Quantum Chemical Modeling of Methanol Oxidation Mechanisms by

Jan 7, 2010 - Institute for Micromanufacturing, Chemical Engineering Program, Louisiana Tech University, Ruston, Louisiana 71272. J. Phys. Chem. A , 2...
0 downloads 0 Views 4MB Size
Download PDF
J. Phys. Chem. A 2010, 114, 1887–1896

1887

Quantum Chemical Modeling of Methanol Oxidation Mechanisms by Methanol Dehydrogenase Enzyme: Effect of Substitution of Calcium by Barium in the Active Site Nagesh B. Idupulapati and Daniela S. Mainardi* Institute for Micromanufacturing, Chemical Engineering Program, Louisiana Tech UniVersity, Ruston, Louisiana 71272 ReceiVed: August 27, 2009; ReVised Manuscript ReceiVed: NoVember 24, 2009

Previous experimental studies have shown that the activation energy for methanol oxidation by naturally occurring Ca2+-containing methanol dehydrogenase (MDH) enzyme is double the methanol activation energy by Ba2+-MDH. However, neither the reason for this difference nor the specific transition states and intermediates involved during the methanol oxidation by Ba2+-MDH have been clearly stated. Hence, an MDH active site model based on the well-documented X-ray crystallographic structure of Ca2+-MDH is selected, where the Ca2+ is replaced by a Ba2+ ion at the active site center, and the addition-elimination (A-E) and hydridetransfer (H-T) methanol oxidation mechanisms, already proposed in the literature for Ca2+-MDH, are tested for Ba2+-MDH at the BLYP/DNP theory level. Changes in the geometries and energy barriers for all the steps are identified, and qualitatively, similar (when compared to Ca2+-MDH) intermediates and transition states associated with each step of the mechanisms are found in the case of Ba2+-MDH. For both the A-E and H-T mechanisms, almost all the free-energy barriers associated with all of the steps are reduced in the presence of Ba2+-MDH, and they are kinetically feasible. The free energy barriers for methanol oxidation by Ba2+MDH, particularly for the rate-limiting steps of both mechanisms, are almost half the corresponding barriers calculated for the case of Ca2+-MDH, which is in agreement with experimental observations. Introduction Methanol dehydrogenase (MDH) is a water-soluble quinoprotein that oxidizes alcohols to their corresponding aldehydes.1 From various X-ray crystallographic studies,2-8 it has been determined that the enzyme is an R2β2 heterotetramer with two subunits R and β. Each heavy subunit (R) has an active site consisting of Ca2+, a pyrroloquinoline quinone (PQQ) cofactor, several amino acids, and also water (W) molecules (Figure 1). According to the structure of Methylobacterium extorquens W3A1,2-5 the water molecules W362, W213, and W615 form hydrogen bonds with Ca2+, PQQ, and GLU177 near the active site,1,9 and the water molecules W130, W131, W134, and W198 form a cluster on the left upper portion of PQQ (Figure 1) and are hydrogen bonded to its oxygen atoms.2-8 The Ca2+ ion coordination with O5, N6, and O10 atoms of PQQ is consistent in all X-ray studies,2-8 showing bond lengths between the ion and those elements in the 2.32-2.47, 2.25-2.77, and 2.30-2.44 Å range, respectively. However, other coordinations between Ca2+ and O12 of GLU177 (2.33-2.74 Å), O11 of ASN 261 (2.27-2.48 Å), O14 of ASP303 (2.37-2.54 Å), and with the oxygen atoms of the W362 (2.51 Å) and W615 (2.62 Å) water molecules (Figure 1) have been the subject of debate in the literature.2-8 Two possible mechanisms for methanol oxidation by the naturally occurring Ca2+-containing MDH have been proposed in the literature; the addition-elimination (A-E) and the hydridetransfer (H-T) mechanisms.1,9 The A-E is considered a threestep mechanism (Figure 2a) that shows a proton (H16) “addition” to ASP303, leading to the formation of a covalent hemiketal intermediate with C5 of PQQ in the first step (Figure 2a). The second step consists of the proton (H16) “elimination” * To whom correspondence should be addressed. E-mail: mainardi@ latech.edu.

Figure 1. X-ray crystal structure of MDH active site. Amino acids labels denote their location in the sequence obtained from the entry 1W6S (Methylobacterium extorquens W3A1) of the Protein Data Bank.5

from ASP303 and transfer to O5 of PQQ, and the final step is characterized by a second proton (H17) transfer from the methoxide to the O4 of PQQ, resulting in the formation of formaldehyde (CH2O).1,9 The H-T mechanism (Figure 2b) involves one proton abstraction and a hydride transfer: H16 and H17 from the methanol molecule to ASP303 and C5 of PQQ, respectively, in the first step, resulting in the formation of formaldehyde. The proton (H16) from ASP303 is transferred

10.1021/jp9083025  2010 American Chemical Society Published on Web 01/07/2010

1888

J. Phys. Chem. A, Vol. 114, No. 4, 2010

Idupulapati and Mainardi

Figure 2. A-E (a) and H-T (b) methanol oxidation mechanisms proposed by Ca2+-containing methanol dehydrogenase enzymes.1,9

to the C5 of PQQ in the second step, and the third and fourth steps involve the proton (H17) transfer from the C5 to O4 of PQQ through ASP303.1,9 It is believed that the presence of this catalytic base (ASP303) at the MDH active site initiates the oxidation reactions of methanol. Irrespective of the mechanism under consideration, the rate-determining step for methanol oxidation by Ca2+-MDH is believed to be the breaking of the Cmet-H17 bond leading to the second proton transfer, which occurs in steps 3 (Figure 2a) and 1 (Figure 2b) of the A-E and H-T mechanisms, respectively.1,9 The proton and electron transfer steps are kinetically and mechanistically separate. The electrons originating from the oxidation of methanol would travel from PQQ toward the natural MDH electron acceptor, cytochrome cL, by means of the surrounding protein environment and by water molecules.5 No perfect agreement has been reached about which methanol oxidation mechanism actually operates (A-E or H-T) at the active site of naturally occurring Ca2+-MDH. Several experimental and theoretical studies were conducted to elucidate the methanol oxidation process that operates under given conditions by Ca2+-MDH. Frank et al.10,11 and Itoh et al.12,13 from their experimental and theoretical studies implied that PQQ systems in organic solution can oxidize methanol, ethanol, and 2-propanol to their corresponding aldehydes via the formation of hemiketal adducts bound to C5 of PQQ, thus suggesting that a covalent PQQ-substrate complex would be possible in favor of the A-E mechanism. On the other hand, Oubrie et al.14,15 from the crystal structure of soluble glucose dehydrogenase at 1.9 Å resolution, found that the carbon atom in glucose was in a good position for hydride transfer to C5 of PQQ, in favor of the H-T mechanism. Quantum mechanics/molecular mechanics and molecular dynamics calculations by Reddy et al.16,17 and Zhang et al.18 also supported the H-T mechanism rather than the A-E mechanism by looking at the conformational variations of the active site residues of the MDH enzyme. Electron paramagnetic resonance studies on substrate binding in ethanol dehydrogenase

by Kay et al.19 indicated a strong coordination of the substrate to the calcium cation, which is energetically very expensive to break (∼50 kcal/mol), thus eliminating the possibility of an addition-elimination process. Leopoldini et al.,20 from their theoretical calculations at the B3LYP theory level, reported energy barriers (or energy heights) for the rate-determining steps for the two proposed A-E and H-T methanol oxidation mechanisms for Ca2+-MDH as 34.6 and 32.3 kcal/mol, respectively. In their calculations, however, only the reactive groups (CH3COO- and NH3COCH3) of the amino acids having coordination with the calcium ion and the PQQ molecule were considered, and the coordinates of hydrogen atoms of each reactive group were fixed during their simulations. No water molecules near the active site were considered by these authors, and also formaldehyde (the final product) was excluded from the calculations performed. The energy barriers that Leopoldini et al.20 found for the methanol oxidation mechanisms by Ca2+-MDH were well above the general kinetic requirements of an enzymatic catalytic process (typical energy barriers should be less than or around 18 kcal/mol). Hence, these authors postulated an alternative mechanism: the “addition-eliminationprotonation” for the methanol oxidation by MDH, which is a modification of the A-E already proposed.1,9 The A-E and H-T mechanisms for methanol oxidation by Ca2+-MDH were explored as part of our previous work21 at the density functional theory level using a model representing the active site of MDH. The model employed consisted of PQQ, Ca2+, three amino acids (ASP303, ASN261, GLU177), and three water molecules (W362, W615, and W213). The model was selected in such a way that the first-shell coordination sphere for the ion was complete and the water molecules near the active site were also taken into account. No constraints were imposed during geometry optimizations, which allowed us to truly characterize the minimum energy and transition state structures following the unconstrained vibrational analysis, as opposed to Leopoldini et al.20 Comparison of the computed positions of

Effect of Substitution of Calcium by Barium the ligands relative to calcium with those in the published X-ray crystallographic structure5 showed that the reference crystal arrangement remained virtually intact. When the protein environment (with a dielectric constant of 4) was applied to the Ca2+-MDH active site model being tested for methanol oxidation, two rate-determining steps for the A-E (steps 1 and 3) with free energy barriers of 18.9 and 20.8 kcal/mol, respectively, and one rate-limiting step for the H-T mechanism (step 1) with a barrier of 19.7 kcal/mol were predicted at the BLYP/DNP density functional theory level by Idupulapati et al.21 Replacing the natural metal ions in the active site of enzymes by other bivalent ions for the reconstitution of apo-enzymes to holo-enzymes (functional enzymes with a cofactor) has also been extensively studied.2,22,23 In the case of MDH, Harris et al.23 replaced Ca2+ by Sr2+ in the active site and observed a 2-fold increase in specific activity toward methanol oxidation due to the improved kinetic properties of the Sr2+-modified methanol dehydrogenase. Goodwin et al.22 have also replaced the Ca2+ ion in the MDH active site with Ba2+ and Sr2+ in order to observe the kinetics of the modified MDH enzyme on methanol oxidation. These authors reported that the activation energy (Ea) for methanol oxidation was less in the case of Ba2+-MDH (Ea ) 3.5 kcal/mol) than that for Sr2+-MDH (Ea ) 7.6 kcal/mol) and Ca2+-MDH (Ea ) 8.5 kcal/mol).22 This result was not expected since the replacement of Ca2+ with Ba2+, a weaker Lewis acid, should decrease the activity of the enzyme and, therefore, its activation energy should be larger. Although the decrease in the activation energy was attributed to a decrease in the affinity for methanol by Ba2+-MDH, the reason why this is happening was not clearly explained.22 Even though cytochrome cL is the physiological electron acceptor for MDH, this enzyme is altered in some way during its isolation, such that ammonia is then required as an activator for methanol oxidation in vitro. Kinetic isotope effects (KIE, the ratio of reaction rate constants with hydrogen and deuterium) are then expected when ammonia is present. Goodwin et al.22 have found that when the hydrogen in the C-H bond of the methanol substrate is replaced by deuterium in the presence of ammonia the activation energy to break that bond is higher (which is the rate-limiting step for both A-E and H-T). These authors extrapolated the deuterium isotope effect at zero ammonia concentration and anticipated the KIE values of 6.1 for Ca2+-MDH and 7.9 for Ba2+-MDH. They also found that the deuterium isotope effect was higher at lower ammonia concentrations; as for example at 500 mM NH4Cl, the isotope effect decreased to about 1.0 for the Ca2+ containing enzymes and to 5.7 for the Ba2+ enzyme. The steps in the mechanism that are affected by the change from protiated substrate to deuterated substrate are those involving the removal of the methyl hydrogen as a hydride (Figure 2). As reported by Goodwin et al.,22 if Ba2+-MDH requires less than half the activation energy for the oxidation of methanol than Ca2+-MDH then Ba2+-MDH enzymes can function as better catalysts; however, to the best of our knowledge, no methanol oxidation mechanism by Ba2+-MDH was suggested in the literature. Hence, a detailed theoretical investigation is carried out in this work on the methanol oxidation mechanisms (A-E and H-T) proposed in the literature using a Ba2+-containing MDH active site model. Comparison of the energetic profiles for the reaction paths with both ions allowed us to assess the possible effects of the metal center substitution in the active site on the structure and energetics.

J. Phys. Chem. A, Vol. 114, No. 4, 2010 1889 Computational Details The generalized gradient approximation (GGA)24,25 within the density functional theory formalism,24,25 as implemented in the DMOL3 module of the Materials Studio software by Accelrys Inc., was used in this work.26 All geometry optimization calculations were performed using the Becke-Lee-Yang-Parr BLYP27 exchange correlation functional and the double numerical with polarization (DNP) basis set, since this was the best set available in DMOL3.26 This basis set considers a polarization d function on heavy atoms and a polarization p function on hydrogen atoms. It compares to the split-valence double-ζ 6-31G** in size; however, it is more accurate than the Gaussian basis sets of the same size.28 The use of the BLYP/DNP theory level has been proven to be adequate for the study of Ca2+and Ba2+-MDH.29 When the BLYP/DNP theory level is used, errors are expected to be to the second decimal place for calculated bond lengths (angstroms) and in the order of 2-5 (kcal/mol) in energies, similar to those expected when using Becke-3 hybrid functionals and the 6-31G** basis set.24,25,30 Harmonic vibrational frequency calculations were performed to ensure that stationary points on the potential energy surface of the systems are, in fact, local minima (all real frequencies) or transition states (only one imaginary frequency). The transition states obtained with DMOL3 may not actually be the ones connecting the intended reactant and product for a particular reaction step, and therefore, intrinsic reaction coordinate (IRC) analysis was performed to thoroughly investigate the reaction paths. When studying enzyme reaction mechanisms, a valid procedure is to use relatively small models for the enzyme active sites and apply quantum chemical methods to study their reaction mechanisms.20,31-37 According to this approach, a homogeneous polarizable medium, with an assumed dielectric constant, is used for treating the neglected enzyme environment and then density functional theory calculations are performed to obtain the reaction energy barriers. The calculated barriers from various studies are often sufficient to confirm or rule out a suggested reaction mechanism and a well-chosen quantum model, representing the enzyme active site, is able to reproduce the chemistry taking place to such a high degree, that it can provide detailed insight and observations about the mechanisms.20,31-37 To estimate the solvation effects of the rest of the enzyme that is not included in the model cluster, a homogeneous polarizable medium was considered with a continuum solvation model known as COSMO38-40 (conductorlike screening model), which is implemented in the DMOL3 module. This model works in such a way that the solute molecule forms a cavity within the dielectric continuum of permittivity that represents the solvent.28 The dielectric constant  was chosen to be 4, which is the standard value used to model protein surroundings.20,31-36 Geometry optimizations were performed in the presence of this solvation model, so that the final energy includes the DMOL3/ COSMO electrostatic energy.28 There are different types of energies reported from our quantum mechanical calculations in this paper. To clearly illustrate their meaning, a minimum energy path (reaction path) for a typical A + B f AB reaction is shown schematically in Figure 3. The solid line in Figure 3 represents the BornOppenheimer energy (i.e., electronic energy + nuclear repulsion) of the system at various configurations along the minimum energy path. The dashed line is obtained by adding the local vibrational zero point energy (ZPE) at each point to the Born-Oppenheimer energy. The physical significance of the quantities shown in Figure 3 are:

1890

J. Phys. Chem. A, Vol. 114, No. 4, 2010

Figure 3. Schematic representation of a minimum energy path (reaction path) for a typical A + B f AB reaction used to illustrate the type of energies reported in this work from quantum mechanical calculations. The solid line represents the Born-Oppenheimer energy (i.e., electronic energy + nuclear repulsion) of the system at various configurations along the minimum energy path. The dashed line is obtained by adding the local vibrational ZPE at each point to the Born-Oppenheimer energy. ∆rV and ∆U0 are the differences between the Born-Oppenheimer energy at 0 K of the transition state and reactants without the ZPE and with the ZPE added, respectively. ∆rU00 is the difference between the Born-Oppenheimer energy at 0K (with ZPE) of the reactant and product.

∆rV ) the classical reaction barrier (energy barrier or reaction height) for the forward reaction. ∆U0 ) the actual (quantum mechanical) reaction barrier for the forward reaction. Since the ZPE at the transition state is generally not the same as that of the reactants, in general ∆rV * ∆U0. This quantity is also called the internal energy of actiVation at absolute zero. ∆rU00 ) the internal energy of reaction at absolute zero. This is also numerically equal to the enthalpy of reaction at absolute zero. The reaction barrier height is a property derived from the Born-Oppenheimer electronic energies and the vibrational zero point energies, which are not affected by temperature, whereas the so-called Arrhenius activation energy, Ea, depends on the temperature.41 Hence, the reaction barrier height ∆U0 is not the same as the Arrhenius activation energy Ea, and their connection may not be straightforward.42,43 In this work, we provide information on the calculated ∆rV, ∆U0, and ∆G0 (that is, the Gibbs Free energy barrier for the forward reaction, i.e., ∆U0 with the thermal corrections to the Gibbs Free energy at 298.15K included) for each step of both the A-E and H-T methanol oxidation mechanisms by MDH, and ∆rU00 for the oxidation mechanisms are also reported. Results and Discussion On the basis of our previous studies,44 a Ba2+-containing MDH active site model was selected to investigate the two proposed A-E and H-T methanol oxidation mechanisms (Figure 2), using quantum chemical calculations. The Ba2+-MDH active site model consisted of the PQQ molecule, Ba2+ ion, aspartic acid (ASP303), glutamic acid (GLU177), asparagine (ASN261), and three water molecules (Figure 4). We have considered the “charge-neutral models” as implemented and tested by Himo et al.45,46 for the amino acids used in our model. According to this approach, charged (unprotonated) amino acids can be modeled by their corresponding neutral (protonated) amino acids. This approach has been tested

Idupulapati and Mainardi

Figure 4. Model representing the Ba2+-containing MDH active site: PQQ molecule, the Ba2+ ion, aspartic acid (ASP303), glutamic acid (GLU177), asparagine (ASN261), and three water molecules (W362, W615, W213).

by Himo et al. and well justified by the fact that proteins and enzymes have low dielectric constants ( ≈ 4), indicating that charge separation, if present, should be very small.45 These authors have shown that, by using these charge-neutral models, accurate enzymatic reaction energies and barriers were yielded.45,46 Therefore, following the charge-neutral models, we have modeled GLU177 and ASN261 as fully protonated amino acids. It is also well-known that during the methanol oxidation mechanism by MDH, neither GLU nor ASN are involved in the redox reactions,9,22 even though some questions were still raised regarding the role of GLU during this oxidation process.16,17 To shed light into the role of GLU during the methanol oxidation by MDH, quantum mechanical calculations at the B3PW91/6311+G** theory level were performed by Idupulapati et al.44 on models representing the active sites of Ca2+- and Ba2+containing MDH in the presence of methanol. Results from those studies indicated the need for the presence of ASP in the active site model for aiding as a base catalyst for methanol oxidation and highlighted the importance of additional neighboring amino acids, such as GLU, in the active site model to be considered for better ion coordination, results that provided extra confidence for fully protonating GLU in this study.44 In the case of aspartate, however, aspartic acid was used but the carboxylic group, the one facing the Ca2+ ion in Figure 4, was kept as such, COO-, since this group is indeed involved in the proposed proton transfer reactions presented in Figure 2. The active site model was then tested upon methanol oxidation following the steps proposed for the A-E and H-T mechanisms (parts a and b of Figure 2 with Ba2+ instead of Ca2+). As discussed earlier in the Computational Details section, to estimate the energetic effects of the protein environment not included in the quantum chemical model chosen, solvation effects were employed on the model with a dielectric constant of 4. Energies of the optimized structures of the reactant, corresponding transition states, intermediates, and products for all of the steps of the mechanisms were obtained, and the energy barriers and corresponding bond distance variations associated with each step are presented and discussed in this section. The initial reactants in both the A-E and H-T methanol oxidation mechanisms by MDH (Figure 2) were the same

Effect of Substitution of Calcium by Barium

J. Phys. Chem. A, Vol. 114, No. 4, 2010 1891

Figure 5. Geometrically optimized structures involved in step 1 for the A-E methanol oxidation mechanism by Ba2+-MDH active site model.

complex. Hence, a methanol molecule was added to the model, and the complete complex was geometrically optimized at the BLYP/DNP theory level. The ground-state geometry of the Ba2+-containing complex agrees well with the conformation of the Ca2+ complex.21 Ba2+ coordinates with the nitrogen (N6) and oxygen (O5 and O10) atoms of PQQ (at 3.16, 2.83, and 3.05 Å, respectively), the carbonyl oxygen atoms of ASN261 (O11 at 2.79 Å), GLU177 (O12 at 2.80 Å), and both oxygen atoms of the water molecules W362 and W615 (at 2.80 Å and 2.85 Å, respectively). Ba2+ also coordinates with O16 of methanol (2.89 Å), which was not observed when Ca2+ is in place of Ba2+ (Figure 5).44 Also, W615 and W213 maintain hydrogen bonds to oxygen atoms of PQQ and GLU177, respectively. There was an increase in bond lengths with respect to the Ba2+ coordination (∼0.3-0.5 Å more than the Ca2+ coordination), which can be mainly due to the ionic size of the element, i.e., the Ba ionic radius (1.34 Å) is larger than that of the Ca ion (0.99 Å), and it is also due to the distortion introduced by the replacement of ions.44 Coordination and binding of these two ions in representative models of MDH active sites and their interaction with methanol were previously studied by us, where the same kind of increase in bond lengths, with respect to Ba2+ and also Ba2+-O16 coordination, was observed.44 For both the Ca2+ and Ba2+ ions in the MDH active site, similar sets of minima and transition state structures were found for the A-E and H-T methanol oxidation mechanisms. However, certain changes in the equilibrium geometry configurations as well as in the energy barriers were noticed for different ions, but qualitatively the structures were the same. A discussion with respect to bond lengths (Tables 1-3) and calculated energies (Table 4) for Ca2+-MDH (added for comparison from ref 21) and Ba2+-MDH active site models is presented in this section for both methanol oxidation mechanisms. From our calculations, it was found that the difference between the Born-Oppenheimer energy at 0 K (with ZPE) of the reactant and final product for

TABLE 1: Selected Ground-State Bond Lengths Corresponding to Step 1 of the Methanol A-E Mechanism by Ca2+-MDH21 and Ba2+-MDH (in Parentheses) Active Site Models distance (Å) O14-H16 C5-O16 ion-O5 W362-O5 W362-O16 ion-O16

reactant

TS1

INT1

1.74 (1.64) 4.10 (4.07) 2.41 (2.83) 3.18 (2.50) 2.91 (2.68) 3.34 (2.89)

1.01 (1.01) 2.78 (2.42) 2.39 (2.71) 2.83 (2.30) 2.62 (1.98) 3.26 (2.64)

0.98 (0.98) 1.54 (1.52) 2.38 (2.42) 2.51 (1.98) 2.54 (1.92) 3.38 (2.49)

TABLE 2: Selected Ground-State Bond Lengths Corresponding to Step 3 of the Methanol A-E Mechanism by Ca2+-MDH21 and Ba2+-MDH (in Parentheses) Active Site Models distance (Å)

INT2

TS3

product

Cmet-H17

1.10 (1.24) 1.89 (1.98) 4.56 (2.98) 2.79 (2.68)

1.47 (1.68) 3.47 (3.92) 3.98 (2.26) 3.42 (2.53)

2.82 (2.02) 4.33 (4.02) 3.86 (2.52) 3.68 (2.98)

C5-O16 W362-O16 ion-O16

the overall methanol oxidation process (∆rU00) is 7.43 and 2.56 kcal/mol for the Ca2+- and Ba2+-MDH active site models, respectively. 1. A-E Mechanism. 1.1. Step 1: Proton Abstraction by ASP303 from Methanol and Nucleophilic “Addition” of Hemiketal Complex to PQQ. As seen from the optimized structure of the reactant (react) species (Figure 5), the active

1892

J. Phys. Chem. A, Vol. 114, No. 4, 2010

Idupulapati and Mainardi

TABLE 3: Selected Ground-State Bond Lengths Corresponding to Step 1 of the Methanol H-T Mechanism by Ca2+-MDH21 and Ba2+-MDH (in Parentheses) Active Site Models parameter (Å) O14-H16 C5-H17 ion-O5 W362-O5 W362-O16

reactant

TS1

INT1

2.00 (1.64) 3.69 (3.57) 2.71 (2.83) 3.18 (2.50) 2.91 (2.68)

1.12 (1.05) 1.22 (1.42) 2.60 (2.42) 2.58 (1.97) 3.02 (2.54)

1.00 (1.00) 1.02 (1.12) 2.54 (2.34) 2.51 (1.88) 3.14 (2.32)

TABLE 4: Calculated BLYP/DNP Energies Associated with the Forward Methanol Oxidation reactions by Ca2+-MDH and Ba2+-MDH (in Parentheses) for Each Step of the A-E and H-T Mechanisms Shown in Figure 2a mechanism/step ∆rV (kcal/mol) ∆U0 (kcal/mol) ∆G0 (kcal/mol) A-E/1 A-E/2 A-E/3 H-T/1 H-T/2 H-T/3 H-T/4

14.54 (9.68) 5.21 (3.36) 15.54 (11.18) 14.39 (8.83) 3.14 (4.08) 6.59 (4.45) 8.30 (4.35)

15.08 (10.93) 6.16 (5.25) 15.70 (13.06) 16.30 (10.09) 4.48 (4.65) 9.05 (5.08) 9.67 (5.60)

18.93 (10.81) 3.76 (2.74) 20.81 (14.38) 19.72 (10.40) 2.89 (3.45) 9.04 (5.46) 6.79 (3.72)

a ∆rV, ∆U0, and ∆G0, as defined in Figure 3, are the differences between the Born-Oppenheimer energy (i.e., electronic energy + nuclear repulsion) at 0 K of the transition states and reactants or intermediates without the ZPE, with the ZPE added, and with ZPE and the thermal corrections to free energies at 298.15 K, respectively.

site was nicely set up for a proton (H16) transfer from methanol to ASP303 and the nucleophilic addition of the CH3O- complex to the C5 of PQQ. The optimum transition state (TS1) for this step was characterized by a single imaginary frequency. The O14-H16 bond length was reduced from 1.64 Å (react) to 1.01 Å (TS1) and to 0.98 Å (INT1), and the C5-O16 bond distance changed from 4.07 Å (react) to 2.42 Å (TS1) and to 1.52 Å (INT1), indicating the formation of Omet-C5 and the shift of a proton from the alcohol OH group of methanol to ASP303. The methoxide (CH3O-) addition to the C5 of PQQ resulted in the formation of the first intermediate (INT1), where the Omet-C5 bond was completely formed (1.52 Å). The changes in the O14-H16 and C5-O16 bond lengths when going from reactant to TS1 to INT1 did not show significant variation in the presence of both ions (Table 1). The free energy barrier (∆G0) for this step 1 was calculated to be 10.8 kcal/mol (18.9 kcal/mol for Ca2+-MDH21) for methanol oxidation by Ba2+-MDH in the presence of the protein environment (Table 4). This decrease in the free energy barrier can be attributed to an increase in charge stabilization by the Ba2+ and W362 as the reaction proceeds to INT1. Once the proton was abstracted by ASP303, the ion coordination to O16 of the methoxide reduced to 2.64 Å at TS1 (2.89 Å at reactant), and there was a strong hydrogen bonding with W362 at this point, thereby stabilizing the developing negative charge. This

ion coordination and hydrogen bonding were not that significant in the case of Ca2+-MDH (Table 1). During this methoxide addition to C5 of PQQ, the C5-O5 bond was polarized through the interaction with Ba2+, and the resulting oxyanion of the carbonyl group (O5) bonded more to the Ba2+, with a bond length of 2.42 Å at INT1 (2.83 Å and 2.71 Å at reactant and TS1, respectively, Table 1). The movement of Ba2+ toward O5 of PQQ during this step was also evident with the increase in the coordination distance of the ion with W615 and the other oxygen (O10) of PQQ (Figure 5). The hydrogen bonding of water molecule (W362) with respect to O5 of PQQ decreased as the reaction proceeded to INT1 (1.98 Å compared to 2.50 and 2.30 Å at reactant and TS1, respectively, Table 1). Hence, the resulting negative charge, as the reaction proceeded from the reactant to INT1, was more electrostabilized by the interaction with Ba2+ and W362, as previously discussed, thereby lowering the energy of the TS1 and INT1 and, therefore, the free energy barrier for this step when compared to the Ca2+MDH case. In this reaction, Ba2+-MDH was less endothermic (by 7.9 kcal/mol) than Ca2+-MDH (14.7 kcal/mol), whereby the free energy of the first intermediate was lowered by 6.8 kcal/mol when Ba2+ was present, compared to the Ca2+case, relative to the reactant energy. 1.2. Step 2: Proton “Elimination” from ASP303 and Transfer to PQQ. Having established that the first proton transfer and nucleophilic addition to PQQ is energetically feasible, the next step (2) was a proton (H16) transfer from ASP303 to the O5 of PQQ. At the transition state for this step (TS2), the proton was shared between the oxygen atoms of PQQ (at 1.23 Å) and the ASP303 (at 1.50 Å), respectively (TS2, Figure 6). The CH3O- complex tilted its orientation with respect to INT1 in such a way that the bond length between O4 of PQQ and H17 of methanol was 2.42 Å compared to 2.68 Å in INT1. The hydrogen bonding of the W362 to the O5 of PQQ was completely lost as the reaction proceeded to INT2, but the bonding with respect to ASP303 remained intact; thereby stabilizing the negative charge on the residue. There is not much change in the free energy barrier for this step in the presence of the protein environment with respect to both ion clusters (∼1.2 kcal/mol). However, the reaction was more exothermic (by 5.7 kcal/mol) for Ca2+-MDH than Ba2+MDH as there was not much change in the energy of the first and second intermediates (Figure 7). 1.3. Step 3: Formation of PQQH2 and Formaldehyde. The proton transfer in step 2 led to the formation of a second intermediate (INT2) along the reaction profile. For this INT2, the CH3 group of the attached methanol is oriented toward the O4 of PQQ, so that the latter was in a good position for breaking the H-CH2 bond (C5-O16 bond increased up to 1.98 Å and 1.52 Å at INT1, Figure 6). The final transition state TS3, which leads to the formation of PQQH2 reduced species and formaldehyde, was located with Cmet-H17 and O4-H17 distances at 1.68 and 1.04 Å, respectively. At TS3, the bond between O16 and C5 of PQQ lengthens up to 3.92 Å, indicating the breakage to form formaldehyde (Figure 6). Water molecules W362 and W615 maintained their coordination with Ba2+, and ASP303 and Ba2+, respectively, throughout this proposed reaction path until the formation of the final product. The free-energy barrier for this step 3 was 14.3 kcal/mol for Ba2+-MDH compared to 20.8 kcal/mol for the Ca2+-MDH case (Figure 7). The reason for the greater charge stabilization of TS3 was due to the coordination of O16 of the resulting formaldehyde with Ba2+ (the distance was less at TS3), and also strong hydrogen bonding with W362, which was not

Effect of Substitution of Calcium by Barium

J. Phys. Chem. A, Vol. 114, No. 4, 2010 1893

Figure 6. Geometrically optimized structures involved in steps 2 and 3 for the A-E methanol oxidation mechanism by Ba2+-MDH active site model.

Figure 7. Potential energy surface (PES) corresponding to the methanol A-E oxidation mechanism by MDH active site models with the two ions. Reactant-relative free energies calculated at the BLYP/DNP theory level are in kcal/mol.

observed when Ca2+ was present (Table 2).21 This reaction was also exothermic in both cases with a difference of only 0.2 kcal/ mol (Figure 7). Also, the cleavage of Cmet-H17, the formation of the O4-H17 bond and the detachment of HCHO to form formaldehyde, was almost completed at the transition state in the presence of Ba2+ (Table 2). In this case, the product formed was almost the same as the transition state obtained, with the particular difference of formaldehyde moving away from the ion (Figure 6). 2. Hydride Transfer Mechanism. 2.1. Step 1: Hydride Transfer to PQQ, Proton Abstraction by ASP303, and Formaldehyde Formation. According to step 1 of the H-T mechanism (Figure 2b), from the initial reactant complex, there should be a direct hydride transfer (H17) from methanol to C5 of PQQ in concert with a proton abstraction (H16) by O14 of ASP303, thus resulting in the formation of formaldehyde. Hence, the first transition state evolved into the first intermediate (INT1), where the formaldehyde produced leaves the active site. From our

calculations at the BLYP/DNP level on Ba2+-MDH, it was found that the O14-H16 bond length evolved from 1.64 Å (reactant) to 1.05 Å (TS1) to 1.00 Å (INT1), showing the binding of the H16 proton to ASP303 (Figure 8). The C5-H17 bond length from the initial reactant was reduced from 3.57 to 1.42 Å and finally to 1.12 Å (Table 3), evidencing the hydride transfer to the C5 of PQQ. The C5 of PQQ changes its hybridization from sp2 to sp3 and establishes a covalent bond with H17 at 1.12 Å. This transition state seems to form earlier in the case of Ba2+MDH than in the presence of Ca2+-MDH by looking at the bond length variations of O14-H16 and C5-H17 from reactant to INT1 (Table 3). The free-energy barrier for TS1 with respect to the reactant was calculated to be 10.4 kcal/mol (19.7 kcal/mol for Ca2+MDH21) for Ba2+-MDH in the presence of the protein environment. Ba2+ polarized the C5-O5 bond, making C5 more positive and O5 more negative at this transition state. The resulting negative charge on the oxyanion O5 of PQQ as the reaction proceeded to INT1 was more electrostabilized by its interaction with Ba2+ and strong hydrogen bonding with W362 (1.88 compared to 2.50 Å at reactant), thereby lowering the energy barrier for this step compared to the case when Ca2+ is present in the MDH active site model (Table 3). Also, the hydrogen bond length of W362 with formaldehyde remained unchanged as the reaction proceeded from the reactant to TS1 to INT1, unlike what happened when Ca2+ was present in the active site, where the bond length increased (Table 3). Like step 1 for A-E, this reaction was also less endothermic (by 9.3 kcal/ mol) for the Ba2+-MDH case than for the Ca2+-MDH case. 2.2. Step 2: Proton Transfer from ASP303 to PQQ. The next step along the reaction profile consisted of a proton transfer from ASP303 to the O5 of PQQ, thus resulting in the formation of a second intermediate (INT2). From our calculations, it was found that the transition state for this step (TS2) showed an intermediate location of H16 on its way from the O14 of ASP303 to the PQQ carbonyl oxygen O5, with H16-O5 )

1894

J. Phys. Chem. A, Vol. 114, No. 4, 2010

Idupulapati and Mainardi

Figure 8. Geometrically optimized structures involved in step 1 of the H-T methanol oxidation mechanism by Ba2+-MDH active site model.

Figure 9. Geometrically optimized structures involved in steps 2, 3, and 4 of the H-T methanol oxidation mechanism by Ba2+-MDH active site model.

1.30 Å and H16-O14 ) 1.26 Å. The hydrogen bonding of W362 with respect to the O5 of PQQ increased as this step proceeded (Figure 9). Since the geometrical characteristics of TS2 were similar to those of INT1, a low free energy barrier was anticipated for this step. There was not much change in the free energy barrier (∼0.6 kcal/mol) for this step in the presence of the protein environment with respect to both ion-containing MDH active site models.

2.3. Steps 3 and 4: Proton Transfer from PQQ to ASP (Step 3) and Back to PQQ (Step 4). The enolization process of transferring the hydrogen (H17) from C5 of PQQ to C4 carbonyl oxygen (O4) involved the proposed steps 3 and 4 of the methanol H-T mechanism by Ba2+-MDH. The third step involved the transfer of a proton (H17) from C5 to O14 of ASP303, emphasizing the role of this amino acid as the base catalyst (Figure 2b). For this step 3, the transition state (TS3) showed an intermediate location of H17 on its way from PQQ

Effect of Substitution of Calcium by Barium

Figure 10. PES corresponding to the methanol H-T oxidation mechanism by MDH active site models with the two ions. Reactant relative free energies calculated at the BLYP/DNP theory level are in kcal/mol.

carbon C5 to the O14 of ASP, with H17-C5 ) 2.83 Å and H17-O14 ) 1.65 Å (Figure 9). The hydrogen bonding of W362 with the O5 of PQQ was completely lost by the end of this step. Through TS4, the final proton was transferred from the O14 of ASP303 to the negatively charged O4 of PQQ. During this final step, the O14-H17 bond length increased from 1.02 (INT3) to 1.40 (TS4) to 1.72 Å (product), and the O4-H17 bond length decreased from 1.67 (INT3) to 1.50 (TS4) to 1.01 Å (product), respectively, indicating that PQQ was finally reduced (Figure 9). Throughout this reaction step, W362 and W615 maintained coordination with Ba2+ and W615, and W213 maintained hydrogen bonding with PQQ and GLU177, respectively. For the Ba2+-MDH case, the free energy barriers are 5.5 and 3.7 kcal/mol for steps 3 and 4 of this mechanism, respectively, which are decreased by 3.6 and 3.1 kcal/mol with respect to the barriers obtained for Ca2+-MDH (Figure 10).

J. Phys. Chem. A, Vol. 114, No. 4, 2010 1895 For the H-T mechanism, step 2 was the fastest step, and step 1 was the kinetically slowest step, which involves a hydride transfer to PQQ and a proton abstraction by ASP303 resulting in the formation of formaldehyde, irrespective of the ion at the MDH active site. The free energy barrier for step 1 was reduced by 9.3 kcal/mol for Ba2+-MDH with respect to that for the Ca2+MDH case. There was a change in the barriers for other steps in the H-T mechanism compared to the Ca2+-case but not as distinctive as step 1. This huge reduction in barriers for the ratedetermining steps for both mechanisms was mainly due to a very good stabilization of transition states corresponding to these steps, with Ba2+ in place instead of Ca2+, in the MDH active site. By comparison of our results from quantum chemical calculations at the BLYP/DNP theory level of this paper to our extensive prior work on Ca2+-MDH,21 it was observed that the free energy barriers of the rate-limiting steps of both A-E and H-T that were explored with the Ba2+-MDH active site model were reduced to almost half of those found for the Ca2+-MDH active site model. These results support the experimental observations by Goodwin et al.22 that Ba2+-MDH oxidizes methanol with less than half the activation energy required for Ca2+-MDH. Acknowledgment. The financial support from the National Science Foundation (NSF) under the NSF CAREER Grant, CTS-0449046, is gratefully acknowledged. Support for computational resources including software and hardware through the Louisiana Board of Regents, Contract No. LEQSF(2007-08)ENH-TR-46, the National Science Foundation Grant No. NSF/ IMR DMR-0414903, and the Louisiana Optical Network Initiative (LONI) are also thankfully acknowledged. Supporting Information Available: Figure depicting the formaldehyde formation as a result of methanol oxidation by Ca2+- and Ba2+-containing MDH enzymes. This material is available free of charge via the Internet at http://pubs.acs.org.

Conclusions Methanol oxidation by a PQQ-containing MDH enzyme was investigated using a model cluster containing Ba2+ instead of Ca2+ using the generalized gradient approximation within the density functional theory formalism at the BLYP/DNP theory level. The two proposed methanol oxidation mechanisms by Ca2+-MDH found in the literature, the A-E and the H-T mechanisms, were investigated for a Ba2+-MDH active site model. Qualitatively, the same kind of intermediates and transition states associated with each step of the proposed mechanisms (when compared to the Ca2+-cluster) were found for the Ba2+-MDH case with changes in geometry conformations and energies, particularly for the rate-determining steps. For both A-E and H-T, almost all energy barriers associated with all the steps were reduced in the presence of Ba2+ in the MDH active site model, and they were kinetically possible when referred to the general kinetic requirements of an enzymatic catalytic process (energy barriers expected to be less than or around 18 kcal/mol20,31-36). For the A-E mechanism, step 2 was the fastest step, irrespective of the ion (Ca2+ or Ba2+) under consideration. The free energy barriers for steps 1 (formation of hemiketal intermediate) and 3 (formaldehyde formation) were 18.9 and 19.7 kcal/mol, respectively, for the Ca2+-MDH case, making both of them ratelimiting; whereas, for the Ba2+-MDH case, the barriers for both steps were reduced by ∼9 kcal/mol, making these reactions more kinetically feasible.

References and Notes (1) Anthony, C. Methanol Dehydrogenase, a PQQ-Containing Quinoprotein Dehydrogenase; Kluwer Academic/ Plenum Publishers: New York., 2000. (2) Afolabi, P. R.; Amaratunga, K.; Majekodunmi, O.; Dales, S. L.; Gill, R.; Thompson, D.; Cooper, J. B.; Wood, S. P.; Goodwin, P. M.; Anthony, C. Biochem. J. 2001, 40, 9799. (3) Ghosh, M.; Anthony, C.; Harlas, K.; Goodwin, M. G.; Blake, C. C. F. J. Struct. 1995, 3, 1771. (4) White, S.; Boyd, G.; Mathews, F. S.; Xia, Z. X.; Dai, W. W.; Zhang, Y. S.; Davidson, V. L. J. Biochem. 1993, 32, 12955. (5) Williams, P. A.; Coates, L.; Mohammed, F.; Gill, R.; Erskine, P. T.; Coker, A.; Wood, S. P.; Anthony, C.; Cooper, J. B. Acta. Crystallogr., Sect. D 2005, D61, 75. (6) Xia, Z. X.; Dai, W. W.; Xiong, J. P.; Hao, Z. P.; Davidson, V. L.; White, S.; Mathews, F. S. J. Biol. Chem. 1992, 267, 22289. (7) Xia, Z. X.; Dai, W. W.; Xiong, J. P.; Hao, Z. P.; Davidson, V. L.; White, S.; Mathews, F. S. J. Mol. Biol. 1996, 259, 480. (8) Xia, Z. X.; Dai, W. W.; Xiong, J. P.; Hao, Z. P.; Davidson, V. L.; White, S.; Mathews, F. S. Biochem. J. 1999, 38, 1214. (9) Anthony, C.; Williams, P. Biochim. Biophys. Acta 2003, 1647, 18. (10) Frank, J.; Dijkstra, M.; Duine, J. A.; Balny, C. Eur. J. Biochem. 1988, 174, 331. (11) Frank, J.; van Krimpen, S. H.; Verwiel, P. E. J.; Jongejan, J. A.; Mulder, A. C. Eur. J. Biochem. 1989, 184, 187. (12) Itoh, S.; Kawakami, H.; Fukuzumi, S. J. Biochem. 1998, 37. (13) Itoh, S.; Kawakami, H.; Fukuzumi, S. J. Mol. Catal., B 2000, 8, 85. (14) Oubrie, A.; Rozeboom, H. J.; Kalk, K. H.; Huizinga, E. J.; Dijkstra, B. W. J. Biol. Chem. 2002, 277, 3727. (15) Oubrie, A.; Rozeboom, H. J.; Kalk, K. H.; Olsthoorn, A. J.; Duine, J. A.; Dijkstra, B. W. EMBO. J 1999, 18, 5187. (16) Reddy, S. Y.; Bruice, T. C. J. Am. Chem. Soc. 2003, 125.

1896

J. Phys. Chem. A, Vol. 114, No. 4, 2010

(17) Reddy, S. Y.; Mathews, F. S.; Zheng, Y. J.; Bruice, T. C. J. Mol. Struct. 2003, 655, 269. (18) Zheng, Y. J.; Bruice, T. C. Proc. Natl. Acad. Sci., U.S.A. 1996, 94. (19) Kay, C. W. M.; Mennenga, B.; Gorisch, H.; Bittl, R. J. Am. Chem. Soc. 2005, 127, 7974. (20) Leopoldini, M.; Russo, N.; Toscano, M. Chem.-Eur. J. 2007, 13, 2109. (21) Idupulapati, N. B.; Mainardi, D. S. Methanol Electro-Oxidation by Methanol Dehydrogenase Enzymatic Catalyst: A Computational Study. In Modern Aspects of Electrochemistry; Balbuena, P. B., Subramanian, V., Eds.; Springer Science: New York, 2009. (22) Goodwin, M. G.; Anthony, C. Biochem. J. 1996, 318, 673. (23) Harris, T. K.; Davidson, V. L. Biochem. J. 1994, 300, 175. (24) Dreizler, R. M.; Gross, E. K. U. Density Functional Theory: An Approach to Quantum Many Body Problem; Springer-Verlag: Berlin, 1990. (25) Koch, W.; Holthausen, M. C. A Chemist’s Guide to Density Functional Theory, 2nd; Wiley-CVH, 2001. (26) Accelrys. Materials Studio, 2006. (27) Lee, C.; Yang, W.; Parr, R. Phys. ReV. B. 1998, 37, 786. (28) Accelrys Inc. DMOL3 User Guide, San Diego, 2003. (29) Idupulapati, N. B.; Mainardi, D. S. Mol. Simul. 2008, 34, 1057. (30) Becke, A. D. J. Chem. Phys. 1988, 88, 2547. (31) Georgieva, P.; Himo, F. Chem. Phys. Lett. 2008, 463, 214.

Idupulapati and Mainardi (32) Hopmann, K. H.; Himo, F. Chem.-Eur. J. 2006, 12, 6898. (33) Hopmann, K. H.; Himo, F. J. Chem. Theory Comput. 2008, 4, 1129. (34) Noodleman, L.; Lovell, T.; Han, W.; Li, J.; Himo, F. Chem. ReV. 2004, 104, 459. (35) Sevastik, R.; Himo, F. Bioorg. Chem. 2007, 35, 444. (36) Velichkova, P.; Himo, F. J. Phys.Chem. B. 2006, 110, 16. (37) Zhen Liao, R.; Yu, J.; Raushel, F. M.; Himo, F. Chem.-Eur. J. 2008, 14, 4287. (38) Cammi, R.; Mennucci, B.; Tomasi, J. J. Phys. Chem. 1999, 103, 9100. (39) Cammi, R.; Mennucci, B.; Tomasi, J. J. Phys.Chem A 2000, 104, 5631. (40) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys. 2002, 43, 117. (41) Hill, C. G. An Introduction to Chemical Engineering Kinetics & Reactor Design; John Wiley & Sons: New York, 1977. (42) Ramachandran, B.; Peterson, K. A. J. Chem. Phys. 2003, 119, 9590. (43) Xie, T.; Bowman, J. M.; Ramachandran, B.; Peterson, K. A. J. Chem. Phys. 2003, 119, 9601. (44) Idupulapati, N. B.; Mainardi, D. S. THEOCHEM 2009, 901, 72. (45) Himo, F. J. Phys. Chem. B. 2002, 106, 7688. (46) Himo, F.; Eriksson, L. A. J. Am. Chem. Soc. 1998, 120, 11449.

JP9083025