Structural and Functional Analysis of a Lytic Polysaccharide Monooxygenase Important for Efficient Utilization of Chitin in Cellvibrio japonicus

Abstract

Cellvibrio japonicusis a Gram-negative soil bacterium that is primarily known for its ability to degrade plant cell wall polysaccharides through utilization of an extensive repertoire of carbohydrate-active enzymes. Several putative chitin-degrading enzymes are also found among these carbohydrate-active enzymes, such as chitinases, chitobiases, and lytic polysaccharide monooxygenases (LPMOs). In this study, we have characterized the chitin-active LPMO,CjLPMO10A, a tri-modular enzyme containing a catalytic family AA10 LPMO module, a family 5 chitin-binding module, and a C-terminal unclassified module of unknown function. Characterization of the latter module revealed tight and specific binding to chitin, thereby unraveling a new family of chitin-binding modules (classified as CBM73). X-ray crystallographic elucidation of theCjLPMO10A catalytic module revealed that the active site of the enzyme combines structural features previously only observed in either cellulose or chitin-active LPMO10s. Analysis of the copper-binding site by EPR showed a signal signature more similar to those observed for cellulose-cleaving LPMOs. The full-length LPMO shows no activity toward cellulose but is able to bind and cleave both α- and β-chitin. Removal of the chitin-binding modules reduced LPMO activity toward α-chitin compared with the full-length enzyme. Interestingly, the full-length enzyme and the individual catalytic LPMO module boosted the activity of an endochitinase equally well, also yielding similar amounts of oxidized products. Finally, gene deletion studies show thatCjLPMO10A is needed byC. japonicusto obtain efficient growth on both purified chitin and crab shell particles.

Keywords: Cellvibrio japonicus; carbohydrate-binding module (CBM); cellulose; chitin; chitinase; electron paramagnetic resonance (EPR); gene knockout; lytic polysaccharide monooxygenase (LPMO); x-ray crystallography.

FIGURE 1.

Sequence analysis of CjLPMO10A. A, domain architecture of CjLPMO10A. The full-length enzyme contains a signal peptide (SP: residues 1–36) that is cleaved off during secretion resulting in the mature enzyme that possesses an N-terminal AA10-type LPMO domain followed by a family 5 chitin-binding module (CBM5) and a C-terminal CBM classified as CBM73 (see text). The three domains are separated by two poly-serine linkers (PSL). The modules and linkers are scaled according to the number of amino acids they contain. B, phylogeny of selected LPMOs from auxiliary activities family 10 built on a previous classification by Book et al. (62), using Phylogeny.fr (64). Enzyme names include additional domains and other relevant sequence motifs, but the phylogeny is based on the catalytic domains only. The indicated clades are as defined in the study by Book et al. (62). Clade I (subclade C and D) contains chitin-oxidizing LPMOs and Clade II (subclade A and B) contains cellulose-oxidizing LPMOs as well as membrane-associated LPMOs with unknown function. Blue colored protein names represent enzymes that have been characterized.

FIGURE 2.

Multiple sequence alignment of CBM5s and CBM73s. The five CBM5 sequences originate from C. japonicus LPMO10A, Chi18C, and Chi18D, Cellvibrio mixus LPMO10, and S. griseus Chi18C. The five CBM73 sequences originate from C. japonicus LPMO10A, Chi18B, Chi18C, Chi18D and C. mixus LPMO10. The pink stars indicate two aromatic residues (YW or WW) for which experiments have shown that they are important for substrate binding by CBM5s (63).

FIGURE 3.

Three-dimensional structure of CjLPMO10A^cd. A, cartoon representation of the CjLPMO10A^cd secondary structure. α-Helices are shown in blue, and β-strands are shown in yellow. The copper ion is shown as an orange sphere, and the coordinating histidine side chains (His-37 and His-136) are shown as gray sticks. The dotted circle shows the protrusion that includes most of the structural diversity in LPMOs. B, surface projection of the substrate binding surface in CjLPMO10A^cd, chain B. The main chain is shown in cartoon representation, and side chains are shown as sticks. The copper ion and selected solvent water molecules are shown as orange and red spheres, respectively. Surface-exposed residues are labeled using the single letter amino acid code, and those discussed in the main text are marked with bold letters (E56, Q78, E79, T133, R197, D202, and E205). Glu-79 and Thr-76 show different rotamers in chain A (yellow sticks) compared with chain B and C (chain B; gray sticks).

FIGURE 4.

Copper active sites of C1-oxidizing LPMO10s. A, superposition of the CjLPMO10A copper active site with chitin-active Enterococcus faecalis LPMO10A (PDB 4ALC (67)) and cellulose-active S. coelicolor LPMO10C (PDB 4OY7 (15)). Carbon atoms in the three enzymes are colored yellow (EfLPMO10A), green (ScLPMO10C), and gray (CjLPMO10A). Side chains are colored according to the following properties: residues fully conserved in all LPMO10s (purple), residues highly conserved among chitin- and cellulose-specific LPMO10s (yellow and green, respectively), and residues that are not conserved (gray). Residues that are either fully or not conserved are numbered according to the CjLPMO10A sequence. Note the similar spatial location of Glu-64 in EfLPMO10A and Glu-205/Glu-217 in the other two enzymes. B, sections of a structural sequence alignment related to the active sites in four chitin-oxidizing LPMO10s (SmLPMO10A, PDB 2BEM; BaLPMO10A, PDB 2YOY; EfLPMO10A, PDB 4ALC; and CjLPMO10A, PDB 5FJQ) and one cellulose-active LPMO10 (ScLPMO10C, PDB 4OY7); the sequence of cellulose-oxidizing T. fusca LPMO10B, for which no structure is known, was added to the structure-based alignment. The top arrows illustrates on which β-strands the residues are found. The green and yellow color indicates similarity to cellulose-active ScLPMO10C and chitin-active EfLPMO10A, respectively. The WDR insertion in CjLPMO10A is located ∼15 Å from the catalytic center, and its side chains are not part of the presumed substrate-binding surface shown in Figs. 3B and 4A. C, active site of CjLPMO10A (chain B). D, active site of the copper-soaked AA10 from Melolontha melolontha entomopoxvirus (PDB 4X27 (68)) which shows 27% overall sequence identity with CjLPMO10A.

FIGURE 5.

Electron paramagnetic resonance spectroscopy of Cu(II)-saturated CjLPMO10A^cd. A, X-band EPR spectra (—) with simulation (- - -) for Cu(II)-saturated CjLPMO10A (top) and Cu(II) in Pipes buffer, pH 6.5. B, comparison of EPR signals of CjLPMO10A, ScLPMO10C (cellulose-active), and SmLPMO10A (chitin-active). The EPR spectra were recorded at 30 K using a microwave power of 0.5 milliwatts.

FIGURE 6.

Binding of the CjLPMO10A domains to chitin and cellulose. A–C show binding of the catalytic LPMO domain (A), the CBM5 domain (B), and the C-terminal domain (C) to chitin and cellulose substrates. The percentage of free protein was determined by measuring the reduction in protein concentration (A₂₈₀) over time, in the absence of an electron donor. The experiments were carried out at 22 °C using 10 mg/ml substrate (α-chitin, β-chitin, Avicel, or filter paper) in 50 mm sodium phosphate buffer, pH 7.0. D, plots of binding data for the CBM5 (■) and the C-terminal (CBM73) domain (●) incubated with α-chitin. P_bound corresponds to bound protein (μmol/g substrate), and P_free corresponds to non-bound protein (μm). Each point represents the average of values obtained in three independent experiments.

FIGURE 7.

Activity of full-length and truncated CjLPMO10A. A, chromatographic analysis of oxidized chito-oligosaccharides from degradation reactions containing 10 mg/ml α-chitin (top chromatograms) or β-chitin (middle chromatograms) after 24 h of incubation with full-length (solid lines) or truncated (dotted lines) CjLPMO10A, with standards for oxidized chito-tetraose (DP4ox) and chito-pentaose (DP5ox). B, MALDI-TOF MS analysis of products generated by the full-length LPMO acting on α-chitin. Prior to the analysis, the samples were saturated with sodium, to simplify the spectra. For each oligomeric product, two major adducts are observed: [DP4–8ox + Na]⁺ and [DP4−8ox−H + 2Na]⁺. All the reactions were carried out in 20 mm Bistris propane buffer, pH 7.2, with 0.5 μm Cu²⁺-saturated LPMO and 1 mm ascorbate as external electron donor.

FIGURE 8.

Synergy between CjLPMO10A and the SmChi18C endochitinase in the degradation of α-chitin. A and B show quantification of GlcNAc and chitobionic acid (GlcNAcGlcNAc1A), respectively, obtained after chitobiase digestion of soluble products from the degradation of 10 mg/ml α-chitin. C, peak areas for chitobionic acid from solubilized material only (gray) and from the complete reaction mixture (black) products after degradation of 2 mg/ml α-chitin using full-length or truncated CjLPMO10A. The total oxidized sugar content was measured after full solubilization of the LPMO-pretreated chitin substrate by the S. marcescens chitinases Chi18A, Chi18C, and chitobiase (CHB). All reactions were carried out at 37 °C in Bistris propane buffer, pH 7.2, in triplicate with 0.5 μm Cu²⁺-saturated LPMO and/or 0.5 μm SmChi18C and in the presence of 1 mm ascorbate as external electron donor.

FIGURE 9.

Growth of C. japonicus wild type and mutants on various substrates. Wild type and mutant strains were grown in MOPS defined medium with glucose (A), N-acetylglucosamine (B), squid pen β-chitin (C), or unprocessed crab shell (α-chitin) (D) as the sole carbon source. Measurement of growth used optical density (OD) at 600 nm. All experiments were run in biological triplicate with error bars showing standard deviation. Strains presented are wild type (closed circles), general secretory protein mutant (closed squares), CjLPMO10A mutant (open diamonds), and CjLPMO10B (open inverted triangles).