MM Study on the Reaction Mechanism of O6-Alkylguanine−DNA

O6-Alkylguanine−DNA alkyltransferase (AGT), also known as O6-methylguanine-DNA methyltransferase (MGMT), is a crucial DNA repairing protein, which ...
2 downloads 0 Views 4MB Size
15296

J. Phys. Chem. B 2010, 114, 15296–15300

QM/MM Study on the Reaction Mechanism of O6-Alkylguanine-DNA Alkyltransferase Qianqian Hou, Likai Du, Jun Gao, Yongjun Liu,* and Chengbu Liu Key Lab of Colloid and Interface Chemistry, Ministry of Education, School of Chemistry and Chemical Engineering, Shandong UniVersity, Jinan, Shandong 250100, China ReceiVed: July 19, 2010; ReVised Manuscript ReceiVed: September 17, 2010

Combined quantum-mechanical/molecular-mechanical (QM/MM) approaches have been applied to investigate the detailed reaction mechanism of human O6-alkylguanine-DNA alkyltransferase (AGT). AGT is a direct DNA repair protein that is capable of repairing alkylated DNA by transferring the methyl group to the thiol group of a cysteine residue (Cys145) in the active site in an irreversible and stoichiometric reaction. Our QM/MM calculations reveal that the methyl group transferring step is expected to occur through two steps, in which the methyl carbocation generating step is the rate-determining step with an energy barrier of 14.4 kcal/mol at the QM/MM B3LYP/6-31G(d,p)//CHARMM22 level of theory. It is different from the previous theoretical studies based on QM calculations by using a cluster model in which the methyl group transferring step is a one-step process with a higher energy barrier. 1. Introduction

SCHEME 1: Proposed Reaction Mechanism of AGT

DNA alkylation can be caused by both endogenous and exogenous DNA-damaging agents,1 such as S-adenosylmethionine in the cell, methylmethane sulfonate (MMS) or N-methylN′-nitro-N-nitrosoguanidine (MNNG) in the environment.2-5 Alkylation adducts frequently occur at the O6 position of guanine, resulting in O6-methylguanine (O6-mG), which if not repaired leads to GC-to-AT transition mutations and cancer.6,7 O6-Alkylguanine-DNA alkyltransferase (AGT), also known as O6-methylguanine-DNA methyltransferase (MGMT), is a crucial DNA repairing protein, which protects against the mutagenic and carcinogenic effects of O6-alkylguanine lesions.8-10 It effectively transfers the methyl group to the thiol group of its active site cysteine residue (Cys145) as the mechanism proposed in Scheme 1. This irreversible and stoichiometric reaction brings about the inactivation of AGT and the restoration of the DNA. After completion, the alkylated enzyme can also recognize the subsequent alkylating agents.11 Recent studies on the AGT-DNA complex have provided the structural details clearly and distinctly. These available crystal structures indicate that AGT binds to DNA following the minor groove which is called the helix-turn-helix (HTH) motif. The alkylated base is flipped out from the DNA into the AGT active site pocket, which has advantages for repairing. It is believed that Cys145 in the active site pocket reacts with the alkylated guanine, which has a very high reactivity due to the interaction with Glu172-His146-water-Cys145 hydrogen bond network.12,13 Previous theoretical studies on the reaction mechanism of AGT have been largely based on calculations using quantum mechanics (QM) methods. One of the recent studies was conducted by P. Georgieva and F. Himo.14 They used a cluster model at the B3LYP/6-31G(d,p) level in the Gaussian 03 program package, in which only the essential residues in the reactive site were included. The transition state for a SN2 methyl transfer was located, and the energy barrier for this step was calculated to be 23 kcal/mol in the polarizable dielectric medium, which is considered to be a few kilocalories per mol * Corresponding author. E-mail: [email protected].

higher than the common enzyme-catalyzed reactions. On the basis of quantum mechanics methods,15 P. K. Shukla and P. C. Mishra concluded that the conversion of cysteine to cysteine thiolate anion is a crucial factor for the catalytic reaction. In another quantum mechanics study, which involved two computational models, the same mechanism was also proved.16 In the first model, Lys165 formed a hydrogen bond to O6-mG, and in the second model, Lys165 is hydrogen-bonded to Tyr114. A little lower barrier energy was gained for the methyl group transferring step by using the former model. In all of the above-mentioned QM studies, the protein environment was considered using solvation calculations in a polarizable dielectric medium having dielectric constant the

10.1021/jp106714m  2010 American Chemical Society Published on Web 11/01/2010

Reaction Mechanism of AGT

J. Phys. Chem. B, Vol. 114, No. 46, 2010 15297

Figure 1. The crystal structure of AGT in complex with methylated DNA (PDB code: 1T38). Ser145 has been mutated to Cys145. The methylated guanine base is shown in a stick; crystal water molecules, in the red dots; and repaired missing residues (36-55), in the red loop.

same as chorobenzene, which has been suggested to represent the protein environment accurately. In this work, we report the first quantum mechanical/molecular mechanical (QM/MM) study of the reaction mechanism of AGT. The basic idea is to divide the whole system into a QM region and a MM region.17-22 In the modeling, the chemically active region that is directly involved with the bond forming and breaking is described by quantum mechanics, and the other, which includes the remaining part of protein and solvent, is described by molecular mechanics. Applying mixed QM/MM methods to determine the pathway of the catalytic mechanism is believed to achieve a better picture of the AGT-DNA complex. 2. Computational Details 2.1. Computational Model. The structure of AGT in complex with the alkylated DNA was constructed following the crystallographic atomic coordinate taken from Protein Data Bank (PDB ID code: 1T38).12 After the missing residues (36-55) were repaired by Modeler23 using a template of another crystal structure of AGT (PDB ID code: 1EH6),24 the mutated residue Ser145 was changed back to Cys145. The obtained model is displayed in Figure 1. Considering the experimental condition, the protonation states of residues were visualized and checked by the VMD program,25 and the missing hydrogen atoms were added via the HBUILD facility in the CHARMM package.26 The system was solvated with 6118 water molecules that formed a water sphere of 30 Å radius centered on the thiolate anion, and then it was neutralized by 24 Na+’s at random positions. These generated a neutral system of 21 809 atoms. The prepared structure was equilibrated with a series of minimizations and a 400 ps MD simulation. All these steps were performed by the CHARMM2227 force field. The obtained structure is shown in Figure 2. 2.2. QM/MM Calculations. In the QM/MM calculations, a QM region contains 61 atoms in the reactive site, including residues Cys145, His146, Ser159, O6-mG of alkylated DNA, and a crystal water molecule, and a MM region contains the remaining 21 748 atoms of the enzyme and water molecules. During the subsequent QM/MM geometry optimization, the QM region and 742 MM atoms within 10 Å of Cys145 were allowed to move, whereas all the other atoms were held frozen for simplification. In the enzyme reaction paths search, the QM part was treated quantum mechanically at the B3LYP/6-31G(d,p)

Figure 2. The structure of the model system treated by the CHARMM package.

level, and the MM part was considered by molecular mechanics using the CHARMM2227 force field. Frequency calculations on the AGT-DNA complex, transition states, intermediates, and final product were carried out at the same levels. The electronic embedding scheme28 was used, where the MM charges were incorporated into the one-electron Hamiltonian of the QM calculations to avoid hyperpolarization of the QM wave function.29,30 Hydrogen link atoms with charge shift model for the QM/MM boundary were adopted in the QM/MM treatment.31 The QM/MM calculations were carried out with the ChemShell package32 integrating Turbomole33 and DL-POLY programs.34 3. Results and Discussion 3.1. Structure of the AGT-DNA Complex. The reported crystal structures of the AGT-DNA complex indicate that the prototypical DNA major groove-binding HTH motif mediates unprecedented minor groove DNA binding.12 The alkylated guanine deoxynucleotide is then flipped out from the base stack into the AGT active site pocket through a 3′-phosphte rotation. A Glu172-His146-water-Cys145 hydrogen bond network is regarded as the active center of the enzyme. In addition, the Tyr114 hydroxyl donates a hydrogen bond to the N3 of O6-mG with bond length of 0.188 nm. The Tyr158 side chain packs against the O6-alkyl groups, providing a hydrophobic environment for small O6-alkyl adducts. Figure 3a shows the structure of the active site from the crystal, and Figure 3b displays the QM/MM optimized structure of the active site. Notable changes can be observed after the mutation of C145S in Figure 3b. On one hand, the orientations of the methyl group attached to the DNA base and the thiol group changed. In Figure 3a, the distance between the carbon atom of the methyl and the oxygen atom of Ser145 is 0.403 nm, whereas in Figure 3b, the distance between that carbon atom and the sulfur atom of Cys145 is only 0.366 nm. Meanwhile, the bond angle of C(O6mG)-O(Ser145)-O(water) in the crystal is 68.86°, and that of C(O6-mG)-S(Cys145)-O(water) is 134.58° in Figure 3b. The structure in the present study shows a little difference from that calculated by F. Himo14 an P. C. Mishra,15 which showed that

15298

J. Phys. Chem. B, Vol. 114, No. 46, 2010

Hou et al.

Figure 4. Energy profile for proton and methyl transfer.

Figure 3. (a) The structure of the active site from the crystal and (b) the optimized structure of the active site by the QM/MM method. Distances are given in 0.1 nm.

cis-O6mG was more stable than trans-O6mG by ∼2.5 kcal/mol.15 It is easy to understand because in their QM calculations, the surrounding residues and solvent molecules were modeled by solvation calculation, and the steric repulsion of the real residues was not included. In the preparation of the initial reactant complex, we have tried different structures, including cis- and trans-O6-mG with respect to N1C6O6C(Me), but the trans conformer was always obtained. On the other hand, after optimization, the relative positions of Cys145, crystal water, and Ser159 also changed a little. We believe that these conformational changes in the active site have great influence on the repair reaction of AGT. 3.2. Reaction Paths. According to the previously proposed reaction mechanism and calculation results,15 we studied the proton-transfer and methyl group transfer processes by QM/ MM calculations. On the basis of our calculations, the total reaction mechanism includes three elementary reactions. The first one is the water-assisted proton transfer from Cys145 to His146, generating the thiolate nucleophile. The next step is the breaking of the methoxyl group to form a methyl carbocation, and the final step is the methyl carbocation bonds to the thiolate anion of Cys145 to finish the repair reaction. The corresponding energy profiles in the reaction path are shown in Figure 4. 3.2.1. Water-Assisted Proton Transfer. The stationary points concerning the proton transfer are shown in Figure 5. As illustrated in Figure 5a, in the reactant complex, there is a Glu172-His146-water-Cys145 hydrogen bond network (Glu172 is not shown), which may mediate the hydrogen transfers. This transfer is considered to be a concerted mechanism in our QM/ MM calculations, in which W1 loses a proton to the N atom of His 146 and simultaneously gains a proton from Cys145. The hydrogen bond distance of N-H has been shortened from 0.156 nm in the reactant to 0.125 nm in TS1. Meanwhile, the distance of O-H has changed from 0.200 nm in the reactant to 0.115 nm in TS1. With the assistance of a water molecule, the calculated energy barrier is 5.0 kcal/mol, which is lower than the result (9.3 kcal/mol) of F. Himo.14 After passing the transition state, the system reaches a thiolate anion intermediate (I1). The hydrogen bonding interaction between the protonated

Figure 5. Optimized structures of reactant (R), transition state (TS1), and intermediate (I1) for proton transfer. Distances are given in 0.1 nm.

histidine and Glu172 stabilizes the intermediate (I1). In intermediate I1, the distance between the S and C atoms has been shortened to 0.368 nm, and the charge of Cys145 thiolate is -0.542, which facilitate the methyl transfer process. 3.2.2. Methyl Group Transfer. After the formation of the thiolate nucleophile in the previous step, the enzyme undergoes a process of methyl group transfer, which favors two steps with barriers of 14.4 and 9.5 kcal/mol, respectively, as shown in Figure 4. The optimized structures of transition states (TS2 and TS3), intermediate (I2), and final product (P) are shown in Figure 6. The sketch map for the methyl transfer reaction is displayed in Figure 7.

Reaction Mechanism of AGT

J. Phys. Chem. B, Vol. 114, No. 46, 2010 15299

Figure 8. Structure of transition state (TS2) and surrounding residues.

Figure 6. Optimized structures of transition state (TS2 and TS3), intermediate (I2), and product (P) for methyl transfer. Distances are given in 0.1 nm.

Figure 7. Sketch map for the methyl transfer reaction.

On the basis of our QM/MM calculations, this methyl group transfer is actually a migration process of methyl carbocation. In TS2, the net charge of the methyl group is only 0.98 and the spin density is 0, implying the characteristic of the methyl carbocation. Frequency calculations give the unique imaginary frequencies of 154i and 155i cm-1 for TS2 and TS3. As shown in the structures of TS2 and I2, the methyl group departs gradually from the O atom of O6-mG until the formation of methyl carbocation. The bond distance of O-C changes from 0.188 nm in TS2 to 0.302 nm in I2.

TS3 corresponds to the reversal of methyl carbocation, which facilitates its approaching to the thiolate anion. The barrier of TS3 is only 9.5 kcal/mol, suggesting the reversing process is facile. It should be noted that if the cluster model was used for studying the methyl transfer process, the intermediate I2 was not found, such as in the report of F. Himo.14 We ascribe it to the different calculating models. In the study of F. Himo,14 for the sake simplify, the cluster model was used, and the protein environment was not considered. But in our QM/MM calculations, all the surrounding residues were included, and the methyl carbocation was stabilized by the protein environment. Figure 8 shows the residues around the reactive site. One can see that some surrounding residues interact directly with the key residues that are responsible for the reaction mechanism. First, the deprotonated carboxyl group of Glu172 stabilizes the structure of the intermediate I1, forming a hydrogen bond to Cys145 and greatly facilitating the proton transfer step. Then, Tyr114 hydroxyl donates a hydrogen bond to the nitrogen atom of O6-mG, which reduces the negative charge on the guanine base. Another part of the catalytic effect comes from the residues surrounding the methyl group, including Arg135, Asn137, Pro140, Tyr158, and Ser159, as suggested experimentally.35-39 Ser159 provides a high degree of stabilization to the methyl transferring step and in particular stabilizes the repaired guanine through a strong hydrogen bond. These residues greatly affect the catalytic paths during the process of methyl transferring that the methyl plane undergoes an inversion, as indicated earlier. This is the main reason that demethylation occurs through two steps. Rather, electrons transfer between these residues, and the reaction center demonstrates a number of reasons for methyl carbocation transferring process. Owing to the QM/MM model, important residues that may stabilize the transition state are included, and the movements of the substrates and residues could be well described. This highlights the value of QM/MM calculations, as compared with QM calculations of enzyme catalytic reactions. 4. Conclusions In the present study, the detailed mechanism of human DNA repair reaction of O6-alkylguanine-DNA alkyltransferase (AGT) was investigated by using combined QM/MM approaches. Reaction paths and activation barriers have been achieved in our study, and calculations show that the methyl-transfer step is a migration process of methyl carbocation through two steps.

15300

J. Phys. Chem. B, Vol. 114, No. 46, 2010

The rate-limiting step is the formation of methyl carbocation with an energy barrier of 14.4 kcal/mol. The residues Glu172, Ser159, and Tyr114 have been confirmed to play key roles in this stepwise process. Compared with the different process described in quantum study, our QM/MM calculations have shown that the protein environment is crucial for this enzymatic catalytic methyl-transfer reaction. Further investigations of this type of enzyme are ongoing. Acknowledgment. This work was supported by the Natural Science Foundation of China (20873075) and Independent Innovation Research Fund of Shandong University (2009JC018). References and Notes (1) Lindahl, T.; Wood, R. D. Science 1999, 286, 1897. (2) Shukla, P. K.; Mishra, P. C.; Suhai, S. Int. J. Quantum Chem. 2007, 107, 1270. (3) Wyatt, M. D.; Pittman, D. L. Chem. Res. Toxicol. 2006, 19, 1580. (4) Mishina, Y.; Duguid, E. M.; He, C. Chem. ReV. 2006, 106, 215. (5) Loechler, E. L.; Green, C. L.; Essigmann, J. M. Proc. Natl.Acad. Sci. U.S.A. 1984, 81, 6271. (6) Hickman, M. J.; Samson, L. D. Mol. Cell 2004, 14, 105. (7) Meikrantz, W.; Bergom, M. A.; Memisoglu, A.; Samson, L. Carcinogenesis 1998, 19, 369. (8) Tubbs, J. L.; Pegg, A. E.; Tainer, J. A. DNA Repair 2007, 6, 1100. (9) Pegg, A. E.; Fang, Q.; Loktionova, N. A. DNA Repair 2007, 6, 1071. (10) Mitra, S. DNA Repair 2007, 6, 1064. (11) Kaina, B.; Fritz, G.; Mitra, S.; Coquerelle, T. Carcinogenesis 1991, 12, 1857. (12) Daniels, D. S.; Woo, T. T.; Luu, K. X.; Noll, D. M.; Clarke, N. D.; Pegg, A. E.; Tainer, J. A. Nat. Struct. Mol. Biol. 2004, 11, 714. (13) Duguid, E. M.; Rice, P. A.; He, C. J. Mol. Biol. 2005, 350, 657. (14) Georgieva, P.; Himo, F. Chem. Phys. Lett. 2008, 463, 214. (15) Shukla, P. K.; Mishra, P. C. Phys. Chem. Chem. Phys. 2009, 11, 8191.

Hou et al. (16) Jena, N. R.; Shukla, P. K.; Jena, H. S.; Mishra, P. C.; Suhai, S. J. Phys. Chem. B 2009, 113, 16285. (17) Senn, H. M.; Thiel, W. Angew. Chem., Int. Ed. 2009, 48, 1198. (18) Gao, J. L.; Ma, S. H.; Major, D. T.; Nam, K.; Pu, J. Z.; Truhlar, D. G. Chem. ReV. 2006, 106, 3188. (19) Cui, Q.; Karplus, M. AdV. Protein Chem. 2003, 66, 315. (20) Gao, J. L.; Truhlar, D. G. Annu. ReV. Phys. Chem. 2002, 53, 467. (21) Field, M. J.; Bash, P. A.; Karplus, M. J. Comput. Chem. 1990, 11, 700. (22) Warshel, A.; Levitt, M. J. Mol. Biol. 1976, 103, 227. (23) Marti-Renom, M. A.; Stuart, A. C.; Fiser, A.; Sanchez, R.; Melo, F.; Sali, A. Annu. ReV. Biophys. Biomol. Struct. 2000, 29, 291. (24) Daniels, D. S.; Mol, C. D.; Arvai, A. S.; Kanugula, S.; Pegg, A. E.; Tainer, J. A. EMBO J. 2000, 19, 1719. (25) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996, 14, 33. (26) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.; Swaminathan, S.; Karplus, M. J. Comput. Chem. 1983, 4, 187. (27) MacKerell, A.; et al. J. Phys. Chem. B 1998, 102, 3586. (28) Bakowies, D.; Thiel, W. J. Phys. Chem. 1996, 100, 10580. (29) Sherwood, P.; de Vries, A. H.; Collins, S. J.; Greatbanks, S. P.; Burton, N. A.; Vincent, M. A.; Hillier, I. H. Faraday Discuss. 1997, 106, 79. (30) Bakowies, D.; Thiel, W. J. Phys. Chem. 1996, 100, 10580. (31) de Vries, A. H.; Sherwood, P.; Collins, S. J.; Rigby, A. M.; Rigutto, M.; Kramer, G. J. J. Phys. Chem. B 1999, 103, 6133. (32) Sherwood, P.; et al. J. Mol. Struct. (THEOCHEM) 2003, 632, 1. (33) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Chem. Phys. Lett. 1989, 162, 165. (34) Smith, W.; Forester, T. R. J. Mol. Graphics Modell. 1996, 14, 136. (35) Rasimas, J. J.; Pegg, A. E.; Fried, M. G. J. Biol. Chem. 2003, 278, 7973. (36) Noll, D. E.; Clarke, N. D. Nucleic Acids Res. 2001, 29, 4025. (37) Edara, S.; Goodtzova, K.; Pegg, A. E. Carcinogenesis 1995, 16, 1637. (38) Morgan, S. E.; Kelley, M. R.; Pieper, R. O. J. Biol. Chem. 1993, 268, 19802. (39) Guengerich, F. P.; Fang, Q.; Liu, L.; Hachey, D. L.; Pegg, A. E. Biochemistry 2003, 42, 10965.

JP106714M