How a Lytic Polysaccharide Monooxygenase Binds Crystalline Chitin

Abstract

Lytic polysaccharide monooxygenases (LPMOs) are major players in biomass conversion, both in Nature and in the biorefining industry. How the monocopper LPMO active site is positioned relative to the crystalline substrate surface to catalyze powerful, but potentially self-destructive, oxidative chemistry is one of the major questions in the field. We have adopted a multidisciplinary approach, combining biochemical, spectroscopic, and molecular modeling methods to study chitin binding by the well-studied LPMO from Serratia marcescens SmAA10A (or CBP21). The orientation of the enzyme on a single-chain substrate was determined by analyzing enzyme cutting patterns. Building on this analysis, molecular dynamics (MD) simulations were performed to study interactions between the LPMO and three different surface topologies of crystalline chitin. The resulting atomistic models showed that most enzyme-substrate interactions involve the polysaccharide chain that is to be cleaved. The models also revealed a constrained active site geometry as well as a tunnel connecting the bulk solvent to the copper site, through which only small molecules such as HO, O, and HO can diffuse. Furthermore, MD simulations, quantum mechanics/molecular mechanics calculations, and electron paramagnetic resonance spectroscopy demonstrate that rearrangement of Cu-coordinating water molecules is necessary when binding the substrate and also provide a rationale for the experimentally observed C1 oxidative regiospecificity of SmAA10A. This study provides a first, experimentally supported, atomistic view of the interactions between an LPMO and crystalline chitin. The confinement of the catalytic center is likely crucially important for controlling the oxidative chemistry performed by LPMOs and will help guide future mechanistic studies.

Characterization and synergistic action of a tetra‐modular lytic polysaccharide monooxygenase from Bacillus cereus

“…It is intriguing that some LPMOs, such as SmLPMO10A, perform seemingly well, without a CBM. For example, in the experimental set-up used here, SmLPMO10A performed better than any of the truncated BcLPMO10A variants, even though residues known to be involved in substrate binding from studies on SmLPMO10A [29,35,40] are highly conserved in BcLPMO10A (Fig. S6).…”

Molecular mechanism of the chitinolytic peroxygenase reaction

“…The starting model for investigating the SmAA10A reaction mechanism was taken from an experimentally informed enzyme-b-chitin model (~150,000 atoms, Figure S4) that had previously been equilibrated for 272 ns. 13 Force field parameters for H2O2 were derived using Paramfit 14 and Gaussian 09 15 (Table S1) while the Cu(I)-histidine-brace force field parameters were taken from a previous study. 13 A water molecule ~3 Å away from the Cu-atom was replaced by a H2O2 molecule and the new model was equilibrated applying the following procedure.…”

“…A snapshot from the QM/MM/MD simulation taken before H2O2 reacted with Cu(I) was minimized using the same AMBER parameters as listed above and by applying the three step QM/MM minimization scheme described previously. 13 The minimized system, containing an intact H2O2 molecule ~3 Å from the copper ion, was truncated from ~150,000 to ~24 000 atoms, only keeping water, NAG units, and amino acid residues closer than 40 Å from the copper ion ( Figure 3A). This initial truncated model was subjected to geometry optimization using ChemShell 22 in combination with ORCA, selecting an extended QM/MM-region ( Figure S5B) and a large active-region (the part of the model that is allowed to move) that allowed the active site environment to relax ( Figure S5C).…”

“…To investigate how accessible the confined reaction cavity is to small molecules such as H2O2, and to assess the energetics associated with the transport process, biased MD simulations (umbrella sampling) were carried out. Two start models were generated from the previously equilibrated SmAA10A-Cu(I)-H2O2-b-chitin complex containing H2O2, one with H2O2 located at the entrance of the tunnel suggested by Bissaro et al, 13 and one with H2O2 in the reaction cavity. A biasing potential of 5.0 kcal mol -1 Å -2 was applied in all simulations, and when necessary (see Figure S23) resampling was carried out using a biasing potential of 10.0 kcal mol -1 Å -2 .…”

“…We have previously revealed the existence of a water tunnel, formed in the enzyme-polysaccharide complex interspace, connecting bulk solvent to the monocopper active site. 13 Here, we investigated the energetics associated with H2O2 diffusion into the active site by MD simulations combined with umbrella sampling, estimating free energy profiles of H2O2 moving in this interspace. Different from our previous work, 13 where the tunnel and its accessibility were assessed primarily using steric criteria , the methodology applied here is founded on assesing non-covalent interactions between H2O2 and the enzyme-substrate complex in a dynamic environment.…”