S.; Kumar, S.; Basu, S. Neurochem. Int. 2011, 58, 914−923. (6) Cherdshewasart, W.; Subtang, S.; Dahlan, W. J. Pharm. Biomed. Anal. 2007, 43, 428−434. (7) Honoré, E. K.; Williams, J. K.; Anthony, M. S.; Clarkson., T. B. Fertil. Steril. 1997, 67, 148−154. (8) MacDonald, R. S.; Guo, J.; Copeland, J.; Browning, J. D., Jr.; Sleper, D.; Rottinghaus, G. E.; Berhow, M. A. J. Nutr. 2005, 135, 1239−1242. (9) Messina, M.; Wu, A. H. Am. J. Clin. Nutr. 2009, 89, 1673S− 1679S. (10) Arora, A.; Nair, M. G.; Strasburg, G. M. Arch. Biochem. Biophys. 1998, 356, 133−141. (11) Kim, S. Y.; Kim, S. J.; Lee, J. Y.; Kim, W. G.; Park, W. S.; Sim, Y. C.; Lee, S. J. J. Am. Coll. Nutr. 2004, 23, 157−162. (12) Hall, E. J. Radiobiology for the Radiologist, 5th ed.; Lippincott Williams & Wilkins: Baltimore, 2000. (13) Wenli, Y.; Yaping, Z.; Bo, S. Food Chem. 2004, 86, 525−529.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. J

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

(14) Pahari, B.; Chakraborty, S.; Chaudhuri, S.; Sengupta, B.; Sengupta, P. K. Chem. Phys. Lipids 2011, 165, 488−496. (15) Chakraborty, S.; Basu, S.; Lahiri., A.; Basak, S. J. Mol. Struct. 2010, 977, 180−188. (16) Erkoc, S.; Erkoc, F.; Keskin, N. J. Mol. Struct: THEOCHEM 2003, 631, 141−146. (17) Trouillas, P.; Marsal, P.; Siri, D.; Lazzaroni, R.; Duroux, J. L. Food Chem. 2006, 97, 679−688. (18) Garcia, J. E. Chem. Phys. Lett. 2004, 388, 274−278. (19) Velkov, Z.; Balabanova, E.; Tadjer, A. J. Mol. Struct: THEOCHEM 2007, 821, 133−138. (20) Queiroz, A. N.; Gomes, B. A. Q.; Moraes, W. M., Jr.; Borges, R. S. Eur. J. Med. Chem. 2009, 44, 1644−1649. (21) Mendoza-Wilson, A. M.; Glossman-Mitnik, D. J. Mol. Struct: THEOCHEM 2006, 761, 97−106. (22) Biswas, P. K.; Gogonea, V. A. J. Chem. Phys. 2005, 123, 164114− 164122. (23) Sadasivam, K.; Kumaresan, R. Mol. Phys. 2011, 109, 839−852. (24) Jorgensen, W. L.; Rives, T. J. Am. Chem. Soc. 1988, 110, 1657− 1666. (25) Berendsen, H. J. C.; Spoel, D. V. D.; Drunen, R. V. Comput. Phys. Commun. 1995, 91, 45−56. (26) Lindahl, E.; Hess, B.; Spoel, D. V. D. J. Mol. Model. 2001, 7, 306−317. (27) Car, R.; Parrinello, M. Phys. Rev. Lett. 1985, 55, 2471−2474. (28) Chalmet, S.; Ruiz-Lopez, M. F. J. Chem. Phys. 2006, 124, 194502−1−194502−6. (29) Martyna, G. J.; Tuckerman, M. E. J. Chem. Phys. 1999, 110, 2810. (30) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (31) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (32) Zhang, J.; Du, F.; Peng, B.; Lu, R.; Gao, H.; Zhou, Z. J. Mol. Struct: THEOCHEM 2010, 955, 1. (33) Russo, N.; Toscano, M.; Uccella, N. J. Agric. Food Chem. 2000, 48, 3232. (34) Fiorucci, S.; Golebiowski, J.; Cabrol-Bass, D.; Antonczak, S. J. Agric. Food Chem. 2007, 55, 903. (35) Leopoldini, M.; Marino, T.; Russo, N.; Toscano, M. Theor. Chem. Acc. 2004, 111, 210. (36) Antonczak, S. J. Mol. Struct.: THEOCHEM 2008, 856, 38. (37) Abolfath, R. M.; Bisaws, P. K.; Rajnarayanam, R.; Brabec, T.; Kodym, R.; Papiez, L. J. Phys. Chem. A 2012, DOI: 10.1021/ jp300258n. (38) Anbazhagan, V.; Kalaiselvan, A.; Jaccob, M.; Venuvanalingam, P.; Renganathan, R. J. Photochem. Photobiol. B: Biol. 2008, 91, 143− 150.

K

dx.doi.org/10.1021/jp303543z | J. Phys. Chem. A XXXX, XXX, XXX−XXX