AA16 Oxidoreductases Boost Cellulose-Active AA9 Lytic Polysaccharide Monooxygenases from Myceliophthora thermophila

Copper-dependent lytic polysaccharide monooxygenases (LPMOs) classified in Auxiliary Activity (AA) families are considered indispensable as synergistic partners for cellulolytic enzymes to saccharify recalcitrant lignocellulosic plant biomass. In this study, we characterized two fungal oxidoreductases from the new AA16 family. We found that MtAA16A from Myceliophthora thermophila and AnAA16A from Aspergillus nidulans did not catalyze the oxidative cleavage of oligo- and polysaccharides. Indeed, the MtAA16A crystal structure showed a fairly LPMO-typical histidine brace active site, but the cellulose-acting LPMO-typical flat aromatic surface parallel to the histidine brace region was lacking. Further, we showed that both AA16 proteins are able to oxidize low-molecular-weight reductants to produce H2O2. The oxidase activity of the AA16s substantially boosted cellulose degradation by four AA9 LPMOs from M. thermophila (MtLPMO9s) but not by three AA9 LPMOs from Neurospora crassa (NcLPMO9s). The interplay with MtLPMO9s is explained by the H2O2-producing capability of the AA16s, which, in the presence of cellulose, allows the MtLPMO9s to optimally drive their peroxygenase activity. Replacement of MtAA16A by glucose oxidase (AnGOX) with the same H2O2-producing activity could only achieve less than 50% of the boosting effect achieved by MtAA16A, and earlier MtLPMO9B inactivation (6 h) was observed. To explain these results, we hypothesized that the delivery of AA16-produced H2O2 to the MtLPMO9s is facilitated by protein–protein interaction. Our findings provide new insights into the functions of copper-dependent enzymes and contribute to a further understanding of the interplay of oxidative enzymes within fungal systems to degrade lignocellulose.


INTRODUCTION
Transition from a fossil-based society to a more sustainable one drives full valorization of lignocellulose-rich agricultural and forestry side-streams for the production of biofuels, biomaterials, and biochemicals. 1 Hereto, enzyme-driven degradation of cellulose and hemicellulose to fermentable monosaccharides is an essential step, 2 in which copperdependent lytic polysaccharide monooxygenases (LPMOs) are key. LPMOs have been shown to oxidatively cleave in particular insoluble substrates, such as cellulose, which synergistically enhances cellulose saccharification by established cellulases. As such LPMOs have become a permanent ingredient in cellulolytic enzyme formulations. 3−5 LPMOs are currently classified as "Auxiliary Activity" (AA) families 9−11 and 13−17 in the Carbohydrate-Active enZymes (CAZy) database (http://www.cazy.org). 6,7 To explore the AA diversity in nature and improve enzyme formulations, new AA families with enigmatic functions need further investigation. 8−10 The recently proposed AA16 family contains so far only one characterized member from Aspergillus aculeatus (AaAA16). 11 AaAA16 has been indicated as C1cellulose-active LPMO, 11 though its activity is much lower compared to accustomed C1-oxidizing AA9 LPMOs ( Figure  1). 12,13 Based on sequence comparison, AA16 members were reported to share common features with other LPMOs, for instance, a copper-dependent active site coordinated by two histidines and a tyrosine (sometimes a phenylalanine residue in AA10 LPMOs). 14 This coordination, also referred to as "histidine brace" (His-brace), is conserved in all LPMOs. 14,15 Although the catalytic mechanism of LPMOs is not fully clear, it is well accepted that the catalytic cycle starts with a so-called "priming reduction" of Cu(II) to Cu(I) by an external electron donor. 16 The external electron donor can be a chemical reductant, such as ascorbic acid (Asc), phenolics (e.g., pyrogallol (Pyg)) including lignin, or a redox enzyme as is well established for cellobiose dehydrogenase (CDH). 17,18 After this priming reduction, the catalytic reaction can follow two routes depending on the cosubstrate, which can either be H 2 O 2 or O 2 ( Figure 1, route 1 or 2). 16,19,20 The mechanistic details of these routes are still under debate as extensively reviewed elsewhere. 21,22 In the absence of a carbohydrate substrate, superoxide (or hydroperoxide) is released, resulting in the production of H 2 O 2 and regeneration of the Cu(II) state ( Figure 1, route 3). 13 H 2 O 2 has been shown to be the preferred cosubstrate over O 2 , as turnover numbers obtained with H 2 O 2 are in certain cases more than three orders of magnitude higher compared to those obtained with O 2 . 16,20,23 On the other hand, a high H 2 O 2 concentration induces oxidative damage of amino acids close to the copper-active site resulting in self-inactivation. Hence, the optimal H 2 O 2 concentration is a balance between activity and inactivation and upholds a delicate equilibrium in LPMO reactions. 16,20,23,24 H 2 O 2 typically results from nonenzymatic or enzymatic routes to drive LPMO reactions. Nonenzymatic H 2 O 2 formation results from metal ions reduced by molecular reductants (e.g., Asc, cysteine) and subsequently reacting with dissolved O 2 , 26,27 while some H 2 O 2 -producing oxidases (e.g., glucose oxidase) also can take up that role. 16 It is noteworthy that, as mentioned above, in absence of a carbohydrate substrate, (reduced) LPMOs show oxidase activity to produce H 2 O 2 . In a recent study, Stepnov and co-workers described that a weak cellulose-binding ScLPMO10C TR (only catalytic domain, without a linker and a carbohydrate binding module) served as a H 2 O 2 producer to enhance oxidative cellulose cleavage by full-length ScLPMO10C. 28 Here, we characterized two new members of the AA16 family. MtAA16A was homologously produced in Myceliophthora thermophila C1, while AnAA16A from Aspergillus nidulans was produced in Pichia pastoris X-33. In contrast to the published AaAA16, both MtAA16A and AnAA16A did not oxidatively cleave carbohydrate substrates. We elucidated the crystal structure of MtAA16A and showed that both AA16s display oxidase activity. Furthermore, we found a substantial boosting effect of the AA16s on various well-characterized MtLPMO9s in degrading cellulose. This boosting effect was absent when the AA16s were combined with three distinct and well-studied AA9 Neurospora crassa (Nc) LPMOs. We discuss possible reasons for this observed (lack of) interplay. In summary, we suggest that AA16 proteins are H 2 O 2 -producing oxidoreductases that may assist LPMOs in degrading lignocellulose.

EXPERIMENTAL SECTION
2.1. Materials. Syringol, pyrogallol, and ammonium acetate were purchased from Sigma-Aldrich (St. Louis, Missouri). Cellobiose, cellotriose, cellotetraose, cellopentaose, and cellohexaose were purchased from Megazyme (Bray, Ireland). Regenerated amorphous cellulose (RAC) was prepared from Avicel PH-101 (Sigma-Aldrich) as described previously. 29 Ascorbic acid (Asc) was purchased from VWR International (Radnor, Pennsylvania). Other aromatic compounds used were purchased from Sigma-Aldrich or VWR International. Other carbohydrate substrates used were purchased from Sigma-Aldrich or Megazyme. Glucose oxidase from Aspergillus niger (AnGOX, 10 000 U g −1 powder) was purchased from Sigma-Aldrich. All water used was produced by a Milli-Q In both H 2 O 2 -and O 2 -dependent routes, the C1-and/or C4carbon position of the carbohydrate (R−H) substrate is hydroxylated (R−OH) and a Cu(I) is ready for the next catalytic cycle. The oxidation at the C1-carbon leads to the formation of δ-lactone, which is converted to an aldonic acid in water. The oxidation at C4-carbon results the generation of 4-ketoaldose, which is in pH-dependent equilibrium with the geminal diol. Adapted from Wang et al. 21  G2QH80) was homologously expressed and produced in a low protease/low (hemi)cellulase-producing M. thermophila C1 strain as described elsewhere. 30,31 MtAA16A was purified by three subsequent chromatographic steps. Crude MtAA16A-rich fermentation broth was filtered and dialyzed against 10 mM potassium phosphate buffer pH 7.6 before chromatographic purification. The dialyzed MtAA16A was purified by anion-exchange chromatography (AEC), followed by size-exclusion chromatography (SEC). Purification settings and elution programs of AEC and SEC have been described previously. 32 The SEC-purified MtAA16A-containing fractions were further purified by cationexchange chromatography (CEC) on an ÄKTA-Micro preparative chromatography system (GE Healthcare). MtAA16A-containing fractions were loaded on a Resource S column (30 × 16 mm internal diameter, GE Healthcare) preequilibrated with 10 mM sodium acetate buffer pH 3.8 (eluent A). The unbound fraction was first removed (one column volume). Eluent B was 10 mM sodium acetate buffer (pH 3.8) containing 500 mM NaCl. Elution (flow rate of 1 mL min −1 ) was performed as follows: from 0 to 30% B in two column volumes; 30% B for one column volume; next, 30−100% B over two column volumes; and finally, 100% B for four column volumes. All fractions were collected and immediately stored on ice. Peak fractions (based on UV absorption at 280 nm) were adjusted to an approximate concentration of 2 mg mL −1 (as determined by the bicinchoninic acid method) and analyzed by sodium dodecyl sulfate-poly(acrylamide) gel electrophoresis (SDS-PAGE), as described previously, 32 to determine the MtAA16A fractions. CEC-purified MtAA16Acontaining fractions were combined and used as the final enzyme stock solution. All CEC-purified MtAA16A fractions were aliquoted into 500 μL size and stored at −80°C.
2.2.2. AnAA16A. The AnAA16A gene (AN0778.2) was amplified directly from the A. nidulans genome and produced in a P. pastoris X-33 strain, as described hereafter. The oligonucleotides AnAA16fw (5′ACAACTAATTATTC-GAAACGATGAAGCACGCTACCACCG3′) and AnAA16rv (5′CCCTGAAAATAAAGATTCTCGCCGTTACCACTTC-CACCAA3′) were used to remove the C-terminal extension region (residues 199−306) and maintain the native signal peptide. The removal of C-terminal extension region prior to protein production has been reported by Filiatrault-Chastel et al. 11 The AnAA16 gene was cloned into a modified pPICZα vector, as previously described. 33 This construction allowed the expression of a recombinant AnAA16A containing a cleavable C-terminal polyhistidine-tag. The P. pastoris X-33 (Invitrogen, Waltham, Massachusetts) was transformed by electroporation using a PmeI-linearized plasmid (pPICZT::A-nAA16A) and selected on yeast extract-peptone-dextrosesorbitol (YPDS)-zeocin plates. The recombinant colonies were randomly picked and grown in a buffered methanol-complex (BMMY) medium, and the gene expression was confirmed by SDS-PAGE analysis of the supernatant content. The transformant showing the highest expression profile was grown in 40 mL of YPD medium overnight and inoculated in four Erlenmeyer flasks containing 0.5 L buffered glycerol-complex medium (BMGY) medium at 30°C and 250 rpm until an OD600 of 2. The yeast cells from each flask were harvested and transferred to 0.1 L BMMY medium and incubated at 30°C and 250 rpm for 72 h. Two percent absolute methanol was added every 24 h to maintain recombinant protein production. The culture supernatant was filtered, and the pH was adjusted to 8.0 using the Tris−HCl buffer. The entire volume was loaded onto a 5 mL HiTrap Chelating HP column (GE Healthcare) connected to an ÄKTA Start system (GE Healthcare) equilibrated with 50 mM Tris/HCl pH 8.0 and 0.3 M NaCl (buffer-A). AnAA16A was eluted using a linear gradient from 0 to 100% of 1 M imidazole within 10 column volumes. The fractions containing the purified enzyme were pooled and concentrated using Amicon Ultra 15 mL centrifugal filters (Merck Millipore) with a cutoff of 10 kDa. The C-terminal His-tag cleavage and removal with tobacco etch virus (TEV) protease was performed according to Kadowaki et al. 33 The nontagged AnAA16A was then saturated with copper by incubating the protein solution with a threefold molar excess of Cu(II)SO 4 for 10 min at room temperature, followed by size-exclusion chromatography on a pre-equilibrated HiLoad 16/60 Sephadex 75 size-exclusion column (GE Healthcare) in 50 mM Tris/HCl buffer pH 8.0 containing 150 mM NaCl. Protein purity was analyzed by SDS-PAGE using Coomassie Brilliant Blue G-250 staining (Sigma-Aldrich) and the concentration was determined using the Bradford method using bovine serum albumin as a standard.
2.3. Expression, Production, and Purification of MtLPMO9s and NcLPMO9s. Six well-characterized AA9 LPMOs were used in this study (Table 1). MtLPMO9E, MtLPMO9H, and MtLPMO9I were homologously expressed in a low protease/low (hemi)cellulase-producing M. thermophila C1 strain 30,31 and purified as described elsewhere. 32,34 The expression and purification of N. crassa LPMOs produced in P. pastoris X-33 (NcLPMO9C, NcLPMO9F, and NcLPMO9M) have been described previously. 13,35 MtLPMO9s and NcLPMO9s were Cu(II)-saturated during  37 with modifications. A pure MtAA16A stock solution (1 mg mL −1 , 500 μL) was incubated with a three-fold molar excess of Cu(II)SO 4 in 50 mM ammonium acetate pH 5.0 for 30 min at 25°C under shaking at 600 rpm (Eppendorf ThermoMixer C, Eppendorf, Hamburg, Germany). Excess Cu(II) was removed by a five-cycle washing-out procedure. For each washing step, 500 μL of Cu(II)-saturated MtAA16A was concentrated 10-fold using Amicon Ultra-0.5 centrifugal filters (Sigma-Aldrich) and subsequently brought back to 500 μL by adding 50 mM ammonium acetate pH 5.0. The final concentration of excess Cu(II) was calculated lower than 0.7 pM. In this study, MtAA16A represents the Cu(II)-saturated form, unless mentioned otherwise. For Cu(II) saturation of AnAA16A, see the previous section. The control sample was prepared in the same way as described above but without MtAA16A, and it is referred to as only Cu(II) sample.

Determination of H 2 O 2 Production by the Amplex Red/Horseradish Peroxidase Assay.
The method for determining H 2 O 2 production was based on a previously reported protocol 13 and performed using a commercial Amplex Red Hydrogen Peroxide/Horseradish Peroxidase (HRP) Assay Kit (catalog number: A22188, Thermo Fisher Scientific, Waltham, Massachusetts). The assay was performed in 96well plates and followed the manufacturer protocol. Each well contained 50 μL of sample including 1 μM AA9 LPMOs with and without 1 μM MtAA16A in the presence of 1 mM Asc in 50 mM ammonium acetate buffer (pH 5.0). Controls were only buffer, boiled MtAA16A, and only Cu(II) sample (described in Section 2.4), all in the presence of 1 mM Asc. In addition, different concentrations of AnGOX (10, 1, 0.1, and 0.01 μg mL −1 ) in the presence of 15 mM glucose and 1 mM Asc were also prepared. All samples were mixed with Amplex Red/HRP working reagents (final concentration was 50 μM Amplex Red reagent, 0.1 U mL −1 HRP, and 50 mM sodium phosphate pH 7.4) in a total volume of 100 μL, after which the measurement was immediately started in a spectrophotometer at 30°C. The reactions were performed in triplicate. The Amplex Red reaction product resorufin was determined by measuring the absorbance at 560 nm every 10 min (5 s shaking prior to each measurement) till 360 min. The slope of the initial linear increase in absorption was used for the calculation of the H 2 O 2 -producing activity. 13 According to the manufacturer, the path length of a 100 μL solution in the 96-well plate is roughly 0.33 cm. An extinction coefficient of resorufin, ε 560 = 58 mM −1 cm −1 , was used to calculate the H 2 O 2 concentration. One unit of enzyme activity (U) is defined as the amount of enzyme that catalyzes the production of 1 μmol H 2 O 2 per min under the assay conditions. 2.6. Incubations of AA9 LPMOs and AA16 Enzymes with RAC. General incubation settings were 50 mM ammonium acetate buffer (pH 5.0), 2 mg mL −1 RAC, and 1 μM AA9 LPMO and/or 1 μM AA16 enzyme. The incubation has been performed in five ways:  All supernatants from the above incubations were collected and stored at −20°C for further analysis. All incubations were performed in duplicate.

Determination of H 2 O 2 by the Ferric-Xylenol
Orange Assay. The level of H 2 O 2 in the supernatants after 2, 4, and 6 h incubation of MtLPMO9B with/without MtAA16A and AnAA16A in the presence or absence of RAC and Pyg was determined using the Peroxide Assay Kit (catalog number: MAK311, Sigma-Aldrich). The assay was performed by following the protocol provided by the manufacturer. First, H 2 O 2 standards (0, 3, 6, 9, 12, 18, 24, and 30 μM) and detection reagent (mixing 1 volume of reagent A with 100 volumes of reagent B) were freshly prepared. Afterwards, 40 μL of undiluted supernatants from the incubations and H 2 O 2 standards were added into separate wells of a 96-well plate. Subsequently, 200 μL of detection reagent was added into wells, and the reactions were incubated for 30 min at room temperature. The absorbance of each sample at 585 nm was determined in a spectrophotometer. The H 2 O 2 levels were calculated based on a calibration curve generated by H 2 O 2 standards. All measurements were performed in duplicate.

HPAEC-PAD Analysis for Oligosaccharide Profiling and Relative Quantification of Products.
All supernatants from the incubations of AA9 LPMO (in the presence and absence of AA16s) with RAC were analyzed by HPAEC. The analysis was performed on an ICS-5000 system (Dionex, Sunnyvale, California) equipped with a CarboPac PA-1 column (2 mm ID × 250 mm; Dionex) in combination with a CarboPac PA guard column (2 mm ID × 50 mm; Dionex). The system was further equipped with pulsed amperometric detection (PAD). Mobile phases were (A) 0.1 M NaOH and (B) 1 M NaOAc in 0.1 M NaOH. The column temperature was set at 20°C. The elution profile applied has previously been described. 12,32 Samples were diluted five times before analysis. For supernatants collected in Section 2.6 (v), the total peak area of released oxidized cello-oligosaccharides was calculated.
2.9. Crystallization, Structure Determination, and Structure Modeling. Prior to crystallization, MtAA16A was treated with endoglycosidase H (Sigma-Aldrich) according to Frandsen et al. 38 with modifications. In brief, a 1 mL MtAA16A (10 mg) solution in 50 mM NaOAc pH 6.0 containing 150 mM NaCl was incubated with 100 μL endoglycosidase H (0.5 U based on the manufactural information) for 16 h at room temperature. Afterward, the incubated sample was exchanged to 20 mM NaOAc buffer pH 5.5. Crystallization was set up with a protein stock solution of 20 mg mL −1 preincubated in sample buffer for at least 1 h with equimolar Cu(II) acetate. Crystallization trials with commercial screens JCSG+ (Qiagen, Hilden, Germany), Index (Hampton Research, Aliso Viejo, California), PEG/Ion (Hampton Research), and Morpheus (Molecular Dimensions, Sheffield, U.K.) were set up with an Oryx-8 crystallization robot (Douglas Instruments, Hungerford, U.K.) using the sitting drop vapor diffusion method in MRC-2-drop 96-well plates at room temperature. The drops had a volume of 0.3 μL consisting of protein stock solution to reservoir in ratios of 3:1 and 1:1. Diffracting crystals/needles were obtained in different conditions, and data were collected at 100 K without additional cryoprotection. Crystals grown from the JCSG+ screen (0.2 M CaOAc, 0.1 M Na−cacodylate pH 6.5, and 40% v/v PEG 300) diffracted well but did not lead to structure determination due to possible twinning. Crystals grown from the Morpheus screen (0.1 M buffer system 1 (pH = 6.5), 30% EDO_P8K, and 0.09 M halogens) 39 led to a preliminary structure determination at 3.1 Å resolution. Optimization of similar Morpheus conditions in MRC MAXI 48-well plates with 1 μL of protein stock and 1 μL of reservoir (prepared by diluting the Morpheus condition with water: 0.1 M buffer system 1, 30% EDO_P8K, divalent cations, 300 μL in a ratio of 9:1) led to a good data set (Table S1). Diffraction tests and collections were carried out at the ID30A-3 beamline 40 at ESRF (Grenoble, France) and BioMAX 41 beamline at MAX IV (Lund, Sweden), and data was processed both through the available automatic pipelines and manually using XDSAPP software 42 of the PReSTO platform or XDS. 43 Molecular replacement was carried out in MOLREP 44 with an AlphaFold 2 45,46 model of MtAA16A obtained through the Colab implementation. 47 A clear solution with three molecules in the asymmetric unit was obtained, which was further refined with REFMAC5 48 and consecutive manual model building by COOT, 49−51 yielding a good quality structure with a maximum resolution of 2.65 Å. Crystallographic statistics are given in Table S1. The structure has been deposited in the Protein Data Bank (PDB) with the accession number 7ZE9. Figures were rendered in PyMOL (v2.0.1 2018, Schrodinger, Inc., New York).

MtAA16A: Molecular Mass, N-Glycosylation, and Methylation of N-Terminal Histidine. Purified
MtAA16A showed a major band at 27 kDa in SDS-PAGE ( Figure S1). Since the predicted molecular mass of MtAA16A based on the amino acid sequence without a signal peptide is 18.4 kDa ( Figure S2), glycosylation of MtAA16A was expected, as also observed with other homologously produced MtLPMO9s. 12 Indeed, after treatment of MtAA16A with (N-acetyl-βglucosaminyl)asparagine amidase (PNGase F), a major band at 19 kDa ( Figure S1) remained, indicating that MtAA16A contained N-glycosylation. The predicted molecular mass of AnAA16A is 19.9 kDa (catalytic domain). Similar to MtAA16A, in the SDS-PAGE experiment, a roughly 30 kDa band was visible, indicative of glycosylation ( Figure S1).
Typical for homologously expressed fungal LPMOs is the methylated N-terminal histidine, of which the methylation is suggested to play a role in protection against auto-oxidation of the copper histidine brace active site. 52 Reversed phase liquid chromatography coupled to mass spectrometry (LC-MS/MS n ) of a tryptic digest revealed that the N-terminal histidine of MtAA16A was indeed methylated (MeHis1; Figure S3). Further identification of peptides in the MtAA16A tryptic digest confirmed that the amino acid sequence of the MtAA16A protein was in accordance with the prediction based on gene annotation.
The amino acid sequence of AnAA16A was also confirmed, as well as the expected nonmethylated N-terminal histidine (data not shown).

MtAA16A
Does Not Oxidatively Cleave Carbohydrates but Oxidizes Syringol-like Compounds. The AA16 family has been suggested to comprise catalytic LPMO-like enzymes, though this suggestion was based on a rather low C1-oxidative cleavage of cellulose observed for only one AA16 candidate (AaAA16). 11 MtAA16A did not oxidatively cleave cellulose, and none of the other carbohydrates were tested, including cellopentaose, cellohexaose, chitin, pectin, hemicelluloses, and combinations thereof (Table S2). Oxidative cleavage was neither observed after renewing Cu(II) saturation of MtAA16A nor by varying the type of the reducing agent or adding H 2 O 2 and also not by increasing substrate or enzyme concentrations (Table S2). We also observed that no oxidized products were released by AnAA16A from cellulosic substates including phosphoric acid swollen cellulose, Avicel PH-101, and cellulose nanocrystals (data not shown). Therefore, we concluded that MtAA16A and AnAA16A have no typical LPMO-like catalytic action toward cellulose and other investigated poly-and oligosaccharides.
Next, we questioned whether MtAA16A and AnAA16A actually are oxidative enzymes or just noncatalytic coppercontaining proteins, similar to Bim1 8 or LaX325. 9 Therefore, we tested if MtAA16A was active in the H 2 O 2 -driven conversion of 2,6-dimethoxyphenol (syringol; Syr). 53 MtLPMO9B, MtLPMO9E, MtLPMO9H, and MtLPMO9I served as the reference, and results are shown in Figure S4. Following the formation of the chromogenic product coerulignone, it allowed the estimation of the specific activity of all MtLPMO9s to range between 0.27 and 0.56 U g −1 .
However, MtAA16A showed a much higher specific activity of

ACS Catalysis pubs.acs.org/acscatalysis
Research Article 4.14 U g −1 . These results indicated that MtAA16A is a copperdependent enzyme able to oxidize syringol in the presence of H 2 O 2 .

MtAA16A Crystal Structure.
To better understand why both AA16s did not catalyze the oxidative cleavage of poly-and oligosaccharides, we determined the crystal structure of MtAA16A. The three-dimensional structure of MtAA16A, the first experimental structure in the AA16 family, shows the typical LPMO fold (Figure 2A). 54 A search with DALI 55,56 revealed high structural similarity with AA9, AA10, and AA11 family members, but MtAA16A has a significantly smaller size than the matched structures (see Figure S5 for comparison). The copper binding site appears to be identical to the one observed in AA9, AA11, AA13, AA15, AA17, and some AA10 members with the His-brace providing three equatorial ligands and an additional Tyr axial ligand (Figures 2B, S6, and Table  S3) with typical distances from the copper to the N ligands (1.8−2.3 Å) and a longer distance to the Tyr−OH (2.7 Å). The methylation of His1 is confirmed in the structure. In the crystal, a carboxylic residue from a neighboring molecule blocks the equatorial position to the copper, expecting to bind water in solution. There is no visible axial ligand, presumably due to photoreduction of the Cu(II) to Cu(I) under X-ray exposure. 38 The θ angles in Table S3 are also consistent with a Cu(I) state. Generally, the geometry of the Cu site looks fully compatible with reactivity, which further supports the demonstrated oxidizing activity on small compounds such as syringol. Second coordination sphere residues include a Gln, which is in the same position as an important and conserved Gln in AA9 ( Figure S6), 57 while the conserved second coordination sphere His from AA9 is substituted by an Asn. Trp149 is able to make π−π stacking interactions with the active site residue Tyr158, which is reminiscent of similar interactions in AA10 and AA11 family members ( Figure S6). 58 Interestingly, the longest molecular axis in AA16 runs in a different direction from the longest axis in the closest DALI hits or LsAA9A, 38 an AA9 LPMO for which information on cello-oligosaccharide binding is available ( Figure 2C). The surface adjacent to the His-brace in MtAA16A lacks the typical flat aromatic features seen in most cellulose-binding LPMOs and exemplified in the complex of LsAA9A with cellohexaose by interaction with a Tyr. At the surface of MtAA16A, another Tyr residue (Tyr27) can be found near the His-brace but lies sideways rather than parallel to the protein surface. Thus, MtAA16A does not seem to possess a likely polysaccharidebinding surface adjacent to the His-brace but rather a small pocket ( Figure 2D), which could be speculated to interact with small aromatics like syringol, on which we have demonstrated activity. Binding experiments ( Figure S7) indicated that, as suggested by the crystal structure, MtAA16A was not able to bind to RAC. In addition, thermal shift assays supported that MtAA16A did not bind to cello-oligosaccharides (DP2−6), while syringol induced significant thermal stabilization ( Figure  S8), which is consistent with binding. However, as this is not a direct binding assay, an alternative possibility could be that the reduction of the active site metal results in stabilization. Distant from the putative substrate-binding surface, Nglycosylation at Asn89 is very obvious in the electron density, and despite treatment with endoglycosidase H, 5−7 glycan units are visible at Asn89, interacting with exposed Phe52, Asn98, and Tyr148. One NAG unit is also visible at Asn126.

MtLPMO9B-Driven Oxidative Cleavage of Cellulose.
We further questioned whether MtAA16A and AnAA16A display oxidase activity, as earlier reported for AaAA16 ( Figure  1, route 3). 8 Indeed, we observed accumulation of H 2 O 2 in MtAA16A-Pyg and AnAA16A-Pyg samples, while accumulation of H 2 O 2 in MtLPMO9B-Pyg samples was absent in the presence of RAC or very low in the absence of RAC ( Figure  3). Accumulation of H 2 O 2 was also absent or lower than 1 μM in incubations with MtLPMO9E, MtLPMO9H, and MtLPMO9I without RAC (data not shown). Apparently, MtLPMO9s are poor H 2 O 2 producers in contrast to NcLPMO9s (Table 2). H 2 O 2 production rates were further assessed and are discussed in Section 3.5. No H 2 O 2 accumulation occurred in the RAC sample in the presence of Pyg when MtLPMO9B was combined with either MtAA16A or AnAA16A (Figure 3).
Based on the above observations and the reported oxidase activity of LPMOs, 16 we hypothesized that MtAA16A produces H 2 O 2 to drive the peroxygenase reaction of MtLPMO9s in cleaving RAC (Figure 1, route 1). Therefore, we incubated both MtLPMO9B and MtAA16A and a mixture of these enzymes, with RAC in the presence or absence of Pyg, and reactions were analyzed by HPAEC-PAD (Figure 3). Both Mtenzymes were free of hydrolytic side activity and, as mentioned in the previous section, MtAA16A did not release (oxidized) oligosaccharides from RAC (Figures 3 and S9). Interestingly, in the combined MtAA16A and MtLPMO9B-RAC incubation ( Figure 3), a pronounced higher amount of nonoxidized (Glc) and oxidized cello-oligosaccharides (GlcOx) was released than in the same incubation with MtLPMO9B alone. Likewise, when the MtLPMO9B-RAC reaction was performed in the presence of AnAA16A, the increase in released products was substantial (Figure 3).
These findings provided support for our hypothesis that, in the presence of a reducing agent, MtAA16A and AnAA16A produce H 2 O 2 that can act as a cosubstrate for MtLPMO9B peroxygenase reactions in cleaving RAC. A comparable scenario has been described by Stepnov and co-workers, who observed that H 2 O 2 was continuously produced in situ by a CBM-truncated ScLPMO10C TR (only catalytic domain) to boost the full-length ScLPMO10C in degrading cellulose. 28

incubations). A shorter incubation (6 h) of RAC + Asc with
NcLPMO9s also showed no significant increase in oxidized products by AA16 addition ( Figure S11). The different boosting effects seen on MtLPMO9s and NcLPMO9s could be due to the different H 2 O 2 -producing abilities of individual LPMOs, as shown by Kittl and coworkers. 13 To determine the H 2 O 2 production rate of MtLPMO9s and NcLPMO9s, the Amplex Red/HRP assay was used instead of the ferric-xylenol orange assay. The former assay measures the H 2 O 2 level continuously (immediate reaction with H 2 O 2 ), while the latter assay determines the steady level of H 2 O 2 in the sample (after 30 min incubation with the reagent) that could lead to the underestimation of the H 2 O 2 -producing rate.
From Table 2, it follows that MtAA16A had a higher H 2 O 2producing activity (74.2 ± 3.3 mU) compared to all four MtLPMO9s. H 2 O 2 -producing activities for the boiled MtAA16A and equivalent amount of Cu(II) only samples were 5.0 ± 1.7 and 2.9 ± 1.3 mU, respectively. MtLPMO9s showed a lower H 2 O 2 -producing activity (MtLPMO9B, 16.6 ± 0.5 mU; MtLPMO9E, 3.2 ± 0.5 mU; MtLPMO9H, 45.4 ± 7.0 mU; MtLPMO9I, 41.3 ± 0.8 mU) compared to NcLPMO9C (301.9 ± 2.6 mU) and NcLPMO9M (510.0 ± 9.6 mU). These results indicated that MtLPMO9s are relatively poor H 2 O 2 producers, and thus, these LPMOs are likely to be boosted by the H 2 O 2 -producing AA16s in situ. Differently, NcLPMO9C and NcLPMO9M were able to produce larger amounts of H 2 O 2 , and it seems that they do not need more H 2 O 2 for their   Figure S14. The signal intensity of each peak and elution profiles in duplicated incubations are comparable, and only one chromatogram is shown here. reaction, as substantiated by the absence of boosting by the AA16s (i.e., for NcLPMO9C and NcLPMO9M). NcLPMO9F showed low H 2 O 2 -producing activity (32.8 ± 0.8 mU). However, previous studies showed its rapid inactivation and low catalytic ability, 13,35,53 which could explain the lack of boosting observed.
We also determined the H 2 O 2 -producing activity of MtAA16A, MtLPMO9s, and NcLPMO9s in the presence of 50 μM Asc (Table S4), the comparable concentration as reported by Kittl and co-workers. 13 From calibration curves, it turned out that the presence of Asc in the Amplex Red/HRP assay led to the underestimation of H 2 O 2 levels; however, 50 μM or 1 mM concentrations of Asc gave no differences in the absorbance at 560 nm ( Figure S12). In the presence of 50 μM Asc, NcLPMO9C and NcLPMO9M still displayed higher H 2 O 2 -producing activity than MtLPMO9s (Table S4), although all enzyme activities were lower than those in the presence of 1 mM Asc (cf. Table 2).

Stepwise Addition of H 2 O 2 is Less
Effective than AA16 Enzyme Supply to Drive the MtLPMO9B Peroxygenase Reaction. Manual stepwise addition of H 2 O 2 to stimulate the catalytic action of AA9 LPMOs to oxidatively cleave cellulose has been shown effective in other studies. 59−61 However, this has not been tested for MtLPMO9s. Hence, we compared such a setup with the AA16 supply for MtLPMO9B-RAC-Asc incubations. H 2 O 2 was added in six equal aliquots at six successive time points during the incubation, versus a single addition of AA16 at the start. MtLPMO9B (+Asc) released oxidized products from RAC (6 h; Figure 5), and the amount of oxidized products was further increased at 16 h ( Figure  S13). Based on the increased amounts of oxidized products formed at 6 h ( Figures 5 and S14), it is concluded that the stepwise addition of H 2 O 2 (0, 10, 25, 50, 100, and 200 μM) boosted the MtLPMO9B action as expected. For the stepwise addition of 50 μM or higher concentration of H 2 O 2 , no additional oxidized products were formed ( Figure S14) after 6 h of incubation, which can be the result of a damaged active site of the MtLPMO9B. 16,23,24 Notably, the addition of either 1 μM MtAA16A or 1 μM AnAA16A to 1 μM MtLPMO9B resulted in approximately a three times higher amount of oxidized products (control reactions are shown in Figure S15) compared to the most optimal H 2 O 2 concentration (50 μM) supplied to the MtLPMO9B-RAC digest ( Figure 5).
To test if H 2 O 2 addition boosted the activity of NcLPMO9s in our experimental setups, H 2 O 2 (six times 0, 10, 25, 50, 100, and 200 μM) was added to NcLPMO9M-RAC (+Asc) digestions. We observed that oxidative cleavage of RAC by NcLPMO9M was visibly boosted by 10 μM H 2 O 2 per addition (60 μM in total) (Figures S16 and S17). In a study by Petrovicá nd co-workers, NcLPMO9A, NcLPMO9C, and NcLPMO9D were shown to release considerably increased amounts of oxidized products at 4 h after manual stepwise addition of 45 μM H 2 O 2 . 61 As described in the previous section, it was anticipated that NcLPMO9s were able to produce sufficient H 2 O 2 to keep the catalytic activity maximum, but these results demonstrated that there was still room to further increase their activity. Apparently, the fact that AA16s boost MtLPMO9s but not NcLPMO9s cannot be merely explained by the in situ H 2 O 2 production.

Glucose Oxidase Is Less Effective than AA16 Enzymes in Boosting MtLPMO9B
Activity. To get more insight into the origin of the boosting effect, we performed time-course incubations of RAC with MtLPMO9B or NcLPMO9C with and without MtAA16A. In addition, we included AnGOX in the incubations with MtLPMO9B or NcLPMO9C to investigate if the same boosting effect as with AA16s could be achieved. AnGOX has been shown to produce H 2 O 2 to drive the peroxygenase reaction of LPMOs. 16 AnGOX (0.12 μg mL −1 ) was dosed based on the comparable H 2 O 2producing activity (76.4 mU) with 1 μM MtAA16A (74.2 ± 3.3 mU) used in the previous experiments. The required AnGOX concentration was calculated by a calibration curve (activity vs concentration) determined by using different concentrations of AnGOX ( Figure S18).
As expected from the previous results, AnGOX boosted MtLPMO9B-RAC degradation till 6 h ( Figure 6A). From   also has been reported in other studies. 16,23 Intriguingly, much higher amounts of oxidized cello-oligosaccharides were generated in the MtLPMO9B-RAC sample with MtAA16A compared to the one without MtAA16A, and even approximately 4 times higher than that in the MtLPMO9B-RAC sample with AnGOX at 6 h. In addition, it was found that the amount of oxidized cello-oligosaccharides was still increasing after 6 h in the MtLPMO9B-RAC sample with MtAA16A, indicating that less inactivation of MtLPMO9B compared to the sample with AnGOX had occurred ( Figure  6A). In other words, MtAA16A boosted MtLPMO9B and somehow also protected MtLPMO9B from the inactivation by H 2 O 2 . These observations strongly indicate, again, that the boosting effect is not only due to the in situ H 2 O 2 production by AA16s.
In contrast to MtLPMO9B-RAC samples, the amount of oxidized products formed in the NcLPMO9C-RAC sample with MtAA16A was only slightly higher than that without MtAA16A, and was equal to the amount in the NcLPMO9C-RAC sample with AnGOX till 6 h ( Figure 6B). From 6−16 h, almost no increase in the formation of oxidized products in all three NcLPMO9C-RAC samples occurred, indicating that NcLPMO9C was inactivated after 6 h.
We attempted to explain the observed different boosting effects on MtLPMO9s and NcLPMO9s by AA16s. We propose the challenging hypothesis that MtAA16A and AnAA16A interact with MtLPMO9s but not with NcLPMO9s. Such an interaction might assist the transmission of H 2 O 2 to the catalytic sites of the MtLPMO9s and stimulate their peroxygenase reaction ( Figure S19).
Attempts to experimentally confirm the hypothetical protein complex by size-exclusion chromatography and mass spectrometric techniques were not successful. It might be because MtAA16A and AnAA16A form weak transient interactions with the MtLPMO9s, which is challenging to study as reviewed by Qin et al. 62 However, of note here is that in a recent study, protein−protein interaction between a cell wall remodeling (CWR) protein CWR-1 AA11 LPMO and a CWR-2 membrane protein was implicated to be important for allorecognition of N. crassa. 63 Figure S19 shows a model of a hypothetical protein−protein interaction between MtLPMO9B and MtAA16A. It should be emphasized that this model, though having a high score, only represents an illustrative model and other factors such as glycosylation location, CBM, and linker were not taken into account. In addition, it cannot be excluded that also other or additional pathways might be valid, such as electron transfer between the AA16 and LPMO active sites.
3.8. Functions of AA16 Oxidoreductases in Nature. As listed in the CAZy database, three putative AA16 proteins have been identified in the genome of M. thermophila, and zero candidates have been found in the genome of N. crassa. 6 This observation may hint at a natural, evolution-driven difference and might relate to our results showing the interplay between the AA16s and MtLPMO9s, being absent for NcLPMO9s. This idea is strengthened by results from a recent study, in which Grieco et al. reported that, in the M. thermophila secretome, one AA9 LPMO (MYCTH_89312; MtLPMO9B) was detected together with another AA16 member (MYCTH_2311254) and one AA3 CDH (MYCTH_81925), when grown on partially delignified sugarcane bagasse. 36 CDH is a well-known electron-donating enzyme for AA9 LPMOs, 13,18,64,65 and we here suggest that the AA16s serve as H 2 O 2 producers, possibly even interacting with other LPMOs. In addition to the AA16s, other H 2 O 2 -producing enzymes are expected to also drive LPMO reactions, such as AA7 oligosaccharide oxidases, 66−69 and such AA7s have also been found to be coexpressed with LPMOs. 70,71 In this study, we were not able to find any carbohydrate substrates for oxidative cleavage by AA16s, though a very intensive substrate screening for AA16s was performed. It still cannot be excluded that AA16s are indeed LPMOs, but the biological substrates remain unknown. So far, only three AA16 members have been studied, which may not represent the complete picture of this family. Looking at the phylogenetic trees ( Figures S20 and S21), AA16s show high sequence variability. Filiatrault-Chastel and co-workers reported that many AA16s have a C-terminal extension, CBM1, or glycosylphosphatidylinositol (GPI) anchors in addition to the catalytic domains. 11 It is still unclear how these additional domains contribute to the AA16 functions in nature.
AA16 sequences were also found in the pathogenetic oomycetes Phytophthora and Pythium species. 11 More recently, AA16s were shown to be the only LPMO family members that coexpressed with the newly discovered AA17 pectin-active LPMOs during the infection of potato leaves by Phytophthora infestans. 7 Though the expression level was lower compared to AA17s, it still indicates that AA16s might play other roles in nature.

CONCLUSIONS
Our study has obtained insights into the catalytic and structural properties of MtAA16A and AnAA16A, members of a new family of CAZy enzymes. Although the crystallographic structure of MtAA16A showed a copper-containing His-brace typical for LPMOs, the adjacent substrate-binding surface differed. In addition, both MtAA16A and AnAA16A did not oxidatively cleave any of the investigated carbohydrates. We showed that both MtAA16A and AnAA16A produced (low levels) H 2 O 2 and stimulated the cellulolytic peroxygenase reaction of MtLPMO9s. No such stimulation was observed with NcLPMO9s, while both MtLPMO9s and NcLPMO9s were boosted by externally added H 2 O 2 . We showed that the strong AA16 boosting effect on MtLPMO9B cannot be achieved using a similar H 2 O 2 -producing activity of AnGOX. We propose that, within a hypothetical protein−protein complex, the formed H 2 O 2 might easily reach the catalytic site of MtLPMO9s, where it serves as a preferred cosubstrate to drive the peroxygenase reaction. Lastly, we discussed the possible functions of AA16s in nature, which deserve further investigation.
Deglycosylation of MtAA16A by PNGase F; incubation of MtAA16A with various carbohydrates; determination of peroxidase activity of MtAA16A and MtLPMO9s by the 2,6-DMP assay; binding of MtAA16A to regenerated amorphous cellulose; thermal shift analysis of MtAA16A interaction with cello-oligosaccharides and syringol; computational modeling of MtLPMO9B−MtAA16A interaction; phylogenetic analysis of the AA16 family; SDS-PAGE of MtAA16A and AnAA16A ( Figure S1); multiple sequence alignment and sequence identity matrices of AaAA16, MtAA16A, and AnAA16A ( Figure  S2); LC-MS analysis of the N-terminal peptide of MtAA16A ( Figure S3); 2,6-DMP activity of MtLPMO9s and MtAA16A ( Figure S4); comparison of the MtAA16A structure with the three highest DALI scoring structures ( Figure S5); comparison of the MtAA16A copper binding site with the three highest DALI scoring structures ( Figure S6); percentage of unbound MtAA16A protein after incubation with RAC ( Figure  S7); thermal denaturation of MtAA16A monitored by nDSF in the presence of potential ligands ( Figure S8 (Table S1); carbohydrates tested for screening MtAA16A activity under different conditions (Table S2); geometry at the MtAA16A copper binding site (Table S3); H 2 O 2 -producing activity of MtAA16A, MtLPMO9s, and NcLPMO9s in the presence of 50 μM Asc (Table S4); and references (PDF) ■ ACKNOWLEDGMENTS