Molecular Dynamics Gives New Insights into the Glucose Tolerance and Inhibition Mechanisms on β-Glucosidases

Abstract

β-Glucosidases are enzymes with high importance for many industrial processes, catalyzing the last and limiting step of the conversion of lignocellulosic material into fermentable sugars for biofuel production. However, β-glucosidases are inhibited by high concentrations of the product (glucose), which limits the biofuel production on an industrial scale. For this reason, the structural mechanisms of tolerance to product inhibition have been the target of several studies. In this study, we performed in silico experiments, such as molecular dynamics (MD) simulations, free energy landscape (FEL) estimate, Poisson-Boltzmann surface area (PBSA), and grid inhomogeneous solvation theory (GIST) seeking a better understanding of the glucose tolerance and inhibition mechanisms of a representative GH1 β-glucosidase and a GH3 one. Our results suggest that the hydrophobic residues Y180, W350, and F349, as well the polar one D238 act in a mechanism for glucose releasing, herein called "slingshot mechanism", dependent also on an allosteric channel (AC). In addition, water activity modulation and the protein loop motions suggest that GH1 β-Glucosidases present an active site more adapted to glucose withdrawal than GH3, in consonance with the GH1s lower product inhibition. The results presented here provide directions on the understanding of the molecular mechanisms governing inhibition and tolerance to the product in β-glucosidases and can be useful for the rational design of optimized enzymes for industrial interests.

Keywords: GH1; GH3; Poisson–Boltzmann surface area; allosteric channel; free energy landscape; glucose tolerance; grid inhomogeneous solvation theory; molecular dynamics simulation; slingshot mechanism; β-Glucosidases.

Figure 1

Two-dimensional root-mean-square deviation (RMSD) plot for the combined five trajectories (60 ns each one) for each one of the starting systems: (A) GH1 (HiBG)–Cellobiose complex, (B) GH1–glucose complex with glucose starting from the crystallographic pose, (C) GH1–glucose complex with glucose starting from the modeled pose, (D) GH1–glucose crystallographic complex (magenta) and the complex starting from the modeled pose (green). (E) GH3 (AaBG)–Cellobiose complex, (F) GH3–glucose complex. RMSD post the alignment of all the frames and considering just the backbone atoms (Cα, C, O, N).

Figure 2

Free energy landscapes (FEL) for cellobiose and glucose at their respective and individual complexes with GH1 and GH3. (A) HiBG complexed with cellobiose, (B) HiBG complexed with glucose. This FEL was depicted considering together the respective GH1–glucose-1 and GH1-glucose-2 sets of simulations. The respective starting points at the simulations GH1–glucose-1 and GH1-glucose-2 are at the regions 3 and 1, (C) AaBG complexed with cellobiose, (D) AaBG complexed with glucose. The distance is relative to the geometric center of each ligand and the catalytic E378 (in GH1) or H121 (in GH3). The angle is formed by these two geometric centers and the geometric center of each protein. (E) Superposition of the ligand poses around all the respective trajectory ensembles for each system. Protein average frames are shown in green (HiBG) and blue (AaBG) cartoons. All frames of ligand positions for cellobiose (orange) and glucose (cyan) were overlapped.

Figure 3

Binding modes for HiBG in complex with glucose of the GH1–glucose FEL (Figure 2B) for the regions 1 (A), 2 (B), 3 (C), 4 (D), 5 (E), and 6 (F). Protein structures are shown in gray, subsite −1 (red sticks), subsite +1/+2 (yellow sticks), D238/W350 (blue sticks), W169 (orange sticks). The poses (selected from the minima at the FEL from Figure 2B) show a probable exit path for glucose from the HiBG active site.

Figure 4

Confront between APBS, and GIST data and the glucose fitting in GH1 and GH3 representative poses at the catalytic cleft. (A,B) APBS respectively for HiBG at the pose representative of the minimum 1 in Figure 2B and AaBG at the pose representative of the minimum 2 in Figure 2C; (C,D) APBS and GIST results respectively for the previous pose of HiBG and AaBG (E,F) APBS and glucose fitting respectively for the previous pose of HiBG and AaBG. APBS scale in red, white, and blue corresponding to ψ values of −20.00:0.00: +20.00, respectively. GIST results are shown in yellow dots for water interacting centers with g^v(O) ≥ 10.00 units of the bulk density and ∆G^v_solv ≥ +3.0 kcal/mol. Glucose is shown in cyan spheres with carbon atoms in cyan, oxygens in red, and hydrogens in white.

Figure 5

Conformations recovered by the FEL profiles of the glucose positioning in GH1 and GH3, colored by the GIST data. (A–F) GIST results and glucose positions occupancy at the respective FEL regions 1–4 and two different samples of region 5 in GH1 in Figure 2B. (G–I) The analog profiles for the FEL respective regions 1–3 in GH3 in Figure 2C. GIST positions are shown in yellow transparent surfaces for water-interacting centers with g^v(O) ≥ 10.00 units of the bulk density and ∆G^v_solv ≥ +3.0 kcal/mol. Glucose, shown in spheres in all the figures, is colored with carbons in cyan, oxygens in red, and hydrogens in white (except for B, which was colored by APBS). Residues involved in the establishment of the hydrophobic cage and regions important for the glucose escaping in GH1 are highlighted. AS: Allosteric site. AC: Allosteric channel. A white asterisk (*) is used to point the site for water exclusion confined by the hydrophobic cage between Y180, F349, and W350. The same figure colored by APBS and GIST is available in the Supplementary Materials (Figure S13).

Figure 6

Structure of the substrate channel of HiGB and the mechanism of glucose release. (A) Five snapshots of the MD simulation that illustrate the slingshot mechanism and residues involved. (B) Structure of the substrate channel organized into four sectors (1–4). A set of hydrophobic residues is located in the entrance of this channel, mainly in the sectors 2, 3, and 4. Circles represent apolar (blue), polar neutral (black), polar positively charged (magenta), and polar negatively charged (green) amino acids. (C–E) Amino acid residues involved in the slingshot mechanism: E167, N236, D238, K257, and N312.

Figure 7

The sequence of molecular events that lead to glucose tolerance at the GH1 enzyme and inhibition at the GH3. (A) Slingshot mechanism influences the glucose tolerance of the HiBG GH1 enzyme. Figure 7A1–8 shows the most probable sequence of events (based on the computational issues here recovered), from the hydrolysis to the final elimination of the non-reducing glucose. (B) Cellobiose cannot so easily escape from the HiBG substrate channel. The sequence of events depicted in B1–4 is similar to the events described for glucose. However, the higher set of synergic interactions involving the protein and the two glucose rings in cellobiose draws back the substrate. (C) At the shallower and more charged pocket from the AaBG GH3 enzyme, the catalytic cleft is more stereochemically fitted to glucose. Besides this, the same constricted cleft tends to retain a higher set of unfavorable water molecules that are more efficiently liberated with the glucose binding. (D) The relatively small cleft of AaBG is less stereochemically suited to cellobiose than to glucose, resulting in higher ligand mobility and less unfavorable water elimination. AC: Allosteric channel. Unfavorable waters (black) are energetically constrained water molecules, favorable waters (yellow) depict water molecules liberated or energetically relaxed, bold countered residue names depict residues involved in interactions at the represented state, transparent residue names are residues not involved in interactions at the represented state.