Hydrolysis and Transglycosylation Transition States of Glycoside

Subscriber access provided by University of Sunderland

B: Biophysics; Physical Chemistry of Biological Systems and Biomolecules

Hydrolysis and Transglycosylation Transition States of Glycoside Hydrolase Family 3 #-Glucosidases Differ in Charge and Puckering Conformation Inacrist Geronimo, Christina Marie Payne, and Mats Sandgren J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.8b07118 • Publication Date (Web): 24 Sep 2018 Downloaded from http://pubs.acs.org on September 27, 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 33 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

The Journal of Physical Chemistry

Hydrolysis and Transglycosylation Transition States of Glycoside Hydrolase Family 3 β-Glucosidases Differ in Charge and Puckering Conformation Inacrist Geronimo†, Christina M. Payne‡*, and Mats Sandgren†* †

Department of Molecular Sciences, Swedish University of Agricultural Sciences, Box 7015, 750 07 Uppsala, Sweden ‡

Department of Chemical and Materials Engineering, University of Kentucky, Lexington, Kentucky 40506-0046, USA

ACS Paragon Plus Environment

1

The Journal of Physical Chemistry 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 33

ABSTRACT

β-glucosidases (βgls) from glycoside hydrolase Family 3 play an important role in biomass degradation by catalyzing cellobiose hydrolysis. However, the hydrolysis rate decreases when the glucose product or another cellobiose competes with water to form oligosaccharides in a reaction called transglycosylation. Both reactions involve proton transfer to the acid/base residue and nucleophilic attack on the glycosyl-enzyme intermediate. To gain a deeper understanding of these competing reactions, quantum mechanics/molecular mechanics calculations were performed. Although both reactions are exothermic and have similar free energy barriers (~18 kcal/mol), the transition state (TS) characteristics are different. The glycosyl-water bond is nearly formed in the hydrolysis TS, leading to reduced ionic character and

4

C1 chair

conformation. The transglycosylation TS is more positively charged and adopts the 4H3 halfchair conformation because bond formation is less advanced. Water interacts solely with the acid/base residue E441, though the long distance between them (2.1 Å) suggests that E441 does not activate water for nucleophilic attack. In comparison, a glucose acceptor has lower deprotonation enthalpy and hydrogen bonds to E441 (1.6 Å), as well as Y204, R169, and R67. Knowledge of these factors relevant to TS formation and stability is valuable for engineering βgls with enhanced hydrolytic activity for industrial applications.


2

Page 3 of 33 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


INTRODUCTION β-glucosidases (βgls) catalyze the hydrolysis of terminal, non-reducing β-D-glucosyl residues to release β-D-glucose.1-3 βgls, particularly from glycoside hydrolase (GH) Family 3 (Figure 1A), are an important component of industrial enzyme cocktails used in the biodegradation of lignocellulosic biomass;4 specifically, these enzymes hydrolyze cellobiose, which inhibits the activity of endoglucanases and cellobiohydrolases.5 GH3 βgls employ the retaining mechanism, wherein both substrate and product share the same stereochemistry (Scheme 1).3 Two amino acid residues, commonly aspartate or glutamate, are involved in catalysis. In the initial glycosylation step, the glycosidic oxygen is protonated by the acid/base residue, while the nucleophilic residue attacks the anomeric carbon, to form the covalently bound glycosyl-enzyme intermediate (GEI).

Figure 1. Active sites of fungal (A) GH3 (Hypocrea jecorina, PDB ID: 3ZYZ6) and (B) GH1 (Humicola insolens, PDB ID: 4MDP7) βgls. Glucose is bound in the -1 site. The acid/base residue and nucleophile (green sticks) are E441 and D236 for GH3 βgl, and E166 and E377 for GH1 βgl. GH1 βgl has a more hydrophobic environment around the acid/base residue and the tyrosine residue near the nucleophile (Y308) can hydrogen bond with the ring oxygen.


3


Page 4 of 33

Scheme 1. Retaining mechanism of βgls. During deglycosylation, hydrolysis occurs if ROH is water, and transglycosylation occurs if ROH is another acceptor, usually a sugar.

Subsequently, the acid/base residue activates a water molecule that carries out another nucleophilic attack on the anomeric carbon to release glucose.8-11 However, another molecule, typically a sugar, at the positive acceptor site(s) (labeled +1 in Scheme 1) may compete with water for transfer to the GEI.12 This competing reaction, called transglycosylation, limits the hydrolytic activity of GH3 βgls at the high substrate loadings typically encountered in industrial biomass conversion applications owing to accumulation of cellobiose and the glucose product.1314

An understanding of the molecular mechanisms underlying both hydrolysis and transglycosylation is critical in developing protein engineering strategies for design of GH3 βgls


4

Page 5 of 33 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


with reduced transglycosidic activity. Mechanistic studies thus far have focused largely on βgls from GH Family 1 (GH1), which also employs the retaining mechanism. Quantum mechanics/molecular mechanics (QM/MM) calculations on a GH1 βgl showed that the hydrolysis transition state (TS) has a dissociative character, that is, the glycosyl-enzyme covalent bond is nearly broken before the C1–O bond with water begins to form.15 This is consistent with the observed large secondary deuterium kinetic isotope effect (kH/kD ≈ 1.1) that indicates a stronger oxocarbenium ion character for the hydrolysis TS than for the glycosylation TS.16-18 Furthermore, a tyrosine residue located close to the nucleophile (e.g., Y308 in Figure 1B) is believed to stabilize the TS through hydrogen bonding with the ring oxygen of the glycosyl moiety.19 The transglycosylation mechanism of GH1 βgls has not been reported, although computational studies of a transglycosylase20 and trans-sialidase21 suggest that it is similar to the hydrolysis mechanism. Active site residues are generally conserved within a family but not across different families; thus, despite the fact that GH1 and GH3 βgls are both retaining GHs, their hydrolysis and transglycosylation TSs likely exhibit different characteristics, limiting the utility of this information in the design of GH3 βgl variants. For instance, the acid/base residue of GH3 βgls has a tight hydrogen bond network, while that of GH1 βgls has a more hydrophobic environment (Figure 1). As shown in a previous constant pH molecular dynamics study, the GH3 βgl acid/base residue has a very low pKa (~2) when the enzyme is in the GEI state, regardless of whether the acceptor molecule is water or another sugar.22 This was attributed to the E441– R125–E128 and E441–R169–E166 hydrogen bond networks, in which the second glutamate residue strengthens the Glu-Arg interaction.22-23 In comparison, the GEI of the GH1 βgls BGLA and BGLB from Bacillus polymyxa have predicted pKas of 6.1 and 5.4, respectively.24 It was


5


Page 6 of 33

hypothesized that the low pKa of GH3 βgls would reduce the ability of the acid/base residue to deprotonate an acceptor, especially water, which has a higher deprotonation enthalpy than sugars.22 Additionally, the position of the tyrosine near the nucleophile in GH3 βgls (e.g., Y204 in Figure 1A) precludes hydrogen bonding with the ring oxygen of the glycosyl moiety; thus, the residue likely does not contribute significantly to TS stabilization unlike the case in GH1 βgls. To elucidate the mechanisms of hydrolysis and transglycosylation in a GH3 βgl and identify the active site residues important to each reaction, QM/MM calculations of Cel3A, a βgl from Hypocrea jecorina (HjCel3A), were performed. Self-consistent-charge density-functional tightbinding (SCC-DFTB25) was used to treat the QM region because it offers the best compromise between computational cost-efficiency and accuracy and, more importantly, provides a good description of carbohydrate ring puckering.26 The potential energy surfaces (PESs) of glycosylation, hydrolysis, and transglycosylation were generated by adiabatic mapping, and the free energy barrier of each reaction step was calculated by umbrella sampling. The nature of the TSs was further examined by Natural Bond Order (NBO)27 analysis using density functional theory (DFT), and the active site residues critical to each step were identified by hydrogen bond analysis.

COMPUTATIONAL DETAILS All simulations were performed using CHARMM version c42b1.28 SCC-DFTB and DFT calculations were done using the sccdftb29 and qchem30 modules, respectively. For the latter, CHARMM was interfaced with Q-Chem 5.0.31 QM/MM preparation. System preparation and equilibration are described in Supporting Information. The equilibrated structure obtained from classical molecular dynamics (MD)


6

Page 7 of 33 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


simulation was used in the QM/MM calculations. The MM region of the protein was described using the CHARMM36 force field,32 while water was modeled using the TIP3P force field.33-34 The QM region, consisting of the D236 and E441 side chain atoms and cellobiose, was treated at the SCC-DFTB level. Hydrogen link atoms, using the divided frontier charge (DIV) scheme,35 were placed between the Cα and Cβ atoms. The size of the QM region, including link atoms, was 63 atoms, and the net charge was -1. QM-QM and QM-MM electrostatic interactions were calculated using Ewald summation (κ = 0.42, kmax = 8) with a 12-Å cutoff.36 A 250-ps QM/MM MD simulation was run prior to adiabatic mapping of glycosylation. The product structure obtained from modeling glycosylation was used as the initial structure for hydrolysis and transglycosylation. To model hydrolysis, the glucose product at the +1 site was removed, and the system was resolvated and equilibrated for 200 ps. A water molecule was included in the QM region for a total of 42 atoms. Adiabatic mapping. A 2D PES of glycosylation was generated with one coordinate representing nucleophilic attack of D236 on the anomeric carbon C1 (RXG) and the other coordinate representing proton transfer from E441 to the glycosidic oxygen O4' (RYG) (Table 1, Figure 2). RXG was scanned from 2.1 to -2.1 Å, while RYG was scanned from 0.9 to -0.9 Å, using an increment of 0.1 Å. A 1D potential energy curve was also generated using the coordinate RG spanning from 3.8 to -1.0 Å with a 0.1-Å increment. For the 2D PES of hydrolysis and transglycosylation, RXH/T represented bond formation between the glycosyl moiety and acceptor (water or glucose) and RYH/T represented proton transfer from the acceptor to E441 (Table 1, Figure 2). A single distance was used for RXH/T and RYH/T, unlike the case in glycosylation, because the glycosyl-aspartate bond d(C1-OD1) was observed to instantly increase from 1.5 to 3 Å after reaching the transition state. RXH/T was scanned from 3.7 to 1.3 Å


7


Page 8 of 33

(0.1 Å increment), while RYH/T was scanned from 2.0 to 0.9 Å (0.05 Å increment). A 1D potential energy curve was also obtained using the sum of RXH/T and RYH/T as coordinate (RH/T), scanning from 5.2 to 2.4 Å in increments of 0.1 Å. Distance restraints were imposed using a harmonic force constant of 5000 kcal/mol/Å2. Where the reaction coordinate was a linear combination of two or three distances, all distances were equally weighted. Geometry minimization at each point was performed using the adopted basis Newton-Raphson method until the average gradient was less than 0.001 kcal/mol/Å. Residues outside the 15-Å radius of the sugar were frozen during minimization. The minimum energy pathway (MEP) in the 2D PES was determined using the Minimum Energy Path Surface Analysis (MEPSA) program.37 Umbrella sampling. The free energy profile of each reaction step was obtained by performing

Table 1. Reaction coordinates used in adiabatic mapping of glycosylation (G), hydrolysis (H), and transglycosylation (T). Reaction coordinatea

Definition

RXG

d(C1-OD1) – d(C1-O4')

RYG

d(HE2-O4') – d(HE2-OE2)

RG

d(C1-OD1) – d(C1-O4') + d(HE2-O4')

RXH

d(C1-OH2)

RYH

d(H2-OE2)

RH

d(C1-OH2) + d(H2-OE2)

RXT

d(C1-O4')

RYT

d(HO4'-OE2)

RT

d(C1-O4') + d(HO4'-OE2)

a

Where the reaction coordinate was a linear combination of two or three distances, all distances were equally weighted.


8

Page 9 of 33 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


umbrella sampling along the 1D reaction coordinate using a harmonic force constant of 200 kcal/mol/Å2. Twenty-five windows were used for glycosylation and 29 windows for hydrolysis and transglycosylation. Each simulation (window) consisted of 50 ps equilibration and 75 ps production/data collection. No atoms were frozen during the equilibration and production simulations. The histograms demonstrate sufficient overlap of the windows for the chosen interval (0.2 Å for glycosylation and 0.1 Å for hydrolysis and transglycosylation) and force constant (Figure S1). The exception is at RH = 3.7–3.8 Å of the hydrolysis reaction coordinate (Figure S1B), which will be explained in Results and Discussion. The weighted histogram analysis method was used to recover the potential of mean force (PMF) from the simulation data.38 Figures S2-S4 show that the PMF curves do not change after 75 ps of production, indicating sampling convergence. Hydrogen bond analysis of the Michaelis complex, TS, and product was performed using a distance cutoff of 2.4 Å and angle cutoff of between 160° and 200° (Tables S1-S9). Atomic charges and Wiberg bond indices (BIs) of selected transition state structures were determined at the B3LYP/6-31+G*39-42 level using NBO 5.043 within the QChem program (Table S10).

RESULTS AND DISCUSSION Glycosylation. Potential energy surface. Glycosylation involves the nucleophilic attack of D236 on the anomeric carbon (C1) of cellobiose and proton transfer from the acid/base catalyst E441 to the glycosidic oxygen (O4') of cellobiose, represented by the reaction coordinates RXG and RYG, respectively. The 2D PES shows that a stepwise mechanism (i.e., either proton transfer or nucleophilic attack occurring first) is a high-energy pathway. In the MEP, proton transfer is nearly complete even before reaching the transition state, wherein the C1–OD1 and C1–O4' bond


9


Page 10 of 33

lengths are almost equal (RXG ~ 0 Å). There is also a low-energy region at positive RYG values and RXG values between -0.6 and -1.0 Å in the 2D PES that corresponds to the transfer of hydrogen from the C2 atom to O4' atom as the glycosidic bond is broken (Figure 2A). Scanning along the reaction coordinate RG reproduced the MEP. QM/MM umbrella sampling was performed along RG to generate the PMF curve of the glycosylation step (Figure S2, Movie

Figure 2. Potential energy surfaces of (A) glycosylation, (B) hydrolysis, and (C) transglycosylation. The minimum energy pathway is represented by a dashed line. The 1D potential energy curve (solid line) obtained using a linear combination of the distances (equally weighted) as a single coordinate (RG, RH, and RT in Table 1) is superimposed on the surface.


10

Page 11 of 33 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


S1). The Michaelis complex, TS, and GEI are located at RG = 3.6, 1.4, and -0.8 Å, respectively (Figure S2). The calculated free energy barrier is 20.5 kcal/mol, making glycosylation the ratedetermining step (vide infra). QM/MM calculation of the glycosylation and deglycosylation (hydrolysis) barriers of a GH1 βgl led to the same conclusion.19 Brønsted plots (log kcat vs. leaving group pKa) of both GH3 and GH1 βgls indicated that, with a poor leaving group (pKa >8) like glucose (pKa = 1244), the rate-determining step is glycosylation.16, 45-46 The calculated barrier is higher than the Gibbs free energy of activation ∆G‡, 16 kcal/mol, derived from the experimental kcat (16.0 ± 0.48 s-1) of HjCel3A with a cellobiose substrate.6 The discrepancy may be attributed to the limited size of the QM region and describing it using a semi-empirical method. SCC-DFTB/MM calculations for a retaining GH Family 18 chitinase also overestimated the experimentally derived ∆G‡. Better agreement was achieved by taking representative structures of the stationary points and performing single-point energy calculations with B3LYP/6-311+G(d,p).47 Formation of the GEI is slightly endothermic (1.2 kcal/mol). Characterization of stationary points. SCC-DFTB compares well with B3LYP in describing the puckered conformations of furanose and pyranose carbohydrate rings despite inaccurate energetics.26 The glycosyl moiety at the -1 site predominantly adopts a conformation somewhere between the stable 4C1 chair and 4H5 half-chair in the Michaelis complex (Figure S5). The crystal structure of a thiocellobiose-bound GH3 β-D-glucan glucohydrolase isoenzyme ExoI (PDB ID: 1IEX), which has the same -1 site residues as HjCel3A, also did not show significant substrate distortion from the 4C1 conformation.48 This suggests that GH3 may be different from other GH families, wherein the Michaelis complex is often characterized by substrate distortion, usually to a boat or skew-boat conformation, as established by X-ray crystallography9,

49

and QM/MM

calculations.50-52 The resulting structural and electronic changes purportedly pre-activate the


11


Page 12 of 33

substrate for proton transfer and nucleophilic attack.51 The conformation transforms to E3 envelope in the TS before relaxing again to 4C1 in the GEI (Figure S5). In comparison, the crystal structure of ExoI with the TS mimic anilinomethyl glucoimidazole (AmGlcIm) (PDB ID: 1X39) showed a 4E conformation, which was postulated to be a precursor to the TS conformation.53 No substrate distortion was observed in the crystal structure of the 2-deoxy-2-fluoro-α-Dglucopyranosyl (2F-DNPG)-ExoI complex (PDB ID: 1IEW), which represents the GEI.48 In the Michaelis complex of HjCel3A, D61 is hydrogen bonded with the 4- and 6-hydroxyl, K158 with the 4-hydroxyl, R125 with the 2- and 3-hydroxyl, D236 with the 2-hydroxyl, and S384 with the 3-hydroxyl (Figure 3A, Table S1). These hydrogen bond interactions are similar to those observed in the thiocellobiose-ExoI crystal structure, except for S384, which is absent in ExoI. The side chain of the positively charged H159 has moved away from its position in the -1 site observed in the crystal structure (Figure 3A, Movie S1);6 this is presumably favorable to the protonated form of E441. The distance between the anomeric carbon and D236 carboxylate oxygen is 3.3 ± 0.1 Å, while that between the glycosidic oxygen and E441 proton is 1.8 ± 0.1 Å (Figure S6). In the TS, the D236 carboxylate oxygen is still 2.5 ± 0.1 Å from the anomeric carbon. However, the calculated BIs show that the HE2–O4' bond is nearly formed (0.64 ± 0.02), while the glycosidic bond is nearly broken (0.18 ± 0.03). The anomeric carbon has a positive charge of 0.5 balanced by the ring oxygen charge, and the overall charge of the -1 site glycosyl is 0.78 (Table S10). This, along with C1–O5 having a partial double bond character (BI = 1.34 ± 0.02), is consistent with the expected oxocarbenium ion-like character of the TS (Scheme 1). As is generally the case with retaining β-glycosidases, the cationic glycosylation TS of GH3 βgl is stabilized by (1) delocalization of lone pair electrons from the ring oxygen, facilitated by the E3 conformation wherein the C2, C1, O5, and C5 atoms are coplanar,11, 52 and (2) hydrogen bond


12

Page 13 of 33 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


Figure 3. Representative structures of the Michaelis complex (MC), transition state (TS), and product of (A) glycosylation, (B) hydrolysis, and (C) transglycosylation. Only the proton transferred from and to E441 is shown for clarity. Atoms treated at the SCC-DFTB level are in green ball-and-stick representation. Hydrogen bond interactions with active site residues (gray sticks) are represented by dashed lines. Interatomic distances between reacting atoms are given in Å. interaction between the 2-hydroxyl, rendered more acidic by the developing positive charge on the ring, and the nucleophile.54 On the other hand, the interaction of the -1 site glycosyl with R125 is broken in the TS as the residue, instead, forms hydrogen bonds with the now negatively


13


Page 14 of 33

charged E441 (Table S2). H159 moves back to its position in the -1 site, but without forming any long-lived hydrogen bond interaction with the glycosyl moiety (Figure 3A, Movie S1). In the AmGlcIm-ExoI crystal structure, the -1 site glycosyl has hydrogen bonds with the corresponding arginine (R158) and histidine (H207) residues, which along with the absent serine residue, could explain the difference in puckering conformation. Hrmova et al. hypothesized that Y253 (Y204 in HjCel3A) might play a role in TS development through its hydrogen bond with the 6-hydroxyl group of the +1 site glycosyl moiety;53 however, such interaction is not observed in the HjCel3A glycosylation TS as Y204 is, instead, hydrogen bonded to W237. The resulting GEI has a C1– OD1 bond distance of 1.52 ± 0.04 Å (Figure S6). At the -1 site, hydrogen bonds with D236, S384, D61, and K158 are observed, but not with R125 and H159 unlike the 2F-DNPG-ExoI crystal structure.48 The glucose product in the +1 site is hydrogen bonded to Y204, R67, and R169 (Figure 3A, Table S3). Hydrolysis and transglycosylation. Potential energy surface. In the subsequent deglycosylation step, the GEI is hydrolyzed to release glucose. The 2D PES shows that the water oxygen forms a bond with the anomeric carbon while the water is hydrogen bonded to E441 (RYH = 1.6 Å). Proton transfer to the E441 carboxylate oxygen begins only once RXH reaches 1.5 Å. The 1D potential energy curve, using the sum of the two distances as reaction coordinate (RH), roughly follows this MEP (Figure 2B). Umbrella sampling simulations along RH show that the hydrogen bond between water and E441 shortens to ~1.4–1.5 Å just before C1–OH2 bond formation at the TS, then lengthens again to ~2.1 Å. This transition occurs at RH = 3.7–3.8 Å and, as shown in Figure S1B, sampling is low at these reaction coordinate values because the system preferentially samples RH = 3.9 Å, where d(C1-OH2) ≈ 2.4 Å and d(H2-OE2) ≈ 1.5 Å, and RH = 3.6 Å, where d(C1-OH2) ≈ 1.5 Å and d(H2-OE2) ≈ 2.1 Å. The water molecule


14

Page 15 of 33 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


subsequently transfers its proton to E441 after the C1–OH2 bond is formed (Movie S2). The Michaelis complex, TS, and product are located at RH = 5.1, 3.6, and 2.4 Å, respectively. Hydrolysis has a free energy barrier of 18.2 kcal/mol and is exothermic by 5.5 kcal/mol (Figure S3). In the presence of another acceptor that can compete with water (in this case, glucose), transglycosylation occurs. The 2D PES shows that, in contrast to the hydrolysis mechanism, proton transfer from the 4-hydroxyl to E441 (RYT) occurs concurrently with bond formation between the O4' atom and anomeric carbon (RXT). A higher energy TS, in which the proton has already been transferred, is obtained from the 1D potential energy curve using RT as reaction coordinate; otherwise, the resulting curve is similar to the MEP (Figure 2C). The umbrella sampling simulations were nevertheless performed using RT since the TS energy was expected to decrease upon equilibration of the system. The distance between the proton (HO4') and E441 carboxylate oxygen is observed to vary from 1.4 to 1.7 Å around the TS, contrary to the initial structures obtained from the 1D potential energy curve (Movie S3). The calculated barrier (18.1 kcal/mol) and location of the stationary points (RT = 5.1, 3.5, and 2.5 Å, for the Michaelis complex, TS, and product, respectively) of transglycosylation are similar to that of hydrolysis. On the other hand, the reaction energy is more positive at -4.7 kcal/mol (Figure S4). Characterization of stationary points. Without glucose in the +1 site and water as the only acceptor, the -1 site glycosyl in the Michaelis complex only has hydrogen bonds at the 3hydroxyl (S384 and H159) (Figure 3B, Table S4). The distance between the water oxygen and anomeric carbon is 3.3 ± 0.1 Å, while the hydrogen bond distance between the water proton and E441 carboxylate oxygen is 1.8 ± 0.1 Å (Figure S6). Short-lived bridging hydrogen bond interactions (

Hydrolysis and Transglycosylation Transition States of Glycoside

Recommend Documents