The Ustilago maydis AA10 LPMO is active on fungal cell wall chitin

ABSTRACT Lytic polysaccharide monooxygenases (LPMOs) can perform oxidative cleavage of glycosidic bonds in carbohydrate polymers (e.g., cellulose, chitin), making them more accessible to hydrolytic enzymes. While most studies have so far mainly explored the role of LPMOs in a (plant) biomass conversion context, alternative roles and paradigms begin to emerge. The AA10 LPMOs are active on chitin and/or cellulose and mostly found in bacteria and in some viruses and archaea. Interestingly, AA10-encoding genes are also encountered in some pathogenic fungi of the Ustilaginomycetes class, such as Ustilago maydis, responsible for corn smut disease. Transcriptomic studies have shown the overexpression of the AA10 gene during the infectious cycle of U. maydis. In fact, U. maydis has a unique AA10 gene that codes for a catalytic domain appended with a C-terminal disordered region. To date, there is no public report on fungal AA10 LPMOs. In this study, we successfully produced the catalytic domain of this LPMO (UmAA10_cd) in Pichia pastoris and carried out its biochemical characterization. Our results show that UmAA10_cd oxidatively cleaves α- and β-chitin with C1 regioselectivity and boosts chitin hydrolysis by a GH18 chitinase from U. maydis (UmGH18A). Using a biologically relevant substrate, we show that UmAA10_cd exhibits enzymatic activity on U. maydis fungal cell wall chitin and promotes its hydrolysis by UmGH18A. These results represent an important step toward the understanding of the role of LPMOs in the fungal cell wall remodeling process during the fungal life cycle. IMPORTANCE Lytic polysaccharide monooxygenases (LPMOs) have been mainly studied in a biotechnological context for the efficient degradation of recalcitrant polysaccharides. Only recently, alternative roles and paradigms begin to emerge. In this study, we provide evidence that the AA10 LPMO from the phytopathogen Ustilago maydis is active against fungal cell wall chitin. Given that chitin-active LPMOs are commonly found in microbes, it is important to consider fungal cell wall as a potential target for this enigmatic class of enzymes.

L ytic polysaccharide monooxygenases (LPMOs) are monocopper enzymes that catalyze the oxidative cleavage of glycosidic bonds in carbohydrate polymers.Their discovery has been a major breakthrough in the understanding of the microbial enzymatic mechanisms involved in the degradation of natural recalcitrant polymers, including cellulose and chitin (1)(2)(3).LPMOs can act at the surface of polysaccharides, in synergy with other oxidoreductases (4) and glycoside hydrolases (GHs), to overcome polymers recalcitrance factors (such as crystallinity), thereby boosting bioconversion yields (5)(6)(7).The unique catalytic properties of LPMOs make them of utmost interest for different types of applications, such as the valorization of lignocellulosic biomass for the production of bioproducts (e.g., biofuels) (8,9) or bio-based materials (10)(11)(12).So far, most studies have explored the role of LPMOs from a biotechnology perspective.Excitingly, studies reported over the past few years have shown that LPMOs are present across nearly all kingdoms of life, encompassing thus a wide range of biological contexts, and pointing at the emergence of new roles and paradigms (13).
LPMOs are classified into eight auxiliary activity (AA) families in the carbohydrateactive enzymes (CAZy) database: AA9-AA11 and AA13-AA17 (13,14).Although their substrate specificity varies among AA families, they are mostly active on cellulose, chitin, xylan, starch, or pectin.The AA10 family is the most taxonomically diverse family with sequences originating from bacteria (15)(16)(17)(18)(19)(20)(21)(22)(23), viruses (24), and archaea (25).Interestingly, the presence of AA10s goes beyond the prokaryotic domain of life, as they have recently been shown to be present in some plants (ferns) (26,27) and in some pathogenic fungi.All AA10 LPMOs hitherto characterized have been found to be active on chitin (C1-oxidizers) and/or cellulose (C1-and C1/C4-oxidizers) (1,3,16).Of note, AA10s are often tested on model substrates such as shrimp/crab chitin or pretreated cellulose, and very little is known on their actual biological substrate.The identification of the latter is crucial to get insight into the biological function of a given LPMO but such approaches are challenging and require integration of the physiology of the living organism.A few attempts have been made along this line.Recent studies have notably shown the role of AA10 LPMOs in the chitin oxidative metabolism of a marine bacterium (28), in the virulence of pathogens (29)(30)(31), or even in the remodeling of bacterial cell wall (32).
Ustilago maydis is a biotrophic parasite (smut fungus) that depends on living tissue for proliferation and development in maize.This ubiquitous pathogen is also a well-estab lished model organism for the study of plant-microbe interactions (33,34).Compared to fungal saprotrophs and some plant pathogens, the U. maydis genome contains a small set of CAZymes (35) with 107 GHs and 23 AAs.While studying CAZymes from U. maydis (36,37), we noted the unusual presence of a unique AA10-encoding gene in its genome (UmAA10), the function of which remains unknown.Interestingly, digging into transcrip tomic data collected during U. maydis plant infection cycle (38), we noted that this UmAA10-encoding gene is overexpressed.Here, we managed to produce UmAA10 in a heterologous system and characterized it with model substrates and a more biologically relevant substrate prepared from U. maydis mycelium.We also probed the concerted action of UmAA10 together with a GH18 chitinase from U. maydis (UmGH18A).Our results, analyzed in light of previously published data in the context of in planta infection, provide hints on the biological function of UmAA10.

Bioinformatics analysis
To get insight into the putative function of UmAA10, we built a phylogenetic tree using 197 AA10 amino acids' sequences from bacteria, eukaryotes (fungi and plant), and viruses (Fig. 1).All fungal AA10 LPMOs (including UmAA10) cluster together within a larger clade of bacterial AA10s, some of which have been biochemically characterized as chitin-active with C1 regioselectivity.This observation could indicate horizontal gene transfer (HGT) events between bacteria and fungi (see Discussion).Another interesting observation is the presence of a predicted intrinsically disordered C-terminal region (dCTR) in UmAA10.dCTRs are regions of unknown function encountered in most LPMO families, and which occur in ~8% of AA10 sequences (39).Interestingly, while most Ustilaginomycetes AA10s display a dCTR, they are distantly related to bacterial AA10-dCTRs.More detailed analysis of these dCTR tails could help in finding common features among fungal AA10s or among AA10s from different species.To the best of our knowledge, none of these microbial AA10s bearing a dCTR have been characterized to date.
As mentioned above, UmAA10 displays a long region (120 residues) predicted to be intrinsically disordered by three different predictors, i.e., MobiDB-lite integrated in InterPro, IUPred2A, and AlphaFold pLDDT (Fig. S1) (41,45,46).The length of this dCTR is close to that of the other dCTRs identified in AA10s (median value = 123 residues; Fig. S1).The dCTR of UmAA10, enriched in Ser (35%), Gly (11%), Ala (10%), Arg (8%), and Thr (8%), is overall hydrophilic with a positively charged patch (Arg-rich) at the C-terminus.Charged residues and distribution of residues of opposite charge along the sequence are one of the main determinants of the conformational properties of intrinsically disordered regions (47)(48)(49).In addition, they could be involved in protein-protein or proteinmembrane electrostatic interactions.Interestingly, the last 47 residues of the UmAA10-dCTR are also predicted by ANCHOR2 (46) to be a disordered binding site.The alignment of the last 50 residues of dCTRs from 30 AA10s revealed some degree of conservation and the presence of a patch of highly positive charged residues at the end of the sequence [see Fig. S2 and reference (50)], which might suggest a common, yet uneluci dated, role across fungal AA10s.

Recombinant production of UmAA10
To investigate the biochemical properties of UmAA10, we attempted its heterologous production in the yeast Pichia pastoris.Unfortunately, even after several attempts, the recombinant production of UmAA10 (in its full-length form) failed, probably due to the disordered nature of the long C-term extension.Therefore, we decided to produce only the catalytic domain (hereafter referred to as UmAA10_cd).Interestingly, we observed upon UmAA10_cd recombinant production a striking difference in cells' sedimentation and cell shape/size (Fig. S3).Addition of ethylenediaminetetraacetic acid (EDTA) to the culture (to potentially inactivate UmAA10_cd by copper chelation) restored a "normal" sedimentation behavior of P. pastoris cells and partially restored cell shape/size (Fig.

S3
).However, the recombinant production yield was not further improved (~0.35 mg of pure protein per liter of culture).After a two-step purification by nickel-affinity and size-exclusion chromatography, UmAA10_cd was purified to homogeneity.It displayed an apparent molecular weight of ~20 kDa on SDS-PAGE (Fig. S4) in good agreement with the theoretical one (19.5 kDa).After copper loading, UmAA10_cd contains ~1.4 copper atom per protein molecule, as determined by inductively coupled plasma mass spectrometry (ICP-MS).

Substrate specificity of UmAA10_cd
Before assessing the substrate specificity of UmAA10_cd, we first tested its ability to reduce O 2 to H 2 O 2 in the presence of an appropriate reducing agent (51).Using ascorbic acid (AscA) as a reductant, we found that UmAA10_cd produces H 2 O 2 at a rate of 3.8 × 10 −4 s −1 , which reflects a relatively average oxidase activity, compared to other LPMOs (4) (Fig. S5).To investigate the substrate specificity of UmAA10_cd, we tested several relevant model substrates of AA10 LPMOs, i.e., α-chitin, β-chitin, crystal line cellulose (Avicel), and phosphoric acid swollen cellulose (PASC), with AscA being consistently used as a reducing agent.We observed the release of several products corresponding to C1-oxidized chito-oligosaccharides (DP2-DP4) from α-and β-chitin (Fig. 3A and B).No activity was detected on cellulose (Fig. S6).These results are in line under stirring (1,000 rpm) at 30°C for 24 h.For each condition, three independent biological replicates were carried out and, for the sake of clarity, only one replicate is shown.
with our bioinformatic predictions (Fig. 1).To deepen our knowledge on UmAA10_cd, we further tested its activity in synergy with a GH18 chitinase.Indeed, many studies have demonstrated the synergistic action of LPMOs and GHs in the degradation of plant polysaccharides (52,53), chitin (1), and other glycans such as peptidoglycan (32).In the context of our study, we hypothesized that UmAA10_cd could work in synergy with one of the GH18 from U. maydis for the degradation of chitin.The genome of U. maydis contains three genes coding for GH18s: UMAG_02758 (UmGH18A), UMAG_10419 (UmGH18B), and UMAG_06190 (UmGH18C) [respectively, cts2, cts1, and cts3 in reference (54)].While UmGH18A and UmGH18C harbor a signal peptide, UmGH18B is predicted to be secreted by an unconventional pathway (55).In this study, we focused our attention on UmGH18A, which was previously shown to be active during yeast growth (54) and upregulated during maize infection (38).After successful production in P. pastoris and purification to homogeneity (Fig. S4), we evaluated its activity toward chitin.UmGH18A displayed optimal activity at 40°C and pH between 4 and 5 (Fig. S7), and exhibited exo-acting hydrolytic activity on α-and β-chitin, releasing chitobiose (CHOS 2) as a major product (Fig. S8).Interestingly, UmGH18A seems to be more active on β-chitin.
It is now clearly established that the preferred co-substrate of LPMOs during the oxidative cleavage of polysaccharides is H 2 O 2 (56,57).LPMOs have also been reported to exhibit a peroxidase activity in the presence of specific organic reductants and H 2 O 2 (58).However, using reaction conditions described by Bissaro et al. (44) and Breslmayr et al. (58) (see Materials and Methods for further details), we could not detect any significant peroxygenase activity (on α-chitin) or peroxidase activity (on 2,6-dimethoxyphenol [DMP]) (data not shown).This result is possibly due to an extra-sensitivity of UmAA10_cd to H 2 O 2 as it has been shown that LPMOs are subject to oxidative inactivation (56,59,60), a phenomenon that depends on various reaction conditions.

Enzymatic assays on a biologically relevant substrate
To go further into the biochemical characterization UmAA10_cd, we aimed at testing it on a more biologically relevant substrate.As plant cell wall is devoid of chitin, the most obvious source of chitin in the environment of U. maydis is its own cell wall, which is remodeled during pathogenesis.We therefore prepared a fraction containing chitin from U. maydis mycelium produced at the filamentous stage (see Materials and Methods).The presence of chitin in the alkali-insoluble fraction was confirmed by compositional analysis and by fluorescence microscopy using the labeled wheat germ agglutinin (WGA) lectin (Fig. S9), a reagent commonly used to study chitin and chitin-like molecules (61).Taking advantage of this U. maydis fungal cell wall fraction (UmFCW), we were able to assay the activity of UmAA10_cd on a more natural form of chitin.As previously shown on model chitin, UmAA10_cd was found to exhibit enzymatic activity on UmFCW with the release of C1-oxidized CHOS.We also provide experimental evidence that UmAA10_cd acts in concert with UmGH18A on UmFCW increasing the overall release of soluble CHOS (Fig. 4).

DISCUSSION
In this study, we demonstrate, using complementary approaches, that UmAA10 is an LPMO that oxidatively cleaves U. maydis FCW extracts containing chitin, and that acts together with a GH18 chitinase from U. maydis (UmGH18A).We would like to emphasize that this study reports the first characterization of a fungal LPMO belonging to the AA10 family and the first biochemical evidence of LPMO activity toward FCW chitin.The appearance of AA10 LPMOs in the fungal kingdom may be ascribed to HGT.In fact, HGT between bacteria and eukaryotes has been documented for cell wall degrading enzymes (62,63) and it is well known that HGT can provide a new function or replace a functional loss in the recipient organism, allowing adaptation to its environment (64).It would therefore be interesting to further investigate these events by focusing on the fungal members of the AA10 family.
On the basis of transcriptomic data from Lanver et al. (38), we know that the gene encoding UmAA10 is overexpressed during advanced stages of maize infection, i.e., between 4 and 12 days post-infection (dpi) (Fig. 5A), when we observe major morpho logical changes that entail a fast and massive rearrangement of U. maydis cell wall with modification of FCW polysaccharides, including chitin.The FCW is a key adaptable scaffold for the survival of the fungus notably during host infection (65).The structural and mechanical properties of the FCW allow the fungus to resist the turgor pressure necessary for polar growth (66) and to withstand environmental stresses (67,65,68).While the mechanisms involved in this cell wall biochemical changes are still poorly known, some CAZymes from U. maydis, such as chitinases (54) and chitin deacetylases (69), have been shown to be involved in FCW modification and virulence.However, to date the role of oxidative enzymes in FCW modification has not been deeply investigated in a chitinolytic context.Of note, we recently characterized two CAZymes from U. maydis that act in synergy on β-glucans, namely a glucanase (UmGH16_1A) active on β-1,3glucans branched with short β-1,6 substitutions and a dehydrogenase (UmAA3_2A) active on β-1,3 and β-1,6-gluco-oligosaccharides released by the former (37).As these substrates are also FCW components, we believe these enzymes could play a role in FCW remodeling.
In a different context, it was recently discovered that a chitin-active AA11 LPMO from Neurospora crassa (CWR-1) is involved in cell-cell allorecognition and cell wall dissolu tion during cell fusion (70).Additionally, another fungal LPMO from the AA9 family (CnCEL1), which harbors a dCTR and is localized at the cell surface, has been suggested to be involved in cell wall integrity and cell cycle of the human pathogen Cryptococcus neoformans (71).Beyond the fungal kingdom, a recent study similarly revealed the involvement of an AA10 LPMO from Streptomyces coelicolor (named as LpmP by the authors) in cell wall remodeling (32).This LPMO is retained in the bacterial cell wall Full-Length Text and facilitates peptidoglycan degradation by lysozyme.Together with our findings, it suggests that the remodeling of microbial cell wall polysaccharides by LPMOs may be widespread in nature.Considering the robustness of FCW chitin (72) and while U. maydis displays no AA11 LPMO and a unique chitin-active AA10 LPMO, we can hypothesize that UmAA10, in addition to U. maydis chitinases, could be involved in the modification of chitin upon cell wall remodeling and morphological changes during maize infection (Fig. 5).Despite the need for functional in cellulo validation in U. maydis, this proposed biological role is supported by the presence of a dCTR that could anchor UmAA10 in the FCW and by the phenotype observed during recombinant production of UmAA10_cd that suggests modification of the yeast cell wall by LPMO.

Conclusion
In the present study, we identified and characterized the catalytic domain of the first fungal LPMO belonging to the AA10 family.This LPMO, which is unique in the genome of Ustilago maydis, was found to be active on model chitin but also on chitin-containing UmFCW extracts.Importantly, we showed that the LPMO activity contributes to chitin hydrolysis by an endogenous GH18 chitinase (UmGH18A).We believe that the acquisi tion by the plant pathogen U. maydis of such unique oxidative activity, possibly by HGT, is not fortuitous and plays an important role in the fungus biology.

Bioinformatic analyses
To build up the phylogenetic tree, 197 AA10 amino acids' sequences (catalytic domain only), including nine sequences from fungi (Ustilaginomycetes) and 46 sequences of characterized AA10 LPMOs, were selected.Multiple sequence alignment was performed using the MAFFT-L-INS-i methods via the MAFFT online platform (74).The phylogenetic tree was generated with IQ-TREE online platform (75), with the maximum likelihood method.The Whelan and Goldman amino acid substitution model was selected (76).Branch support was calculated by 1,000 Bootstrap repetitions (values displayed in percent on the tree).The tree was visualized in iTOL (77) and edited in Illustrator CC 2017.The C-terminal regions of unknown function were extracted and analyzed with the disorder predictor IUPred2A (46).The program returns a score between 0 and 1, which represents the probability for each residue of being part of a disordered region (score ≥ 0.5).UmAA10 C-terminal region was further analyzed with MobidB-lite 3.0 using default parameters and the AlphaFold pLDDT score [values ≤ 68.8; see reference (78)].The properties of the dCTRs were derived from their primary sequence through the toolkit localCIDER (79).

Cloning and production of recombinant enzymes
The proteins were produced using the in-house 3PE platform (Pichia pastoris pro tein express; www.platform3pe.com/)as described in reference (80).The nucleotide sequence coding for the AA10 of Ustilago maydis (Genbank ID XP_011391789.1)was codon optimized and synthesized (GenScript, Piscataway, USA) for expression in Pichia pastoris (syn.Komagataella phaffii).The region corresponding to the native signal sequence was kept, and upon production of the catalytic domain only (UmAA10_cd), the C-terminal extension region was removed (at position corresponding to between Gly203 and Gly204).The genes encoding Serratia marcescens GH20 (SmGH20) (Genbank ID Q54468.1) and U. maydis GH18A (UmGH18A) (Genbank ID XP_011389118.1)were synthesized after codon optimization for P. pastoris (GenScript) using the full-length native sequence.Each gene was inserted into the expression vector pPICZαA (Invitrogen, Cergy-Pontoise, France) with a C-terminal (His) 6 -tag.PmeI-linearized pPICZαA recombi nant plasmids were used to transform by electroporation competent P. pastoris cells (SuperMan5 for UmAA10_cd and SmGH20, and X33 for UmGH18A).Zeocin-resistant transformants were then screened for protein production.The best-producing trans formants were grown in 2 L of buffered glycerol-complex medium (BMGY) (10 g/L yeast extract, 20 g/L peptone, 3.4 g/L yeast nitrogen base without amino acids, 10 g/L ammonium sulfate, 10 g/L glycerol, potassium phosphate buffer pH 6.0, biotin 0.4 mg/L) in flasks at 30°C in an orbital shaker (200 rpm) for 16 h to an optical density ( OD 600 ) of 2-6.Expression was induced by transferring cells into 400 mL of buffered methanol-complex medium (BMMY) (10 g/L yeast extract, 20 g/L peptone, 3.4 g/L yeast nitrogen base without amino acids, 10 g/L, ammonium sulfate, potassium phosphate buffer pH 6.0, biotin 0.4 mg/L).The medium was supplemented (initially and everyday) with 1% (vol/vol) methanol at 16°C for UmAA10_cd and 3% (vol/vol) methanol at 20°C for SmGH20 and UmGH18A in an orbital shaker (200 rpm) for another 3 days.The cells were harvested by centrifugation (5,000 rpm, 10 min at 4°C), and just before purification, the pH of the supernatant was adjusted to 7.8 with sodium hydroxide and was filtered at 0.45 µm (Millipore, Burlington, Massachusetts, USA).

ICP-MS analysis
Copper content was analyzed using ICP-MS as described in reference (81).The samples were mineralized, then diluted in ultrapure water, and analyzed by an ICAP Q apparatus (Thermo Electron, Les Ullis, France).The copper concentration was determined using Plasmalab (Thermo Electron) software, at m/z = 63 with an accuracy of ±5%.

H 2 O 2 consumption and production assay
The ability of UmAA10_cd to produce H 2 O 2 (oxidase assay) was determined by absorbance measurement, using the Amplex red assay (Thermo Fisher Scientific) and HRP, as described by Kittl et al. (51).The reaction mixture (100 µL in 96-well microplate) contained sodium phosphate buffer (pH 7.0, 50 mM), HRP (0.1 mg/mL), Amplex red (100 µM), and UmAA10_cd (1 µM).The reaction, incubated at 23°C, was initiated by adding ascorbic acid (50 µM) and the release of resorufin was monitored at 575 nm during 40 min.Two control reactions were carried out in the same condition: one without UmAA10_cd and another without ascorbic acid.The stoichiometry of the resorufin release in regard to the H 2 O 2 produced is 2:1.A standard curve of H 2 O 2 was performed under the same conditions.H 2 O 2 consumption was measured using two different assays.The first one was carried out following the protocol described by Breslmayr et al. (58).This assay is based on the oxidation of 2,6-DMP to coerulignone by the enzyme in the presence of H 2 O 2 (peroxidase reaction).The stoichiometry of the reaction is 2 H 2 O 2 consumed or 2 molecules of 2,6-DMP oxidized per coerulignone produced.Reactions were run in a 100 µL mixture, containing borate buffer (50 mM, pH 7.5), UmAA10_cd (1 µM), 2,6-DMP (500 µM), and H 2 O 2 (100 µM).The release of coerulignone was monitored at 469 nm in a plate reader (ε469 = 53,200 M −1 /cm).The standard curve obtained was used to determine H 2 O 2 consumption kinetic parameters.In the second reaction, H 2 O 2 consumption was measured in the presence of the polysaccharide as described by Bissaro et al. (56), with a monitoring by the method described by Kittl et al. (51).Reactions were conducted in a 600 µL mixture containing sodium phosphate buffer (50 mM, pH 6.0), α-chitin as substrate (10 g/L), EDTA (50 µM), and UmAA10_cd (50 nM).After 10 min of incubation (30°C, 1,000 rpm), H 2 O 2 (100 µM) was added to the mixture and the reaction was initiated by adding ascorbic acid (20 µM).Seventy-five microliters of the samples was taken from the mixture at 0, 5, 10, 15, 20, 30, and 60 min of reaction, filtered using a 96-well filtration plate (0.22 µm filters, Merck Millipore, Ireland and 25 µL of each filtered sample was added to 75 µL of oxidase reagent as described above.A H 2 O 2 standard curve was realized in the condition of the oxidase assay.

Enzymatic characterization of UmGH18A
Temperature and pH optima for UmGH18A were determined as described previously (82) with slight modifications.The optimal pH was determined using various buffers (50 mM), within the pH range of 3-9.The buffers used were sodium citrate (pH 3, 4, 5, and 6), sodium acetate (pH 4, 5, and 5.5), sodium phosphate (pH 6, 7, and 8), and Tris-HCl (pH 7.2, 8, and 9).While, for determining optimum temperature, the chitinase activity assay was performed under the optimum pH, within a temperature range of 20°C-70°C.In both experiments, the reaction mixture contained colloidal chitin (10 g/L) as the substrate and purified UmGH18A (1µM), incubated at the respective temperature under stirring at 800rpm for 1h.Chitinase activity was determined by analyzing the presence of reducing sugars using Schales' assay as described previously (82).
Time course degradation of crystalline chitin substrates was performed by incubating UmGH18A (1 µM) with α-or β-chitin (10 g/L) under the optimal conditions and shaking at 1,000 rpm.Aliquots were collected at different time points between 1 and 24 h and filtered using a 96-well filter plate (0.45 µm filters; Merck Millipore, USA) operated by a Millipore vacuum manifold.The filtered samples were then mixed with an equal volume of 70% acetonitrile and analyzed using high-performance liquid chromatography (HPLC).The products obtained were separated on Shim-pack GIST NH2 column (5 µm, 4.6 × 250 mm, Shimadzu, Japan), through isocratic elution using 70% acetonitrile, with a 0.7 mL/min flow rate.Throughout the analysis, the column oven temperature was maintained at 45°C, and the products were detected at 210 nm.Quantification of the CHOS was performed essentially as described earlier (83).

Ustilago maydis growth conditions and cell wall preparation
The U. maydis strain [521/FGSC 9021; see reference (33)], which was provided by the CIRM-CF collection (strain CIRM-BRFM1093) (85) was grown in 100 mL of yeast extract peptone dextose (YPD) medium (10 g/L yeast extract, 20 g/L peptone, 20 g/L dextrose) for 48 h at 28°C in 250-mL baffled Erlenmeyer flasks under orbital agitation (150 rpm).Cells were then harvested, washed once in H 2 O by centrifugation (1,500g, 10 min), and stored at −80°C in 20% glycerol at 10 7 cells/mL.To produce material for sequen tial extraction (Fig. S9), eight Roux flasks containing 180 mL of YPD medium were inoculated at 10 5 cells/mL and incubated for 17 days at 28°C.The resulting mycelium was then harvested and washed three times with H 2 O by filtration on Miracloth and lyophilized.Five grams of this material was resuspended in 500 mL H 2 O, homogenized using Ultra-Turrax (13,500 rpm, 2 min), and boiled for 3 h.After centrifugation (6,000g, 10 min), the supernatant was discarded and the pellet was resuspended in 500 mL of NaOH (1.25 M) for 4 h at 60°C.After another centrifugation step, the alkali insoluble residue was washed three times using centrifugation (8,000g, 20 min) in 1 L of H 2 O.The pellet referred to as UmFCW, which contains chitin, was lyophilized, weighted (967 mg), resuspended in H 2 O in 1:1 ratio (wt/vol), and stored at 4°C until further use.

Compositional analysis of fungal cell wall polysaccharides
Identification and quantification of polysaccharide neutral sugars were performed by gas-liquid chromatography (GC) after sulfuric acid degradation as described in reference (86).Briefly, 5 mg of dried alkali insoluble fraction was dispersed in 2 N trifluoroacetic acid and then hydrolyzed 90 min at 121°C.Neutral monosaccharides were converted to alditol acetates and analyzed on a TG-225 GC Column (30 × 0.32 mm ID) using TRACE Ultra Gas Chromatograph (Thermo Scientific; temperature 205°C, carrier gas H 2 ).Standard sugars' solution and inositol as internal standard were used for calibration.
Glucosamine residues were quantified after acid hydrolysis [adapted from reference (87)] and HPAEC-PAD analyses [adapted from reference (88)].Briefly, 5 mg of dried alkali insoluble fraction was dispersed in acetic acid 1% and hydrolyzed in concentrated HCl 10 M at 105°C during 6 h.After dilution in milli-Q water, glucosamine residues were quantified by HPAEC-PAD (ICS-6000, Thermo Scientific) using a CarboPac PA20 column (2 × 250 mm, Thermo Scientific), thermostated at 30°C.An isocratic elution of 1.7 mM sodium acetate (NaOAc) in 1 mM NaOH was used at a 0.25 mL/min flow rate.Standard glucosamine solutions were used for calibration.

FIG 1
FIG1 Phylogenetic tree of AA10s.The phylogenetic tree was generated based on a multiple alignment of 197 sequences (AA10 catalytic domain only).The predicted substrate specificity is based on the presence of characterized AA10 LPMOs (in red) within each clade.The substrate specificity of clades highlighted in gray has not been determined yet.AA10 sequences with a predicted dCTR are highlighted with a black star.

FIG 2
FIG 2 Structural analysis of UmAA10.(A) Overview of the predicted 3D structure (AlphaFold2) of UmAA10 and of the crystal structure of SmAA10A [PDB 2BEM; see reference (15)].The catalytic domain of the predicted UmAA10 structure is shown in salmon and the C-terminal extension in black.(B) UmAA10 predicted substrate-binding residues (shown as sticks) based on those identified in SmAA10A.UmAA10 residues (labeled in black) differing from the corresponding residues occurring in SmAA10A (labeled in orange) are marked by a red star.

FIG 4
FIG 4 Activity of UmAA10_cd on U. maydis fungal cell wall chitin.(A) The graph shows high-performance anion-exchange chromatography-pulsed amperomet ric detection chromatograms of oxidized CHOS released from UmFCW (10 g/L) by either UmGH18A (0.1 µM) or UmAA10_cd (1 µM) and AscA (1 mM), alone or combined.(B) Quantification of total soluble products (from the reaction described in panel A).A GH20 chitobiase from Serratia marcescens (SmGH20) was used to degrade the mixtures of CHOS to GlcNAc (CHOS 1) and GlcNAc-GlcNAc1A (CHOS 2 ox ), as described by Loose et al. (17).All reactions were carried out in sodium phosphate buffer (50 mM, pH 6.0) under stirring (1,000 rpm) at 30°C for 24 h.Data points show average values and error bars correspond to standard deviations from three independent biological replicates.

FIG 5
FIG 5 Putative function of UmAA10 in vivo.(A) Normalized expression of genes coding for UmAA10 and UmGH18A during corn infection, illustrated with data adapted from Lanver et al. (38).Under each bar graph, is presented the corresponding U. maydis infection cycle stage.dpi: day post-infection.(B) Schematic view of the proposed cooperation between UmGH18A and UmAA10 in vivo.The figure was created using BioRender.com.