Mechanistic Insights into the Reaction of Chlorination of Tryptophan Catalyzed by Tryptophan 7-Halogenase

Tryptophan 7-halogenase catalyzes chlorination of free tryptophan to 7-chlorotryptophan, which is the first step in the antibiotic pyrrolnitrin biosynthesis. Many biologically and pharmaceutically active natural products contain chlorine and thus, an understanding of the mechanism of its introduction into organic molecules is important. Whilst enzyme-catalyzed chlorination is accomplished with ease, it remains a difficult task for the chemists. Therefore, utilizing enzymes in the synthesis of chlorinated organic compounds is important, and providing atomistic mechanistic insights about the reaction mechanism of tryptophan 7-halogenase is vital and timely. In this work, we examined a mechanism for the reaction of tryptophan chlorination, performed by tryptophan 7-halogenase, by calculating potential energy and free energy surfaces using two different Combined Quantum Mechanical/Molecular Mechanical (QM/MM) methods both employing Density Functional Theory (DFT) for the QM region. Both computational strategies agree on the nature of the rate-limiting step and provided close results for the reaction barriers of the two reaction steps. The calculations for both the potential energy and the free energy profiles showed very similar geometric features and hydrogen bonding interactions for the characterized stationary points.

. (A) Modified X-ray crystallographic structure of tryptophan 7-halogenase, shown in new cartoon representation. Line-represents the 10 Å distance between FAD (cofactor) binding site and tryptophan (substrate) binding site, HOCl (chlorinating agent) shown in licorice representation; carbon atoms -in cyan, nitrogen atoms -in blue, oxygen atoms -in red, and hydrogen atoms -in white. 1 (B) Zoom in view of the active site, representing the main actors of catalysis -the substrate tryptophan, the chlorinating agent -HOCl and K79 and E346 residues. Coordinates taken from PDB entry 2AR8. transition states and products 12 . However, the development of new tools and methodologies allowed for increased number of QM/MM reaction path free energy studies at ab initio and DFT level of theory for the QM region 13 . Recently, a multiscale study combining hybrid QM/MM-MD, multiple steering MD, and the Minh-Adib estimator reproduced experimental thermodynamics data of Diels-Alder reaction, the activation free energy barrier, and the effect of different solvent environments such as water and methanol 14 was revealed.
To complement the missing, but vitally important, knowledge about the reaction mechanism of the tryptophan 7-halogenase, we applied two different computational approaches in studying the two chemical steps of the enzymatic chlorination of tryptophan. In particular, we focused on QM/MM potential energy and QM/MM free energy reaction paths of the most important part of the catalytic cycle of the proposed reaction mechanism of tryptophan 7-halogenase -the process of tryptophan chlorination by HOCl. We applied two different QM/ MM methods: one to reveal the potential energy reaction path, and another to gain insight about the free energy changes along the reaction; both of them applying Density Functional Theory level for the quantum part of the system. Experimental data suggests that HOCl is formed, and then migrates from its site of formation in the FAD binding site to the substrate's tryptophan binding site where the reaction of chlorination takes place 4 . Exploration of potential energy and free energy profiles were carried out by means of QM/MM potential energy reaction path approach at B3LYP/6-31 + G(d)-CHARMM27 level, and Multiple Steering Molecular Dynamics method using the QM/MM-MD approach through Amber-PUPIL-NWChem (B3LYP/6-31 G) interface 15,16 , respectively. Here, we explored the common trends and specificities for the potential and free energy reaction profiles, the structures of the stationary points and revealed key interactions between the reactants and the enzyme active site, which may contribute to the catalytic efficiency of the process.

Methods
Choice of X-ray crystal structure for modelling. A number of X-ray crystallographic structures of tryptophan 7-halogenase are available in the Protein Data Bank (PDB) 17 . The X-ray crystallographic structure with PDB ID: 2AR8 4 isolated from Pseudomonas fluorescens with a resolution of 2.20 Å, was chosen for modelling. The structure contains the following components -a chloride ion (Cl − ), a flavin adenine dinucleotide (FAD) cofactor and the reaction product, 7-chlorotryptophan. This structure was chosen for modelling because it is the only structure that contains 7-chlorotryptophan, which was used to model the Reactant Complex (RC). In order to reproduce the reactant complex structure, the product 7-chlorotryptophan was modified to the substrate tryptophan and HOCl. Variations in the system setup depended on the particular simulation type applied to the system, namely: potential energy path or free energy path modelling, and are detailed in the supporting information (SI).

QM/MM Potential Energy Reaction Path Calculations.
In all QM/MM calculations the protein was 'divided' into two parts -a quantum mechanical (QM) part and a molecular mechanical (MM) part. The quantum mechanical (QM) part consisted of HOCl, the substrate tryptophan, and the side chains of lysine 79 (K79) and glutamate 346 (E346) ( Figure S1 in SI), that are proposed to play key roles in the reaction mechanism 4,7-11 of the enzyme. The molecular mechanical (MM) part involved the rest of the system. This resulted in 57 atoms in total for the quantum mechanical region which had a net charge of zero. Most QM/MM systems have the QM and MM regions in the same molecule and thus, it is necessary to "cut" at least one covalent bond to define the two regions. In cases like this it is necessary to fill the unsaturated valence of the QM atom. Therefore, for our system of interest at the points where the QM/MM boundary bisects a covalent bond, the valences of the QM atoms were saturated using the link-pair formalism 18 . Two hydrogen link atoms were added to the side chains of E346 and K79, respectively.
Potential energy reaction path modelling was performed with the adiabatic mapping method 19,20 at B3LYP/6-31 + G(d)-CHARMM27 using the QoMMMA 21 interface. The adiabatic mapping method has been shown to elucidate the reaction mechanism of many enzymes including FAAH 22 , COX-1 23 and chorismate mutase 24 . The B3LYP functional has been largely applied for studying aromatic halogenation 25,26 and in large number of QM/ MM studies of enzyme-catalyzed reaction mechanisms 9,23,24,27 . The QM calculations were performed using the Jaguar software 28 and the MM atoms in the system were treated by CHARMM27 all-atom force field 29 implemented in the Tinker MM code 30,31 . The MM atoms were optimized during each step of the QM optimization and the polarization effects were accounted for by including the point charges of all MM atoms in the QM Hamiltonian. The transition states were determined approximately, as the highest energy points along the respective reaction profile. The reaction coordinate for the formation of the Wheland intermediate (RC1, Fig. 3) was chosen as: bond breaking between the O and Cl atoms of HOCl, and the formation of a bond between the Cl atom of HOCl and the 7 th C atom from the indole ring of tryptophan: The reaction coordinate used for describing the proton abstraction from the Wheland intermediate by the oxygen of E346 (RC2, Fig. 3) was defined as a linear combination of the following distances: the distance between QM/MM-MD Free Energy Reaction Path Calculations. The free energy reaction profile along the respective reaction of interest was obtained by statistical treatment of the work function, along the chosen reaction coordinate, using different trajectories via the Multiple Steered Molecular Dynamics (MSMD) method, which relates the system's non-equilibrium dynamics to its equilibrium properties 32,33 . The MSMD method considers a system as a subject to external time-dependent perturbation [λ = λ(t)], described by the Hamiltonian H (r, λ). Where ∆G(λ) is the free energy change and W(λ) is the external work performed on the system, as it evolves from an initial to a final state (λ 0 → λ). The Jarzynski relationship 34 and the Crooks fluctuation theorem 35 allow for relating the non-equilibrium work distribution and the equilibrium free energy differences. This allows for recovery of arbitrary equilibrium ensemble averages from measurements of driven non-equilibrium processes.
In particular, the Crooks's path-ensemble average theorem 36 relates the forward average of an arbitrary functional F(λ) of the phase space trajectory, to its work-weighted average in the reverse process, The forward average <···> F, is an average over all trajectories (path-ensemble average) generated in the forward process, wherein the external parameter λ, is driving the system from the initial to the final state (e.g., reactants to products in a pulling experiment). Similarly, <···> R defines the average in the reverse direction. ∆F is the free energy difference between the equilibrium states, corresponding to the final and initial points. This result offers a means of including of the reverse process in the forward path ensemble, when their arbitrary functional is reweighted by e β(W−ΔF) . Based on those theorems, Minh and Adib 37 , proposed a bidirectional PMF estimator that includes trajectories from the forward and reverse perturbation along the reaction coordinate, where G 0 (λ) is the potential of mean force of the unperturbed system and ∆F t = F t − F 0 is the free energy difference between the equilibrium states defined at time t and 0, respectively. V(λ;t) is the time-dependent potential acting on a collective coordinate (e.g., harmonic potential along the reaction coordinate). The Mihn-Adib PMF estimator has been used previously in several works such as reconstructing free energy path of DNA stretching experiments 38 and describing RNA/peptide binding/unbinding processes 39 . The use of Jarzynski's equality was validated on several systems, including deca-alanine stretching 40,41 , most of them involving classical MD. The methodology has been compared to existing biased MD techniques such as umbrella sampling 42 and targeted MD 43 , and the results were comparable. Potential energy path simulations have been widely applied by means of hybrid QM/MM approach using semi-empirical methods (e.g. PM3, AM1) 44 . However, a few potential of mean force calculations have been performed using DFT QM/MM-MD methods in large systems for simulating processes such as unfolding and ligand diffusion in proteins 45 , reaction mechanism involving small molecules [46][47][48] and some specific chemical reactions of biochemical interest 49 because of their computational cost. Steered-MD and QM/MM calculations have been used to elucidate enzymatic reactions 50 including covalent enzyme-inhibitor reactions 51 .
The free energy reaction profiles were generated by applying hybrid quantum mechanical/molecular mechanical -molecular dynamics (QM/MM-MD) calculations conducted by the AMBER 52 -PUPIL 53 -NWChem 54 interface 15,16 and ff03.r1 55,56 Amber force field, used to describe the MM part of the simulated system. The B3LYP/6-31 G functional was used to describe the quantum mechanical (QM) part of the system. Specific parameters for the cofactor FAD were obtained from the R.E.D. project 57 . The QM part of the simulated system chosen for the free energy path scanning was very similar to the one applied in the potential energy reaction path simulations; it includes HOCl and the side chain of K79. The only difference was that it also included the whole E346 residue and only the side chain of the substrate tryptophan ( Figure S2 in the SI). The link-pair formalism was used to treat the covalent bonds that cross the QM/MM boundary and thus, four hydrogen link atoms were added to tryptophan, K79, L345, and S347, respectively. A classical MD simulation was performed for 5 ns and fifteen different structures were extracted from the classical production trajectory, as initial structures, for generating the free energy paths (see SI). All the fifteen initial structures were relaxed under the hybrid QM/MM-MD approach during 0.5 ps (1000 steps), before pulling through the reaction coordinate. Distance constraints were applied to HOCl, K79 and tryptophan. For each simulation, a harmonic potential with a force constant of 5000 kcal mol −1 Å −2 was applied. The center of the potential was moved along the reaction coordinate during all 3000 steps of the MD simulation. The step size was 0.5 fs. A reverse reaction was simulated for the fifteen final structures obtained by the pulling from reactants to products. Free energy paths were studied using two different reaction coordinates (RC) depending on the reaction step namely: the RC1 reaction coordinate was used for simulating the reaction of formation of the Wheland intermediate and the RC2 reaction coordinate was used for simulating its consequent deprotonation. RC1 was defined as a lineal combination of distances involved in the formation of two bondsnamely the bond between the hydrogen atom of K79 and the oxygen atom of HOCl, and the bond between the chlorine atom of HOCl, and the 7 th carbon atom of the indole ring of tryptophan: For modelling the reaction of deprotonation of the Wheland intermediate, a reaction coordinate that describes the formation of a bond between the oxygen atom of the carboxyl group of E346 and the hydrogen atom of the Wheland intermediate was chosen. In this process, which involves transfer of a proton from the Wheland intermediate carbon to E346 oxygen, a mid-plane reaction coordinate was used (RC2, Fig. 4). Taking into account that C is the Wheland intermediate carbon, to which the hydrogen atom is initially bound to, and O is the E346 oxygen atom, to which the proton will be transferred to, the mid-plane reaction coordinate, RC2 was defined as the ratio r/R, where R is the length of the distance between C and O and r is the length of the projection of the C-H distance onto distance R. RC2 reaction coordinate is dimensionless since it is a ratio between two distances. Mid-plane reaction coordinate is one of the most appropriate approaches for reproducing reversible pathways in studying proton transfer processes, and is the least susceptible one to hysteresis 58 .
The free energy reaction profiles were obtained using the RC1 reaction coordinate values in the range from −4.7 Å to −2.6 Å and the dimensionless RC2 reaction coordinate values ranging from 0.3 to 0.8, in simulating the forward reactions of the two reaction steps of the simulated mechanism. When the mid-plane reaction coordinate, RC2 has a value smaller than 0.5 (RC2 <0.5), the hydrogen is still bound to the C7 atom of the Wheland intermediate, and when RC2 equals 0.5 (RC2 = 0.5), the hydrogen is halfway between the carbon of the Wheland intermediate, and the oxygen of E346.

Results and Discussion
Potential Energy Reaction Paths. The starting structure for QM/MM potential energy adiabatic mapping calculations was the enzyme-substrate (ES) complex/reactant complex (RC) structure, minimized at the B3LYP/6-31+G(d)-CHARMM27 level. The calculated potential energy barrier for the reaction of the Wheland intermediate formation was 4.0 kcal/mol ( Figure S2) and the reaction was exothermic with reaction energy of −8.6 kcal/mol. Two hydrogen bonds were observed in the optimized reactant complex structure (RC1) for the first step of the simulated reaction. The first one, between the K79 hydrogen and the HOCl oxygen (distance 1.6 Å) ( Table 1, Fig. 5a), supports the experimentally proposed role of K79 in activating HOCl, for the consequent electrophilic aromatic chlorination of tryptophan 4 . The second one is between the hydrogen atom of the NH group of the indole ring of tryptophan, and the oxygen of the carboxyl group of E346 (distance 1.7 Å). In addition, the HOCl chlorine atom is well positioned for making an electrophilic attack on the 7 th carbon of the indole ring of tryptophan (distance 2.5 Å). Indeed, an additional hydrogen bond between HOCl and the oxygen of the E346 peptide bond (out of the quantum region and therefore not shown in Fig. 5) at distance of 1.7 Å helps for this positioning. The last two hydrogen bonds were also found in the transition state (TS1) and the product complex (PC1) structures for the first simulated step of the reaction. An additional hydrogen bond was found in the PC1 structure (Fig. 5c), where it stabilizes a water molecule.
The transition state structure (TS1), shows advancement in the bond formation between HOCl chlorine and the 7 th carbon atom of tryptophan, and the bond formation between the hydrogen atom of K79 and the oxygen of HOCl (Table 1, Fig. 5b) that finally leads to formation of 7-chlorotryptophan and a water molecule. The product complex structure (PC1) had both the Wheland intermediate, and the water molecule completely formed (Table 1, Fig. 5c). In the Wheland intermediate structure, the bond between the chlorine and the carbon atom at 7 th position in the indole ring of tryptophan is formed, and the aromatic hydrogen at position 7 of the indole ring, points toward E346. This atomic arrangement is consistent with the reaction of deprotonation of the Wheland intermediate, performed by the carboxylate group of E346.
Deprotonation of the Wheland intermediate is the second step in the regiospecific chlorination of tryptophan, leading to formation of the final product 7-chlorotryptophan. An oxygen atom from the carboxyl group of E346 is thought to abstract the proton 4 . The structure of the Wheland intermediate (PC1 in the first step of the reaction mechanism) was used as a starting structure for the reaction modelling of the second step (Fig. 5c). The activation   Fig. 5e).
Free Energy Reaction Paths. The calculated free energy reaction paths (Potential of Mean Force, PMF) for both reaction steps of the enzymatic chlorination of tryptophan, are shown in Fig. 6. The bidirectional PMF estimator of Minh-Adib (Equation 2) was used to reconstruct the unbiased PMF along the two RCs and to optimally combine the forward and backward profiles thus, accelerating the convergence of the free energy. The PMF using unidirectional Hummer-Szabo estimator is getting more biased, as the states are further away from the starting equilibrium state, showing a clear overestimation of PMF. Obtaining PMF, using MSMD methodology at DFT level, requires high pulling velocity due to its large computational cost. This leads to growing error with increasing the distance from its original equilibrium state. It has been shown that bidirectional PMF estimators improve resultant PMF profiles by estimating the optimal free energy difference using data from forward and backward pulling 37 . Therefore, the free energy differences between reactants, transition states and product complexes were derived from Minh-Adib estimator profile, and do not present the biasing on the PMF profile, because of the use of forward and backwards reaction profiles ( Table 2).
The free energy barrier for the first reaction step is about 3.0 kcal/mol, resulting in an exothermic reaction with a free energy of −2.4 kcal/mol. Importantly, in the first reaction step HOCl is activated for electrophilic attack by two strong hydrogen bonds with K79 and E346 residues. The transition state structure keeps a weaker hydrogen bond with E346 and forms a new strong hydrogen bond with K79 that remains in the product complex structure thus, helping to stabilize both the transition state and final product complex structures (Table 3 Fig. 7).
The calculated free energy path for the second step of the reaction mechanism -the reaction of deprotonation of the Wheland intermediate by the E346 oxygen, shows an activation energy barrier of 6.0 kcal/mol, and reaction energy of −20.4 kcal/mol. Thus, the second reaction step is the rate-limiting one, which is in agreement with the potential energy path calculations (9.4 kcal/mol). The free reaction energy of −20.4 kcal/mol is in agreement with the potential reaction energy of −24.3 kcal/mol and suggests about an exothermic reaction. The reaction free energy for the whole reaction of chlorination of tryptophan is −22.8 kcal/mol.

Structure and Charge Analysis of Reaction Complexes. Key interactions, corresponding to reactants,
transition states, and products for both reaction steps of the mechanism, averaged through several snapshots of the hybrid QM/MM-MD, are presented in Table 3, Fig. 7. For example, there is a strong hydrogen bond between the K79 hydrogen and the oxygen of HOCl (1.5 Å), in the reactant complex (RC1) for the reaction of formation of the Wheland intermediate. Indeed, this hydrogen bond facilitates the process of chlorine transfer from HOCl to tryptophan, and the formation of a water molecule. An additional hydrogen bond of 1.9 Å between HOCl and the carboxyl group of E346 is found to stabilize the reactive complex, and is found to disappear after the first reaction step. A similar hydrogen bonding network, stabilizing the RC, TS, and PC structures was also found in   Fig. 8. These interactions facilitate the electrophilic attack of Cl to the C atom at 7 th position of the indole ring of tryptophan, with an observed charge redistribution/reorganization of ~0.4 a.u. (Fig. 9). The Cl atom is a subject of a major atomic displacement between the chlorine and the C atom at the 7 th position of the indole ring of tryptophan (RC1, Fig. 6) along bond formation, starting from 2.7 Å in the reactant complex, going up to 2.0 Å in the product complex, and located at 2.3 Å in the transition state structure (Table 3) In order to study charge redistribution amongst the main chemical species participating in the two reaction steps of the simulated reaction mechanism, a few species/fragments, that are part of the quantum region (QM) of the simulated system, were defined ( Fig. 9) as follows: deprotonated K79 residue (K79(deprotonated)), a proton from the protonated amino group (NH 3 + ) of K79 (H(K79) + ), hydroxyl group (OH − ) of HOCl (OH(HOCl)), chlorine atom (Cl) of HOCl, tryptophan (Trp), and deprotonated E346 residue (E346( − )), respectively. The charge on these fragments was obtained by adding the Mulliken atomic charges at each fragment for each trajectory snapshot. The global charges for each fragment, of the two reaction steps, as a function of the RC, and averaged from all forward pullings are given in Fig. 9. As RC1 increases, there is a slight polarization, in direction of tryptophan, due to charge transfer mainly from K79 deprotonation, as seen by the decrease in the positive charge from K79 (deprotonated) fragment, and the corresponding charge increase on the tryptophan fragment. Similarly, as RC2 increases, there is a slight positive charge transfer (~0.4 a.u.) from the tryptophan fragment to E346 fragment. Thus, there is a net redistribution of ~0.4 a.u of positive charge transferred from K79 to E346 during the whole process of chlorination of tryptophan. The K79 fragment is the one that is the most involved in the charge transfer to tryptophan (with about 0.4 a.u.), but there are other contributions too. HOCl molecule and in particular its OH fragment, also contribute a little to the charge transfer (with about 0.1 a.u.).

Similarities and Differences between the Potential Energy and Free Energy Reaction Paths.
Comparison between the activation and the reaction energies of the two steps of the simulated reaction mechanism calculated at the different levels of theory was performed.
There is a very close agreement between the activation energies calculated for the potential energy path and the free energy path for the first step of the simulated mechanism -the reaction of Wheland intermediate formation within 1 kcal/mol (Table 4). A difference of about 6 kcal/mol for the reaction energy for the same reaction was found between the free energy and the potential energy pathways, with the potential energy path providing a more exothermic reaction. This could be attributed to the formation of a higher energy Wheland intermediate in the free energy reaction path simulation. In the free energy PC1 structure, a water molecule is found to participate in a bridge interaction between the chlorine atom of the Wheland intermediate and the nitrogen atom of the amino group of K79. In fact, a more elongated C-Cl bond (d C-Cl = 2.0 Å), and a stronger  Table 3. Active site interactions (Å/degrees) in the RC, TS and PC structures (Å/degrees) for the different reactive species -tryptophan, HOCl, E346 and K79, from the free energy reaction paths scans for the two reaction steps of chlorination of tryptophan, (±standard deviation).
hydrogen bond between a water molecule and K79 amino group (d N-H = 1.6 Å, Table 3, Fig. 7c) were observed, in respect to the same structure in the potential energy path where the bond distances are 1.8 Å and 1.8 Å, respectively (Table 1, Fig. 5). The calculated activation and reaction energies, for both the potential energy and the free energy path calculations, for the proton abstraction of the Wheland intermediate reaction step performed by the oxygen atom of E346, are presented in Table 5.
The calculated activation energy of the deprotonation reaction from the free energy calculation is lower by 3.4 kcal/mol than the one calculated from the potential energy scan. The free energy simulations indicate less exothermic reactions, as compared to the potential energy path. These energetic alterations can be due either to methodological differences between the two and/or due to entropic contribution. Despite the small difference in  the basis set used for the QM subsystem between potential energy path and free energy path simulations, there are important methodological differences such as the applied periodic boundary conditions, contribution of the conformational entropy, and the fact that the full protein system is taken into account, that are considered only in the free energy path simulations. In the QM/MM potential energy scans the solvated system was truncated at 25 Å from the CZ atom of tryptophan whilst, in the free energy scans a larger system (the whole protein in a solvent box) was used. Such differences could be an origin of some deviations in the results.
The potential energy surface and the free energy surface for the simulated reaction mechanism of chlorination of tryptophan are given in Fig. 10.
Both the potential energy and free energy calculations show that the rate-limiting step appears to be the second one -the proton abstraction (Tables 4 and 5). Entropy considerations and methodology differences lead to a reduction of the energy barrier and also to a higher energy Wheland intermediate state. In fact, the introduction of kinetic energy in the two simulation steps of the calculated reaction pathways leads to a substantially more unstable arenium cation complex with a lower activation energy, that facilitates the final proton abstraction step when compared to the potential energy path which always finds/favors the most stable conformation. As a consequence, the overall reaction free energy is −22.8 kcal/mol and the global PMF maximum (the activation energy) is 3.6 kcal/mol, calculated in respect to the initial reactant complex energy.
In this first computational study of the reaction mechanism of tryptophan 7-halogenase we applied two different computational QM/MM strategies with DFT treatment of the QM part to explore the feasibility of a proposed catalytic mechanism of chlorination of tryptophan, to characterize structures of the stationary points (reactant complexes, transition states, product complexes) and to determine the energetics of the process. Despite all the differences in the two QM/MM methodologies, the structures of the stationary points and the observed hydrogen bonding networks characterized in the two methods are very similar. The rate-limiting step appears to be the same -the deprotonation of the Wheland intermediate. The activation barriers for the two reaction steps are also similar. The information from the potential energy surface with minima energy structures was complemented by the dynamic information that is derived through the free energy simulations. The provided atomistic and electronic structural insights in the reaction mechanism of enzymatic tryptophan chlorination would contribute for molecular design of halogenase enzymes with desired activity and regioselective properties for production of biologically and pharmaceutically active natural products that contain a chlorine atom as well as for further implementation of enzymes in the synthesis of chlorinated organic compounds.

Conclusions
The enzymatic mechanism of regioselective aromatic electrophilic chlorination of tryptophan by hypochlorous acid in tryptophan 7-halogenase was investigated for first time using two different QM/MM potential energy and QM/MM-MD free energy methodologies at DFT level of theory for the QM parts. The stationary points (reactant complexes, transition states and product complexes) along the reaction paths were characterized with emphasis on the most important interactions. For the Wheland intermediate formation the potential energy and free energy path calculations showed nearly the same energy barriers, however, the free energy path profile yields a higher energetic state of the Wheland intermediate. Both methods, however, provide similar hydrogen bonding interactions in the active site that stabilize the stationary points along this reaction step. The calculations for the second step, deprotonation of the Wheland intermediate, revealed that one of the oxygens of the carboxyl group of E346 plays the role of proton acceptor, while the other oxygen is involved in a hydrogen bond interaction with the nearby NH group of tryptophan. In the reaction of chlorination of tryptophan, a net charge transfer of ~+0.4 a.u, was observed from K79 to E346, and the potential energy paths and free energy paths calculations indicate that the proton abstraction step is the rate-limiting one. Both reaction profiles show a similar global maximum of energy close to ~4 kcal/mol; meanwhile the reaction energy differs with values of −32.9 kcal/mol and   −22.8 kcal/mol for the potential energy path and free energy path profiles, respectively. Calculations confirm the experimentally suggested roles of K79 and E346 in activating HOCl, assuring the regioselectivity of the process and provide atomistic insight about their roles. The insights from this first computational study of the reaction mechanism of tryptophan 7-halogenase would be valuable for a design of halogenating enzymes with desired properties with broad applications in chemical biology, pharmaceutical industry and organic synthesis.