Discovery of processive catalysis by an exo-hydrolase with a pocket-shaped active site

Substrates associate and products dissociate from enzyme catalytic sites rapidly, which hampers investigations of their trajectories. The high-resolution structure of the native Hordeum exo-hydrolase HvExoI isolated from seedlings reveals that non-covalently trapped glucose forms a stable enzyme-product complex. Here, we report that the alkyl β-d-glucoside and methyl 6-thio-β-gentiobioside substrate analogues perfused in crystalline HvExoI bind across the catalytic site after they displace glucose, while methyl 2-thio-β-sophoroside attaches nearby. Structural analyses and multi-scale molecular modelling of nanoscale reactant movements in HvExoI reveal that upon productive binding of incoming substrates, the glucose product modifies its binding patterns and evokes the formation of a transient lateral cavity, which serves as a conduit for glucose departure to allow for the next catalytic round. This path enables substrate-product assisted processive catalysis through multiple hydrolytic events without HvExoI losing contact with oligo- or polymeric substrates. We anticipate that such enzyme plasticity could be prevalent among exo-hydrolases.

of G6SG-OMe aligned with the indole ring of Trp286 and Trp434 such that the intra-ring oxygen of the reducing-end Glc moiety (hydrophilic face) faced Trp286, while that of the non-reducing-end moiety (hydrophobic face) pointed to Trp434. The loop carrying Glu491 to Asn498 adopted a new position, compared to that in ligand-free recombinant HvExoI. This productive binding of G6SG-OMe was emphasised by the presence of the hydrogen bond (H-bond) of 2.7 Å formed between C2-OH of non-reducing Glc and Oδ2 of Asp285 (catalytic nucleophile), and by the water-mediated contact formed between Oε2 of Glu491 (acid/base) and S1 of G6SG-OMe.

Reciprocal docking of β-D-glucopyranosyl-(1,2)-D-glucose (G2OG) or Glc to obtain ternary
HvExoI:Glc:G2OG complexes. G2OG or Glc was docked in the HvExoI:Glc complex to obtain ternary HvExoI:Glc:G2O complexes 1 and 2, which we considered to be structural intermediates during the Glc displacement route. To obtain complex 1, Glc was docked in the -1 subsite of the HvExoI:G2OG complex, derived from the crystal structure of HvExoI in complex with G2SG-OMe bound with the non-reducing glucosyl moiety in the +1 subsite, while the reducing moiety remained solvent exposed at the putative +2 subsite (score 58) (Supplementary Fig. 6; right-top panel). In this stable complex after 90 ns of MD simulations ( Supplementary Fig. 5), Tyr253 changed its conformation, although to the other side compared to that in complex 1, making the exposed lateral cavity next to the -1 subsite shallower (Supplementary Fig. 6; right-bottom panel).

Supplementary Note 4
Exploration of alternative binding sites by MD simulations and docking. We reasoned that incoming substrates could potentially bind to different sites than those at the +1 and putative +2 subsites, described in complexes 1 and 2 above, and to trigger Glc displacement form the -1 subsite through the +1 subsite, which could only then be occupied by a new substrate. To investigate this possibility, MD simulations of binary HvExoI:Glc complexes revealed that Glc stayed bound in the active site for the entire 200 ns simulation time ( Supplementary Fig. 5), while Trp434 changed its conformation that affected the space available in the +1 subsite ( Supplementary Fig. 7b, d). Most of the time (>75%) Trp434 tilted perpendicularly with respect to Trp286 that barely moved along simulations, and partially blocked the space assigned to Glc' (PDB 3WLH), or to Glc2 in recombinant HvExoI perfused with Glc (PDB 3WLO) ( Supplementary Fig. 7b, d; Fig. 3). The CA-CB-CG-CD1 dihedral angle of Trp434 at approximately -50° indicated the closed conformation of Trp434, which could establish various interactions. One of most stable ones was the H-bond between HE2 of Trp434 and O6 of Glc, formed 15 ns from the simulation onset that was retained for 100 ns.
The second, a minor conformation observed for Trp434 with the CA-CB-CG-CD1 dihedral angle of around 50° ( Supplementary Fig. 7c, e), corresponded to that seen in the crystal structure (PDB 3WLH), with the indole ring parallel to that of Trp286 and HE2 of Trp434 interacting with the catalytic acid/base Glu491 ( Supplementary Fig. 7c, e). Next, docking of G2OG and β-Dglucopyranosyl-(1,3)-D-glucose (G3OG) ( Supplementary Fig. 6, 8, complexes 3-6) performed on the MD simulation snapshot described above ( Supplementary Fig. 7b, d) aimed to shed light on identifying substrate binding sites other than those of the +1 and putative +2 subsites. These docking solutions described below, identified a solvent-exposed lateral cavity for potential substrate binding ( Supplementary Fig. 6, 8).
Complex 3: This complex presents an open and the most frequent solvent exposed conformation of Trp434, providing this residue was set to be flexible during docking calculations. G3OG bound (score 68) with its non-reducing-end in the lateral cavity and the reducing-end in the +1 subsite (Supplementary Fig. 8; left-top panel). However, within first 4 ns of the MD simulation, G3OG positioned its non-reducing-end in the +1 subsite and interacted with solvent exposed Trp434. After around 40 ns, Trp434 changed its conformation to that observed in the native structure with trapped Glc (PDB 3WLH), while G3OG bound with its non-reducing-end in the +1 subsite and the reducingend solvent exposed, as seen in the structure with G2SG-OMe (PDB 6MD6) (Supplementary Fig.   8; left-middle panel). After around 2 ns, Tyr253 changed its conformation to that observed in complex 1. These re-arrangements remained stable up to 125 ns of the simulation (Supplementary However, during equilibration G2OG abandoned this lateral cavity and remained bound at the protein surface near Trp434, which blocked the +1 subsite, and for 100 ns interacted with Glu36 (Supplementary Fig. 8; right-middle panel). After 104 ns Trp434 changed its conformation to that observed in crystal structures, while G2OG entered the +1 subsite with its non-reducing-end first adopting a binding mode analogous to that observed in the crystal structure with G2SG-OMe (PDB 6MD6), and the Tyr253 sidechain rotated as seen in complexes 1 and 3. This system remained stable up to 150 ns of the simulation run (Supplementary Fig. 8; right-bottom panel).
Complex 5: The Trp434 sidechain presents an open and solvent exposed conformation. This docking solution (score 64) was selected as G2OG bound next to the lateral cavity, thus leaving the +1 subsite unoccupied, so that it could be used as the potential displacement route. However, this was not a stable binding mode of G2OG, which during the first 10 ns of simulation was translated into the +1 subsite and subsequently rapidly moved into the solvent ( Supplementary Fig. 5). Trp434 ended up blocking the +1 subsite, as it was seen during HvExoI:Glc simulations ( Supplementary   Fig. 7b, d).
Complex 6: Although Trp434 could change its conformation in this complex, which is one of the lowest energy solutions (score 75), Trp434 partially blocked the +1 subsite. G3OG bound initially to the lateral cavity, however further progression of MD simulations revealed that this was not a stable binding mode, as G3OG moved away in less than 10 ns ( Supplementary Fig. 5). Trp434 remained blocking the +1 subsite for the next 40 ns of the MD simulation run. a, Native HvExoI with trapped Glc. The Glc molecule (Glc and Glc', carbons: yellow sticks) at 0.5 occupancy oscillates between the -1 and +1 subsites. b, Native HvExoI perfused with 3dGlc. Glc (carbons: yellow sticks) and 3dGlc (carbons: orange sticks) at 1.0 occupancies are bound in -1 and +1 subsites. c, Native HvExoI perfused with octyl-O-Glc. Octyl-O-Glc (carbons: orange sticks) at 1.0 occupancy is bound across -1 and +1 subsites. d, Native HvExoI perfused with PEG (n=5-10). Two PEG molecules (PEG 1 in alternate conformations at occupancies 0.5, PEG 2 at occupancy 1.0) (carbons: orange sticks) are bound at the +1 and putative +2 subsites. Molecular surface morphologies are coloured by electrostatic potentials: white, neutral; blue, +5 kT·e -1 ; red, -5 kT·e -1 . Grey, red, and blue represent carbon, oxygen and nitrogen atoms, respectively. Water molecules are shown as red or magenta (alternate water molecules) spheres in complexes with Glc and PEG molecules. Residues are marked in top left panel only.  a, Conservation of surface residues of the HvExoI structure (left) (PDB 3WLH), and a detail of the active site with 11 residues (right: sticks coloured by confidence interval colour -CIC). b, Conservation of surface residues of the HvExoI:G3OG complex (left) (cf. Fig. 6h), and a detail of the lateral cavity with 14 residues (right; sticks coloured by CIC). c, Conservation scores 4 of residues in the active site (left), and in the lateral cavity (right). Scales of conservation represent the lower and upper bounds of intervals, where burgundy and turquois are the extremes of conservation on the scale 9=conserved and 1=variable, respectively. The HvExoI structure (PDB 3WLH) was used as a search parameter to identify 500 sequences with 35-95% sequence identity to the HvExo1 sequence, amongst them putative pro-and eukaryotic entries.  Separations from the Glc geometric centre in the -1 subsite are given. The analysis was performed with the FindHBond option of UCSFChimera, which uses a set of geometric criteria that consider different donor-acceptor combinations and are based on a survey of small-molecule crystal structures 5 .

Donor
Acceptor Hydrogen