Main

Wood is the most abundant organic source of biomass on Earth, with an annual production of about 5.64 × 1010 tons of carbon1. Its widespread nature has allowed humans to use it in many contexts, most notably as a building material, owing to its exceptional mechanical properties and resistance to decay. In bio-based industries, the use of wood is taking on a new importance, as it constitutes the most promising source for advanced biofuels and plant-derived products. Notwithstanding its potential, however, the cost-effective conversion of woody feedstocks is limited by a single key factor, the recalcitrance of the lignocellulosic matrix to degradation by enzyme cocktails2. To overcome this recalcitrance, biorefineries utilize energy-demanding pretreatment processes to solubilize the inaccessible biomass components before enzymatic saccharification. The recalcitrant fraction reflects its heteroxylan content, which is known to be particularly resistant to xylanases due to extensive decoration and because these xylans can adopt a flat conformation, with their chains solidly adhering via hydrogen bonds to the surface of cellulose microfibrils3,4. Finding sustainable means to overcome this resistance to degradation is one of the main challenges faced by modern biorefineries. Indeed, the xylan problem is so severe that consideration is being given to engineering energy crops modified to contain fewer recalcitrant xylans5.

In nature, fungi play a vital role in the terrestrial carbon cycle and dominate wood decomposition in boreal forests6. Wood-decaying basidiomycetes classified as white-rot and brown-rot fungi naturally degrade cellulose and hemicelluloses using a large diversity of carbohydrate-active enzymes (CAZymes; http://www.cazy.org)7 and Fenton-type chemistry8. In this context, the understanding of plant cell wall deconstruction was recently overturned by the discovery of LPMO enzymes, which cleave polysaccharides through an oxidative, as opposed to hydrolytic, mechanism9,10,11. Such is their importance, that industrial enzyme mixtures for the conversion of agricultural residues to biofuels now incorporate cellulose-active LPMOs12, helping biorefineries move toward environmental and economic sustainability. Despite the considerable efficiencies that LPMOs have brought to biomass degradation, industrial enzyme cocktails are still unable to degrade woody biomass completely, and there is a major need to identify new enzymes capable of effecting this breakdown. From this perspective, there are three fungal LPMO families (termed AA9, AA11 and AA13 in the CAZy classification)7 that were discovered from genome sequences by virtue of their modular structure, whereby the catalytic LPMO domain is sometimes appended to known substrate-targeting carbohydrate-binding modules (CBMs). Each fungal LPMO family is associated with the oxidative cleavage of distinct polysaccharides, with AA9 acting mainly on cellulose and xyloglucan10, AA11 on chitin13 and AA13 on starch14,15; a solely xylan-acting LPMO is conspicuous by its absence.

Using comparative post-genomic approaches among fungal wood decayers, we identified the existence of a previously unknown family of LPMOs. This new family, to be termed AA14 in the CAZy classification, differs phylogenetically and structurally from the previous AA9, AA10, AA11 and AA13 families. The first characterized members from the white-rot basidiomycete fungus P. coccineus target xylan chains covering wood cellulose fibers, thus unlocking the enzymatic degradation of wood biomass.

Results

Discovery of the AA14 family among fungal wood decayers

The white-rot basidiomycete P. coccineus is an efficient degrader of both hardwood and softwood16. While studying the effect of different types of biomass on P. coccineus growth using transcriptomics and secretomics, we identified a gene encoding a protein of unknown function that was highly upregulated on pine and poplar as compared to the maltose control condition16. The corresponding protein (JGI ID 1372210; GenBank ID KY769370) was secreted only during growth on pine and poplar, suggesting a role in wood decay. A BLAST search against public sequence databases identified more than 300 proteins with significant similarity to KY769370 from P. coccineus, of which many were from well-known saprotrophic fungi. Sequence alignment revealed a conserved N-terminal histidine (Supplementary Fig. 1), commensurate with a copper-binding histidine brace active site10, a hallmark of known LPMOs. A phylogenetic analysis showed that the newly identified sequences strongly cluster together with high bootstrap values and are very distant from AA9, AA10, AA11 and AA13 sequences (Supplementary Fig. 2), thereby defining a new LPMO family designated AA14 in the CAZy database. AA14 members are found in all well-known white-rot (Pleurotus ostreatus, Phanerochaete chrysosporium and Trametes versicolor) and brown-rot (Serpula lacrymans, Coniophora puteana and Postia placenta) basidiomycetes, as well as in some wood-inhabitant ascomycetes within the Xylariaceae and Hypocreaceae families. A slight gene family expansion is observed in wood-decaying basidiomycetes (the average number per species is 3.35 in basidiomycetes and 1.28 in ascomycetes; Fig. 1; Supplementary Data Set 1). None of the AA14 members identified in fungal genomes harbors a CBM, explaining why this family was not previously discovered together with AA11 and AA13 through the 'module walk' approach13,15.

Figure 1: Phylogeny of the AA14 family of LPMOs.
figure 1

A phylogenetic tree of 283 fungi was analyzed for AA14 members. The number of AA14 genes in fungal species (listed in Supplementary Data Set 1) is represented by red bars (the scale is indicated at the bottom of the figure). When available, the mode of wood decay (brown rot or white rot) is specified next to the leaves of the tree. Pictures illustrate the taxonomical diversity of wood-decaying fungi displaying AA14 LPMOs; Genus (order) from top to bottom: Xylaria (Xylariales), Trichoderma (Hypocreales), Nectria (Hypocreales), Tremella (Tremellales), Pycnoporus (Polyporales), Ganoderma (Polyporales), Serpula (Boletales), Gymnopilus (Agaricales). Photo credit: C. Lechat (AscoFrance) and A. Favel (CIRM-CF).

Full size image

Expression and biochemical characterization of PcAA14

Two P. coccineus proteins, PcAA14A (#KY769369) and PcAA14B (#KY769370), displaying 65% sequence identity were produced to high yield in Pichia pastoris, purified to homogeneity and biochemically characterized (Supplementary Table 1; Supplementary Figs. 3 and 4). We confirmed the correct processing of the native signal peptide, which exposed the N-terminal histidine residue at position 1 in the mature polypeptide chain (Supplementary Table 1). MS analyses revealed that both proteins contained ∼1 copper atom per protein molecule, and treatment with EDTA led to partial apo forms (∼0.1 copper atom per protein molecule). PcAA14A and PcAA14B were both able to produce hydrogen peroxide in the presence of ascorbate, cysteine or gallate as electron donors (Supplementary Table 2).

Crystal structure of PcAA14

The structure of PcAA14B was solved by multiple-wavelength anomalous dispersion data recorded at the gadolinium edge, and refined at 3.0 Å resolution. The core of the protein folds into a largely antiparallel immunoglobulin-like β-sandwich (Fig. 2a), a fold globally similar to those seen in LPMOs from other families. The active site of PcAA14B constituted by His1, His99 and Tyr176, forming the canonical histidine brace that is exposed at the surface (Fig. 2b). In contrast to the flat substrate-binding surfaces observed in AA9 LPMOs17, the surface of PcAA14B has a rippled shape with a clamp formed by two prominent surface loops (Supplementary Fig. 5). Both loops are located in the N-terminal half of PcAA14B, and are equivalent to the L2 and L3 loop regions in AA9 LPMOs. Conventionally, the N-terminal part of AA9 LPMOs upstream of the L2 loop region makes up a β-strand segment (single β-strand or a β-hairpin). No equivalent β-strands are found in the PcAA14B structure, which, in contrast, forms loop segments immediately after the N-terminal His residue (Supplementary Fig. 5). The PcAA14B structure also reveals a cysteine (Cys67–Cys90) in the L3-equivalent region, which borders an extension not present in AA9 LPMOs (Supplementary Fig. 5). It is highly interesting to note that the two loops making up the clamp in PcAA14B correspond to modified L2 and L3 loop regions, as these have been shown to be involved in LPMO–substrate interactions17. For AA9 LPMOs a conserved Tyr has been shown to be involved in substrate interactions at the active site surface17. Interestingly, PcAA14B possesses an equally conserved tyrosine residue, Tyr240, at the edge of the substrate-binding surface, albeit located on a different loop region, which could potentially make substrate interactions. Overall the crystal structure of PcAA14B reveals novel features within its putative substrate binding site, which may suggest differences in terms of substrate specificity compared to known LPMOs.

Figure 2: Structure of AA14 LPMO and organization of the copper active site.
figure 2

(a) Overall three-dimensional structure of PcAA14B (ribbon depiction with active site residues shown as sticks under transparent surface). (b) Active site (histidine brace) overlay of Cu-LsAA9A (magenta) and PcAA14B (gold) with the copper atom represented by a sphere. (c) Continuous wave X-band EPR spectrum (9.3 GHz, 165 K) with simulation (red) of PcAA14A. More data are presented in Supplementary Figures 5 and 6.

Full size image

EPR spectroscopic analysis of the copper site of PcAA14

Multi-frequency electron paramagnetic resonance (EPR) analysis was carried out on both PcAA14A and PcAA14B to determine the nature of the copper active site (Fig. 2c; Supplementary Fig. 6). The spin-Hamiltonian parameters (Supplementary Table 3) displayed axial parameters (gx ≈gy < gz) with a d(x2−y2) singly occupied molecular orbital (SOMO), placing the copper active site squarely within a type 2 Peisach–Blumberg classification18. Simulations required the addition of two (I = 1) nitrogen atoms (coupling in the range of 30 to 36 MHz), as would be expected from the coordinating histidine side chains. Overall, these spin-Hamiltonian parameters are similar to those obtained for AA9 LPMOs, confirming the presence of the copper(II) ion within the histidine brace coordination environment19. These data support the hypothesis that PcAA14s display LPMO characteristics and that copper is their native metal cofactor.

Substrate specificity of PcAA14

Activity assays were initially carried out with PcAA14A and PcAA14B on a wide range of polysaccharides including cellulose and xylans in the presence of ascorbic acid, which is widely used as electron donor for LPMOs. Using standardized methods previously employed to characterize AA9 LPMOs20, we were unable to detect any activity on these polysaccharides (Supplementary Fig. 7). Next, we performed saccharification assays on pretreated biomass including poplar, pine and wheat straw using a Trichoderma reesei CL847 cocktail composed of mainly cellulases and xylanases21. A boost of glucose release from poplar and pine was observed upon addition of either of the AA14 enzymes to the cocktail (Fig. 3a). When the reactions were conducted in the absence of a reductant, the boost effect was maintained (Supplementary Fig. 8), suggesting that one of the components from the biomass (for example, lignin) may act as an electron donor22. This improvement in glucose release was dose dependent, yielding up to ∼100% increase on pretreated softwood (Fig. 3b). However, no significant boost was observed on wheat straw (Supplementary Fig. 8), which differs in terms of hemicellulose composition in comparison to wood, indicating that AA14 enzymes specifically target one of the components of woody biomass. In a finding that has important consequences for biorefinery use of woody biomass as feedstock, the T. reesei CL847 cocktail enriched in AA9 LPMO acting on cellulose was also boosted by PcAA14A, suggesting that AA9 and AA14 enzymes may act on different regions within the lignocellulosic matrix (Supplementary Fig. 8). Because AA14 members do not harbor any CBM module, we artificially attached a fungal CBM1 module targeting crystalline cellulose to PcAA14A. The resulting modular PcAA14A-CBM1 enzyme performed less efficiently than the catalytic module alone (Supplementary Fig. 8), suggesting that AA14 enzymes may not require specific binding to the flat crystalline cellulose surface.

Figure 3: Contribution of PcAA14 enzymes to the saccharification of biomass.
figure 3

(a) Glucose release upon saccharification of pretreated pine and poplar by the CL847 Trichoderma reesei enzyme cocktail in the presence of PcAA14A or PcAA14B and ascorbic acid. Glucose was quantified using ionic chromatography. More saccharification assays are presented in Supplementary Figure 8. d.m., dry matter. (b) Dose effect of the addition of PcAA14A on the saccharification of pretreated pine. Concentrations of PcAA14A were 0.1 μM (+), 0.5 μM (++) and 1 μM (+++). Error bars indicate s.e.m. from triplicate independent experiments. Data points are shown as dots.

Full size image

To discern which polymer was attacked by AA14 enzymes, we used birchwood cellulosic fibers consisting of 79% cellulose and 21% xylan as a substrate. After incubation with PcAA14A or PcAA14B, wood fibers were disrupted (Fig. 4a), uncovering cellulose structures visualized at different scales using transmission electron microscopy and atomic force microscopy (Supplementary Fig. 9). These observations suggest a weakening of the cohesive forces that link the wood fibers together in a manner similar to that previously described for AA9 enzymes23. Samples treated with AA14 enzymes were further analyzed using solid-state cross-polarization magic angle spinning 13C nuclear magnetic resonance (13C CP/MAS NMR). The impact of AA14 enzymes on the fibers was different to that recently observed for AA9 LPMOs23. In the case of PcAA14 enzymes, no meaningful change was observed on cellulose signals (Fig. 4b; Supplementary Fig. 10). Interestingly, however, significant changes in signal areas corresponding to hemicelluloses located at 101 p.p.m. and 82 p.p.m. were observed when the NMR spectra were deconvoluted in the C1 and the C4 regions (Supplementary Fig. 10). These results suggest that AA14 enzymes act on xylans bound to cellulose, which have a rigidity and a conformation similar to that of the underlying cellulose chains4. The specific attack of PcAA14 on xylan substrates differentiates this new class of enzymes from all other LPMOs24,25, none of which have previously been reported to oxidize xylan in such a selective and efficient manner.

Figure 4: Enzymatic activity of PcAA14 LPMOs.
figure 4

(a) Morphology of birchwood cellulosic fibers treated with PcAA14A orPcAA14B LPMOs and untreated (reference). Images were recorded after dispersion and are representative of the samples analyzed. (b) Solid state 13C CP/MAS NMR analysis of LPMO-treated cellulosic fibers. The differences in hemicellulose content in enzyme-treated fibers were calculated from the C1 and C4 region deconvolution of NMR spectra and are indicated in Supplementary Figure 10. (c) Assays in the presence of a GH11 xylanase were performed on birchwood cellulose fibers. Xylobiose (X2) and xylotriose (X3) were quantified by ionic chromatography. Error bars indicate s.e.m. from triplicate independent experiments. (d) MS identification of the X3 oxidized species detected at 429.13 m/z generated from birchwood cellulosic fibers by PcAA14A in synergy with a GH11 xylanase. The fragmentation pattern corresponds to a C1 oxidized species with an aldonic acid at the reducing end. ▿, water losses; □, H2CO losses. The structure represents a C1 oxidized xylotriose deduced from the spectrum. An expanded view of the spectrum is provided in Supplementary Figure 12.

Full size image

To further substantiate the idea that AA14 enzymes act on xylan bound to cellulose, we performed synergy assays of AA14 enzymes in combination with a fungal GH11 xylanase using birchwood cellulosic fibers. Addition of PcAA14A to a GH11 xylanase significantly increased the release of xylooligomers from birchwood cellulosic fibers by 40% (Fig. 4c; Supplementary Fig. 11). Additionally, no improvement of xylan conversion was observed on birchwood cellulosic fibers when the xylanase was combined with a cellulose-acting AA9 LPMO (Fig. 4c).

We further investigated the nature of soluble products generated after synergistic action of PcAA14A and the GH11 xylanase. Using ionic chromatography, we found that a range of oligosaccharides eluted at similar retention times to C1-oxidized oligosaccharides (Supplementary Fig. 11). MS analyses performed on the same samples allowed the identification of several putative oxidative species with masses corresponding to C1-oxidized xylotriose (X3ox), C1-oxidized xylotetraose (X4ox) and nonoxidized xylooligosaccharides substituted with glucuronic acid (X3MeGlcA, X4MeGlcA and X5MeGlcA) (Supplementary Fig. 12). The structure of the C1-oxidized xylotriose with an aldonic acid on the reducing end (Fig. 4d) was confirmed by fragmentation of the species observed at 429 m/z by tandem MS (MS/MS) (Supplementary Fig. 12). The identification of oxidative products demonstrates that AA14 enzymes are LPMOs.

Discussion

Our findings that xylans are susceptible to AA14 oxidative cleavage only when they are adsorbed onto crystalline cellulose and not when they are in solution are supported by reports showing that xylans exist in different contexts within the cell wall4,26. Recalcitrant xylans bound to cellulose microfibrils display a two-fold screw axis conformation aligned parallel to the cellulose chain direction4 that is compatible with the proper orientation of the carbohydrate H1 and H4 atoms with respect to the LPMO catalytic center22. Unraveling the substrate specificity of AA14s has been challenging, as these enzymes are not active on xylans in solution, most probably due to the three-fold helical screw conformation of the substrate27. Using multidisciplinary approaches, we reveal that AA14 LPMOs probably specifically target the protective shield made by heteroxylans that cover cellulose microfibrils in wood. The conformation of xylan in this context contributes to wood recalcitrance, and glycoside hydrolases are not able to access such a sterically restricted substrate. The cleavage of these rare motifs by AA14 LPMOs unlocks the accessibility of xylan and cellulose chains to glycoside hydrolases, therefore improving the overall saccharification of woody biomass. These results not only greatly enhance our knowledge of wood superstructure, but also allow us to better understand and exploit biomass deconstruction by fungal saprotrophs.

Methods

Transcriptomics and secretomics of Pycnoporus sp.

Transcriptomic and proteomic data of 3-d-old cultures of Pycnoporus coccineus BRFM 310 and Pycnoporus sanguineus BRFM 1264 grown on cellulose (Avicel), wheat straw, pine and aspen are described in refs. 16,28.

Bioinformatic analysis of AA14 LPMOs.

P. coccineus AA14 sequences (Genbank IDs KY769369 and KY769370) were compared to the NCBI nonredundant sequence database using BlastP29 in February 2016. Blast searches conducted with AA14 did not retrieve AA9s, AA10s, AA11s or AA13s with significant scores, and vice versa. MUSCLE30 was used to perform multiple alignments. To avoid interference from the presence or the absence of additional residues, the signal peptides and C-terminal extensions were removed. Bioinformatic analyses were performed on 286 fungal genomes sequenced and shared by JGI collaborators. Protein clusters are available, thanks to the JGI (https://goo.gl/ZAa2NX), for each of these fungi. A phylogenetic tree has been inferred using 100 cleaned and merged alignments of proteins from selected clusters of proteins. Those clusters are present, as much as possible, in all fungi in one copy in order to maximize the score ∑1/n (with n being the number of copy in the genome). Sequences from clusters were aligned with Mafft31 and trimmed with Gblocks32, and a phylogenetic tree was built with concatenation of alignments with Fasttree33. The tree is displayed with Dendroscope34 and Bio::phylo35.

Production of P. coccineus AA14 LPMOs.

The sequences corresponding to PcAA14A (Genbank ID KY769369) and PcAA14B (Genbank ID KY769370) genes from P. coccineus BRFM310 were synthesized after codon optimization for expression in P. pastoris (GenScript, Piscataway, USA). The region corresponding to the native signal sequence was kept, and the C-terminal extension region was removed. Synthesized genes were further inserted into a modified pPICZαA vector (Invitrogen, Cergy-Pontoise, France) using BstBI and XbaI restriction sites in frame with the (His)6-tag located at the C terminus of recombinant proteins. Fusion of PcAA14A with CBM1 was carried out using the CBM1 domain of PaLPMO9E, which was added to PcAA14A at the end of the catalytic module using the linker sequence of PaLPMO9E20. Constructs without the (His)6-tag sequence were also designed by adding a stop codon at the end of the AA14 catalytic module. P. pastoris strain X33 and the pPICZαA vector are components of the P. pastoris Easy Select Expression System (Invitrogen); all media and protocols are described in the manufacturer's manual (Invitrogen).

Transformation of competent P. pastoris X33 was performed by electroporation with PmeI-linearized pPICZαA recombinant plasmids, and zeocin-resistant P. pastoris transformants were screened for protein production as described in ref. 36. The best-producing transformants were grown in 2 L of BMGY medium containing 1 mL L−1Pichia trace minerals 4 (PTM4) salts in shaken flasks at 30 °C in an orbital shaker (200 r.p.m.) to an OD600 of 2 to 6. Cells were then transferred to 400 mL of BMMY medium containing 1 ml L−1 of PTM4 salts at 20 °C in an orbital shaker (200 r.p.m.) for 3 d, with supplementation of 3% (v/v) methanol every day.

Bioreactor productions were carried out in 1.3-l New Brunswick BioFlo 115 fermentors (Eppendorf, Hamburg, Germany) following the P. pastoris fermentation process guidelines (Invitrogen). Recombinant enzymes were secreted up to ∼1 g L−1 (Supplementary Fig. 13).

Purification of PcAA14 LPMOs.

The culture supernatants were recovered by pelleting the cells by centrifugation at 2,700g for 5 min at 4 °C and filtered on 0.45-μm filters (Millipore, Molsheim, France). For (His)6-tagged enzymes, the pH was adjusted to 7.8 and the supernatants were loaded onto 5 mL His-Trap HP columns (GE healthcare, Buc, France) connected to an ÄKTAxpress system (GE healthcare). Prior to loading, the columns were equilibrated in 50 mM Tris–HCl, pH 7.8, and 150 mM NaCl (buffer A). The loaded columns were then washed with 5 column volumes (CV) of 10 mM imidazole in buffer A, before the elution step with 5 CV of 150 mM imidazole in buffer A. Fractions containing the protein were pooled and concentrated with a 3-kDa VivaSpin concentrator (Sartorius, Palaiseau, France) before being loaded onto a HiLoad 16/600 Superdex 75 Prep Grade column (GE Healthcare) and separated in 50 mM sodium acetate buffer pH 5.2. Gel-filtration analysis showed that both PcAA14 proteins are monomeric in solution. For enzymes without (His)6-tag, salts contained in the culture media were diluted ten-fold in 20 mM Tris–HCl pH 8, and culture supernatants were then concentrated with a Pellicon-2 10-kDa cutoff cassette (Millipore) to a volume of approximately 200 mL and loaded onto a 20-mL HighPrep DEAE column (GE Healthcare). Proteins were eluted using a linear gradient of 1 M NaCl (0 to 700 mM in 200 mL). Fractions were then analyzed by SDS–PAGE, and those containing the recombinant protein were pooled and concentrated. The concentrated proteins were then incubated with one-fold molar equivalent of CuSO4 overnight before separation on a HiLoad 16/600 Superdex 75 Prep Grade column in 50 mM sodium acetate buffer pH 5.2.

Biochemical analysis of AA14 LPMOs.

Concentration of purified proteins was determined by using the Bradford assay (Bio-Rad, Marnes-la-Coquette, France) or with a nanodrop ND-2000 device with calculated molecular mass and molar extinction coefficients derived from the sequences. Proteins were loaded onto 10% SDS–PAGE gels (Thermo Fisher Scientific, IL, USA) that were stained with Coomassie Blue. The molecular mass under denaturating conditions was determined with reference standard proteins (Page Ruler Prestained Protein Ladder, Thermo Fisher Scientific). Native IEF was carried out in the Bio-Rad gel system, using pI (isoelectric point) standards ranging from 4.45 to 8.2 (Bio-Rad).

N-terminal amino acid sequence determination.

The N-terminal amino acid sequences of purified PcAA14A and PcAA14B were determined according to the Edman degradation. Samples were electroblotted onto a polyvinylidene difluoride membrane (iBlot, Life Technologies). Analyses were carried out on a Procise Sequencing System (Thermo fisher).

Matrix-assisted laser desorption ionization–mass spectrometry.

Matrix-assisted laser desorption ionization mass spectra analyses were performed on a Microflex II mass spectrometer (Bruker Daltonics). One microliter of matrix (10 mg of 2,5-dihydroxybenzoic acid in 1 mL of CH3CN/H2O 50/50 (v/v), 0.1% formic acid (v/v)) was added to 1μL of intact PcAA14A or PcAA14B protein sample (100 pmoles) in the same solution. Then, mixtures were allowed to dry at room temperature (18–22 °C). Data acquisition was operated using the Flex control software. External mass calibration was carried out on Peptide calibration standard (Bruker Daltonics).

Deglycosylation assays.

To remove N-linked glycans, purified enzymes were treated with Endo Hf (New England BioLabs, Ipswich, MA) under denaturing conditions according to the manufacturer's instructions. Briefly, 10 μg of protein were incubated in 0.5% SDS and 40 mM DTT and heated for 10 min at 100 °C for complete denaturation. Denaturated samples were subsequently incubated with 1,500 units of EndoHf in 50 mM sodium acetate pH 6.0 for 1 h at 37 °C. Deglycosylated and control samples were analyzed by SDS–PAGE.

Amplex Red assay.

A fluorimetric assay based on Amplex Red and horseradish peroxidase was used as described previously37. The reaction (total volume 100 μL at 30 °C for 30 min) was measured in 50 mM sodium acetate buffer, pH 6.0, containing 50 μM Amplex Red (Sigma-Aldrich, Saint-Quentin Fallavier, France), 7.1 U mL−1 horseradish peroxidase, 0.2 to 4 μM enzyme, and 50 μM reductant, i.e. ascorbate, p-coumaric acid, caffeic acid, cinapic acid, vanillic acid, menadione, L-cysteine, tannic acid, syringic acid, gallic acid, 3-hydroxyanthranilic acid (3-HAA) and epigallocatechin gallate in water, and fluorescence was detected using an excitation wavelength of 560 nm and an emission wavelength of 595 nm using a Tecan Infinite M200 plate reader (Tecan, Männedorf, Switzerland). The specific activity was counted from H2O2 calibration curve, and the slope (13,227 counts μmol−1) was used to convert the fluorimeters' readout (counts min−1) into enzyme activity.

ICP–MS analysis.

To obtain apo enzymes, 100 mM EDTA treatment was performed overnight. Prior to the analysis, samples were mineralized in a mixture containing 2/3 of nitric acid (Sigma-Aldrich, 65% Purissime) and 1/3 of hydrochloric acid (Fluka, 37%, Trace Select) at 120 °C. The residues were diluted in ultra-pure water (2 mL) before ICP–MS analysis. The ICP–MS instrument was an ICAP Q (Thermo Electron, Les Ullis, France), equipped with a collision cell. The calibration curve was obtained by dilution of a certified multi-element solution (Sigma-Aldrich). Copper concentrations were determined using Plasmalab software (Thermo Electron), at a mass of interest m/z = 63.

Saccharification assays.

Wheat straw, pine and poplar biomass were pretreated under acidic conditions. Sugar composition was determined using the alditol acetate method38. Wheat straw consisted of 51.98 ± 2.02% (w/v) glucose, 5.70 ± 0.23% (w/v) xylose and 0.46 ± 0.04% (w/v) arabinose. Pine consisted of 43.25 ± 1.34% (w/v) glucose, 0.24 ± 0.01% (w/v) xylose and 0.15 ± 0.02% (w/v) arabinose. Poplar consisted of 50.85 ± 0.91% (w/v) glucose, 0.39 ± 0.01% (w/v) xylose and 0.07 ± 0.01% (w/v) arabinose. The enzymatic treatments were carried out in sodium acetate buffer (50 mM, pH 5.2) in a final volume of 1 mL at 0.5% consistency (w dry matter/v). The LPMO treatment was carried out sequentially with a CL847 T. reesei enzyme cocktail21 provided by IFPEN (Rueil-Malmaison, France). Each PcAA14 enzyme was added to the substrate at a concentration of between 0.1 and 1 μM in the presence or absence of 1 mM ascorbic acid for 72 h, followed by addition of 1 mg g−1 dry matter substrate of commercial cellulases from T. reesei for 24 h. Enzymatic treatments were performed in 2-mL tubes incubated at 45 °C and 850 r.p.m. in a rotary shaker (Infors AG, Switzerland). Then, samples were centrifuged at 14,000g for 5 min at 4 °C, and the soluble fraction was heated for 10 min at 100 °C to stop the enzymatic reaction. Glucose was quantified by high-performance anion exchange chromatography coupled with amperometric detection (HPAEC–PAD) as described in ref. 20.

Polysaccharides cleavage assays.

Avicel was purchased from Sigma-Aldrich, and lichenan (from Icelandic moss), curdlan, starch, barley β-1,3/1,4-glucan, konjac glucomannan, wheat arabinoxylan, and tamarind xyloglucan were purchased from Megazyme (Wicklow, Ireland). PASC (phosphoric acid swollen cellulose) was prepared from Avicel as described previously20 in 50 mM sodium acetate buffer, pH 5.2. A similar protocol was used to prepare swollen squid pen chitin, provided by D. Gillet (Mahtani Chitosan, India). Glucuronoxylans were extracted from birchwood as described previously39.

All the cleavage assays contained between 0.5 and 1 μM of PcAA14s in the presence of 1 mM ascorbate and 0.1% (w/v) polysaccharides. The enzyme reactions were performed in 2-mL tubes and incubated in a thermomixer (Eppendorf, Montesson, France) at 45 °C and 850 r.p.m. After 16 h of incubation, samples were heated for 10 min at 100 °C to stop the enzymatic reaction and then centrifuged at 14,000g for 15 min at 4 °C to separate the soluble fraction from the remaining insoluble fraction before determination of soluble products using HPAEC as described above with oligosaccharide standards (Megazyme).

Microscopy.

Aqueous dispersions of Kraft birchwood cellulosic fibers (kindly provided by S. Tapin, FCBA, Grenoble, France) were adjusted to pH 5.2 with acetate buffer (50 mM) in a final reaction volume of 5 mL. Each PcAA14 enzyme was added to the fibers at a final concentration of 20 mg g−1 in the presence of 1 mM of ascorbic acid. Enzymatic incubation was performed at 40 °C under mild agitation for 48 h. Samples were then dispersed by a Polytron PT 2100 homogenizer (Kinematica AG, Germany) for 3 min, and ultrasonicated by means of a QSonica Q700 sonicator (20 kHz, QSonica LLC., Newtown, USA) at 350 W ultrasound power for 3 min as described previously23. The reference sample was submitted to the same treatment but it did not contain the PcAA14 enzyme. Birchwood cellulose fibers (reference and PcAA14-treated) were deposited onto a glass slide and observed by a BX51 polarizing microscope (Olympus France S.A.S.) with a 4× objective. Images were captured by a U-CMAD3 camera (Olympus Japan). For the atomic force microscopy (AFM) experiments, samples were deposited onto mica substrates from fiber solutions at 0.1 g L−1, and allowed to dry overnight. Topographical images on mica were registered by a Nanoscope III-A AFM (Brukernano, Santa Barbara, US). The images were collected in tapping mode under ambient air conditions (temperature and relative humidity) using a monolithic silicon tip (RFESP, Brukernano) with a spring constant of 3 N m−1, and a nominal frequency of 75 kHz. Image processing was performed with the WSxM 5.0 software. For transmission electron microscopy (TEM) experiments, fiber solutions at 0.1 g L−1 in water were deposited on freshly glow-discharged carbon-coated electron microscope grids (200 mesh, Delta Microscopies, France) and the excess of water was removed by blotting. The sample was then immediately negatively stained with uranyl acetate solution (2%, w/v) for 2 min and dried after blotting. The grids were observed with a Jeol JEM 1230 TEM at 80 kV.

NMR spectroscopy.

Solid-state 13C NMR experiments were performed on a Bruker Avance III 400 spectrometer operating at a 13C frequency of 100.62 MHz using a 4 mm double-resonance (H/X) magic angle spinning (MAS) probe. Samples were dialyzed against ultrapure water (MWCO 12-14000) for 7 d to remove buffer, ascorbate and released soluble sugars. Experiments were conducted at room temperature (18–22 °C) at a MAS frequency of 9 kHz using a cross-polarization sequence (CP/MAS). The 13C chemical shift was referenced using the carbonyl signal of glycine at 176.03 p.p.m. The cross-polarization pulse sequence parameters were: 3.2 μs proton 90° pulse, 2.50 ms contact time at 67.5 kHz, and 10 s recycle time. Typically, the accumulation of 5,120 scans was used. All spectra obtained were processed and analyzed using Bruker TopSpin version 3.2. To determine the crystallinity and the general cellulose's morphology of the C1 and C4 region of the samples, we used the sophisticated approach40 described in detail in our previous work23. For the C1 region, this approach used three Lorentzian lines for the crystalline part (Cr (Iα) and Cr (Iβ)) and one Gaussian line for the less ordered cellulose (paracrystalline cellulose, PCr). For the C4 region, four lines for the crystalline part corresponding to crystalline and PCr cellulose and three Gaussian lines for the amorphous part (accessible surfaces, AS, and inaccessible surface, IAS) were used. The cellulosic fibers contained xylan, which was considered in the spectral decomposition: in the C1 region with one line at 101.4 p.p.m. and in the C4 region with one broad line centered at 81.6 p.p.m.

Synergy assays with xylanase.

Assays were run on the birchwood cellulose fibers used in microscopy and NMR experiments. Fibers were grinded (<0.18 mm particle size) and hydrated in water under stirring for 48 h before enzymatic assays. One milliliter reaction volumes containing 0.5% (w/v) birchwood fibers were incubated with 1 μM of PcAA14s and 0.1 μM of GH11 xylanase M4 (Aspergillus niger) from Megazyme (reference E-XYAN4) in 10 mM sodium acetate, pH 5.2, alone or supplemented with 1 mM L-cysteine. Prior to the reaction, the GH11 xylanase was buffer exchanged with 10 mM sodium acetate pH 5.2 using a PD-10 column (GE Healthcare) to remove any trace of ammonium sulfate. Enzymatic reactions were performed in 2-mL tubes and incubated in a thermomixer (Eppendorf, Montesson, France) at 45 °C and 850 r.p.m. for 24 h. Samples were then centrifuged at 14,000g for 5 min at 4 °C to separate the soluble fraction from the remaining insoluble fraction. Proteins were removed from the soluble oligosaccharides fraction by filtering the supernatants using Nanosep 3K Omega centrifugal devices (Pall corporation). Soluble oligosaccharides generated were analyzed by HPAEC as described above and mass spectrometry (see below) using non-oxidized xylooligosaccharides (Megazyme) as standards. Corresponding C1-oxidized standards (from DP2 to DP4) were produced from nonoxidized xylooligosaccharides by using purified PaCDHB prepared as described previously20. All assays were carried out in triplicate.

Electrospray mass spectrometry (ESI–MS and MS/MS).

Experiments were performed on a Synapt G2Si high-definition mass spectrometer (Waters Corp., Manchester, UK) equipped with an Electrospray ion (ESI) source. Two types of mass measurements were performed on the samples: first, a mass profile was done on a mass range of 300–2,000 m/z (M/S). Ions of interest were further isolated and fragmented by collision-induced dissociation in the transfer cell of the instrument (MS/MS). In these experiments, ion mobility (IM) was activated to reduce interference from sample impurities. IM was performed in a traveling-wave ion mobility (TWIM) cell. The gas flows were held at 180 mL min−1 He in the helium cell and at 90 mL min−1 N2 in the mobility cell. The IM traveling wave height was set to 40 V, and its wave velocity was set to 480 m s−1 for positive ionization mode and 500 m s−1 for negative ionization mode. Samples were diluted ten-fold in MeOH/H2O (1:1, v/v) and infused at a flow rate of 5 μL min−1 in the instrument. The instrument was operated in positive or negative polarity, as well as in 'sensitivity' mode.

Crystallization, data collection, structure determination and refinement.

All crystallization experiments were carried out at 20 °C by the sitting-drop vapor-diffusion method using 96-well crystallization plates (Swissci) and a Mosquito Crystal (TTP Labtech) crystallization robot. Reservoirs consisted of 40 μL of commercial screens and crystallization drops were prepared by mixing 100 nL reservoir solution with 100, 200 and 300 nL of protein solution. An initial hit was obtained after 1 week from a condition of the AmSO4 screen (Qiagen) consisting of 2.4 M (NH4)2SO4 and 0.1 M citric acid pH 4.0. This condition was further optimized to obtain diffraction-grade crystals by mixing protein solution at 28 mg mL−1 with precipitant solution consisting of 2.4 M (NH4)2SO4 and 0.1 M citric acid pH 4.4 at a volume ratio of 3:1. PcAA14B crystals grew to dimensions of 0.15 × 0.15 × 0.05 mm in 1 week. Crystals belong to space group P41212 with cell axes 204 × 204 × 110 Å and two molecules per asymmetric unit.

Crystals of PcAA14B were soaked for 5 min in a solution in which 2.4 M (NH4)2SO4 of the mother liquor was replaced by 2.4 M Li2SO4 for the sake of cryoprotection before flash-cooling in liquid nitrogen. As X-ray fluorescence scans on native crystals did not reveal a significant presence of copper within the crystals, a heavy atom derivative was prepared by soaking the crystals in reservoir solution supplemented with 55 mM of the gadolinium complex gadoteridol before cryo-cooling. Native diffraction data were collected on beamline ID23-1, while a MAD data set was collected on beamline ID30B at the European Synchrotron Radiation Facility (ESRF), Grenoble, France, at wavelengths of 1.711 and 1.698 Å for peak/inflection and remote energies, respectively. Data were indexed and integrated in space group P41212 using XDS41, and subsequent processing steps were performed with the CCP4 software suite42. Determination of the Gd3+ substructure and subsequent phasing combined with solvent flattening were carried out with SHELXC/D/E42, leading to a pseudo-free correlation coefficient of 71.8%. Starting from experimental phases, an initial model comprising 526 residues (out of 584) was automatically built with Buccaneer43 and manually completed with Coot44. This initial model was used for rigid body refinement followed by restrained refinement against native data with the program Refmac45. A random set of 5% of reflections was set aside for cross-validation purposes. Model quality was assessed with internal modules of Coot44 and using the Molprobity server46. Figures representing structural renderings were generated with the PyMOL Molecular Graphics System (DeLano, W.L. The PyMOL Molecular Graphics on http://www.pymol.org/). Atomic coordinates and structure factors have been deposited within the Protein Data Bank (http://www.rcsb.org)47. Data collection and refinement statistics are summarized in Supplementary Table 4.

Electron paramagnetic resonance (EPR).

Continuous wave (cw) X-band frozen solution EPR spectra of a 0.2 to 0.3 mM solution of Cu(II)-PcAA14A and PcAA14B, prepared and copper loaded as described above, in 10% v/v glycerol at pH 5.2 (50 mM sodium acetate buffer) and 165 K were acquired on a Bruker EMX spectrometer operating at ∼9.30 GHz, with modulation amplitude of 4 G, modulation frequency of 100 kHz and microwave power of 10.02 mW (4 scans). Both enzymes showed identical EPR spectra. Cw Q-band frozen solution spectra of 1.0 mM solution of Cu(II)- PcAA14A at pH 5.2 (50 mM sodium acetate buffer) and 113 K were acquired on a Jeol JES-X320 spectrometer operating at ∼34.7 GHz, with modulation width 1 mT and microwave power of 1.0 mW (8 scans).

Spectral simulations were carried out using EasySpin 5.0.3 (ref. 48). Simulation parameters are given in Supplementary Table 3. gz and |Az| values were determined accurately from the absorptions at low field. It was assumed that g and A tensors were axially coincident. Accurate determination of the gx, gy, |Ax| and |Ay| was obtained by simultaneous fitting of both X and Q band spectra. The superhyperfine coupling values for the nitrogen atoms could not be determined accurately, although it was noted that satisfactory simulation could only be achieved with the addition of two nitrogen atoms with coupling in the range 30–36 MHz.

Statistics.

For all statistics, n = 3 values from independent experiments were used to calculate the s.e.m. For all representative results, experiments were repeated at least two times and at least 20 images were collected for microscopy analyses.

Life Sciences Reporting Summary.

Further information on experimental design and reagents is available in the Life Sciences Reporting Summary.

Data availability.

All data generated or analyzed during this study are included in this published article (and its Supplementary Information files). PcAA14A and PcAA14B sequences were deposited in GenBank under accession numbers KY769369 and KY769370, respectively. The X-ray structure of PcAA14B was deposited in the Protein Data Bank with accession number 5NO7. Raw EPR data are available on request through the Research Data York (DOI: 10.15124/8758d712-1e67-467e-b0f0-f0dd99f0232a).

Additional information

Any supplementary information, chemical compound information and source data are available in the online version of the paper. Reprints and permissions information is available online at http://www.nature.com/reprints/index.html. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Correspondence and requests for materials should be addressed to J.-G.B.