Computational remodeling of an enzyme conformational landscape for altered substrate selectivity

Abstract

Structural plasticity of enzymes dictates their function. Yet, our ability to rationally remodel enzyme conformational landscapes to tailor catalytic properties remains limited. Here, we report a computational procedure for tuning conformational landscapes that is based on multistate design of hinge-mediated domain motions. Using this method, we redesign the conformational landscape of a natural aminotransferase to preferentially stabilize a less populated but reactive conformation and thereby increase catalytic efficiency with a non-native substrate, resulting in altered substrate selectivity. Steady-state kinetics of designed variants reveals activity increases with the non-native substrate of approximately 100-fold and selectivity switches of up to 1900-fold. Structural analyses by room-temperature X-ray crystallography and multitemperature nuclear magnetic resonance spectroscopy confirm that conformational equilibria favor the target conformation. Our computational approach opens the door to targeted alterations of conformational states and equilibria, which should facilitate the design of biocatalysts with customized activity and selectivity.

Fig. 1. AAT conformational landscape.

a E. coli AAT is a 90 kDa homodimer that undergoes a conformational change from an open (green, PDB ID: 1ARS) to closed (dark blue, PDB ID: 1ART) state upon substrate binding. This conformational transition involves rotation of a small moving domain (colored) relative to a fixed domain (white and gray for chains A and B, respectively), which causes a 2.4 ± 0.4 Å displacement (mean C_α distance ± s.d.) of the helix formed by residues K355–F365 (indicated by an asterisk). The PLP cofactor bound at the active site is shown as spheres (salmon). b Hinge movement analysis of chain A reveals a 7.1-degree rotation of the moving domain relative to the fixed domain along an axis between two planes (dotted line). Hinge-bending residues and PLP are shown as orange spheres and white sticks, respectively. c Superposition of HEX structures in the absence (yellow, PDB ID: 1AHE) and presence (blue, PDB ID: 1AHY) of bound inhibitor with that of the WT closed state (white, PDB ID: 1ART) show that this mutant is closed in both cases. C_α displacement (mean ± s.d.) of residues K355–F365 (asterisk) is indicated. d These results demonstrate that the six mutations (mut.) of HEX remodel its conformational landscape to favor the closed conformation.

Fig. 2. Computational remodeling of AAT conformational landscape by multistate design.

To remodel the AAT conformational landscape, we followed a 5-step process: (1) identification of hinge-bending residues involved in transition between open (green) and closed (dark blue) conformational states; (2) generation of structural ensembles approximating backbone flexibility to model open and closed states; (3) optimization of rotamers for mutant sequences on both open- and closed-state ensembles; (4) calculation of energy differences between conformational states (ΔE = E_closed‒E_open) to predict equilibrium of each mutant, and (5) combinatorial library design using ΔE values to generate Closed, Open_Low, and Open_High libraries for experimental testing. Designed residues V35, K37, T43, and N64 correspond to V39, K41, T47, and N69 in the previously published crystal structures of wild-type AAT (PDB ID: 1ARS and 1ART).

Fig. 3. Crystal structures.

Overlay of AAT structures (Chain A) in the presence and absence of maleate show transition from open to closed states upon inhibitor binding for WT, VFCS, and AIFS, but not for HEX, VFIT, and VFIY, which already adopt the closed conformation in the absence of bound inhibitor. Average displacements of helix formed by residues K355–F365 upon maleate binding are reported as the average pairwise distance of corresponding C_α atoms for the 11 residues comprising this helix (mean ± s.d.).

Fig. 4. Conformational landscape analysis by NMR.

a To evaluate conformational equilibrium of AAT variants, we introduced the fluorinated amino acid 4-trifluoromethyl-l-phenylalanine at position F217 (orange), which is located closer to hinge-bending and designed residues (green) than to the bound maleate inhibitor (blue). Crystal structure shown is that of wild-type (WT) AAT at 278 K (PDB ID: 8E9K). b ¹⁹F NMR spectra of AAT variants in the absence of ligand show dynamic equilibrium between 278 K and 308 K for HEX, VFIT, VFIY, and AIFS, confirming that these proteins are undergoing exchange. This is not the case for WT and Open_Low library mutant VFCS, who both adopt predominantly the open conformation within this temperature range. For HEX and Closed library mutants VFIT and VFIY, the open conformation is enriched as temperature increases. For AIFS, spectra suggest that this Open_High library mutant samples conformations distinct from those sampled by the other variants.