Higher Flexibility of Glu-172 Explains the Unusual ... - ACS Publications

Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/IC

Higher Flexibility of Glu-172 Explains the Unusual Stereospecificity of Glyoxalase I Sonia Jafari,†,‡ Nadia Kazemi,† Ulf Ryde,‡ and Mehdi Irani*,† †

Department of Chemistry, University of Kurdistan, P.O. Box 66175-416, Sanandaj, Iran Department of Theoretical Chemistry, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden

‡

S Supporting Information *

ABSTRACT: Despite many studies during the latest two decades, the reason for the unusual stereospecificity of glyoxalase I (GlxI) is still unknown. This metalloenzyme converts both enantiomers of its natural substrate to only one enantiomer of its product. In addition, GlxI catalyzes reactions involving some substrate and product analogues with a stereospecificity similar to that of its natural substrate reaction. For example, the enzyme exchanges the pro-S, but not the proR, hydroxymethyl proton of glutathiohydroxyacetone (HOCSG) with a deuterium from D2O. To find some clues to the unusual stereospecificity of GlxI, we have studied the stereospecific proton exchange of the hydroxymethyl proton of HOC-SG by this enzyme. We employed density functional theory and molecular dynamics (MD) simulations to study the proton exchange mechanism and origin of the stereospecificity. The results show that a rigid cluster model with the same flexibility for the two active-site glutamate residues cannot explain the unusual stereospecificity of GlxI. However, using a cluster model with full flexibility of Glu-172 or a larger model with the entire glutamates, extending the backbone into the neighboring residues, the results showed that there is no way for HOC-SG to exchange its protons if the alcoholic proton is directed toward Glu-99. However, if the hydroxymethyl proton instead is directed toward the more flexible Glu-172, we find a catalytic reaction mechanism for the exchange of the HS proton by a deuterium, in accordance with experimental findings. Thus, our results indicate that the special stereospecificity of GlxI is caused by the more flexible environment of Glu-172 in comparison to that of Glu-99. This higher flexibility of Glu-172 is also confirmed by MD simulations. We propose a reaction mechanism for the stereospecific proton exchange of the hydroxymethyl proton of HOC-SG by GlxI with an overall energy barrier of 15 kcal/mol.

1. INTRODUCTION Glyoxalase I (EC 4.4.1.5, lactoylglutathione lyase; GlxI) is a metalloenzyme present in most organisms from bacteria to human. It is one of the two members of the glyoxalase system. The system consists of two enzymes, GlxI and glyoxalase II (GlxII). These enzymes detoxify methylglyoxal (MG) that is produced in normal cell metabolism. This system is vital for organisms, because MG is a highly toxic metabolite. It is a reactive dicarbonyl compound endogenously produced primarily as a byproduct of glycolysis, but also as a product of ketone body metabolism and of threonine catabolism and in the degradation of glycated proteins.1 The glyoxalase system has been implicated in chemoresistance in some human tumors, and it has therefore been suggested to use GlxI inhibitors as anticancer drugs.2 GlxI catalyzes the conversion of hemithioacetals of MG and a variety of aromatic and aliphatic α-ketoaldehydes to αhydroxythioesters, and after that, GlxII splits the produced lactoylglutathione into glutathione (H-SG) and D-lactate (Scheme 1).3 The hemiacetal substrates of GlxI are formed from a nonenzymatic reaction of a thiol (either H-SG, or © XXXX American Chemical Society

trypanothione, depending on the organism, e.g. H-SG for human GlxI) and MG.4 Interestingly, GlxI converts both enantiomers of the substrate into the same S-D-lactoylglutathione product.5 For catalytic activity, GlxI requires a divalent metal ion, which varies depending on the organism. It is Zn(II) for human GlxI.6−8 However, GlxI of Escherichia coli shows its maximum catalytic activity in the presence of Ni(II), whereas it is inactive in the presence of Zn(II) and has a reduced activity with Co(II), Cd(II), and Mn(II).9,10 Aronsson et al. showed that GlxI from both human and yeast contains a Zn(II) ion in a stoichiometry of about 1 mol/mol of enzyme subunit and that removal of the Zn(II) ion gave an inactive enzyme.11 Crystal structures of human GlxI show that the active-site Zn(II) ion is coordinated by His-126, Gln-33, Glu-99, Glu-172, and two water molecules.12 It is generally accepted that the Glu-172 or Glu-99 residue of the active site initiates the catalytic reaction of GlxI by abstraction of the H1 atom from Received: December 21, 2017

A

DOI: 10.1021/acs.inorgchem.7b03215 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Scheme 1. Glyoxalase Pathway

Scheme 2. Initial Step of the GlxI Reaction with the (a) S and (b) R Substrates

the S or R substrate, respectively (the naming of the atoms is shown in Scheme 2). This proton abstraction generates an enediolate intermediate, a key structure in the reaction pathway12−15 (cf. Scheme 2). However, experiments show that the proton is always added to the si face of C2 and not to its re face, resulting in only one enantiomer of the product.5 This shows that the two active-site glutamate residues play different roles in the reaction, probably owing to their different environments in the enzyme (see ref 16 for a detailed description of the reaction mechanism). GlxI can also catalyze reactions involving some substrate analogues. For example, it converts both enantiomers of the two thiohemiacetals 1 and 3 in Scheme 3 to only the S enantiomer of the thioesters 2 and 4, respectively.17,18 It was also demonstrated by NMR spectroscopy that the enzyme exclusively catalyzes the exchange of the pro-S, but not the proR, hydroxymethyl proton of the product analogue glutathiohydroxyacetone (HOC-SG) with a deuterium from the D2O solvent (Scheme 4). On the other hand, the enzyme also specifically mediates the exchange of the pro-S, but not pro-R, hydroxymethyl deuterium of [2H]-HOC-SG with a proton from H2O solvent.19 The rate constant for solvent exchange is about 1 s−1.19 This value is similar to the solvent exchange rate of (S)-D-lactoylglutathione (the normal product of the enzyme) by GlxI.20 On the basis of experimental results, Creighton and Hamilton concluded that a cis enediolate intermediate is formed during the reaction of normal substrates, provided that

Scheme 3. Conversion of Both Isomers of Thiohemiacetals 1 and 3 to Only the S Enantiomer of the Thioesters 2 and 4, by GlxI

Scheme 4. Stereospecific Deuterium Exchange of HOC-SG by GlxI

the substrate analogues are processed in the same way as for the normal substrates.15 B


Article

Inorganic Chemistry

Figure 1. Optimized structures of M1 (a, b), M2 (c, d), and M3 (e, f) in the HOto99 (a, c, e) and HOto172 (b, d, f) conformations. HOC-SG, the Zn atom, and HOH-404 are shown by a ball-and-stick representation and the amino acids by tubes. Fixed atoms are marked with asterisks. Selected distances are shown in Å.

use the same reaction mechanism for the S and the R

We have in a previous study used the quantum mechanical cluster (QM-cluster) approach to compare several alternative reaction mechanisms for both enantiomers of the natural substrate of human GlxI.16 The calculations gave support to the previously proposed mechanism involving an enediolate intermediate.13−15 However, we found that the enzyme can

enantiomers of the substrate, but with exchanged roles of the two active-site glutamate residues (Glu-99 and Glu-172). From the results, we concluded that the only possibility for the stereospecificity of GlxI is a difference in the electrostatic C


Article

Inorganic Chemistry surroundings and flexibilities of the glutamate residues in the active site, owing to their neighboring residues in the protein. Thus, despite all previous studies during the latest two decades, the reason for the special stereospecificity of GlxI is still unclear.5,12−19 A study of the reactions of substrate analogues may give new clues to solve the old problem of GlxI stereospecificity. Therefore, we focus in this work on the stereospecific proton exchange of the hydroxymethyl proton of HOC-SG (Scheme 4). We investigate the mechanism of the enzymatic exchange of HS by a deuterium atom and why HR cannot be exchanged by a deuterium. We employ the QMcluster methodology, which has been successfully used to elucidate a large number of different enzymatic reaction mechanisms.16,21−25 We also use molecular dynamics (MD) simulations to study the flexibility of the active site.

All QM-cluster calculations were performed using the density functional B3LYP,26 implemented in the Gaussian software, versions 0927 and 16.28 The structures of the reactants, transition states, intermediates, and products were optimized using the 6-31+G(d) basis set for the H, C, N, O, and S atoms and the LANL2DZ29 pseudo potential and basis set for the Zn ion. More accurate energies were calculated with single-point calculations on the optimized structures using the larger 6-311++G(2d,2p) basis set for all atoms. To consider the surroundings, solvation effects were evaluated at the B3LYP/631+G(d)/LANL2DZ level of theory by performing single-point calculations using the CPCM solvation model.30 The solvation cavity was built up using the UFF radii, and the dielectric constant was set to 4. Natural bond orbital (NBO) analysis31,32 was used to calculate atomic charges on the optimized structures. It was performed at the same level of theory as the single-point energy calculations. Frequencies of the stationary states on the potential energy surface were calculated to obtain zero-point energies. The frequency calculations were performed at the same level of theory as for the geometry optimizations. The final energy of each stationary point and all energies on the potential surfaces were obtained by including the corresponding zero-point energy and the electrostatic part of the solvation energy as a correction to the electronic energy calculated from the higher-level single-point calculations. 2.2. MD Simulations. The MD simulations were based on the same crystal structure (1QIN).12 The entire dimeric enzyme and all crystal water molecules were included in the calculations. However, the inhibitor was replaced by HOC-SG. The protonation states of all residues were determined from a detailed study of the hydrogen-bond pattern and the solvent accessibility. They were checked by the PROPKA software.33 The two subunits of the dimeric protein were treated in the same way. All Cys residues were assumed to be protonated. The active site histidine residue (His-126) was protonated on the ND1 atom, His-102 was protonated on NE2, and another histidine residue (His-115) was protonated on both ND1 and NE2. The two aspartic acid residues Asp-165 and Asp-167 were assumed to be protonated, and the others were negatively charged. All Arg, Lys, and Glu residues were assumed to be charged. All protonation states are in agreement with the predictions of PROPKA. The only exception is Glu-99, which PROPKA predicts to be protonated. However, since it is coordinated to the Zn ion, we treated it in its deprotonated form. In addition, His-126 coordinates to the Zn ion by the NE2 atom; therefore, we protonated it on ND1. All MD simulations were performed using the GPU-accelerated pmemd code34−36 of AMBER 16.37 The protein and HOC-SG were described with the Amber ff14SB38 and GAFF39 force fields, respectively. Water molecules were described by the TIP3P model.40 Two different sets of simulations were performed in this study. One involved the HOto99 conformation and the other one the HOto172 conformation of the substrate. For the Zn sites, restrained electrostatic potential charges were employed,41 fitted to electrostatic potentials calculated, and sampled with the Mertz−Kollman scheme.42 The calculations were performed on the QM-optimized cluster models of HOto99 and HOto172 conformations, shown in Figure S1 in the Supporting Information. The geometries were truncated from the starting crystal structure.12 These calculations were performed using the Turbomole software (version 7.1)43 at the TPSS/def2-SV(P) level of theory.44,45 The optimization processes were sped up by expanding the Coulomb interactions in an auxiliary basis set, the resolution of identity approximation.46,47 The calculated charges are given in Tables S1 and S2. The metal sites were treated by a nonbonded model with restraints between the metal and the ligands.48 This is necessary to keep the structure of the Zn cluster intact. For the restraints, we used the metal−ligand distances observed in the crystal structure (averaged over the two subunits) and the force constants given in Table S3. The latter were obtained from TPSS/def2-SV(P) frequency calculations of the optimized models in Figure S1, calculated by the method of Seminario49,50 and averaged over interactions of the same type.

2. MODELING AND COMPUTATIONAL DETAILS 2.1. QM-Cluster Models and Calculations. To study the desired reaction with QM-cluster approach, we need a model of the active site. The experimental data come from yeast GlxI, but there is no crystal structure for GlxI from this organism. Therefore, we had to use a crystal structure for GlxI from another species to model the active site. Aronsson et al. showed that GlxI from both human and yeast contains one zinc ion per active site.11 Therefore, we used the crystal structure of native human GlxI to model the active site (Protein Data Bank entry 1QIN).12 The model consisted of the zinc atom and its firstcoordination-shell amino acids (Gln-33, Glu-99, Glu-172, and His126), as well as a model of HOC-SG. The glutamates were represented by propionate, glutamine was represented by propanamide, and histidine was represented by methylimidazole. The inhibitor in the crystal structure (S-[N-hydroxy-N-(p-iodophenyl)carbamoyl]glutathione) was modified to a model of HOC-SG: the p-iodophenyl group was replaced by a hydrogen atom, the N atom next to the iodophenyl group by a carbon atom, and the −SG group by a −CH2SH group. Hydrogen atoms were added manually. To maintain the overall structure of the active site, carbon atoms bound to H atoms that truncated the active-site model were fixed at their positions in the crystal structure during the geometry optimizations. We also employed an extended QM model. In this model, the glutamine, histidine, and substrate were the same as in the first model. However, the complete glutamate residues were included in the model (including the backbone). In addition, −COCH3 and −NHCH3 from the neighboring residues (Leu-98, Leu-100, Ile-171, and Ile-173) were also included, giving a CH3−CONH−CHCH2CH2COO−CONH− CH3 model of each glutamate. For glutamine and histidine, the same atoms were fixed in the two models. However, for the glutamates two atoms in the ending − H3 groups were fixed (those corresponding to CB and N or C). We denote the two models as M1 and M2, respectively. An investigation of the crystal structure showed that none of the neighboring residues have any direct interaction with the active site. However, the OG1 atom of Thr-101 makes a water-mediated (HOH404) hydrogen bond to Glu-99. Therefore, we also extended M2 with Thr-101 and HOH-404 in a third model (M3). The threonine was modeled as isopropyl alcohol. In addition, we enlarged the model of HOC-SG by replacing the −SG group by −CH2SCH3 in M3 (instead of −CH2SH in M1 and M2). The same atoms were fixed in M2 and M3, and the CA atom was fixed for Thr-101. The three models are illustrated in Figure 1, the fixed atoms being marked with asterisks. From Figure 1, it can be seen that the HR and HS atoms are directed toward Glu-99 and Glu-172, respectively. However, the alcoholic proton of HOC-SG (HO) can be directed toward either Glu-172 or Glu-99 (the atom names are defined in Figure 1). Therefore, we performed calculations for two different states, depending on the direction of HO. We denote these two conformations HOto99 and HOto172, when HO is directed toward Glu-99 and Glu-172, respectively, and they are both shown in Figure 1. D


Article

Inorganic Chemistry Scheme 5. Possible Proton Transfers in the Re states of M3 in (a) HOto99 and (b) HOto172 Conformations

Figure 2. IM1 states for model M3: (a) HR to Glu-99 in HOto99; (b) HS to Glu-172 in HOto99; (c) HR to Glu-99 in HOto172; (d) HS to Glu-172 in HOto172. Selected distances are shown in Å. The root-mean-square deviation (RMSD) and root-mean-square fluctuation (RMSF) was calculated with the AMBER cpptraj module,56 analyzing trajectories with coordinates saved every 10 ps and using the crystal structure as the reference. The reported values are averages over these 10000 sets of coordinates.

The setup of the MD simulations is similar to that in our recent work.51 The enzyme was solvated in a periodic truncated octahedral box of TIP3P water molecules, extending at least 12 Å from the solute using the leap program in the Amber suite.37 The final system contained 38465 atoms. After solvation, we performed 1000 cycles of minimization, with the heavy atoms of the protein restrained toward the starting structure with a force constant of 100 kcal/(mol Å2). This was followed by a 20 ps constant-volume and a 20 ps constantpressure equilibration with the same restraints, but with a force constant of 50 kcal/(mol Å2). Finally, the system was equilibrated for 1 ns without any restraints, followed by a 100 ns production simulation, during which coordinates were sampled every 10 ps. The temperature was kept constant at 300 K using Langevin dynamics with a collision frequency of 2 ps−1.52 The pressure was kept constant at 1 atm using Berendsen’s weak coupling isotropic algorithm with a relaxation time of 1 ps.53 Long-range electrostatics were handled by particle-mesh Ewald summation54 with a fourth-order B spline interpolation and a tolerance of 10−5. The cutoff radius for LennardJones interactions was set to 8 Å. All bonds involving hydrogen atoms were constrained to their equilibrium values using the SHAKE algorithm,55 allowing for a time step of 2 fs during the simulations.

3. RESULTS AND DISCUSSION Among the protons of the substrate analogue, only three of them (HR, HS, and HO) can directly participate in the catalytic reaction and only one of them (HO) can be exchanged noncatalytically by deuterium (atom names are shown in Figure 1). In other words, HO is an alcoholic proton and can be exchanged by a deuterium from D2O in the reaction medium. On the other hand, HR and HS are connected to an sp3 carbon atom (C1) and cannot be exchanged by a solvent deuterium noncatalytically. Therefore, we suppose that HO has been exchanged by a deuterium atom before entering the enzyme active site and, after that, the enzyme exchanges HS by this deuterium. E


Article

Inorganic Chemistry

Glu-99.12 Therefore, we improved the model by giving more flexibility to the Glu-172 residue, by releasing the fixed atom of this residue. The results for this flexible M1 model showed that there is no way for HOC-SG to exchange its protons if the alcoholic proton is directed toward Glu-99. However, if the alcoholic proton instead is directed toward the more flexible Glu-172 residue, we find a reaction path for exchange of the HS proton by deuterium, in accordance with the experimental findings (i.e., that the product will always have the deuterium in the S position).19 The flexible M1 model showed that Glu-172 flexibility is important to accurately describe the reaction mechanism. This flexibility is better accounted for in the M2 and M3 models, in which the two active-site glutamate groups are extended. Therefore, we focus on the results for the M2 and M3 models, which are more reliable. The results for the rigid and flexible M1 models are collected in the Supporting Information. 3.1. Results with M3 QM-Cluster Model. As described in the Modeling and Computational Details section and shown in Figure 1, the M2 model included the entire glutamate residues and also the backbone of the previous and following residues. We also tried a slightly larger model (M3) in which HOH-404 and Thr-101 are also added to M2. We found that models M2 and M3 gave similar reaction paths, although the energy profiles are slightly different (the barriers of M3 are lower than those of M2), indicating that the crystal water and the threonine residue do not change the reaction path but lower the energy barrier. Therefore, we will focus on the largest model (M3). The M2 optimized structures, selected distances of the optimized structures, and energy profiles are described in the Supporting Information. The optimized structures of the Re states of M3 are shown in Figure 1e,f. In contrast to the results obtained with model M1, the HO atom has moved from O1 to O5 of Glu-172 in Re of the HOto172 conformation. This proton exchange stabilizes the Re state of HOto172 by 2.0 kcal/mol with respect to Re of HOto99. The abstraction of HO by Glu-172 in Re of HOto172 implies a higher basicity for Glu-172. This is in accordance with our M1 results that the higher flexibility of Glu-172 and its displacement from the Zn ion make it a stronger base than Glu99 (cf. the Supporting Information). This was also suggested by Cameron et al.,12 who reported that the displacement of Glu172 from the Zn ion upon substrate binding results in a pKa shift of its carboxylate group to better match that of the substrate, which is lowered by the Zn interaction. Starting from the Re states, there are three possible proton transfers for the HOto99 conformation (HO and HR to Glu-99 and HS to Glu-172; these are represented by arrows in Scheme 5a) but only two possibilities for HOto172 (HR to Glu-99 and HS to Glu-172; these are represented by arrows in Scheme 5b). Our calculations showed that HO cannot move to Glu-99 in the first step for the HOto99 conformation (the proton transfer shown by a red arrow in Scheme 5a). In conclusion,

Figure 3. Calculated energy profiles of transferring HR to Glu-99 and HS to Glu-172 in model M3. The energy profiles were obtained from relaxed potential energy scans with a single fixed distance.

Scheme 6. Possible Proton Transfers from the Second IM1 of HOto99

Scheme 7. Possible Proton Transfer from the Second IM1 of HOto172

We used three different QM-cluster models to study the proton exchange mechanism. The first model (M1; Figure 1a,b) is the smallest one, and the models of the most important residues (Glu-99 and Glu-172) in M1 are small; this makes it very rigid. Our results showed that this rigid model cannot perform the reaction (some of the proton transfers could not be performed or had activation energies that were too high; a full account of the M1 results are given in the Supporting Information). Therefore, we tried to investigate the reaction mechanisms, considering the experimental fact that there is a flexible loop including residues 152−159 over the active site in the crystal structure of GlxI, which is closer to Glu-172 than to

Scheme 8. Proton Transfers from (a) the First IM1 and (b) IM2 of HOto172

F


Article

Inorganic Chemistry

Figure 4. (a) IM2 and (b) Pr states of model M3. Selected distances are shown in Å.

Figure 5. Full energy profile for model M3. The energy profiles were obtained from relaxed potential energy scans with a single fixed distance.

transferring HO to the glutamates cannot be the first step in the proton exchange mechanism. On the other hand, the other four proton-transfer reactions turned out to be possible (these are shown by blue arrows in Scheme 5). We denote the corresponding products of these proton transfers as IM1. The optimized structures of the IM1 states and the energy barriers for production of these states are shown in Figures 2 and 3, respectively. For the HOto99 conformation we found that the height of the energy profiles of the proton abstractions shown by blue arrows in Scheme 5a depends on the direction of HO (cf. Figure 3 for the profiles). The energy barrier is high (23.8 kcal/mol) when HO and the

abstracted proton are on the same side of the O1−C1−C2−O2 plane (HR to Glu-99; the proton transfer is shown by a dark blue arrow in Scheme 5a), whereas it is low (10.2 kcal/mol) when they are on opposite sides (HS to Glu-172; the proton transfer shown by a light blue arrow in Scheme 5a). However, for the HOto172 conformation of M3 the barriers of these reactions are almost the same. As shown in Figure 3, the energy barriers of HS to Glu-172 (the proton transfer shown by a dark blue arrow in Scheme 5b) and HR to Glu-99 (the proton transfer shown by a light blue arrow in Scheme 5b) are 14.9 and 15.1 kcal/mol, respectively. G


Article

Inorganic Chemistry Table 1. Selected Distances of Re, IM1 of HR to Glu-99 in HOto172, and Pr in the Flexible M1, M2, and M3 cluster model M1-Fa state Zn−O1 Zn−O2 Zn−O3 Zn−O4 Zn−O5 Zn−O6 C1−C2 C1−O1 C2−O2 C1−HS C1−HR HR−O4 HR−O1 HS−O6 HS−O1 HO−O1 HO−O4 HO−O6 HO−C1 CA−CAc CB−CBc CG−CGc Zn−CA-99d Zn−CA-172e Zn−CB-99f Zn−CB-172g

M1-Fa b

Re 2.24 2.27 2.06 3.43 2.16 3.40 1.52 1.41 1.23 1.10 1.09 2.50 2.06 3.30 2.07 1.01 4.19 1.62 2.01

IM1′ 2.07 2.35 2.14 2.29 4.28 4.22 1.53 1.39 1.22 1.11 1.11 3.00 2.08 2.82 2.07 1.58 4.78 1.03 2.40

9.65 8.90

11.49 10.18

4.81 4.94

4.73 7.05

M2

M3

M1-Fa

M2

M3

M1-Fa

M2

M3

Re 2.05 2.29 1.97 3.07 3.65 4.16 1.52 1.40 1.23 1.10 1.10 2.47 2.07 3.05 2.07 1.50 4.80 2.42 2.35 12.22 9.53 9.51 6.10 6.59 4.58 5.17

Re 2.05 2.28 1.98 3.60 3.52 4.29 1.53 1.40 1.23 1.10 1.11 4.22 2.08 2.81 2.07 1.54 6.00 2.40 2.38 12.12 9.41 9.65 6.05 6.66 4.52 5.20

IM1 2.26 2.05 2.39 3.81 2.12 3.34 1.37 1.41 1.30 1.09 1.96 1.01 2.39 4.06 2.13 1.00 4.17 1.71 2.01

IM1 2.33 2.07 2.23 3.82 2.06 3.47 1.38 1.41 1.30 1.09 1.90 1.02 2.46 3.85 2.13 1.03 4.29 1.57 2.02 11.88 9.04 8.91 6.33 6.16 4.84 4.65

IM1 2.35 2.06 2.23 3.82 2.08 3.50 1.38 1.41 1.29 1.09 1.84 1.03 2.37 3.77 2.13 1.03 4.21 1.57 2.01 11.93 9.10 8.91 6.31 6.20 4.83 4.67

Pr 2.22 2.26 2.18 3.41 2.06 3.43 1.52 1.42 1.23 1.10 2.02 1.62 1.01 3.94 2.08 2.05 4.42 2.52 1.09

Pr 2.23 2.26 2.15 3.38 2.06 3.53 1.52 1.42 1.23 1.10 2.02 1.62 1.01 3.96 2.08 2.05 4.42 2.53 1.09 11.91 9.10 8.87 6.36 6.19 4.84 4.67

Pr 2.32 2.23 2.09 3.43 2.08 3.55 1.52 1.41 1.23 1.10 2.01 1.50 1.03 3.89 2.07 2.06 4.22 2.49 1.09 11.87 9.04 8.81 6.24 6.20 4.72 4.68

9.86 9.11

5.11 4.88

9.37 8.88

5.00 4.81

a M1-F refers to the flexible M1 model. bWe used IM′ for the reactant state of model M1 in which HO is moved on Glu-172. Actually, in Re of the HOto172 conformation of M2 and M3 HO has moved from O1 to O5 of Glu-172, therefore giving directly the IM1′ state. cDistances between the corresponding atoms in Glu-99 and Glu-172. dDistance between Zn and CA of Glu-99. eDistance between Zn and CA of Glu-172. fDistance between Zn and CB of Glu-99. gDistance between Zn and CB of Glu-172.

addition, the increased negative charge on O2 (from −0.61 in Re to −0.83 in IM1 of HOto99 and from −0.61 in Re to −0.81 in IM1 of HOto172) leads to a stronger coordination of O2 to the Zn ion (the Zn−O2 distances are 2.28 and ∼2.07 Å for Re and IM1, respectively). Considering the barriers in Figure 3, we discard the highenergy proton abstraction (HR by Glu-99; the proton transfer is shown by a dark blue arrow in Scheme 5a) and concentrate on the low energy abstraction (HS by Glu-172; the proton transfer is shown by a light blue arrow in Scheme 5a) in HOto99 but we study both possible proton abstractions in the HOto172 conformation (HS by Glu-172 and HR by Glu-99). In the HOto99 conformation of M3, after HS is moved to Glu-172 (cf. Figure 2b for the resulting IM1), the next possible proton transfer is to move HO to O4 of Glu-99 (the straight arrow in Scheme 6). Our results show that this transfer is not possible, because the proton returns to the starting point after releasing any bond constraints. We also tried to directly move HS to O1 in IM1 (the curved arrow in Scheme 6). However, we found that this transfer is also not possible. Thus, we did not find any possible reaction path for the HOto99 conformation. The same result was obtained also with model M2. As was mentioned before, according to the barriers summarized in Figure 3, we study the reactions from both IM1 states of the HOto172 conformation (the structures shown in Figure 2c,d; it is notable that HO has returned from Glu-172 to O1 in both IM1 states of this conformation). In the second

Thus, the barrier for the transfer of HR to Glu-99 in HOto172 is higher than the barrier of HS to Glu-172 in HOto99 (15.1 vs 10.2 kcal/mol; the barriers correspond to the transfers shown by the light blue arrows in Scheme 5). The reason for this is that Glu-172 is already protonated in Re of HOto172 (cf. Figure 1f), and this gives a negative charge on the substrate. Therefore, abstraction of the second proton (HR) by Glu-99 will impose a double negative charge to the substrate, which increases the energy. In addition, the abstracting base in the HOto99 conformation is Glu-172, which is a stronger base than Glu99. In contrast, the barrier of HR to Glu-99 in the HOto99 conformation is higher than the barrier of HS to Glu-172 in HOto172 (23.8 vs 14.9 kcal/mol; the barriers correspond to the transfers shown by the dark blue arrows in Scheme 5). This again confirms the higher basicity of Glu-172, which is due to its different environment inside the enzyme. The different environments of the glutamates will be discussed in MD Results. Previous calculations have suggested that the role of the Zn ion in the reaction of the normal substrate was to electrostatically stabilize the enediolate intermediate, thereby lowering the energy of proton transfer and or stabilizing the developing negative charge on O2 of the enediolate intermediate.16,57 The ion plays the same role in the present reaction. On going from Re to IM1, the C1−C2 single bond is shortened from 1.53 to 1.38 Å and the C2O2 double bond is elongated from 1.22 to 1.29 Å (cf. Table S5 and S6) in both conformations. In H


Article

Inorganic Chemistry

Figure 6. Superimposition of (a) Re on Re, (b) Re on IM1′, (c) IM1 on IM1, and (d) Pr on Pr of M3 and M1. M1 and M3 states are shown by tube and ball and stick representations, respectively. Zn atoms are shown by balls in both models. The extra atoms of model M3 are omitted.

abstraction by the glutamates but a stereospecific proton delivery to the si face of C2 (the face which is directed to Glu172). Our results show that both glutamates can abstract protons (HS or HR), but deuterium can only be added by Glu172 to only one face of C1. We could not find any reaction path from the HOto99 conformation in the QM-cluster models. In fact, the HOto99 conformation of the reactant does not seem to be stable in the active site of the enzyme. In a detailed discussion in our previous study,16 on the basis of the experimental observation that the binding an inhibitor displaces Glu-172 but not Glu-99 from the active-site Zn ion12 and the theoretical results of Richter and Krauss,13 which showed that if the two glutamate ligands are free to move during the optimization a protonated Glu-172 will dissociate from the zinc ion, we concluded that in the crystal structure Glu-172 abstracted a proton from the substrate and dissociated from Zn. Our current results show that this partially unstable conformation (HOto99) does not lead to any reaction path. 3.3. Comparison of the Results of the Various QMCluster Models. The results coming from the flexible M1 model (discussed in the Supporting Information) show that Glu-172 flexibility needs to be accounted for in order to

IM1 of HOto172 (IM1 of HS to Glu-172; Figure 2d) we moved HO to O5 of Glu-172 (the transfer shown by a red arrow in Scheme 7), but HS returned to C1, giving the Re state of HOto172. Thus, there is no reaction path from the second IM1 of HOto172. In the first IM1 of HOto172 (IM1 of HR to Glu-99; Figure 2c), we moved HO to Glu-172, but it returned to the starting point on O1 (the transfer shown by a blue arrow in Scheme 8a). However, we found a path from the first IM1 when we tried to directly transfer HR to O1 (the transfer shown by a green arrow in Scheme 8a). In this transfer, HO moved simultaneously to Glu-172, producing IM2 (cf. Figure 4a or Scheme 8b). Finally, we transferred HO to C1 in IM2 (the transfer shown by a green arrow in Scheme 8b), producing Pr (Figure 4b). The energy profile for the reaction path (moving first HR to Glu-99, then HR to O1, and finally HO to C1) in the HOto172 conformation is shown in Figure 5. This is the only reaction path found for both models M2 and M3. It exchanges HS, but not HR, with HO (which may be exchanged by a deuterium before entering the active site). Exchanging HS by a deuterium is in accordance with the experimental observation.19 These findings are also in line with the results of Landro et al.,5 who showed by 1H NMR analysis a nonstereospecific proton I


Article

Inorganic Chemistry Scheme 9. Proposed Mechanism for Proton Exchange of HOC-SG by GlxI

Table 2. Mass Weighted RMSD and RMSF Values (in Å) of the Two Active-Site Glutamate Residues for the HOto99 and HOto172 Conformations of HOC-SGa HOto99

HOto172

active site 1 RMSD RMSF a

active site 2

active site 1

active site 2

Glu-99A

Glu-172B

Glu-99B

Glu-172A

Glu-99A

Glu-172B

Glu-99B

Glu-172A

0.29 7.29

0.85 7.46

0.38 7.13

0.87 7.96

0.31 11.09

0.89 11.80

0.35 6.72

0.94 7.83

Glu-99A and Glu-172B belong to active site 1, and Glu-99B and Glu-172A belong to active site 2.

Glu-172, giving directly the IM1′ state). According to the results summarized in Table 1 and Figure 6, we can conclude that the geometries and the important prediction from the flexible M1 (the higher flexibility of Glu-172) are reasonable. In addition, the flexibility of Glu-172 with respect to Glu-99 is confirmed by the MD simulations, discussed in MD Results. Other important results of model M1 are that (i) Glu-172 is a stronger base than Glu-99 and (ii) the HOto99 conformation does not led to any proton exchange reaction. Both of these results are confirmed by the results of the larger cluster models. On the other hand, some details of the predicted reaction profile from the flexible M1 model (Figure S9) are different from those obtained with M2 and M3. Therefore, our proposed mechanism (Scheme 9) and the energy profile (Figure 5) are based on the results of the largest model, which should be more reliable. From the results of the QM-cluster models, we can conclude that the proton-exchange mechanism depends on the different actions of the active-site glutamates; in particular, Glu-172 is

accurately describe the reaction mechanism. This is one of the most important results in this study. However, M1 is very small and the optimized geometries in the absence of any constraint in Glu-172 might not be accurate. To check the accuracy of the geometries obtained with the flexible M1 model, we compared some selected distances of three states of this model (Re of HOto172, IM1 of HR to Glu-99 in HOto172, and Pr; these states are the common states of the models present in the reaction profile given in Figure 5) with the same states in M2 and M3 in Table 1. In addition, we superimposed the corresponding states of M1 and M3 (cf. Figure 6). According to Table 1, there is no significant difference between the distances of the various states in different cluster models. According to the superimposed structures in Figure 6, the IM1 and Pr states of M1 and M3 fit nicely, whereas the carboxylate groups of both of the glutamates have somewhat different positions in Re and IM′ (we use IM′ for the reactant state of M1 in which HO is moved on Glu-172; in Re of the HOto172 conformation of M2 and M3 HO has moved from O1 to O5 of J


Article

Inorganic Chemistry

Figure 7. Calculated RMSF values of chains A and B of GlxI in the conformations (a) HOto99 and (b) HOto172.

more flexible and more basic than Glu-99. To confirm this suggestion, we studied the flexibility of the glutamate residues in MD simulations. The results are discussed in MD Results. 3.2. MD Results. We have run two 100 ns MD simulations of GlxI with the substrate analogue bound in either the HOto99

or HOto172 conformations. To analyze the flexibility of the two active-site glutamate residues, we calculated the mass-weighted RMSD and RMSF values (with the crystal structure as the reference) of these two residues during the simulations. The results in Table 2 show that the RMSD values of Glu-172 in K


Article

Inorganic Chemistry

whole enzyme (15.0 vs 11.9 Å) in chain A, but it is the most flexible part of chain B (cf. Figure 7b). Residues 63−67 have the highest RMSF in chain A in the simulation of the HOto172 state. We show the position of these flexible loops (residues 63−67, 135−145, and 152−159) along with the two glutamate residues in Figure 8. According to Figure 8, loops 135−145 and 152−159 are much closer to Glu-172 than to Glu-99. However, loop 63−67 is almost at the same distance from both glutamates. These show that the Glu-172 environment is more flexible than the Glu-99 environment. As stated by Cameron et al.12 and as our QM-cluster results show, Glu-172 is a stronger base than Glu-99. The higher basicity of Glu-172 cannot be due to a direct induction effect from the neighboring residues, because the neighboring residues of both Glu-99 and Glu-172 are similar (they are all hydrophobic residues: Leu-98, Leu-100, Ile-171, and Ile-173). The most important cause for the higher basicity of Glu-172 could be that it dissociates from the Zn ion, as our QM-cluster results show (these are discussed in section 3.1; compare the Re states of M3 in Figure 1e,f) and as was suggested by Cameron et al.12 In other words, dissociation from a positive ion (the Zn ion) induces more negative charge and therefore more basicity to the carboxylate group of Glu-172. As our MD results show, Glu-172 has a more flexible environment than Glu-99. This flexibility allows dissociation of Glu-172 and consequenctly increases its basicity. We also checked this hypothesis by comparing average RMSF values of the backbone atoms of the four nearest neighbors of the glutamates (i.e., the backbone of residues 95−103 for Glu-99 and residues 168−176 for Glu-172). The results are summarized in Table 3. The side chains of the glutamates are connected to the backbone, and the more flexible backbone allows the side chain to dissociate from the Zn ion. According to Table 3, the neighboring backbone of Glu-172 is more flexible than that of Glu-99 in both simulations and both active sites.

Figure 8. Flexible loops around the glutamates inside GlxI. Loop 152− 159 is shown by tubes to distinguish it from the other loops.

both active sites (we simulated the full dimeric protein) and with both conformations of the substrate are 2−3 times larger than those of Glu-99 (0.85−0.94 Å, in comparison to 0.29− 0.38 Å). The RMSF values of Glu-172 in both active sites and with both conformations of the substrate are also larger than those of Glu-99. These show that Glu-172 is much more flexible than Glu-99. This can be understood by looking at the position of the glutamates in the crystal structure, as is shown in Figure 8. It can be seen that both residues are buried in the protein (both residues have a vanishing solvent-accessible surface area) but that Glu-172 is somewhat closer to the surface of the protein and Glu-99 is closer to the center of the protein. It is quite natural that a residue nearer the surface is more flexible than a buried residue. To check the active site glutamate environments, we calculated RMSF values of the whole enzyme in both conformations of the substrate and summarized them in Figure 7. According to the RMSF results shown in Figure 7, residues 135−145 constitute the part of the enzyme with the highest RMSF in the HOto99 conformation. The experimentally suggested flexible loop (residues 152−159) is not the most flexible part of the enzyme in the HOto99 conformation. However, the average RMSF value of this loop is more than the average RMSF of the whole enzyme (11.3 vs 9.8 Å) in the HOto99 conformation (cf. Figure 7a). The results are somewhat different in the HOto172 conformation. The average RMSF of the loop 152−159 is higher than the average RMSF of the

4. CONCLUSION To gain some understanding of stereospecific deuterium exchange of HOC-SG by GlxI, we have performed DFT calculations on active-site QM-cluster models. With the minimum M1 model, the height of the energy barriers for proton abstraction from C1 depends on the direction of HO: the reaction barrier is high if HO and the abstracted proton are on the same side of the O1−C1−C2−O2 plane, but it is lower if HO and the abstracted proton are on opposite sides. However, this effect vanished in the HOto172 conformation with the larger M3 model. The small model with the same flexibility for the two active-site glutamate residues could not explain the unusual stereospecificity of GlxI. However, if we do not constrain any atoms in Glu-172 or use a larger model, including the entire glutamate residues as well as the backbone of the surrounding residues (models M2 and M3), we find a catalytic reaction mechanism for the exchange of the HS proton by a deuterium as was observed by experiments.19 This is possible only if the substrate HO is directed toward the more

Table 3. Average RMSF Values (in Å) of the Backbone of Four Nearest Neighbors of the Glutamates HOto99

HOto172

active site 1 RMSF

active site 2

active site 1

active site 2

Glu-99A

Glu-172B

Glu-99B

Glu-172A

Glu-99A

Glu-172B

Glu-99B

Glu-172A

7.80

8.00

7.52

8.64

11.27

12.00

7.38

9.14

L


Inorganic Chemistry

■

flexible Glu-172 residue, whereas no such reaction is possible if HO is directed toward Glu-99. The reason for this is that Glu172 is more flexible and more basic than Glu-99. These suggestions were confirmed by MD simulations, which showed that Glu-172 is 2−3 times more flexible than Glu-99. This flexibility allows it to temporarily dissociate from the active-site Zn ion. We found that models M2 and M3 gave a similar reaction path, although the energy profiles are slightly different (the barriers of M3 are lower than those of M2), indicating that the crystal water HOH-404 and the Thr-101 residue do not change the reaction path but lower the energy barrier. On the basis of the DFT calculations with the largest M3 QM-cluster model, we propose the mechanism shown in Scheme 9 for the proton exchange of HOC-SG by GlxI. First, HOC-SG exchanges noncatalytically its alcoholic proton with a deuterium from the solvent, before binding. Next, the deuterium is abstracted by Glu-172 and HR moves to Glu-99. After that, HR is transferred from Glu-99 to O1, leading to the simultaneous transfer of the deuterium from O1 to Glu-172. Then, the deuterium moves from Glu-172 to C1, producing the proton-exchanged product. Finally, the product dissociates from the active site, allowing the enzyme to start a new catalytic cycle. According to this mechanism, the product will always have the deuterium in the S position, in accordance with the experimental findings. It is the higher flexibility of Glu-172 that explains the unusual stereospecificity of GlxI.

■

REFERENCES

(1) Kalapos, M. P. Methylglyoxal in Living Organisms - Chemistry, Biochemistry, Toxicology and Biological Implications. Toxicol. Lett. 1999, 110 (3), 145−175. (2) Antognelli, C.; Baldracchini, F.; Talesa, V. N.; Coslanlini, E.; Zucchi, A.; Mearini, E. Overexpression of Glyoxalase System Enzymes in Human Kidney Tnmor. Cancer J. 2006, 12 (3), 222−228. (3) Thornalley, P. J. The Glyoxalase System: New Developments towards Functional Characterization of a Metabolic Pathway Fundamental to Biological Life. Biochem. J. 1990, 269 (1), 1−11. (4) Sousa Silva, M.; Ferreira, A. E. N.; Gomes, R.; Tomás, A. M.; Ponces Freire, A.; Cordeiro, C. The Glyoxalase Pathway in Protozoan Parasites. Int. J. Med. Microbiol. 2012, 302 (4−5), 225−229. (5) Landro, J. A.; Brush, E. J.; Kozarich, J. W. Isomerization of (R)and (S)-Glutathiolactaldehydes by Glyoxalase I: The Case for Dichotomous Stereochemical Behavior in a Single Active Site. Biochemistry 1992, 31 (26), 6069−6077. (6) Om, M.; Mannervik, B. Optimized Heterologous Expression of the Human Zinc Enzyme Glyoxalase I. Biochem. J. 1996, 314, 463− 467. (7) Saint-Jean, A. P.; Phillips, K. R.; Creighton, D. J.; Stone, M. J. Active Monomeric and Dimeric Forms of Pseudomonas Putida Glyoxalase I: Evidence for 3D Domain Swapping. Biochemistry 1998, 37 (29), 10345−10353. (8) Clugston, S. L.; Daub, E.; Kinach, R.; Miedema, D.; Barnard, J. F. J.; Honek, J. F. Isolation and Sequencing of a Gene Coding for Glyoxalase I Activity from Salmonella Typhimurium and Comparison with Other Glyoxalase I Sequences. Gene 1997, 186 (1), 103−111. (9) Clugston, S. L.; Barnard, J. F. J.; Kinach, R.; Miedema, D.; Ruman, R.; Daub, E.; Honek, J. F. Overproduction and Characterization of a Dimeric Non-Zinc Glyoxalase I from Escherichia Coli: Evidence for Optimal Activation by Nickel Ions. Biochemistry 1998, 37 (24), 8754−8763. (10) He, M. M.; Clugston, S. L.; Honek, J. F.; Matthews, B. W. Determination of the Structure of Escherichia Coli Glyoxalase I Suggests a Structural Basis for Differential Metal Activation. Biochemistry 2000, 39 (30), 8719−8727. (11) Aronsson, A. C.; Marmstål, E.; Mannervik, B. Glyoxalase I, a Zinc Metalloenzyme of Mammals and Yeast. Biochem. Biophys. Res. Commun. 1978, 81 (4), 1235−1240. (12) Cameron, A. D.; Ridderström, M.; Olin, B.; Kavarana, M. J.; Creighton, D. J.; Mannervik, B. Reaction Mechanism of Glyoxalase I Explored by an X-Ray Crystallographic Analysis of the Human Enzyme in Complex with a Transition State Analogue. Biochemistry 1999, 38 (41), 13480−13490. (13) Richter, U.; Krauss, M. Active Site Structure and Mechanism of Human Glyoxalase I-an Ab Initio Theoretical Study. J. Am. Chem. Soc. 2001, 123 (19), 6973−6982. (14) Himo, F.; Siegbahn, P. E. M. Catalytic Mechanism of Glyoxalase I: A Theoretical Study. J. Am. Chem. Soc. 2001, 123 (42), 10280− 10289. (15) Creighton, D. J.; Hamilton, D. S. Brief History of Glyoxalase I and What We Have Learned about Metal Ion-Dependent, EnzymeCatalyzed Isomerizations. Arch. Biochem. Biophys. 2001, 387 (1), 1−10. (16) Jafari, S.; Ryde, U.; Irani, M. Catalytic Mechanism of Human Glyoxalase I Studied by Quantum-Mechanical Cluster Calculations. J. Mol. Catal. B: Enzym. 2016, 131, 18−30. (17) Chari, R. V.; Kozarich, J. W. Deuterium Isotope Effects on the Product Partitioning of Fluoromethylglyoxal by Glyoxalase I. Proof of a Proton Transfer Mechanism. J. Biol. Chem. 1981, 256 (19), 9785− 9788. (18) Kozarich, J. W.; Chari, R. V. J. (Glutathiomethyl) Glyoxal: Mirror-Image Catalysis by Glyoxalase I. J. Am. Chem. Soc. 1982, 104 (9), 2655−2657. (19) Chari, R. V. J.; Kozarich, J. W. Glutathiohydroxyacetone: Proton NMR Determination of the Stereochemistry of Proton Exchange by Glyoxalase I. Evidence for a Cis-Enediol Intermediate Based on Mirror-Image Catalysis. J. Am. Chem. Soc. 1983, 105 (24), 7169−7171.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b03215. MM charges, force constants and restraints used for the Zn clusters, full discussion of M1 results, selected distances of the various states in M1, M2 and M3, illustration of stationary states and energy profiles of M2, dispersion correction effect on the first reaction step in Figure 5, and Cartesian coordinates of stationary points in the QM-cluster models (PDF)

■

Article

AUTHOR INFORMATION

Corresponding Author

*M.I.: e-mail, [email protected]; tel, +98−9128018046; fax, +98−8733624133. ORCID

Ulf Ryde: 0000-0001-7653-8489 Mehdi Irani: 0000-0002-7409-7760 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This investigation has been supported by grants from the Swedish research council (project 2014-5540) and from the University of Kurdistan (project 4/7224). The computations were performed on computer resources provided by the Swedish National Infrastructure for Computing (SNIC) at Lunarc at Lund University and at HPC2N at Umeå University. M


Article

Inorganic Chemistry

Residues in Empirical P K a Predictions. J. Chem. Theory Comput. 2011, 7 (2), 525−537. (34) Le Grand, S.; Götz, A. W.; Walker, R. C. SPFP: Speed without Compromise - A Mixed Precision Model for GPU Accelerated Molecular Dynamics Simulations. Comput. Phys. Commun. 2013, 184 (2), 374−380. (35) Salomon-Ferrer, R.; Götz, A. W.; Poole, D.; Le Grand, S.; Walker, R. C. Routine Microsecond Molecular Dynamics Simulations with AMBER on GPUs. 2. Explicit Solvent Particle Mesh Ewald. J. Chem. Theory Comput. 2013, 9 (9), 3878−3888. (36) Götz, A. W.; Williamson, M. J.; Xu, D.; Poole, D.; Le Grand, S.; Walker, R. C. Routine Microsecond Molecular Dynamics Simulations with AMBER on GPUs. 1. Generalized Born. J. Chem. Theory Comput. 2012, 8 (5), 1542−1555. (37) Case, D. A.; Betz, R. M.; Cerutti, D. S.; Cheatham, T. E.; Darden, T. A.; Duke, R. E.; Giese, T. J.; Gohlke, H.; Goetz, A. W.; Homeyer, N.; Izadi, S.; Janowski, P.; Kaus, J.; Kovalenko, A.; Lee, T. S.; LeGrand, S.; Li, P.; Lin, C.; Luchko, T.; Luo, R.; Madej, B.; Mermelstein, D.; Merz, K. M.; Monard, G.; Nguyen, H.; Nguyen, H. T.; Omelyan, I.; Onufriev, D. R.; Roe, D. R.; Roitberg, A.; Sagui, C.; Simmerling, C. L.; Botello-Smith, W. M.; Swails, J.; Walker, R. C.; Wang, J.; Wolf, R. M.; Wu, X.; Xiao, L.; Kollman, P. A. AMBER 2016; University of California, San Francisco, CA, 2016. (38) Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.; Hauser, K. E.; Simmerling, C. ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from ff99SB. J. Chem. Theory Comput. 2015, 11 (8), 3696−3713. (39) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004, 25 (9), 1157−1174. (40) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79 (2), 926−935. (41) Bayly, C. I.; Cieplak, P.; Cornell, W.; Kollman, P. A. A WellBehaved Electrostatic Potential Based Method Using Charge Restraints for Deriving Atomic Charges: The RESP Model. J. Phys. Chem. 1993, 97 (40), 10269−10280. (42) Besler, B. H.; Merz, K. M.; Kollman, P. A. Atomic Charges Derived from Semiempirical Methods. J. Comput. Chem. 1990, 11 (4), 431−439. (43) Furche, F.; Ahlrichs, R.; Hättig, C.; Klopper, W.; Sierka, M.; Weigend, F. Turbomole. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2014, 4 (2), 91−100. (44) Schäfer, A.; Horn, H.; Ahlrichs, R. Fully Optimized Contracted Gaussian-Basis Sets for Atoms Li to Kr. J. Chem. Phys. 1992, 97 (4), 2571−2577. (45) Tao, J.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Climbing the Density Functional Ladder: Nonempirical Meta−Generalized Gradient Approximation Designed for Molecules and Solids. Phys. Rev. Lett. 2003, 91 (14), 146401. (46) Eichkorn, K.; Treutler, O.; Ö hm, H.; Häser, M.; Ahlrichs, R. Auxiliary Basis Sets to Approximate Coulomb Potentials. Chem. Phys. Lett. 1995, 240 (4), 283−289. (47) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Auxiliary Basis Sets for Main Row Atoms and Transition Metals and Their Use to Approximate Coulomb Potentials. Theor. Chem. Acc. 1997, 97 (1− 4), 119−124. (48) Hu, L.; Ryde, U. Comparison of Methods to Obtain Force-Field Parameters for Metal Sites. J. Chem. Theory Comput. 2011, 7 (8), 2452−2463. (49) Seminario, J. M. Calculation of Intramolecular Force Fields from Second-Derivative Tensors. Int. J. Quantum Chem. 1996, 60 (7), 1271−1277. (50) Nilsson, K.; Lecerof, D.; Sigfridsson, E.; Ryde, U. An Automatic Method to Generate Force-Field Parameters for Hetero-Compounds. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2003, 59 (2), 274−289. (51) Cao, L.; Caldararu, O.; Ryde, U. Protonation States of Homocitrate and Nearby Residues in Nitrogenase Studied by

(20) Sellin, S.; Rosevear, P. R.; Mannervik, B.; Mildvan, A. S. Nuclear Relaxation Studies of the Role of the Essential Metal in Glyoxalase I. J. Biol. Chem. 1982, 257 (17), 10023−10029. (21) Daver, H.; Das, B.; Nordlander, E.; Himo, F. Theoretical Study of Phosphodiester Hydrolysis and Transesterification Catalyzed by an Unsymmetric Biomimetic Dizinc Complex. Inorg. Chem. 2016, 55 (4), 1872−1882. (22) Hopmann, K. H. Full Reaction Mechanism of Nitrile Hydratase: A Cyclic Intermediate and an Unexpected Disulfide Switch. Inorg. Chem. 2014, 53 (6), 2760−2762. (23) Roos, K.; Siegbahn, P. E. M. Activation of Dimanganese Class Ib Ribonucleotide Reductase by Hydrogen Peroxide: Mechanistic Insights from Density Functional Theory. Inorg. Chem. 2013, 52 (8), 4173−4184. (24) Dong, G.; Ryde, U. O2 Activation in Salicylate 1,2-Dioxygenase: A QM/MM Study Reveals the Role of His162. Inorg. Chem. 2016, 55 (22), 11727−11735. (25) Sheng, X.; Himo, F. Theoretical Study of Enzyme Promiscuity: Mechanisms of Hydration and Carboxylation Activities of Phenolic Acid Decarboxylase. ACS Catal. 2017, 7, 1733−1741. (26) Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti Correlation-Energy Formula into a Functional of the Electron Density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37 (2), 785−789. (27) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, P. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision E.01 ; Gaussian, Inc.: Wallingford, CT, 2009. (28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J. V.; Izmaylov, A. F.; Sonnenberg, J. L.; Ding, F.; Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson, T.; Ranasinghe, D.; Zakrzewski, V. G.; Gao, J.; Rega, N.; Zheng, G.; Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Throssell, K.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J. J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Millam, J. M.; Klene, M.; Adamo, C.; Cammi, R.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J. Gaussian 16; Wallingford, CT,2016. (29) Hay, P. J.; Wadt, W. R. Ab Initio Effective Core Potentials for Molecular Calculations. Potentials for the Transition Metal Atoms Sc to Hg. J. Chem. Phys. 1985, 82 (1), 270. (30) Barone, V.; Cossi, M. Quantum Calculation of Molecular Energies and Energy Gradients in Solution by a Conductor Solvent Model. J. Phys. Chem. A 1998, 102 (11), 1995−2001. (31) Reed, A. E.; Weinhold, F. Natural Bond Orbital Analysis of nearHartree-Fock Water Dimer. J. Chem. Phys. 1983, 78 (6), 4066−4073. (32) Reed, A. E.; Curtiss, L. a.; Weinhold, F. Intermolecular Interactions from a Natural Bond Orbital, Donor-Acceptor Viewpoint. Chem. Rev. (Washington, DC, U. S.) 1988, 88 (6), 899−926. (33) Olsson, M. H. M.; Søndergaard, C. R.; Rostkowski, M.; Jensen, J. H. PROPKA3: Consistent Treatment of Internal and Surface N


Article

Inorganic Chemistry Computational Methods and Quantum Refinement. J. Phys. Chem. B 2017, 121 (35), 8242−8262. (52) Wu, X.; Brooks, B. R. Self-Guided Langevin Dynamics Simulation Method. Chem. Phys. Lett. 2003, 381 (3−4), 512−518. (53) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola, A.; Haak, J. R. Molecular Dynamics with Coupling to an External Bath. J. Chem. Phys. 1984, 81 (8), 3684−3690. (54) Darden, T.; York, D.; Pedersen, L. Particle Mesh Ewald: An N· log(N) Method for Ewald Sums in Large Systems. J. Chem. Phys. 1993, 98 (12), 10089−10092. (55) Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. Numerical Integration of the Cartesian Equations of Motion of a System with Constraints: Molecular Dynamics of N-Alkanes. J. Comput. Phys. 1977, 23 (3), 327−341. (56) Roe, D. R.; Cheatham, T. E. PTRAJ and CPPTRAJ: Software for Processing and Analysis of Molecular Dynamics Trajectory Data. J. Chem. Theory Comput. 2013, 9 (7), 3084−3095. (57) Feierberg, I.; Cameron, A. D.; Åqvist, J. Energetics of the Proposed Rate-Determining Step of the Glyoxalase I Reaction. FEBS Lett. 1999, 453 (1−2), 90−94.

O


Higher Flexibility of Glu-172 Explains the Unusual ... - ACS Publications

Recommend Documents