Scholarly article on topic ' Linking Electrostatic Effects and Protein Motions in Enzymatic Catalysis. A Theoretical Analysis of Catechol O -Methyltransferase '

Linking Electrostatic Effects and Protein Motions in Enzymatic Catalysis. A Theoretical Analysis of Catechol O -Methyltransferase Academic research paper on "Chemical sciences"

0
0
Share paper
Academic journal
The Journal of Physical Chemistry B
OECD Field of science
Keywords
{""}

Academic research paper on topic " Linking Electrostatic Effects and Protein Motions in Enzymatic Catalysis. A Theoretical Analysis of Catechol O -Methyltransferase "

PHYSICAL CHEMISTRY

Subscriber access provided by University of Sydney Library

Article

Linking Electrostatic Effects and Protein Motions in Enzymatic Catalysis. A Theoretical Analysis of Catechol O-Methyltransferase

Rafael Garcia-Meseguer, Kirill Zinovjev, Maite Roca, J. Javier Ruiz-Pernia, and Inaki Tunon

J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/jp505746x • Publication Date (Web): 27 Aug 2014 Downloaded from http://pubs.acs.org on August 31, 2014

Just Accepted

"Just Accepted" manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides "Just Accepted" as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. "Just Accepted" manuscripts appear in full in PDF format accompanied by an HTML abstract. "Just Accepted" manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). "Just Accepted" is an optional service offered to authors. Therefore, the "Just Accepted" Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the "Just Accepted" Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these "Just Accepted" manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036

Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

10 11 12

20 21 22

Linking Electrostatic Effects and Protein Motions in Enzymatic Catalysis. A Theoretical Analysis of Catechol O-Methyltransferase

Rafael García-Meseguera, Kirill Zinovjeva, Maite Rocaa, Javier J. Ruiz-

Pernía*b, Iñaki Tuñón*a

a Departament de Química Física, Universitat de Valencia, 46100 Burjassot,

b Departament de Química Física i Analítica, Universitat Jaume I, 12071

Castelló, Spain

* Corresponding Authors:

ignacio.tunon@uv.es pernia@qfa.uji.es

here is the reaction catalyzed by Catechol O-Methyltransferase, a methyl transfer

3 Abstract

5 The role of protein motions in enzymatic catalysis is the subject of a hot scientific

7 debate. We here propose the use of an explicit solvent coordinate to analyze the

g impact of environmental motions during the reaction process. The example analyzed

12 reaction from S-adenosylmethionine (SAM) to the nucleophilic oxygen atom of

14 catecholate. This reaction proceeds from a charged reactant to a neutral product and

16 then a large electrostatic coupling with the environment could be expected. By means

18 of a two-dimensional Free Energy Surface we show that a large fraction of the

1g environmental motions needed to attain the Transition State happens during the first

21 stages of the reaction because most of the environmental motions are slower than

23 changes in the substrate. The incorporation of the solvent coordinate in the definition

25 of the Transition State improves the transmission coefficient and the committor

histogram in solution while the changes are much less significant in the enzyme. The

28 equilibrium solvation approach seems then to work better in the enzyme than in

30 aqueous solution because the enzyme provides a preorganized environment where

32 the reaction takes place.

37 Keywords: QM/MM, Transition State Theory, Reaction Coordinate, Transmission

39 coefficient, Committor, Dynamical Effects

10 11 12

20 21 22

1. Introduction

Prof W. L. Jorgensen has pioneered the development and application of computational methodologies for the calculation of free energy changes associated to chemical transformations taking place in condensed phases. Molecular simulations were used to calculate free energy differences related to solvation,1 ion-pairing,2 acid-base3 and conformational equilibria4 or protein-ligand binding.5,6 Prof. Jorgensen also pioneered the use of free energy profiles to rationalize environmental effects on the rate of chemical reactions with seminal works on the SN2 process in solution.7,8 According to these works, the rate constant diminution observed for the identity Cl- + CH3Cl reaction in aqueous solution relative to the gas-phase can be explained as the consequence of an increase in the activation free energy due to the weaker solvation interactions established at the Transition State (TS). This interpretation relies on the validity of the equilibrium solvation assumption to describe the TS of the chemical reaction, a hypothesis frequently invoked in applications of Transition State Theory (TST).9,10 In this regard, a correct definition of the TS or, equivalently, an adequate selection of the reaction coordinate, can be crucial to ensure the correct use of TST to determine the rate constant of a chemical reaction. In their study, Jorgensen and coworkers employed the antisymmetric combination of the distances of the broken and formed bonds as the reaction coordinate, assuming then that all the solvent degrees of freedom equilibrates at each value of this coordinate. Solvent dynamics is composed by a mixture of different kind of motions (including hindered translations and rotations) and not all of them are fast enough to be considered in equilibrium with the antisymmetric stretching reaction coordinate. The influence on non-equilibrium solvation effects on the reaction rate for the SN2 reaction in water was addressed some years after by Hynes and coworkers11,12 and Warshel and coworkers13 that found that the fate of a trajectory is largely dependent on the solvent configuration at the TS. This effect can be incorporated into TST either by the calculation of the transmission coefficient (k) associated to the reaction coordinate or by changing to a new reaction coordinate incorporating not only solute degrees of freedom but also information about the solvent. Although the transmission coefficient corrections calculated for the Sn2 reaction in water (k ~ 0.5) only have a modest quantitative effect on the reaction

reaction systems strongly coupled to the surroundings

not rigid scaffolds where catalysis takes place but that they must be flexible enough to

3 rate constant, these studies show that solvent participation in the reaction process

5 may be different from that obtained from the equilibrium picture, especially for

9 The role of environmental motions during the chemical reaction is at the heart of the

11 debate in the field of enzymatic chemistry. It is a well-accepted fact that proteins are

14 permit the evolution among different conformations along the catalytic cycle.14,15

16 However, the question of how molecular motions within the protein structure may

18 influence the rate constant of the chemical step has been the subject of many

19 is 36

20 contradictory and sometimes controversial studies during the last years. - Even

21 accepting that the enzymatic active site is preorganized to optimize the electrostatic

23 interaction with the TS,37 changes during the chemical step do not merely affect the

25 chemical system but also the protein. The protein environment is reorganized during

27 the evolution of the system from the reactant state (RS) to the TS, which implies the

28 participation of protein motions in the chemical transformation.21,26,38,39 Different

30 kinds of protein motions participating in the chemical transformation have been

32 identified from molecular simulations27,30,39-41 or invoked to explain experimental

34 observations such as Kinetic Isotope Effects and consequences of mutations on the

rate constant. These motions include both vibrational motions, , taking place in a

37 timescale similar to the barrier crossing event, to very slow conformational

38 29,42

39 changes29,42 that could be considered as frozen during barrier crossing.

The main question to solve with respect to the role of protein motions in enzymatic

43 catalysis is to determine if the equilibrium treatment of protein motions provides a

45 reasonable description for enzymatic reactions34 or if an explicit dynamical treatment

46 33,43

47 is required for at least some of these motions.33,43 It has been recently claimed that the

equilibrium assumption inherent to traditional TST can results in TS ensembles

50 significantly different to those obtained by means of Transition Path Sampling

52 techniques, where this assumption is not made.36 Obviously the validity of the

54 equilibrium approach is intimately linked to the nature of the reaction coordinate

56 chosen to represent the advance of the system along the chemical process. Empirical

57 Valence Bond (EVB) simulations44 typically use the energy gap coordinate, where both

10 11 12

20 21 22

solute/substrate and solvent/protein motions are incorporated. On the other hand, Quantum Mechanics/Molecular Mechanics (QM/MM) simulations that employ semiempirical, Density Functional or ab initio hamiltonians are usually based on the selection of a reaction coordinate defined in terms of geometrical valence coordinates of the solute/substrate, such as bond distances or combinations of them. These last coordinates are usually faster than some environmental motions that could be active during the reaction progress. Even so, values of transmission coefficients of enzymatic reactions calculated for these kind of reaction coordinates usually range between 0.5 and 0.925,45 indicating a minor participation of protein motions during barrier crossing. However, these findings do not discard a more active participation of the environment in other stages of the chemical process.

In order to address the problem of the participation of environmental motions in enzymatic catalysis we proposed to explore the reaction using a reaction coordinate (%(q,s)) based on a combination of solute (q) and solvent (s) degrees of freedom.35 For the solute coordinate a simple valence coordinate can be used, while for the solvent we propose a collective variable expressed as a function of the electrostatic potential created by the environment on key atoms of the chemical system.46 The analysis of the Free Energy Surfaces (FESs) traced as a function of the two coordinates shows that the reaction progress must be expressed as a combination of both coordinates and that the equilibrium approach fails to reproduce the minimum free energy path found on this surface.35 However, for the enzymatic examples analyzed so far, barrier crossing takes place essentially along the solute coordinate and the equilibrium picture provides a reasonable estimation of the TS ensemble and thus also of the activation free energy and the rate constant. 35 In fact, deviations from this equilibrium approach are found to be larger for reaction in aqueous solution because a larger reorganization of the environment is required to reach the TS.35,46

In this work we analyze the performance of the electrostatic potential coordinate to study the role of environmental motions in the case of very strong electrostatic solute-solvent coupling. We present a theoretical analysis of the methyl transfer reaction from 5-adenosylmethionine (SAM) to the nucleophilic hydroxylate oxygen atom of catechol catalyzed by Catechol O-MethylTransferase (COMT, EC 2.1.1.6).47 This

3 reaction (see Figure 1) is a direct bimolecula r Sn2 process that can be formally

5 considered as an inverse Menshutkin reaction where ionic reactants proceed to

neutral products48-54 and then important reorganization effects could take place in the

Calculation of the transmission coefficient and the friction kernel associated with a

8 surroundings of the chemical system. Previous theoretical works on this system

10 showed that the enzymatic active site provides an electrostatic environment more

11 41,55

12 adequate for the reaction than in aqueous solution,41,55 where solute-solvent

14 interactions provide TS destabilization with respect to charged reactant state.

17 simple valence coordinate for the methyl transfer showed a stronger coupling of

19 solvent motions at the TS in solution than in the enzymatic environment.39 However

21 these analyses did not explicitly address the role of protein motions during the whole

24 Kinetic Isotope Effects (KIEs)48,57 in rat liver COMT and the effects of site-specific active

26 site mutations on the rate constant and KIEs in Human COMT32 were interpreted as the

28 signature of the participation of environmental motions in the reaction that could

30 favor the compression of the methyl donor-acceptor distance in the enzyme with

31 respect to the in solution process. However, QM/MM simulations were able to

33 reproduce the inverse KIEs without observing any reduction in the donor-acceptor

35 distance at the enzymatic TS determined within the equilibrium solvation

37 approximation.58 Simulations presented in this paper allow a deeper discussion on the

reaction process. The experimental observation of an inverse secondary a-deuterium

role of environmental motions for a reaction where the energy strongly depend on

40 coupled solute-solvent terms (strong coupling), both in solution and in the enzyme.

42 Analysis of the FESs obtained as a function of the electrostatic (solvent) and valence

44 (solute) coordinates show the participation of protein motions at different stages of

the reaction advance. In general, environmental motions are required to prepare a

47 good electrostatic environment where the reaction can proceed. These motions

49 precede solute motions and barrier crossing takes place mostly along the valence

51 coordinate, especially in the enzymatic reaction. The equilibrium-solvation TS

53 ensemble characterized along the solute coordinate is thus a good approximation to a

56 (environmental) degrees of freedom

more rigorous TS ensemble characterized with explicit consideration of other

10 11 12

20 21 22

Figure 1. (A) Schematic representation of the methylation reaction of catecholate by SAM. (B) Transition State structure obtained for the enzymatic reaction showing the interactions of the substrate with the Mg2+ ion and with Lys144.

2. Methodology

The initial structure for the simulations carried out in this work was taken from our previous works on rat liver Catechol O-MethylTransferase and the counterpart process in aqueous solution.39'41'55'58'59 Briefly, the system was prepared starting from the x-ray structure with Protein Data Bank code 1VID, that contains the ternary complex of the enzyme with the S-adenosylmethionine (SAM) cofactor and an inhibitor 3,5-dinitrocatechol.60 Both nitro groups were removed to prepare the natural substrate which was ionized by proton transfer to neutral Lys144.52 The whole system was placed inside a pre-equilibrated cubic box of water molecules of side 79.5 A, deleting all water molecules with oxygen atoms found at 2.8 A from any non-hydrogen atom of the protein. Two sodium ions where added to the enzymatic system to keep it

3 electroneutral. To study the reaction in solution we followed the same procedure

5 placing the whole cofactor SAM and the substrate in a prequilibrated water box of side

9 We used a QM/MM computational scheme in which S-adenosylmethionine (SAM) and

10 61 11 catecholate (63 atoms), are described using the AM1 semiempirical Hamiltonian. This

55.8 A.

QM level overestimates the free energy barrier of the reaction but provides

MM and QM-MM interactions. While the treatment of long-range electrostatic

14 reasonable geometries for the stationary structures, with a slight underestimation of

16 the distance between the carbon atom of the transferred methyl group to the

18 nucleophilic oxygen atom of cathecholate (O1 in Figure 1).59,62 The rest of the system

20 form the MM subsystem described by means of the AA-OPLS force field for the

21 enzyme63 and a modified version of the TIP3P potential for water molecules.64,65

23 Except when indicated, all the simulations were carried out in the NVT ensemble using

25 the Langevin-Verlet integrator with a time step of 1 fs and a target temperature of 300

27 K. A group-based switched cutoff radius between 12 and 14 A was employed for all

30 QM/MM interactions could be improved using Ewald summation66 we found that the

32 use of an electrostatic cutoff was enough to reproduce the catalytic effect, this is the

34 lowering in the activation free energy when passing from aqueous solution to the

36 enzyme. Thus, this treatment is enough for a comparative analysis between both

40 In this work we obtained FESs using two different coordinates, a solute coordinate (q)

44 W(q, s) = C' — kBTln J p(xN) S(q(xN) — q0)S(s(xN) — s0)dxN (1)

46 where p(xN) is the probability density of finding the system at the configuration xN ,

48 kB the Boltzmann constant and T the absolute temperature In this case, the solute

50 coordinate is the antisymmetric combination of the distances of the SAM sulfur atom

52 and Catecholate oxygen to the carbon atom of the transferred methyl group

environments.

and a solvent coordinate (s). The FES can be expressed as:

(q = d(SC) — d(01C)), see Figure 1. The solvent coordinate was the antisymetric

55 combination of the electrostatic potential created by the environment on sulfur and

57 oxygen atoms:

10 11 12

20 21 22

s = V01(xN)-Vs(xN)=Z<y=

i = 1\x1-

xj~xOl\

i=1 \x,-xs\

where the sums run on the M sites of the environment with point charges Q,. This is a collective coordinate that involves all the environmental atoms with an electrostatic influence on the donor or acceptor atoms. It must be noticed that, by construction, the solute and the solvent coordinates are not orthogonal, as they involve the coordinates of some common atoms (the methyl donor and acceptor). However, we checked that at the TSs, these coordinates are in practice almost orthogonal. The averaged angle between them, determined from the gradients in the mass-weighted Cartesian space, were 90.3 ± 0.5 and 95.4 ± 0.4 degrees, in aqueous solution and in the enzyme, respectively.

The FESs that correspond to the reaction in aqueous solution and in the enzyme were obtained using umbrella sampling,67 applying harmonic biasing potentials to the solute and the solvent coordinates:

V, = \Kq{q - q0)2; Vs = ^(s - s0)2

With these constraints the system preferentially explores the most probable configurations around the reference coordinate values q0 and s0. These values are changed to sample the range of values of interest. The joint probability distributions of the two coordinates were obtained by means of the weighted histogram analysis method (WHAM).68 A total of 2840 simulation windows were employed to trace the FES in aqueous solution, while for the enzyme 1420 windows were needed. The difference comes from the fact that the reaction takes place spanning a wider range of values of the solvent coordinate in aqueous solution than in the enzyme, as shown below. Each simulation window consisted in 5 ps of equilibration and 45 ps of production. The force constants used to keep the system at the reference values of the solute and solvent coordinates were 2500 kJ-mol"1 A-2 and 0.01 kJ_1-mol-|e|2, respectively. These values were shown to provide a good control of the coordinates.46

The TS ensembles for the reaction in solution (TSw) and in the enzyme (TSe) were characterized on a mono-dimensional surface traced as a function of the solute coordinate (TSw(q) and TSe(q)) and on the two-dimensional surface, function of the

3 solute and solvent coordinates (TSw(q,s) and TSe(q,s)). The approximate positions of

5 the TS ensembles were first estimated from the corresponding FESs. Then, 100 ps long

QM/MM simulations were performed with a biasing potential added to restrain the

fTS = H~1(p~1C~1 - (f(R))b -K-n (4)

where f3=1/kBT, H is the 2x2 Hessian matrix associated to the 2D-FES, C the

8 system around the saddle point. Assuming that the region of the FES accessible to such

10 a simulation can be well described within the harmonic approximation, we calculated

12 the precise position of the saddle point (^TS) on the FESs as:

17 covariance matrix, K the diagonal matrix of force constants, (%(R))b the average over

19 the biased simulation and the reference position of the bias. The Hessian matrix is

21 obtained as:

26 the simulation to sample the TS ensemble can be constructed

H = P1C1-K (5)

Once the Hessian matrices and TS positions are known a biasing potential restraining

29 VTS(f(R))=^-((m — fTS)-V-y (6)

32 where KTS is the force constant, that must be as large as possible to keep the system

close to the TS, and v_ is the eigenvector of the Hessian that corresponds to the

35 negative eigenvalue. With this bias, sets of 100 TS structures were obtained from MD

37 simulations, saving those structures were VTS < 0.5-^bT and that were separated by at

39 least 1 ps of simulation. From each TS structure 50 trajectories with random initial

velocities taken from a Maxwell-Boltzmann distribution at T=300K were integrated

42 both forward and backward in time in the NVE ensemble using the Velocity-Verlet

44 integrator to obtain the probability to commit to the reactants and products basins.

3. Results

49 Free Energy Surfaces

52 The AM1/MM FESs corresponding to the methyl transfer in aqueous solution and in

53 the active site of COMT as a function of the solute (q) and solvent (s) coordinates are

55 presented in Figure 2. The solute coordinate evolves from negative values at the

57 reactant state (when the methyl group is bonded to the sulfur atom of the cofactor) to

10 11 12

20 21 22

positive values at the products state (when the methyl group is bonded to the nucleophile oxygen atom of catecholate). The solvent coordinate changes from positive values at the reactant state to smaller values at the product state. The electrostatic potential created by the environment at the reactant state on the nucleophile atom of catecholate is positive while the potential on the sulfur atom of SAM is negative as a consequence of the charge held by these atoms. The electrostatic potential on these two atoms diminishes at the product state where charge separation has been annihilated. Average reactive trajectories initiated at the TS have been projected on the FESs and represented with a continuous line. The position of the TS ensembles on the two-dimensional surfaces (TS(q,s)) are represented by means of a black dashed line, while the TSs obtained in the mono-dimensional or equilibrium solvation description (TS(q)) corresponds to the vertical red dashed line. It should be noticed that the average trajectory does not necessarily cross exactly at the saddle point on the FES because of the asymmetries of the free energy landscape in the directions orthogonal to the reaction advance.

Both FES show a saddle point at very similar values of the solute and solvent coordinates. In the enzyme this saddle point is located at qj5 = 0.13 A andsj5 = 67 kcal ■ mol_1 ■ |e|_1; while in aqueous solutions the values found for the two coordinates are q^5 = 0.08 A and s^5 = 78 kcal ■ mol_1 ■ |e|_1. Larger differences appear at the reactant region. In aqueous solution the FES does not display a minimum at the reactant region, at least for the range of values explored for the solute coordinate. Using a very similar computational approach a solvent-separated ion pair was located for the catecholate-trimethylsulfonium ion system in aqueous solution at a solute coordinate of about -4.2 Â.55 Because our primary interest in this work is the description of the TS we did not extend the exploration of the FES in solution up to larger values (in absolute terms) of the solute coordinate. Instead, the enzymatic FES shows a free energy minimum for a contact distance between the cofactor and the substrate, with q;?5 « -1.1 A. Interestingly, while the solvent coordinate takes very similar values at the TS in solution and in the enzyme, large differences appear at the reactant regions. In the enzyme, the value of the solvent coordinate in this region

3 is SeS ~ 85 kcal ■ mol_1 ■ lei-1 , close to the value found for the TS. However, the

5 value of the solvent coordinate in solution is s^S > 140 kcal ■ mol 1 ■ lei 1.

The different behavior of the environment in the catalyzed and non-catalyzed reaction

9 illustrates the concept of enzyme preorganization:37,70 the electrostatic properties of

11 the active site at the Michaelis complex of COMT are already close to those needed to

reach the reaction TS. When going from the reactant complex to the TS in aqueous

14 solution, the solvent environment must suffer an important reorganization around the

16 two reacting fragments (SAM and catecholate). This reorganization can be quantified

18 from the differences in the solvent coordinate between the TS and the reactant state

20 &Si. In aqueous solution this magnitude is Asw > 60 kcal ■ mol-1 ■ lei-1. Instead, in

22 the enzyme, the reorganization is much smaller Ase~ 18 kcal ■ mol-1 ■ |e|-1(note the

23 different scales used for the solvent coordinates in Figure 2 for the enzyme and the

25 solution reaction). This larger reorganization implies a free energy penalty reflected in

27 the larger barrier for the reaction in solution. An interesting feature derived from

both FESs is that the environmental reorganization, the changes in the solvent

30 coordinate, mostly precedes the changes taking place along the solute coordinate.

32 According to the average reactive trajectories presented in Figure 2 (depicted as the

34 black continuous line on the FESs) most of the solvent reorganization takes place in the

first stages of the reaction. The timing for the evolution of the solute and solvent

37 coordinates is clearer for the reaction in aqueous solution where the change in the

39 solvent coordinate is substantially larger than in the enzyme.

It is interesting to stress the role played by the electrostatic fluctuations of the

43 environment to reduce the free energy barrier. According to the FESs presented in

45 Figure 2 the reaction would present a larger free energy barrier in a frozen

47 environment, especially in aqueous solution. Motions of the solvent/protein facilitate

49 the reaction providing a more adequate electrostatic environment for the progress of

the reaction, reducing the free energy barrier. However, the largest fraction of this

52 barrier is still associated to changes along the solute coordinate. Then, the chemical

54 process should not be described as environmental motions or vibrations leading to a

56 barrierless methyl transfer scenario. The reaction clearly involves fluctuations of the

10 11 12

20 21 22

solute and the solvent degrees of freedom, but the former are in fact the determining step to cross the barrier.

Finally, it is interesting to comment that the qualitative picture obtained from the analysis of the evolution of the chemical process in terms of the solute and the solvent coordinates and the differences observed between the catalyzed reaction and the counterpart process in aqueous solution are quite similar to those obtained for other enzymatic reactions: Chalcone Isomerase,46 Haloalkane Dehalogenase35 and Dihydrofolate Reductase.71 In terms of the electrostatic potential, the environmental reorganization is drastically reduced in the enzyme with respect to the uncatalyzed reaction, which is translated into a small participation of the solvent in the reaction coordinate. This picture was useful to interpret the experimental kinetic results of

isotopically substituted versions of Dihydrofolate Reductase.

Figure 2. Free Energy Surfaces (isocontour lines are drawn each 2 kcal-mol- ) corresponding to the enzymatic reaction (left) and in solution (right). Averaged reactive trajectories are shown as a continuous line. TS surfaces obtained using the solute or the solute and the solvent coordinates are indicated by dashed lines in black and red, respectively.

3 Characterization of the Solvent Coordinate

6 We have characterized the average frequencies of the solute and solvent coordinates

7 at the TS from the free energy landscape and effective masses derived from the

9 equipartition principle. The average frequencies for the solvent coordinate in aqueous

11 solution and in the enzyme are quite similar 235 and 225 cm , respectively. These

values are considerably smaller than the average frequency associated to the solute

14 coordinate, about 750 and 770 cm-1 in solution and in the enzyme. Thus, motion along

16 the solvent coordinate is, in average, slower than the motion along the solute

18 coordinate. As described above, in an averaged reactive trajectory solvent motions

20 dominate during the first stages. However, one must take into account that the

21 electrostatic potential is a collective coordinate that involves many fundamental

23 motions or fluctuations of the environment: from fast stretching motions to slow

25 conformational or rotational motions and then time-separation from solute coordinate

27 motions is not complete. In the case of the enzyme the positively charged Lys144 and

28 the Mg2+ ion contribute significantly to the electrostatic potential coordinate and thus

30 their motions are reflected in the spectrum, while in aqueous solution first solvation

32 shell water molecules contribute more significantly.

34 Figure 3 shows the autocorrelation function of the solvent coordinate ((8s(t) ■

36 5s(0))) computed from 50 ps TS trajectories in solution and in COMT where the solute

38 coordinate was kept frozen at q = qTS using a Lagrange-multiplier based SHAKE

40 algorithm. The time evolution of the autocorrelation function shows a slow decay in

the picosecond timescale and also a fine structure reflecting the participation of faster

43 components. Interestingly, a slower decay is observed in solution than in the enzyme.

45 While, in principle, much slower motions could happen in the enzyme (associated for

47 example to global conformational changes) than in solution, the participation of slow

49 motions in the electrostatic coordinate seems to be more important in solution.

51 The Fourier spectra of the solvent coordinate in both media is also presented in Figure

53 3. The electrostatic coordinate is the result of different motions but most of them are

slower than the solute reaction coordinate. In solution we observed a small fast

56 contribution due to the O-H stretching of water molecules hydrogen bonded to the

58 nucleophile oxygen atom. A larger signal is observed in the enzyme assigned to the N-

10 11 12

20 21 22

H stretching of Lys144, a positively charged residue that is hydrogen bonded to the nucleophilic atom (see Figure 1). Because the reaction proceeds with annihilation of the charge on the substrate, one cannot attribute TS stabilization effects (relative to the reactants) to this hydrogen bonding interaction. It is important to remind that catalysis is relative to the process in solution and that rate acceleration can be attained if the TS is less destabilized with respect to the reactants in the enzyme than in solution.56

Some low intensity signals appear in both environments in the region between 14001800 cm-1, corresponding to heavy atoms stretching in the protein and water molecule bending. However, the most intense signals are found below 50 cm-1, which correspond to motions taking place in the picosecond timescale. It must be taken into account that the slower components of the solvent coordinate must precede the motion along the solute coordinate. Only those components of the solvent coordinate presenting higher frequencies could couple to the solute coordinate during the passage through the dividing surface. The enzyme shows more important participation of motions in the 100-500 cm-1 range than the aqueous solution, while in aqueous solution some intense signals appear in the 600-700 cm-1 range, a frequency similar to that of the solute coordinate. Hydrogen bonds and other electrostatic interactions between the environment and the reacting system appear in these regions of the spectra.

Time (ps)

2000 Frequency (cm-1)

Figure 3. Autocorrelation function (up) and Fourier transform (down) determined from simulations at q=qTS for the system in aqueous solution (red line) and in COMT (black line). The insert focus on the lower frequencies region of the spectra.

Characterization of TS ensembles

The TS ensembles defined using only the solute coordinate (TS(q)) or on the 2D-surface defined by the solute and the solvent coordinates (TS(q,s)) are different, as observed in Figure 2. In the 2D-picture, the explicit consideration of the solvent coordinate produces a rotation of the TS plane. However, it must be taken into account that when the TS is defined exclusively in terms of the solute coordinate then the solvent is in equilibrium. Thus TS(q) corresponds to a free energy minimum along the solvent coordinate. Then, except for cases with a very large participation of the solvent coordinate in the barrier crossing and/or with very small curvatures around the saddle point, the average description obtained for the two TS ensembles (TS(q) and TS(q,s)) can be quite similar. From a dynamical perspective, the slower components of the solvent reorganization precede changes in the faster solute coordinate, providing an equilibrium-like solvation at the TS. This similarity between both TS ensembles is in

10 11 12

20 21 22

fact observed for catecholate methylation in aqueous solution and in COMT. Geometrical parameters given in Table 1 are statistically indistinguishable. The averaged geometrical parameters corresponding to the broken and formed bonds, the donor-acceptor distance and the attacking angle are similar, with differences below the standard deviations. Slightly larger changes are observed between the TS ensembles in aqueous solution, in particular for the methyl attacking angle (S-C-O1), although the dispersion associated to the structures of the TS ensembles prevent any quantitative conclusion from this observation. We also provide some distances characterizing the interaction of the chemical system to the environment. In the enzyme the distance of the hydroxyl oxygen atom of catechol (O2) to the Magnesium ion present in the enzymatic active site and the distance of the nucleophilic oxygen O1 to the ammonium group of Lys144 are equivalent in both TS ensembles. In solution, the average distance of the O1 atom to the closest water molecule (O1-Ow) is also equivalent in both ensembles within their standard deviation.

In passing, it is interesting to note that we did not find any evidence for a significant difference in the distances between the methyl donor and acceptor in the enzymatic TS with respect to the solution one. The experimental observation of a much more inverse secondary a-deuterium KIEs for the COMT catalyzed reaction than for the uncatalyzed methyl transfer was originally interpreted as the consequence of a tighter transition state in the enzymatic reaction, with significantly smaller donor-acceptor distance.49 However, QM/MM calculations based in equilibrium simulations along the solute coordinate reproduced the experimental observation without evidence for compression.58,73 While it could be argued that the sampling resulting from those simulations could be biased to produce similar results irrespectively of the environment, the new simulations presented here show that the averaged distances are still the same, within their standard deviations, in both environments. The explicit inclusion of environmental motions in the definition of the TS ensemble does not affect this conclusion, suggesting that the inverse secondary a-deuterium KIES in COMT32,48,57 can be explained by the preferential equilibrium solvation provided by the environment.58 The origin of the larger increase of the force constants associated to motions of the a-hydrogens at the enzymatic TS is not due to changes in the

interactions between the reacting fragments but in the interactions established with the environments. This finding illustrates the necessity of adequate computational models to assign the physical origin of complex experimental observations. If the environment is neglected the more inverse KIE must result from the compression of the donor-acceptor distance but, if the environment is included, alternative (and correct) explanations arise.

It is also interesting to compare the TS geometries in solution and in the enzyme in the context of the Hammond postulate.74 In our simulations a smaller barrier in the enzyme corresponds to a more advanced TS in terms of methyl-donor and methyl-acceptor distances. This behavior can be rationalized considering that the polarity of the solute decreases as the reaction advances and thus catalysis is the result of a smaller destabilization of the TS in the enzyme than in solution. This scenario for enzymatic reactions was discussed in ref 56.

Table 1. Average geometrical parameters (distances in A, angles in degrees) and transmission coefficients (k) obtained for the TS ensembles characterized using the solute (q) or the solute and solvent coordinates (q,s)

Aqueous Solution COMT

TSw(q) TSw(q,s) TSe(q) TSe(q,s)

D(S-C) 2.12 ±0.04 2.12 ±0.04 2.17 ±0.05 2.16 ±0.05

D(C-O1) 2.03 ±0.04 2.03 ±0.05 1.99 ±0.05 1.98 ±0.04

D(S-O1) 4.14 ±0.08 4.11 ±0.08 4.12 ±0.09 4.11 ±0.08

A(S-C-O1) 172 ±4 167 ±6 166 ±5 166 ±6

D(Mg-O2) — — 2.23 ±0.09 2.24 ±0.10

D(O1-NsLys144) — — 2.74 ±0.09 2.75 ±0.09

D(O1-Ow) 2.76 ±0.07 2.73 ±0.08 — —

K 0.49 ±0.05 0.70 ±0.05 0.76 ±0.05 0.72 ±0.05

The TS ensembles were also characterized analyzing a set of 5000 trajectories initiated from TS structures selected for each of the four ensembles (50 trajectories started from 100 TS structures for each ensemble, using different velocities taken from a

10 11 12

20 21 22

Maxwell-Boltzmann distribution). From these trajectories we computed the transmission coefficient as:11

where it has been assumed that all the trajectories have the same probability because they were obtained from an equilibrium distribution. Vi is the initial velocity associated to the reaction coordinate and Qi is equal to 1 for reactant to product trajectories, 0 for reactant to reactant or product to product and -1 for product to reactant trajectories. The values of the transmission coefficient for the four TS ensembles and their standard deviation are provided in Table 1. Standard deviations where obtained according to ref 12 for 100 independent trajectories (the number of initial structures) The values of the transmission coefficient are similar to those obtained in previous works of the same system using slightly different simulation conditions.39,41 It is interesting to note that the transmission coefficients calculated from the enzymatic TS ensembles optimized along one (TSe(q)) and two (TSe(q,s)) coordinates are practically indistinguishable, while the transmission coefficient in water increases when passing from the TSw(q) to the TSw(q,s). This effect illustrates the fact that the participation of solvent motions during barrier crossing is significantly more important in aqueous solution than in the enzyme. When solvent motions are incorporated into the definition of the TS ensemble through the electrostatic coordinate, the transmission coefficients are very similar for the catalyzed and the uncatalyzed reaction. This coincidence runs against the proposals about a larger role of dynamical contributions during barrier crossing in enzymatic reactions. Note that the transmission coefficient found for the TS(q,s) ensembles is not unity. First, the existence of a no-recrossing TS hypersurface is not guaranteed even in simple systems, because recrossings could emerge not only from a non-optimal choice of the reaction coordinate but also from non-linear dynamical effects. 75,76 Second, we have optimized the TS in one and two coordinates, without considering the complete multidimensional configuration space. Other solute and solvent coordinates could have noticeable participation in the accurate determination of the TS ensemble. For example, the methyl inversion was

shown to be strongly coupled to the antisymmetric coordinate.

3 A more stringent test about the quality of the TS ensembles is provided by the analysis

5 of the committor histograms. These histograms, presented in Figure 4, provide the

probability that free trajectories initiated from a given TS structure reach the products

8 basin before the reactants one. For each of the 100 structures selected for the four TS

10 ensembles given in Table 1, the probability of ending in products for each structure

12 was determined from the set of 50 trajectories. If the TS ensemble is correctly defined,

14 the committor histogram should be a unimodal distribution peaked around 0.5. The

broadness observed for the committor distributions presented in Figure 4 can be due

17 to at least three factors: i) an incomplete optimization of the reaction coordinate (the

19 gradient of the committor function around the TS is usually high); ii) an insufficient

21 sampling in terms of initial structures and/or trajectories and iii) dynamical effects.

23 According to Figure 4 the explicit consideration of the solvent coordinate in the

25 definition of the TS surface in aqueous solution clearly improves the distribution of

27 probabilities to commit to the product basins. Then, a more rigorous picture of the TS

can be obtained when the solvent coordinate is explicitly considered in aqueous

30 solution. The TS ensemble obtained using exclusively the solute coordinate shows a

32 bimodal distribution that indicates the presence of a small barrier along the solvent

34 coordinate. This barrier can be estimated from the ratio of the transmission

coefficients obtained including or not the solvent coordinate. This ratio equals to 1.4

37 and can be translated into a free energy difference of only 0.2 kcal-mol-1 at 298K,

39 below the expected statistical error in the evaluation of activation free energies.

In the enzyme, the consideration of the solvent coordinate in the TS definition

43 moderately improves the committor histogram. This result correlates with the changes

45 observed in the transmission coefficients given in Table 1, which are almost equivalent

47 in the two enzymatic TS ensembles. Participation of environmental motions in the

reaction coordinate at the TS seems to be less important than in solution. The larger

50 reorganization needed in solution to reach the TS, reflected in the larger variation

52 observed for the electrostatic coordinate along the reaction (see Figure 2), results in a

54 larger coupling between both coordinates in this medium. It must be emphasized that

56 the solvent coordinate is composed of both fast and slow motions and only those

motions presenting timescales similar to those of the solute coordinate can couple

10 11 12

20 21 22

during barrier crossing. Slower motions must precede the changes in the solute coordinate, which is reflected in the average reactive trajectory presented in Figure 2.

c 0.15

O" <u 0.10

TSe(q)

|°.l5 3

£0.10 u.

0.05 0.00

0.6 0.8 p(P)

0.6 0.8 p(P)

0.25 0.20

= 0.15

0.05 0.00

0.25 0.20

= 0.15 '0.10 0.05 0.00

0.6 0.8 p(P)

0.6 0.8 p(P)

Figure 4. Commitor histograms obtained for the four TS ensembles analyzed from the probabilities that trajectories initiated from a given trajectory commits to products (p(P)). The graphs on the left correspond to the enzymatic reaction and those on the right to the aqueous solution reaction (mono-and two-dimensional definitions up and down, respectively).

4. Conclusions

We have presented a QM/MM computational analysis of the reaction of catechol O-methylation in aqueous solution and in the active site of COMT. FESs have been obtained as a function of a solute coordinate (the antisymmetric combination of the distances of the methyl group to the donor and the acceptor atoms) and a solvent coordinate (defined as the electrostatic potential created by the environment on the methyl donor and acceptor atoms).

From the analysis of the FESs, shown in Figure 2, and the corresponding TS ensembles we can conclude that the participation of environmental motions in the reaction coordinate at the TS is more important for the reaction in solution than in the enzyme

!u0.10

3 because a larger reorganization of the environment is needed in solution than in the

5 enzyme. The enzymatic active site is largely prepared to accommodate the reaction at

the Michaelis complex. When analyzing the changes taking place in the environment

8 during the chemical process it can be observed that most of the environmental

10 motions participating in the electrostatic reorganization are slower than the solute

12 reaction coordinate. Then a large fraction of the changes in the solvent coordinate

14 takes place during the first stages of the process. However, some of these motions can

17 crossing. This is reflected in a slight rotation of the TS surface obtained in the two-

19 dimensional description with respect to the equilibrium treatment (see Figure 2).

21 However, our results clearly show that the use of the equilibrium approximation

happen in similar time scales and can couple to the solute coordinate during barrier

results in a very good definition of the TS ensemble in both environments and that a

24 non-equilibrium treatment of the environment does not significantly change the TS

26 ensembles (see the average geometrical parameters given in Table 1).

The effect of an explicit consideration of a solvent coordinate in the definition of the

histogram obtained using only the solute coordinate shows that a small free energy

30 TS ensemble can be appreciated calculating the transmission coefficients and the

32 committor histograms. The inclusion of the solvent coordinate improves the

34 transmission coefficient in aqueous solution by a factor of 1.4. The committor

37 barrier is present along the solvent coordinate (about 0.2 kcal-mol-1 at 298 K). The

39 changes observed in the transmission coefficients and committor histograms

41 calculated for the enzymatic reaction are significantly less significant than in aqueous

43 solution. This observation agrees with a smaller participation of environmental

motions in the catalyzed reaction than in the uncatalyzed one. The equilibrium

46 solvation approach works even better for the enzymatic reaction than for the reaction

48 in solution.

3 5. Acknowledgements

6 The authors gratefully acknowledge financial support from FEDER funds and the

Ministerio de Economía y Competitividad (project CTQ2012-36253-C03) and

9 Generalitat Valencia project GV/2012/053. R. G-M and K. Z. acknowledge a fellowship

11 from Ministerio de Educación (FPI and FPU, respectively). The authors acknowledge

13 computational facilities of the Servei d'Informatica de la Universitat de Valencia in the

15 Tirant supercomputer.

20 21 22

3 References

5 (1) Jorgensen, W. L.; Ravimohan, C. Monte Carlo Simulation of Differences in Free

6 Energies of Hydration. J. Chem. Phys. 1985, 83, 3050-3054.

7 (2) Jorgensen, W. L.; Buckner, J. K.; Huston, S. E.; Rossky, P. J. Hydration and

8 Energetics for Tert-Butyl Chloride Ion Pairs in Aqueous Solution. J. Am. Chem. Soc. 1987, 109,

9 1891-1899.

11 (3) Gushurst, A. J.; Jorgensen, W. L. Computer-Assisted Mechanistic Evaluation of

12 Organic Reactions. 12. pKa Predictions for Organic Compounds in Me2SO. J. Org. Chem. 1986, ,|3 51, 3513-3522.

14 (4) Jorgensen, W. L. Monte Carlo Simulation of n-Butane in Water. Conformational

15 Evidence for the Hydrophobic Effect. J. Chem. Phys. 1982, 77, 5757-5765.

16 (5) Smith, R. H.; Jorgensen, W. L.; Tirado-Rives, J.; Lamb, M. L.; Janssen, P. A. J.;

17 Michejda, C. J.; Kroeger Smith, M. B. Prediction of Binding Affinities for TIBO Inhibitors of HIV-1

18 Reverse Transcriptase Using Monte Carlo Simulations in a Linear Response Method. J. Med.

19 Chem. 1998, 41, 5272-5286.

20 (6) Jorgensen, W. L. Rusting of the Lock and Key Model for Protein-Ligand Binding.

21 Science 1991, 254, 954-955.

(7) Chandrasekhar, J.; Smith, S. F.; Jorgensen, W. L. SN2 Reaction Profiles in the

23 Gas Phase and Aqueous Solution. J. Am. Chem. Soc. 1984, 106, 3049-3050.

25 (8) Chandrasekhar, J.; Smith, S. F.; Jorgensen, W. L. Theoretical Examination of the

26 SN2 Reaction Involving Chloride Ion and Methyl Chloride in the Gas Phase and Aqueous

27 Solution. J. Am. Chem. Soc. 1985, 107, 154-163.

28 (9) Garrett, B. C.; Truhlar, D. G. Criterion of Minimum State Density in the

29 Transition State Theory of Bimolecular Reactions. J. Chem. Phys. 1979, 70, 1593-1598.

30 (10) Eyring, H. The Theory of Absolute Reaction Rates. J. Chem. Soc. Faraday Trans.

31 1938, 34, 41-48.

32 (11) Bergsma, J. P.; Gertner, B. J.; Wilson, K. R.; Hynes, J. T. Molecular Dynamics of a

33 Model SN2 Reaction in Water. J. Chem. Phys. 1987, 86, 1356-1376.

34 (12) Gertner, B. J.; Wilson, K. R.; Hynes, J. T. Nonequilibrium Solvation Effects on

35 Reaction Rates for Model SN2 Reactions in Water. J. Chem. Phys. 1989, 90, 3537-3558.

37 (13) Hwang, J. K.; King, G.; Creighton, S.; Warshel, A. Simulation of Free Energy

OQ Relationships and Dynamics of SN2 Reactions in Aqueous Solution. J. Am. Chem. Soc. 1988,

39 110, 5297-5311.

40 (14) Hammes, G. G.; Benkovic, S. J.; Hammes-Schiffer, S. Flexibility, Diversity, and

41 Cooperativity: Pillars of Enzyme Catalysis. Biochemistry 2011, 50, 10422-10430.

42 (15) Henzler-Wildman, K. A.; Lei, M.; Thai, V.; Kerns, S. J.; Karplus, M.; Kern, D. A

43 Hierarchy of Timescales in Protein Dynamics is Linked to Enzyme Catalysis. Nature 2007, 450,

44 913-916.

45 (16) Warshel, A. Dynamics of Enzymatic Reactions. Proc. Natl. A. Sci. USA 1984, 81,

46 444-448.

47 (17) Cameron, C. E.; Benkovic, S. J. Evidence for a Functional Role of the Dynamics

48 of Glycine-121 of Escherichia coli Dihydrofolate Reductase Obtained from Kinetic Analysis of a

50 Site-Directed Mutantt. Biochemistry 1997, 36, 15792-15800.

c. (18) Radkiewicz, J. L.; Brooks, C. L. Protein Dynamics in Enzymatic Catalysis:

52 Exploration of Dihydrofolate Reductase. J. Am. Chem. Soc. 1999, 122, 225-231.

53 (19) Antoniou, D.; Schwartz, S. D. Internal Enzyme Motions as a Source of Catalytic

54 Activity: Rate-Promoting Vibrations and Hydrogen Tunneling. J. Phys. Chem. B 2001, 105, 555355 5558.

56 (20) Benkovic, S. J.; Hammes-Schiffer, S. A Perspective on Enzyme Catalysis. Science

57 2003, 301, 1196-1202.

10 11 12

20 21 22

(21) Garcia-Viloca, M.; Gao, J.; Karplus, M.; Truhlar, D. G. How Enzymes Work: Analysis by Modern Rate Theory and Computer Simulations. Science 2004, 303, 186-195.

(22) Antoniou, D.; Basner, J.; Núñez, S.; Schwartz, S. D. Computational and Theoretical Methods to Explore the Relation between Enzyme Dynamics and Catalysis. Chem. Rev. 2006, 106, 3170-3187.

(23) Gao, J.; Ma, S.; Major, D. T.; Nam, K.; Pu, J.; Truhlar, D. G. Mechanisms and Free Energies of Enzymatic Reactions. Chem. Rev. 2006, 106, 3188-3209.

(24) Olsson, M. H. M.; Parson, W. W.; Warshel, A. Dynamical Contributions to Enzyme Catalysis: Critical Tests of A Popular Hypothesis. Chem. Rev. 2006, 106, 1737-1756.

(25) Pu, J.; Gao, J.; Truhlar, D. G. Multidimensional Tunneling, Recrossing, and the Transmission Coefficient for Enzymatic Reactions. Chem. Rev. 2006, 106, 3140-3169.

(26) Kamerlin, S. C. L.; Warshel, A. At the Dawn of the 21st Century: Is Dynamics the Missing link for Understanding Enzyme Catalysis? Proteins 2010, 78, 1339-1375.

(27) Nashine, V. C.; Hammes-Schiffer, S.; Benkovic, S. J. Coupled Motions in Enzyme Catalysis. Curr. Opin. Chem. Biol. 2010, 14, 644-651.

(28) Adamczyk, A. J.; Cao, J.; Kamerlin, S. C. L.; Warshel, A. Catalysis by Dihydrofolate Reductase and Other Enzymes Arises from Electrostatic Preorganization, not Conformational Motions. Proc. Natl. A. Sci. USA 2011, 108, 14115-14120.

(29) Bhabha, G.; Lee, J.; Ekiert, D. C.; Gam, J.; Wilson, I. A.; Dyson, H. J.; Benkovic, S. J.; Wright, P. E. A Dynamic Knockout Reveals That Conformational Fluctuations Influence the Chemical Step of Enzyme Catalysis. Science 2011, 332, 234-238.

(30) Boekelheide, N.; Salomón-Ferrer, R.; Miller, T. F. Dynamics and Dissipation in Enzyme Catalysis. Proc. Natl. A. Sci. USA 2011, 108, 16159-16163.

(31) Ramanathan, A.; Agarwal, P. K. Evolutionarily Conserved Linkage between Enzyme Fold, Flexibility, and Catalysis. PLoS Biol 2011, 9, e1001193.

(32) Zhang, J.; Klinman, J. P. Enzymatic Methyl Transfer: Role of an Active Site Residue in Generating Active Site Compaction That Correlates with Catalytic Efficiency. J. Am. Chem. Soc. 2011, 133, 17134-17137.

(33) Hay, S.; Scrutton, N. S. Good Vibrations in Enzyme-Catalysed Reactions. Nat. Chem. 2012, 4, 161-168.

(34) Glowacki, D. R.; Harvey, J. N.; Mulholland, A. J. Taking Ockham's Razor to Enzyme Dynamics and Catalysis. Nat. Chem. 2012, 4, 169-176.

(35) Garcia-Meseguer, R.; Marti, S.; Ruiz-Pernia, J. J.; Moliner, V.; Tuñon, I. Studying the Role of Protein Dynamics in an SN2 Enzyme Reaction Using Free-Energy Surfaces and Solvent Coordinates. Nat. Chem. 2013, 5, 566-571.

(36) Doron, D.; Kohen, A.; Nam, K.; Major, D. T. How Accurate Are Transition States from Simulations of Enzymatic Reactions? J. Chem. Theory Comput. 2014, 10, 1863-1871.

(37) Warshel, A.; Sharma, P. K.; Kato, M.; Xiang, Y.; Liu, H. B.; Olsson, M. H. M. Electrostatic Basis for Enzyme Catalysis. Chem. Rev. 2006, 106, 3210-3235.

(38) Karplus, M.; McCammon, J. A. Dynamics of Proteins: Elements and Function. Annu. Rev. Biochem. 1983, 52, 263-300.

(39) Roca, M.; Moliner, V.; Tuñón, I.; Hynes, J. T. Coupling Between Protein and Reaction Dynamics in Enzymatic Processes: Application of Grote-Hynes Theory to Catechol O-Methyltransferase. J. Am. Chem. Soc. 2006, 128, 6186-6193.

(40) Garcia-Viloca, M.; Truhlar, D. G.; Gao, J. Reaction-Path Energetics and Kinetics of the Hydride Transfer Reaction Catalyzed by Dihydrofolate Reductaset. Biochemistry 2003, 42,13558-13575.

(41) Roca, M.; Andrés, J.; Moliner, V.; Tuñón, I.; Bertrán, J. On the Nature of the Transition State in Catechol O-Methyltransferase. A Complementary Study Based on Molecular Dynamics and Potential Energy Surface Explorations. J. Am. Chem. Soc. 2005, 127, 1064810655.

(45) Ruiz-Pernia, J. J.; Tunon, I.; Moliner, V.; Hynes, J. T.; Roca, M. Dynamic effects on Reaction Rates in a Michael Addition Catalyzed by Chalcone Isomerase. Beyond the Frozen

3 (42) Whittier, S. K.; Hengge, A. C.; Loria, J. P. Conformational Motions Regulate

4 Phosphoryl Transfer in Related Protein Tyrosine Phosphatases. Science 2013, 341, 899-903.

5 (43) Kosugi, T.; Hayashi, S. Crucial Role of Protein Flexibility in Formation of a Stable

6 Reaction Transition State in an a-Amylase Catalysis. J. Am. Chem. Soc. 2012, 134, 7045-7055.

7 (44) Warshel, A.; Weiss, R. M. An Empirical Valence Bond Approach for Comparing

8 Reactions in Solutions and in Enzymes. J. Am. Chem. Soc. 1980, 102, 6218-6226.

12 Environment Approach. J. Am. Chem. Soc. 2008, 130, 7477-7488.

13 (46) Ruiz-Pernía, J. J.; Martí, S.; Moliner, V.; Tuñón, I. A Novel Strategy to Study

14 Electrostatic Effects in Chemical Reactions: Differences between the Role of Solvent and the

15 Active Site of Chalcone Isomerase in a Michael Addition. J. Chem. Theory Comput. 2012, 8,

16 1532-1535.

17 (47) Takusagawa, F.; Fujioka, M.; Spies, A.; Schowen, R. L. In Comprehensive

18 Biological Catalyzis; Sinnott, M., Ed.; Academic Press: San Diego, CA,: 1998; Vol. 1, p 1-30.

19 (48) Hegazi, M. F.; Borchardt, R. T.; Schowen, R. L. .alpha.-Deuterium and Carbon-

20 13 Isotope Effects for Methyl Transfer Catalyzed by Catechol O-Methyltransferase. SN2-Like

21 Transition State. J. Am. Chem. Soc. 1979, 101, 4359-4365.

23 (49) Rodgers, J.; Femec, D. A.; Schowen, R. L. Isotopic Mapping of Transition-State Structural Features Associated With Enzymic Catalysis of Methyl Transfer. J. Am. Chem. Soc.

24 1982, 104, 3263-3268.

26 (50) Woodard, R. W.; Tsai, M. D.; Floss, H. G.; Crooks, P. A.; Coward, J. K.

27 Stereochemical Course of the Transmethylation Catalyzed by Catechol O-Methyltransferase. J.

28 Biol. Chem. 1980, 255, 9124-9127.

29 (51) Knipe, J. O.; Vasquez, P. J.; Coward, J. K. Methylase Models: Studies on

30 General-Base vs. Nucleophilic Catalysis in the Intramolecular Alkylation of Phenols. J. Am.

31 Chem. Soc. 1982, 104, 3202-3209.

32 (52) Zheng, Y.; Bruice, T. C. A Theoretical Examination of the Factors Controlling the

33 Catalytic Efficiency of a Transmethylation Enzyme: Catechol O-Methyltransferase. J. Am.

34 Chem. Soc. 1997, 119, 8137-8145.

35 (53) Kahn, K.; Bruice, T. C. Transition-State and Ground-State Structures and Their

Interaction with the Active-Site Residues in Catechol O-Methyltransferase. J. Am. Chem. Soc. 1999, 122, 46-51.

(54) Lau, E. Y.; Bruice, T. C. Comparison of the Dynamics for Ground-State and

40 Transition-State Structures in the Active Site of Catechol O-Methyltransferase. J. Am. Chem.

41 Soc. 2000, 122, 7165-7171.

42 (55) Roca, M.; Martí, S.; Andrés, J.; Moliner, V.; Tuñón, I.; Bertrán, J.; Williams, I. H.

43 Theoretical Modeling of Enzyme Catalytic Power: Analysis of "Cratic" and Electrostatic Factors

44 in Catechol O-Methyltransferase. J. Am. Chem. Soc. 2003, 125, 7726-7737.

45 (56) Tuñón, I.; Hynes, J. T. A Simple Model for Barrier Frequencies for Enzymatic

46 Reactions. Chem. Phys. Chem. 2011, 12, 184-190.

47 (57) Gray, C. H.; Coward, J. K.; Schowen, K. B.; Schowen, R. L. .alpha.-Deuterium and

48 Carbon-13 Isotope Effects for a Simple, Intermolecular Sulfur-to-Oxygen Methyl-Transfer

49 Reaction. Transition-State Structures and Isotope Effects in Transmethylation and

51 Transalkylation. J. Am. Chem. Soc. 1979, 101, 4351-4358.

52 (58) Ruggiero, G. D.; Williams, I. H.; Roca, M.; Moliner, V.; Tuñón, I. QM/MM

53 Determination of Kinetic Isotope Effects for COMT-Catalyzed Methyl Transfer Does Not

54 Support Compression Hypothesis. J. Am. Chem. Soc. 2004, 126, 8634-8635.

55 (59) Roca, M.; Moliner, V.; Ruiz-Pernía, J. J.; Silla, E.; Tuñón, I. Activation Free

56 Energy of Catechol O-Methyltransferase. Corrections to the Potential of Mean Force. J. Phys.

57 Chem. B 2005, 110, 503-509.

3 (60) Vidgren, J.; Svensson, L. A.; Liljas, A. Crystal Structure of Catechol O-

4 Methyltransferase. Nature 1994, 368, 354-358.

5 (61) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. Development and

6 Use of Quantum Mechanical Molecular Models. 76. AMI: a New General Purpose Quantum

7 Mechanical Molecular Model. J. Am. Chem. Soc. 1985, 107, 3902-3909.

8 (62) Ruiz-Pernia, J. J.; Silla, E.; Tunon, I.; Marti, S. Hybrid Quantum

9 Mechanics/Molecular Mechanics Simulations with Two-Dimensional Interpolated Corrections:

11 Application to Enzymatic Processes. J. Phys. Chem. B 2006, 110, 17663-17670.

12 (63) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. Development and Testing of

13 the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids.

14 J. Am. Chem. Soc. 1996, 118, 11225-11236.

15 (64) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L.

16 Comparison of Simple Potential Functions For Simulating Liquid Water. J. Chem. Phys. 1983,

17 79, 926-935.

18 (65) Neria, E.; Fischer, S.; Karplus, M. Simulation of Activation Free Energies in

19 Molecular Systems. J. Chem. Phys. 1996, 105, 1902-1921.

20 (66) Nam, K.; Gao, J.; York, D. M. An Efficient Linear-Scaling Ewald Method for Long-

21 Range Electrostatic Interactions in Combined QM/MM Calculations. J. Chem. Theory Comput.

22 2004, 1, 2-13.

24 (67) Torrie, G. M.; Valleau, J. P. Nonphysical Sampling Distributions in Monte Carlo

25 Free-Energy Estimation: Umbrella Sampling. J. Comput. Phys. 1977, 23, 187 - 199.

26 (68) Kumar, S.; Rosenberg, J. M.; Bouzida, D.; Swendsen, R. H.; Kollman, P. A. The

27 Weighted Histogram Analysis Method for Free-Energy Calculations on Biomolecules. I. The

28 Method. J. Comput. Chem. 1992, 13, 1011-1021.

29 (69) Kästner, J. Umbrella Integration in Two or More Reaction Coordinates. J. Chem.

30 Phys. 2009, 131, 034109.

31 (70) Warshel, A. Electrostatic Origin of the Catalytic Power of Enzymes and the Role

32 of Preorganized Active Sites. J. Biol. Chem. 1998, 273, 27035-27038.

33 (71) Ruiz-Pernia, J. J.; Luk, L. Y. P.; Garcia-Meseguer, R.; Marti, S.; Loveridge, E. J.;

34 Tunon, I.; Moliner, V.; Allemann, R. K. Increased Dynamic Effects in a Catalytically

Compromised Variant of Escherichia coli Dihydrofolate Reductase. J. Am. Chem. Soc. 2013,

36 135, 18689-18696.

38 (72) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. Numerical Integration of the

39 Cartesian Equations of Motion of a System with Constraints: Molecular Dynamics of n-Alkanes.

40 J. Comput. Phys. 1977, 23, 327-341.

41 (73) Kanaan, N.; Ruiz-Pernia, J. J.; Williams, I. H. QM/MM Simulations for Methyl

42 Transfer in Solution and Catalysed by COMT: Ensemble-Averaging of Kinetic Isotope Effects.

43 Chem. Commun. 2008, 6114-6116.

44 (74) Hammond, G. S. A Correlation of Reaction Rates. J. Am. Chem. Soc. 1955, 77,

45 334-338.

46 (75) Mullen, R. G.; Shea, J. E.; Peters, B. Transmission Coefficients, Committors, and

47 Solvent Coordinates in Ion-Pair Dissociation. J. Chem. Theory Comput. 2014, 10, 659-667.

48 (76) Mullen, R. G.; Shea, J. E.; Peters, B. Communication: An Existence Test for

50 Dividing Surfaces Without Recrossing. J. Chem. Phys. 2014, 140, 041104.

ACS Paragon Plus Environment