Subscriber access provided by Kaohsiung Medical University
Article
Computational Study of Mycobacterium smegmatis Acyl transferase Reaction Mechanism and Specificity Masoud Kazemi, Xiang Sheng, Wolfgang Kroutil, and Fahmi Himo ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.8b03360 • Publication Date (Web): 09 Oct 2018 Downloaded from http://pubs.acs.org on October 9, 2018
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 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 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.
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.
Page 1 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Computational Study of Mycobacterium smegmatis Acyl Transferase Reaction Mechanism and Specificity
Masoud Kazemia, Xiang Shenga, Wolfgang Kroutilb and Fahmi Himoa
a Department
of Organic Chemistry, Arrhenius Laboratory, Stockholm University, SE-10691,
Stockholm, Sweden. b
Department of Chemistry, University of Graz, Heinrichstrasse 28, A-8010 Graz, Austria.
Corresponding author:
[email protected] 1 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 23
ABSTRACT The acyl transferase from Mycobacterium smegmatis (MsAcT) catalyzes the acyl transfer between a range of primary and secondary alcohols, whereby its outstanding ability is to perform this reaction in aqueous solution. Therefore MsAcT opens different options for acylation reactions enabling alternatives for many conventionally hydrolytic enzymes used in biocatalysis. Nevertheless, hydrolysis is still a major side reaction of this enzyme. To provide a detailed understanding of the competition between hydrolysis and transesterification reactions, a combination of density functional theory and free energy perturbation methods have been employed. The relative binding free energies and the energy profiles of the chemical steps involved in the reaction were calculated for a number of substrates. The calculations show that the enzyme active site exhibits a higher affinity for substrates with an aromatic ring. The ratedetermining step corresponds to the collapse of a negatively charged tetrahedral intermediate in the substrate acylation half-reaction. The intrinsic barriers of the transesterification and hydrolysis half-reactions are calculated to be of similar heights, suggesting that the determining factor in the MsAcT specificity is the higher binding affinity of the active site for the alcohol substrates relative to water. Finally, the influence of the acyl donor on the MsAcT-catalyzed reaction is also investigated by considering different esters in the calculations.
KEYWORDS: Acylation, transesterification, enzymology, density functional theory, free energy perturbation, transition state, reaction mechanism
2 ACS Paragon Plus Environment
Page 3 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
I. INTRODUCTION Biocatalytic acyl transfer catalyzed e.g. by lipases is a broadly applied technique to perform kinetic and dynamic resolution as well as to prepare products of the daily life [1]. These reactions are preferentially performed in an organic solvent in the absence of water to circumvent the hydrolysis of the acyl donor. Recently, selected reports have demonstrated the acyl transfer in water/buffer as solvent [2-9]. One enzyme suitable for acyl transfer in buffer for preparative transformation of non-natural substrates turned out to be the acyl transferase from Mycobacterium smegmatis (MsAcT) [9], which has been described to catalyze a series of transesterification and perhydrolysis reactions (Scheme 1) [2,10,11]. Various substrates have been tested whereby it turned out that the enzyme shows higher activity towards primary alcohols and amines, although some secondary compounds have also been transformed. The enzyme shows good enantioselectivity for a selected number of propargylic or nitrile aliphatic substrates [2].
Scheme 1. Transesterification reaction catalyzed by MsAcT. In the present study, we have considered 1a, 1d, and 1e as acyl donors, and 2b as the acyl acceptor. The competing hydrolysis reaction involving water (2c) as the acyl acceptor is also investigated. MsAcT is composed of eight identical subunits and shares common features with the SGNH superfamily of serine hydrolases, including the catalytic triad and the oxyanion hole (Figure 1) [9]. The active site entrance channel is made up of three adjacent subunits and exhibits hydrophobic character, which has been suggested to possibly contribute to the ability of this enzyme in catalyzing the transesterification reaction in water [2,9]. 3 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 23
Although the transesterification reaction is clearly observed in buffer, hydrolysis is still a major
side
reaction.
Therefore,
the
understanding
of
the
parameters
determining
transesterification versus hydrolysis is of high importance to improve this enzyme and increase its biocatalytic applicability. The goal of the present study is to investigate the mechanisms of both the transesterification and hydrolysis reactions of this enzyme, and to elucidate the sources of discrimination between the two. To this end, we constructed a large model of the enzyme active site (consisting of ca 270 atoms) and used a combination of density functional theory (DFT) and free energy perturbation (FEP) methods to calculate the energy profiles associated with the two reactions.
Figure 1. X-ray crystal structure of the MsAcT active site with the covalently bound inhibitor (PDB code 2Q0S). The catalytic triad is depicted in orange and the inhibitor is highlighted in cyan.
4 ACS Paragon Plus Environment
Page 5 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
II. COMPUTATIONAL METHODS II.A. Quantum Chemical Details The B3LYP-D3 hybrid DFT functional was used [12,13] for all quantum chemical (QC) calculations, which were performed with the Gaussian 09 program [14]. All geometries were optimized using the 6-31G(d,p) basis set. The effect of protein environment, beyond the cluster model, was approximated by single-point energy calculation using the SMD solvation model [15] with a dielectric constant of ε = 4, in line with previous studies [16]. Zero-point energies (ZPE) were obtained by analytical frequency calculations on the optimized geometries. Both the ZPE and solvation calculations were performed at the same level of theory as the geometry optimization. To obtain more accurate energies, single-point calculations were performed with the large 6-311+G(2d,2p) basis set. Thus, the final results are obtained by correcting the energies calculated with the large basis set with the ZPE and SMD solvation corrections. II.B. Quantum Chemical Active Site Model A cluster model of the active site (Figure 2) was constructed based on the crystal structure of the inhibitor-bound Mycobacterium smegmatis acyl transferase (MsAcT) (PDB code 2Q0S) [9]. The model includes the residues that comprise the catalytic triad (Ser11, Asp192, and His195) and the binding cavity (Asp10, Lue12, Trp16, Ala55, Ser54, Asn94, Thr93, Lys97, Val125, Phe150, Ile153, Phe154, Phe174, Ile194). The residues Leu12, Thr93, and Ile194 compose a small pocket, while Ala55, Ser54, Lys97, Val125, Phe150, Ile153, Phe154, Phe174 comprise a larger pocket that is connected to the entrance channel. In addition, the backbone amides of His195, Ser11 and Ala55 were also included in the model. The amide of His195 is involved in hydrogen bonding with the catalytic Asp192, and the amide groups of Ser11 and Ala55, together with the side chain of Asn94, form the oxyanion hole. Three crystallographic water molecules were also included in the model (W1, W2, and W3 in Figure 2). These water molecules do not participate in the chemical transformation. However, W1 appears to form a hydrogen bond with the catalytic Asp192 stabilizing the negative charge of this residue, and W2 is within hydrogen bonding distance of Asp10 and Thr93. W3 forms a hydrogen bond with Ser54 and helps to keep the orientation of this residue in the cluster model. To maintain the integrity of the active site model, 5 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 23
26 atoms were kept fixed to the crystallographic positions during geometry optimizations. In most cases, carbon atoms were fixed; however, for some residues hydrogen atoms were also fixed (Figure 2).
Figure 2. A) Optimized structure of the reactant state of enzyme acylation half-reaction with 1a (highlighted with green carbons) as acyl donor, and B) a schematic drawing of the groups included in the cluster model. The catalytic triad is highlighted with orange carbons. The fixed atoms are indicated with asterisks. For clarity, nonpolar hydrogen atoms are omitted except for the residues Trp16, Phe150, Phe154, Phe174, and Asp192 for which Cα atoms were converted to H. The model contains four ionizable residues, namely Asp10, Lys97, Asp192, and His195. His195 and Asp192 are parts of the catalytic triad in which His195 initially accepts a proton from Ser11, thus activating this residue for the nucleophilic attack, and Asp192 participates in electrostatic stabilization of His195. Therefore, His195 and Asp192 were modeled as neutral and negatively charged, respectively. Asp10 is located beneath the oxyanion hole and is in hydrogen bonding with W2. Since this residue is buried, it was chosen to be neutral in the quantum chemical model. In the crystal structure, Lys97 forms hydrogen bonds to two water molecules and to the side chain of Asn94, indicating that Lys97 is protonated. The positively charged Lys97 also appears to be stabilized by the neighboring Asp62, which is not included in the 6 ACS Paragon Plus Environment
Page 7 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
model. As such, the residues Lys97 and Asp62 can be considered as an ion pair. In the model, the neutral form of the Lys97 was chosen, since Asp62 is not included in the cluster model. Overall, the final model contains 255 atoms, not including the substrate (269 and 276 atoms with bound 1a and 2b, respectively), and has a total charge of -1. II.C. MD Simulations Molecular dynamics (MD) simulations were carried out with the OPLS-AA force field [17] and TIP3P water model using the Q software [18]. A spherical boundary condition with a radius of 28 Å was employed for MD simulation of different substrates in the enzyme and water. The simulation sphere of the enzyme was centered on the hydroxyl oxygen of Ser11 and the residues outside the sphere were excluded from calculations (see SI for a figure). The simulation sphere is thus comprised of 10375 atoms without the substrate. Water molecules near the boundary were treated according to the SCAAS model [18,19]. Direct non-bonded interactions were calculated with a 10 Å cutoff. The electrostatic interactions beyond the cutoff were calculated with the local reaction field multipole expansion method [20] for all atoms except the substrate, for which no cutoff was applied. The SHAKE algorithm [21] was used to constrain the bond and angle of the solvent molecules during the MD simulations. The ionizable residues close to the simulation boundary and protein surface lack electrostatic screening [22]. Thus, to avoid artifacts, these residues were neutralized. The charge states of the other residues were chosen by visual inspection of the structure to be compatible with pH=7 and the ionization states of the active site residues were kept as in the DFT calculations, except for Lys97. As mentioned above, Lys97 is located in the proximity of Asp62. Since the simulation sphere includes both these residues, Lys97 and Asp62 were modeled in their ionized forms. The crystal structure of the inhibitor-bound enzyme was converted to the acylated form, in which the acetyl group is covalently bound to Ser11. Each substrate was docked manually into the active site. The crystallographic waters were kept when the substrate-enzyme complexes were solvated. The enzyme-substrate complexes were initially equilibrated in eight steps where the systems were gradually heated from 1K to 298K and subsequently simulated for 3 ns with a 1 fs time-step. During the simulations, 5.0 kcal/molÅ2 distance restraints were used between the hydroxyl group of the substrate and active site residues Ser11 and His195. These small restraints were imposed to keep the substrates in the active site, especially for the smaller ones. The 7 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 23
contribution of the restraints to the potential energy is negligible and nearly constant for the substrates, 0.7 kcal/mol averaged over substrates. Their effect is expected to cancel since the same restraints were used for all substrates. From the trajectories of the initial simulation of each substrate, 10 structures were extracted at equal intervals and the structures were used as the starting points to repeat free energy perturbation (FEP) calculations for each substrate. II.D. Free Energy Perturbation Calculations FEP calculations were performed to obtain more accurate estimates of the relative binding free energies of the substrates. These calculations were also done with OPLS-AA force field [17] and TIP3P water model and were started from the previously extracted structures, which were equilibrated for 1 ns. The FEP calculations were carried out with a similar scheme as used in the recent studies of alanine scanning of G-protein coupled receptors [23,24]. That is, instead of directly transforming ligands to each other to obtain the relative binding free energy, a reference state was chosen and all ligands were transformed to this reference state. In the present study, methanol was chosen as the reference state and larger alcohols were transformed into methanol. The water transformation was performed from methanol to water. In these calculations, each transformation was divided into sub-perturbations (5 to 8 depending on the size of alcohol, see Supporting Information for details) in which the most distant atom groups from the hydroxyl were annihilated in succession. In the annihilation process, the partial charges of each atom group were set to zero, the soft core potential was then used for these atoms, and finally the Van der Waals parameters of the atom group were set to zero. In the second last sub-perturbation, where the annihilation process was complete, the bonded terms and Van der Waals parameters were set to those of methanol, and in the final sub-perturbation the partial charges of methanol were turned on. Each sub-perturbation was carried out with 51 FEP windows of 30 ps each. This amounts to 7.7-12 ns simulation time for each FEP calculation, which was repeated 10 times using the structures obtained from the MD simulations (see above). The binding free energies of different alcohols relative to each other were then obtained by simply combining the final free energies of each ligand transformation into the reference state.
8 ACS Paragon Plus Environment
Page 9 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
III. RESULTS AND DISCUSSION As discussed in the Introduction, the similarity of MsAcT with other enzymes in the SGNH superfamily of serine hydrolases suggests that the reaction mechanism of MsAcT should in principle be similar to these related enzymes, involving acid-base catalysis via formation of a tetrahedral intermediate (TI) (Scheme 2).[25] The catalytic triad in this case would be comprised of residues Ser11, Asp192, and His195, and the oxyanion hole consisting of the backbones of Ser11 and Ala55 and the side chain of Asn94 [9]. The catalytic cycle of the MsAcT-catalyzed transesterification involves thus two halfreactions. The enzyme is first acylated by an acyl donor, followed by acyl group transfer from the enzyme to an acyl acceptor (substrate). His198 HN Asp192 His198
N
O
O
R
TS1 N
HN Asp192
O
N
O O R
Asn94 H O
Ser11
H O
Ser11
O H H H N N Ala55 Ser11 INT1
O H H H N N Asn94 N Ala55 Ser11 E R-Ac R'
O
O O
N O
H O
R
O
O
H H H N N Asn94 N Ala55 Ser11 E-Ac R-OH R'
2nd half-reaction: Acyl transfer
His198
His198
O
Asp192
Ser11
1st half-reaction: Enzyme acylation
O R
HN Asp192
O
His198
TS2
O
HN
O R'
O
E R'-Ac N HO
E-Ac R'-OH
Ser11
Asp192 INT4
O H H N N Asn94 N Ala55 Ser11 H
TS4
O
HN
O R'
N
H
O
O
O R'
HN H O
N O
OH
Ser11
O H H H N N Asn94 N Ala55 Ser11
His198
Asp192
O
R
OH
Ser11
TS3
O H
H H N N Asn94 N Ala55 Ser11
Scheme 2. Mechanism of MsAcT-catalyzed transesterification. The catalytic cycle can be divided into two half-reactions, involving enzyme acylation and acyl transfer to the substrate. 9 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 23
In the present study we chose a number of acyl donors and acceptors to investigate the detailed mechanism of this reaction (Scheme 1). To establish the transesterification mechanism, ethyl acetate 1a was used as the acyl donor in the acylation half-reaction, and in the second halfreaction we chose benzyl alcohol 2b as the acyl acceptor, since this compound was shown to be converted efficiently to the corresponding ester [2]. The results of these investigations are discussed in Section III.A. As mentioned in the Introduction, one of the obstacles in using MsAcT for synthesis is the hydrolysis side reaction in which water acts as the acyl acceptor, resulting in the formation of carboxylic acids. The hydrolysis reaction not only reduces the efficiency of the transesterification, but also deactivates the enzyme due to the low pH [2]. In the light of this, it is important to develop some understanding for the sources of competition between the transesterification and hydrolysis. Therefore, the second half-reaction was also studied with water (2c) as the acyl acceptor (section III.B). Finally, since MsAcT catalyzes transesterification with a variety of acyl donors, the reaction was also investigated employing phenyl acetate 1d and vinyl acetate 1e as acyl donors for comparison (section III.C).
III.A. Transesterification The mechanism of the transesterification reaction was first studied with 1a as acyl donor in the first half-reaction and 2b as acceptor in the second half-reaction. In the optimized Michaelis complex of the first half-reaction (Figure 2 and E∙1a in Figure 3), the carbonyl oxygen of 1a forms two hydrogen bonds with the backbone amide of Ala55 and the side chain of Asn94, which are parts of the oxyanion hole, indicating a role of these residues in the initial binding and positioning of acyl donors. The methyl substituent of 1a points in the direction of the small pocket in the active site (formed by the side chains of residues Leu12, Thr93, and Ile194), while the leaving group points toward the active site entrance channel (Figure 2).
10 ACS Paragon Plus Environment
Page 11 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Figure 3. Optimized geometry of various intermediates in the acylation half-reaction. For clarity, only a small part of the model is shown in the figure. For full model, see Figure 1. The first step in the enzyme acylation half-reaction involves nucleophilic attack of Ser11 on the carbonyl carbon of 1a (TS1 in Figure 3), forming a negatively charged tetrahedral intermediate (INT1 in Figure 3). In TS1, the hydroxyl group of Ser11 is activated by a proton transfer to His198, which occurs concurrently with the nucleophilic attack. The energy of TS1 is calculated to be +8.7 kcal/mol and INT1 is found to be 2.4 kcal/mol higher than the E∙1a complex (Figure 4). The negative charge of the TI is stabilized by three hydrogen bonds to the oxyanion hole, while the positive charge of the protonated His198 is stabilized by the carboxylate of the neighboring Asp192. Subsequently, the TI collapses, resulting in the formation of acylated enzyme in complex with ethanol 2a (called E-Ac∙2a). This occurs via a concerted transition state in which the C−O bond breaks and a proton is transferred from His198 to the oxygen of the leaving group (TS2 in Figure 3). The energy of TS2 is lower than TS1 and is found to be only 3.0 kcal/mol higher than INT1 (Figure 4). In the E-Ac∙2a complex, the carbonyl oxygen of the acetyl group forms hydrogen bonds to the peptide bonds of Ala55 and Ser11, and to the side chain of Asn94, while the ethanol is in hydrogen bonding with His195 (Figure 3). The energy of E-Ac∙2a complex is calculated to be -2.2 kcal/mol relative to the E∙1a complex (Figure 4).
11 ACS Paragon Plus Environment
ACS Catalysis
Hydrolysis Relative Energy (kcal/mol)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Transesterification
9.8
8.7 TS1 0.0 E∙1a
Page 12 of 23
2.4 INT1
TS3w 8.3 INT4w
5.4 TS2 2 -2.2 E-Ac∙2a 2
2
2
14.8 TS4w 10.7 TS4
2.2 0.8 TS3 0.4 E-Ac∙2c INT4 -5.6 E-Ac∙2b
5.9 E∙1c -1.0 E∙1b -7.0 E-His⁺∙AcO⁻
Figure 4. Calculated energy profiles of transesterification and hydrolysis with ethyl acetate 1a as acyl donor. The acyl acceptors for the transesterification and hydrolysis reactions are benzyl alcohol 2b and water 2c, respectively. Note that the energies of the chemical steps are calculated with DFT, while the energies of the ligand exchange steps are calculated with FEP. For the transesterification to occur the product of the first half-reaction, ethanol 2a, has to exit the active site and the substrate of the second half-reaction, benzyl alcohol 2b, has to bind. The energy associated with this ligand exchange step was initially calculated with cluster model according to equation (1), ΔΔE = E(En-Ac∙2b) + Eaq(2a) − E(En-Ac∙2a) − Eaq(2b)
(1)
where the energies are potential energies including ZPE and implicit solvation, and Eaq refers to the energy of the substrate in aqueous solution. Using equation (1), the binding energy of 2b relative to 2a is calculated to be -9.3 kcal/mol. Since the active site is mainly composed of hydrophobic residues, the obtained favorable binding energy of 2b seems to be reasonable. The magnitude of the relative binding energy, however, appears to be overestimated. This could be due to the limited size of the active site model employed here, and/or the inherent inaccuracy of the implicit solvation model used to describe the substrates in the solution. To obtain a better estimate, the relative binding free energies were instead calculated using the more accurate and more computationally demanding free energy perturbation (FEP) method (Table 1). These 12 ACS Paragon Plus Environment
Page 13 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
calculations confirm that the MsAcT active site has a higher affinity for alcohols with hydrophobic substituents. Accordingly, the active site residues Trp16, Phe150, Phe154, and Phe174 interact favorably with bulky alcohols, such as 2b and 2d. Using the FEP results, the binding free energy of 2b relative to 2a is now calculated to be -3.4 kcal/mol, and the E-Ac∙2b complex lies thus 5.6 kcal/mol lower than the E∙1a complex (Figure 4). In the second half-reaction, the acyl group is transferred from the enzyme to the substrate 2b. Once bound to the active site (Figure 5A), the first step is the nucleophilic attack of the hydroxyl group of 2b on the carbonyl carbon of the acylated Ser11, which, similarly to the acylation halfreaction, results in the formation of a negatively charged TI (INT4 in Scheme 2). In the transition state of this step (TS3 in Figure 5C), His198 acts as a general base and activates the hydroxyl group of the alcohol by accepting a proton from this group. The energies of TS3 and INT4 relative to the E-Ac∙2b complex are calculated to be +7.8 kcal/mol and +6.0 kcal/mol, respectively (Figure 4). The final step of transesterification involves the collapse of INT4, which results in the formation of phenyl acetate 1b. The transition state of this step (TS4 in Figure 5C) involves also a proton transfer from His198 to Ser11. The energy of TS4 is +16.3 kcal/mol relative to the E-Ac∙2b complex, and the energy of enzyme-product complex (E∙1b) lies 4.6 kcal/mol higher than E-Ac∙2b (Figure 4). According to these results, thus, TS4 is the rate-limiting step of the overall mechanism with a barrier of +16.3 kcal/mol, relative to the EAc∙2b complex (Figure 4). Table 1. Relative binding energies of the alcohol series calculated using free energy perturbation and the cluster approach. The energies are given in kcal/mol relative to ethanol. Alcohol ΔΔG(FEP) ΔΔE(Cluster model) Ethanol (2a)
0.0
0.0
Benzyl alcohol (2b)
-3.4 ± 0.3
-9.3
Phenol (2d)
-2.8 ± 0.2
-8.6
Vinyl alcohol (2e)
-0.4 ± 0.2
-1.2
Acetaldehyde
+1.3 ± 0.2
+4.6
Water (2c)
+3.0 ± 0.4
+4.4a
a
Two water molecules are assumed to bind to the active site in these calculations.
13 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 23
To close the catalytic cycle, benzyl acetate 1b should be replaced with a new ethyl acetate 1a. Considering that the overall transesterification reaction of 1a with 2b has been measured to be close to thermoneutral [26], and that the energy of the E∙1b complex is calculated to be 1.0 kcal/mol lower than E∙1a (Figure 4), the ligand exchange step can be estimated to be close to thermoneutral and, assuming that the ligand binding/release is not rate-limiting, it will not influence the kinetics of the reaction.
Figure 5. Optimized geometries of the reactants of the second half-reaction with 2b (A) and 2c (B) as acyl acceptors. The corresponding geometries of the transition states are shown in panels C and D, respectively. 14 ACS Paragon Plus Environment
Page 15 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
III.B. Hydrolysis The hydrolysis reaction is similar to the transesterification except that a water molecule, rather than alcohol, acts as the acyl acceptor. To investigate this reaction, we replaced the product of acylation half-reaction (2a) with two water molecules (W4 and W5 in Figure 5B). While W4 is the substrate of hydrolysis (2c), W5 is only a spectator and was added because the active site cavity is large enough to accommodate two water molecules in the absence of alcohols. Also in the MD simulations it was observed that multiple water molecules enter the active site. The energy associated with replacing ethanol 2a with water 2c was calculated with both the cluster model and FEP method, and was found to be quite similar, +4.4 kcal/mol and +3.0 kcal/mol, with the two methods respectively (Table 1). Consistently with the transesterification reaction, the FEP result was chosen to construct the reaction energy profile. It should be noted that in the FEP calculations, the actual perturbation is performed on only one water molecule, but since the MD simulation is carried out with explicit waters, the second water molecule enters the active site spontaneously. Using the FEP results, the calculated relative binding free energy of this step brings thus the energy of the E-Ac∙2c complex to +0.8 kcal/mol, relative to the E∙1a complex (Figure 4). The transition states of the hydrolysis reaction are very similar to those of the transesterification reaction (Figure 5D). The energies of TS3w and TS4w are calculated to be +9.0 kcal/mol and +14.0 kcal/mol, respectively, relative to the E-Ac∙2c complex (Figure 4). Interestingly, the product (acetic acid) is initially in the protonated form, but since it has a lower pKa than His195, it is found to transfer a proton to this residue, which results in a lowering of the energy of the enzyme-product complex by 12.9 kcal/mol (Figure 4). To close the catalytic cycle, the energy associated with the release of acetic acid 1c and the binding of a new ethyl acetate 1a can be estimated from the experimental value of the overall hydrolysis reaction energy, which has been measured to be -1.6 kcal/mol [27]. Since the enzymeproduct complex E-His⁺∙AcO⁻ is calculated to be 7.0 kcal/mol lower than the enzyme-reactant complex E∙1a, the final ligand exchange step can thus be estimated to be endothermic by 5.4 kcal/mol. Overall, these calculations suggest that the rate-limiting step for the hydrolysis pathway is the TI decomposition in the second half-reaction, similarly to transesterification. The overall 15 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 23
activation barrier for this pathway is predicted to be +20.4 kcal/mol, which corresponds to the energy difference between the E-Ac∙2b complex and TS4w (Figure 4). The calculations indicate thus that MsAcT favors the transesterification reaction compared to hydrolysis, which is consistent with the experimental observations. The calculated energy difference between the rate-limiting TSs of the hydrolysis and transesterification reactions is 4.1 kcal/mol (TS4 vs. TS4w). Since the intrinsic activation barrier of the second-half reaction of hydrolysis (+14.0 kcal/mol between E-Ac·2c and TS4w) is calculated to be lower than that of the transesterification half-reaction (+16.3 kcal/mol between E-Ac·2b and TS4), the results show that the differentiation stems mainly from the preferential binding of hydrophobic substrates compared to water. In this case, the difference in binding free energy between 2b and water is 6.4 kcal/mol. The calculated energy difference between the barriers appears, however, to be overestimated considering the fact that the hydrolysis reaction is an observed side reaction. This overestimation is partly due to the fact that the concentration effect is not taken into account in the calculations. Namely, under experimental conditions, the concentration of water is 55 M, which is much higher than that of the alcohol, typically 0.1 M [2]. Including the concentrations in the energies would shift the binding equilibrium towards the hydrolysis by ca 4 kcal/mol (RT∙ln[55/0.1]).
III.C. Effect of Acyl Donors An interesting observation in the MsAcT-catalyzed transesterification is that the choice of acyl donor affects the conversion rate, and it has been shown that phenyl acetate (1d) and vinyl acetate (1e) are better acyl donors than 1a [2]. To investigate the influence of acyl donors on the MsAcT-catalyzed reactions, we here recalculated the acylation half-reaction also with 1d and 1e as acyl donors (Figure 6). The full energy profiles of the hydrolysis and transesterification pathways with these acyl donors were obtained using the FEP calculations to estimate the relative binding free energies (see Supporting Information).
16 ACS Paragon Plus Environment
Page 17 of 23
8.7 Relative Energy (kcal/mol)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
6.57.5 TS1 E∙1a E∙1d E∙1e
5.4 2.5
1.6
2.4 -7.4 INT1
-6.1 TS2
E-Ac∙2a -2.2 -9.8 E-Ac∙2e -12.8 -10.8 E-Ac∙2d E-Ac∙MeCHO
Figure 6. DFT-calculated energetics of the acylation half-reaction with phenyl acetate (1d) and vinyl acetate (1e) as acyl donors. For comparison the energetics of ethyl acetate (1a) are also included. Overall, the acylation mechanisms with these acyl donors are very similar to that of 1a (see optimized geometries in Figure 7). However, one major difference is that the acylation halfreactions with these esters become largely exothermic, implying that thermodynamic equilibrium favors the acylated enzyme (Figure 6). One reason for this could be that the corresponding alcohols of these acyl donors are more acidic (pKa ca. 10) than that of ethyl acetate (pKa ca. 16), thus providing a better leaving group for the acylation half-reaction. Interestingly, the TI intermediate of the 1e half-reaction was found to be ca. 10 kcal/mol more stable than that of 1a, despite that the optimized structures of these intermediates are almost identical (see superposition of structures in Supporting Information). To analyze this, we further investigated these half-reactions by constructing a smaller model that only includes the catalytic triad and the oxyanion hole. These calculations resulted in a similar trend, although the calculated energy barriers are higher (see Supporting Information for details). The calculated energy difference between 1a and 1e in the small model was found to be 7.4 kcal/mol, which suggests that the difference is intrinsic to the substrates and does not stem from the choice of the active site model. The product of the acylation half-reaction with 1e (vinyl alcohol) can undergo tautomerization, forming acetaldehyde [28]. In this half-reaction, acetaldehyde bound to the enzyme was found to be more stable than the corresponding alcohol by 3.0 kcal/mol (Figure 6). 17 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 18 of 23
Figure 7. Optimized structures of transition states of the acylation half-reaction with phenyl acetate 1d (A) and vinyl acetate 1e (B) as acyl donor. Overall, the calculations show, in agreement with the experiments, that 1d and 1e are better acyl donors than 1a since the reaction equilibrium is shifted toward the acylated enzyme. Furthermore, according to the calculations, the acylation half-reaction with 1d and 1e is faster than the one with 1a, as can be judged from the corresponding activation barriers. However, the faster reaction rates of these acyl donors do not provide any kinetic advantage since the overall activation energy of transesterification is not dictated by the acylation half-reaction.
IV. CONCLUDING REMARKS MsAcT is an attractive enzyme in biocatalysis because of its ability to promote acyl transfer reaction in aqueous solution. The hydrolysis side reaction, however, renders MsAcT not ideal for synthetic purpose yet, since a significant amount of the donor gets degraded. We have in the present study investigated the competition between the hydrolysis and transesterification by using water and benzyl alcohol as acyl acceptors. The calculations are consistent with the experimental findings and indicate that MsAcT favors transesterification reaction. The rate-determining step for both transesterification and hydrolysis reactions is found to be the collapse of the negatively charged tetrahedral intermediate in the second half-reactions, and the overall activation energy of the transesterification reaction is calculated to be lower than that of hydrolysis. The main reason is found to be the higher affinity of MsAcT for the alcohols compared to water, which is a consequence of the hydrophobic nature of the active site.
18 ACS Paragon Plus Environment
Page 19 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Calculations on the first half-reaction with phenyl acetate and vinyl acetate indicate that these acyl donors shift the equilibrium towards acylated enzyme. This is possibly due to the lower pKa of the leaving group (the corresponding alcohol) in these molecules. Since the rate-limiting steps in the transesterification and hydrolysis reactions are not affected by the first half-reactions, the more active substrates have no effects on the enzyme specificity. The insights gained by the current calculations will be valuable to direct future efforts in order to redesign this enzyme for better biocatalytic applicability.
SUPPORTING INFORMATION MD model of the active site, detailed energies of the FEP calculations, calculated full energy profiles of the 1d and 1e as acyl donors, superimposed structures of 1a and 1e as acyl donors, calculated energy profiles using small model of the active site, absolute energies and energy corrections of the DFT calculations, and Cartesian coordinates of stationary points.
ACKNOWLEDGEMENTS FH acknowledges the Swedish Research Council for financial support. We thank Prof. Jeremy Harvey for valuable discussions.
REFERENCES [1] a) Faber, K. Biotransformations in Organic Chemistry, 7th ed. Springer, Berlin, 2018; b) Bornscheuer, U. T.; Kazlauskas, R. J. Hydrolases in Organic Synthesis, 2nd ed. Wiley-VCH, Weinheim, 2006; c) Paravidino, M.; Böhm, P.; Gröger, H.; Hanefeld, U. Enzyme Catalysis in Organic Synthesis, 3rd ed.; Drauz , K., Gröger, H., May, O. Eds.; Wiley-VCH, Weinheim, 2012, pp. 249–362. [2] de Leeuw, N.; Torrelo, G.; Bisterfeld, C.; Resch, V.; Mestrom, L.; Straulino, E.; van der Weel, L.; Hanefeld, U. Ester Synthesis in Water: Mycobacterium smegmatis Acyl Transferase for Kinetic Resolutions. Adv. Synth. Catal. 2018, 360, 242–249. [3] Li, Y.; Zhang, W.; Zhang, H.; Tian, W.; Wu, L.; Wang, S.; Zheng, M.; Zhang, J.; Sun, C.; Deng, Z.; Sun, Y.; Qu, X.; Zhou, J. Structural Basis of a Broadly Selective Acyltransferase from the Polyketide Synthase of Splenocin. Angew. Chem. Int. Ed. 2018, 130, 5925–5929. 19 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 23
[4] Subileau, M.; Jan, A. H.; Drone, J.; Rutyna, C.; Perrier, V.; Dubreucq, E. What makes a lipase a valuable acyltransferase in water abundant medium? Catal. Sci. Technol. 2017, 7, 2566– 2578. [5] Scherkus C.; Schmidt S.; Bornscheuer U. T.; Gröger, H.; Kara, S.; Liese, A. A Fed‐Batch Synthetic Strategy for a Three‐Step Enzymatic Synthesis of Poly‐ϵ‐caprolactone. ChemCatChem. 2016, 8, 3446–3452. [6] Schmidt, S.; Scherkus, C.; Muschiol, J.; Menyes, U.; Winkler, T.; Hummel, W.; Gröger, H.; Liese, A.; Herz, H.-G.; Bornscheuer, U. T. An Enzyme Cascade Synthesis of ε-Caprolactone and its Oligomers. Angew.Chem. Int. Ed. 2015, 54, 2784–2787. [7] Sattler, J. H.; Fuchs, M.; Mutti, F. G.; Grischek, B.; Engel, P.; Pfeffer, J.; Woodley, J. M.; Kroutil, W. Introducing an in situ capping strategy in systems biocatalysis to access 6‐aminohexanoic acid. Angew. Chem. Int. Ed. 2014, 53, 14153–14157. [8] Schmidt, S.; Büchsenschütz, H. C.; Scherkus, C.; Liese, A.; Gröger, H.; Bornscheuer, U. T. Biocatalytic Access to Chiral Polyesters by an Artificial Enzyme Cascade Synthesis. ChemCatChem. 2015, 7, 3951–3955. [9] Mathews, I.; Soltis, M.; Saldajeno, M.; Ganshaw, G.; Sala, R.; Weyler, W.; Cervin, M.A.; Whited, G.; Bott, R. Structure of a novel enzyme that catalyzes acyl transfer to alcohols in aqueous conditions. Biochemistry. 2007, 46, 8969–8979. [10] Wiermans, L.; Hofzumahaus, S.; Schotten, C.; Weigand, L.; Schallmey, M.; Schallmey, A.; Domínguez de María, P. Transesterifications and Peracid‐Assisted Oxidations in Aqueous Media Catalyzed by Mycobacterium smegmatis Acyl Transferase. ChemCatChem. 2013, 5, 3719–3724. [11] Land, H.; Hendil-Forssell, P.; Martinelle, M.; Berglund, P. One-pot biocatalytic amine transaminase/acyl transferase cascade for aqueous formation of amides from aldehydes or ketones. Catal. Sci. Technol. 2016, 6, 2897–2900. [12] a) Becke, A. D. Density functional Thermochemistry. III. The Role of Exact Exchange. J. Chem. Phys. 1993. 98, 5648–5652. b) Lee, C.; Yang, W.; Parr, R. G. Development of the ColleSalvetti Correlationenergy Formula into a Functional of the Electron Density. Phys. Rev. B. 1988, 37, 785–789. [13] a) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H−Pu. J. Chem. Phys. 2010, 132, 154104. b) Grimme, S.; Ehrlich, S.; Goerigk, L. Effect of the 20 ACS Paragon Plus Environment
Page 21 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Damping Function in Dispersion Corrected Density Functional Theory. J. Comput. Chem. 2011, 32, 1456–1465. [14] Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision D.01; Gaussian, Inc., Wallingford, CT, 2016. [15] Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal Solvation Model Based on Solute Electron Density and on A Continuum Model of the Solvent Defined by the Bulk Dielectric Constant and Atomic Surface Tensions. J. Phys. Chem. B. 2009, 113, 6378–6396. [16] a) Sevastik, R.; Himo, F. Quantum Chemical Modeling of Enzymatic Reactions: The Case of 4-Oxalocrotonate Tautomerase Bioorg. Chem. 2007, 35, 444−457. b) Hopmann, K. H.; Himo, F. Quantum Chemical Modeling of the Dehalogenation Reaction of Haloalcohol Dehalogenase. J. Chem. Theory Comput. 2008, 4, 1129−1137. c) Georgieva, P.; Himo, F. Quantum Chemical Modeling of Enzymatic Reactions: The Case of Histone Lysine Methyltransferase. J. Comput. Chem. 2010, 31, 1707−1714. d) Liao, R.-Z.; Yu, J.-G.; Himo, F. Quantum Chemical Modeling of Enzymatic Reactions: The Case of Decarboxylation. J. Chem. Theory Comput. 2011, 7,1494−1501. [17] Jorgensen, W.L.; Maxwell, DS.; Tirado-Rives J. Development and testing of the OPLS AllAtom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 1996. 118, 11225–11236.
21 ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 22 of 23
[18] Marelius, J.; Kolmodin, K.; Feierberg, I.; Åqvist J. Q: A molecular dynamics program for free energy calculations and empirical valence bond simulations in biomolecular systems. J. Mol. Graph. Model. 1998, 16, 213–225. [19] King G.; Warshel, A. A surface constrained all-atom solvent model for effective simulations of polar solutions. J. Chem. Phys. 1989, 91, 3647–3661. [20] Lee, F. S.; Warshel, A. A local reaction field method for fast evaluation of long-range electrostatic interactions in molecular simulations. J. Chem. Phys. 1992, 97, 3100–3107. [21] Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. Numerical integration of the Cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 1977, 23, 327–341. [22] Åqvist, J. Calculation of absolute binding free energies for charged ligands and effects of long‐range electrostatic interactions. J. Comput. Chem. 1996, 17, 1587–1597. [23] Keränen, H.; Gutiérrez-de-Terán, H.; Åqvist, J. Structural and energetic effects of A2A adenosine receptor mutations on agonist and antagonist binding. PloS one. 2014, 9, e108492. [24] Boukharta, L.; Gutiérrez-de-Terán, H.; Åqvist, J. Computational prediction of alanine scanning and ligand binding energetics in G-protein coupled receptors. PLoS Comput. Bio. 2014, 10, e1003585. [25] Frey, P. A.; Hegeman, A. D. Enzymatic reaction mechanisms, 7th ed. Oxford University Press. 2007. [26] Fehlandt, P. R.; Adkins, H. Replacement Series of the Alkyl Groups as Determined by Alcoholysis of Esters. J. Am. Chem. Soc. 1935, 57, 193–195. [27] Jencks, W. P.; Gilchrist, M. The free energies of hydrolysis of some esters and thiol esters of acetic acid. J. Am. Chem. Soc. 1964. 86, 4651–4654. [28] Jacobsen, E. E.; Hoff, B. H.; Anthonsen, T. Enantiopure derivatives of 1,2-alkanediols: Substrate requirements of lipase B from Candida antarctica. Chirality. 2000, 12, 654–659.
22 ACS Paragon Plus Environment
Page 23 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Table of Contents Graphics
O
OH
O
H 2O
H
O O
O
HO OH
O
23 ACS Paragon Plus Environment