Kinetics of H2O2-driven catalysis by a lytic polysaccharide monooxygenase from the fungus Trichoderma reesei

Owing to their ability to break glycosidic bonds in recalcitrant crystalline polysaccharides such as cellulose, the catalysis effected by lytic polysaccharide monooxygenases (LPMOs) is of major interest. Kinetics of these reductant-dependent, monocopper enzymes is complicated by the insoluble nature of the cellulose substrate and parallel, enzyme-dependent, and enzyme-independent side reactions between the reductant and oxygen-containing cosubstrates. Here, we provide kinetic characterization of cellulose peroxygenase (oxidative cleavage of glycosidic bonds in cellulose) and reductant peroxidase (oxidation of the reductant) activities of the LPMO TrAA9A of the cellulose-degrading model fungus Trichoderma reesei. The catalytic efficiency (kcat/Km(H2O2)) of the cellulose peroxygenase reaction (kcat = 8.5 s−1, and Km(H2O2)=30μM) was an order of magnitude higher than that of the reductant (ascorbic acid) peroxidase reaction. The turnover of H2O2 in the ascorbic acid peroxidase reaction followed the ping-pong mechanism and led to irreversible inactivation of the enzyme with a probability of 0.0072. Using theoretical analysis, we suggest a relationship between the half-life of LPMO, the values of kinetic parameters, and the concentrations of the reactants.

Owing to their ability to break glycosidic bonds in recalcitrant crystalline polysaccharides such as cellulose, the catalysis effected by lytic polysaccharide monooxygenases (LPMOs) is of major interest. Kinetics of these reductant-dependent, monocopper enzymes is complicated by the insoluble nature of the cellulose substrate and parallel, enzyme-dependent, and enzyme-independent side reactions between the reductant and oxygen-containing cosubstrates. Here, we provide kinetic characterization of cellulose peroxygenase (oxidative cleavage of glycosidic bonds in cellulose) and reductant peroxidase (oxidation of the reductant) activities of the LPMO TrAA9A of the cellulose-degrading model fungus Trichoderma reesei. The catalytic efficiency ðk cat =K mðH2O2Þ Þ of the cellulose peroxygenase reaction (k cat = 8.5 s −1 , and K mðH2O2Þ ¼ 30 μM) was an order of magnitude higher than that of the reductant (ascorbic acid) peroxidase reaction. The turnover of H 2 O 2 in the ascorbic acid peroxidase reaction followed the ping-pong mechanism and led to irreversible inactivation of the enzyme with a probability of 0.0072. Using theoretical analysis, we suggest a relationship between the half-life of LPMO, the values of kinetic parameters, and the concentrations of the reactants.
Lytic polysaccharide monooxygenases (LPMOs) are monocopper enzymes that catalyze oxidative cleavage of glycosidic bonds in various polysaccharides (1)(2)(3)(4)(5). LPMOs are widespread in nature, and they are classified within several auxiliary activity families in the database of carbohydrate-active enzymes (6). Although structurally and biochemically well characterized (7,8), the kinetic data of LPMOs are scarce. Besides polysaccharide substrate and oxygen-containing cosubstrate (O 2 /H 2 O 2 ), LPMOs need a reductant for their activity (9)(10)(11)(12). In H 2 O 2 -driven oxidative cleavage of glycosidic bond (polysaccharide peroxygenase reaction), the reductant is needed only for the initial priming reduction of LPMO-Cu(II) resting state to a catalytically active LPMO-Cu(I) (12). LPMO-Cu(I) activates H 2 O 2 leading to the hydroxylation and cleavage of glycosidic bond ( Fig. 1) (12). When the LPMO active site is free from polysaccharide, the LPMO-Cu(I) can be reoxidized by H 2 O 2 to LPMO-Cu(II). The latter results in the stoichiometric oxidation of the reductant and is referred to as the reductant peroxidase reaction (Fig. 1). Kinetic studies of LPMOs are complicated by the insoluble nature of the substrate, various side reactions such as reductant oxidase/peroxidase activity of LPMOs, and inactivation of LPMOs in reductant peroxidase reaction (13). Furthermore, an enzyme-independent oxidation of the reductant by O 2 often leads to the formation of H 2 O 2 (14,15). These multiple parallel side reactions not only complicate the analysis of the kinetic data but may also lead to wrong conclusions regarding the nature of the cosubstrate as exemplified by the LPMO of the bacterium Serratia marcescens (SmAA10A), which was first identified as a monooxygenase (1) but later turned out to be a peroxygenase (16). An analogous example is provided by a fosfomycin-producing non-heme iron epoxidase that was initially identified as an oxidase (17) but turned out to be a peroxidase (18). Despite the major interest, to date, a detailed kinetic characterization is available only for the O 2 -driven oxidation of soluble cellohexaose by an LPMO from the fungus Myceliophthora thermophila (MtPMO9E) (19), the H 2 O 2 -driven oxidation of insoluble chitin by SmAA10A (20) and soluble chitooligosaccharides by AfAA11B from the fungus Aspergillus fumigatus (21). In a recent study, several LPMOs were also characterized in terms of catalytic efficiency ðk cat =K mðH 2 O 2 Þ Þ of the cellulolytic peroxygenase reaction (22). Since the discovery of the peroxygenase activity in 2017 (16), a number of further studies have reported much higher rates with H 2 O 2 in the oxidation of both, the substrate (15,(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) and the reductant (28,(31)(32)(33)(34)(35). The results of computational studies also suggest that H 2 O 2 is a relevant cosubstrate of LPMOs (33,36). Because of the great potential in various applications, the activity of LPMOs on cellulose is of particular interest (37). Enzymatic degradation of lignocellulose takes place in the complex redox-active environment (38,39) that can provide LPMOs with electrons and the H 2 O 2 cosubstrate (26). However, the lack of the in-depth knowledge on the kinetics of LPMO catalysis hampers rational finetuning of the reaction conditions toward optimal activity and stability of the LPMOs.
Here, we used a uniformly 14 C-labeled cellulose substrate that enabled kinetic characterization of the H 2 O 2 -driven cellulose peroxygenase reaction by the LPMO (TrAA9A) of the model fungus Trichoderma reesei (also known as Hypocrea jecorina). The catalytic efficiency of the cellulose peroxygenase reaction (2.9 × 10 5 M −1 s −1 ) was an order of magnitude higher than that of the reductant (ascorbate, AscA) peroxidase reaction. The AscA peroxidase reaction also led to the irreversible inactivation of TrAA9A with the probability of 0.0072.

Kinetics of cellulolytic peroxygenase reaction
The use of a 14 C-labeled polysaccharide substrate enables sensitive detection of soluble LPMO products (20). Here, we provide a kinetic characterization of the cellulolytic peroxygenase reaction of an LPMO using uniformly 14 C-labeled bacterial microcrystalline cellulose (BMCC) as a substrate and TrAA9A as a model LPMO. TrAA9A has been shown to release both C1 and C4 oxidized sugar products from cellulose (40)(41)(42). Before we describe the results, we note that, provided with the value of the stoichiometry showing the amount of the released soluble products per molecule of H 2 O 2 consumed, the regioselectivity of oxidation and the distribution of products between the soluble and the insoluble fractions are not important in deriving the values of kinetic parameters (20).
In the presence of the reductant, ascorbic acid (AscA), and H 2 O 2 , TrAA9A released 14 C-labeled soluble products from BMCC (Fig. S1). No release of radioactivity was observed in the absence of added H 2 O 2 (Fig. S1), suggesting that the possible TrAA9A monooxygenase reaction and the production of H 2 O 2 in TrAA9A-dependent (AscA oxidase) and TrAA9A-independent reactions are insignificant under our study conditions. The rate of the release of soluble products (expressed in glucose equivalents, Glc eq ) increased with the increasing concentration of AscA (Fig. S2). The apparent half-saturating concentration of AscA was around 0.1 mM, and based on these results, we choose 1 mM AscA in the further experiments of BMCC degradation.
Progress curves of the release of soluble Glc eq from BMCC by TrAA9A at different [H 2 O 2 ] are shown in Figure 2A. Single exponential function was used as a first approximation in the analysis of the progress curves (20).
In Equation 1, [Glc eq ] is the concentration of soluble products (μM), [Glc eq ] max is the maximum concentration of Glc eq that is released under given experiment conditions, and k obs is the observed first-order rate constant (s −1 ). [Glc eq ] max first increased with increasing [H 2 O 2 ] but started to decrease at H 2 O 2 concentrations above 50 μM (Fig. 2B). On the other hand, the initial decrease of k obs with [H 2 O 2 ] was followed by its increase at [H 2 O 2 ] above 50 μM (Fig. 2C). These kinetic signatures are similar to those observed at the H 2 O 2 -driven degradation of chitin by SmAA10A and suggest that the irreversible inactivation of LPMO takes place in parallel with the polysaccharide peroxygenase reaction (20 (20).
The dependency of the initial rates of Glc eq formation on [H 2 O 2 ] was consistent with the Michaelis-Menten equation (Fig. 2D). The Michaelis-Menten curves measured at BMCC concentrations of 0.5, 1.0, and 1.5 g l −1 overlapped within experiment scatter, indicating that the enzyme was saturated with cellulose ( Fig. S4A). Measurements of the concentration of the cellulose-free TrAA9A also suggested strong binding to BMCC (Fig. S5) in the presence of 1 mM AscA. The binding was significantly weaker in the absence of AscA (Fig. S5). High affinity to BMCC does not allow precise measurement of initial rates at the subsaturating BMCC concentrations (i.e., very low), and thus, we cannot assess the dependency of apparent V max and K mðH 2 O 2 Þ values on [BMCC] (Fig. S4  concentrations of 0.5, 1.0, and 1.5 g l −1 (Fig. 2D), we estimated the true V max and K mðH 2 O 2 Þ values of 1.28 ± 0.05 μM Glc eq s −1 and 30 ± 5 μM, respectively. Considering the total concentration of TrAA9A of 0.05 μM, V max translates to the k cat value of 25.6 ± 1.0 soluble Glc eq s −1 . To find the k cat value for the cellulolytic peroxygenase reaction, one must know the value of the stoichiometry coefficient (n), which shows an average number of soluble products (in Glc eq ) released per molecule of H 2 O 2 consumed in the peroxygenase reaction.  (Fig. 2B), and based on these figures, we estimated the value of n = 2.90 ± 0.05 soluble Glc eq /H 2 O 2 . We also used an alternative approach for the determination of the value of n, where H 2 O 2 was in situ produced by the glucose/glucose oxidase (GO) reaction with the rate of 1.2 ± 0.1 μM min −1 (Fig. S6A).
The time curves of Glc eq formation were independent of [TrAA9A], indicating that the rate is limited by the GO reaction but deviated from linearity with the effect being larger at lower [BMCC] (Fig. S6B). However, at the shortest incubation time (10 min), the rate of Glc eq formation was independent of [BMCC], and based on these data, we estimated the value of n = 3.1 ± 0.2 soluble Glc eq /H 2 O 2 (Fig. S6B).
Combining the n values measured using two different approaches results in an average n value of 3.0 ± 0.15 soluble Glc eq /H 2 O 2 for the TrAA9A/BMCC system. Provided with the value of n, we can now calculate the k cat value of 8.5 ± 0.4 s −1 and k cat =K mðH 2 O 2 Þ value of 290,000 ± 50,000 M −1 s −1 for the cellulolytic peroxygenase reaction of TrAA9A (Table 1).

Kinetics of reductant peroxidase reaction
Decreasing [Glc eq ] max values of the peroxygenase reaction with decreasing [BMCC] already at H 2 O 2 loads below the K mðH 2 O 2 Þ of the peroxygenase reaction (see results with 20 μM H 2 O 2 in Fig. 2B) indicates that H 2 O 2 is also consumed in other reaction(s). The reductant peroxidase reaction of LPMOs is a well-known side reaction that involves the reoxidation of reduced LPMO-Cu(I) by H 2 O 2 (11,33,34). Because the peroxidase reaction takes place with the population of LPMOs with the active site free from the substrate, its contribution is expected to increase with the decreasing concentration of the polysaccharide substrate. Importantly, the peroxidase reaction also leads to the irreversible inactivation of LPMO (16). Therefore, we also performed an in-depth kinetic characterization of the AscA peroxidase reaction of TrAA9A. Figure 3A shows the time curves of AscA (100 μM) oxidation (followed by the absorbance at 265 nm,  decreased exponentially to a nonzero plateau value (Fig. 3B). Supplementation of the reaction at this point with TrAA9A resulted in a new burst of the AscA oxidation, whereas the supplementation of the reaction with AscA or H 2 O 2 had no effect on the further rate of AscA oxidation (Fig. 3B). These results indicate that TrAA9A has been inactivated before the complete consumption of AscA and/or H 2 O 2 . Equation 2 was used as the first approximation in the analysis of the time curves of AscA oxidation.
Δ[AscA] max is the maximum concentration of AscA consumed in the reaction, k app is the apparent first-order rate constant, and [AscA] ∞ is the remaining concentration of 290,000 ± 50,000 26,900 ± 3000 Probability of inactivation (P i ) 0.0072 ± 0.0003

± 23
a All reactions were made at pH 5 and 25 C. b Because of insufficient saturation with AscA, these figures must be treated with caution.
[AscA] (μM)   (Fig. 3D). This figure translates to the probability of irreversible inactivation P i = 0.0072 ± 0.0003 of TrAA9A in the ascorbate peroxidase reaction (Table 1). We also analyzed the time curves of AscA peroxidase reaction after compensating them for the AscA oxidized in the absence of added H 2 O 2 (Fig. S9). After compensation, Δ[AscA] max measured in the experiments with 10 μM H 2 O 2 was reduced to 9.1 ± 0.3, and the estimate of AscA oxidized per molecule of TrAA9A was 130 ± 11 (P i = 0.0077, Fig. S9).
The initial rates of AscA oxidation were consistent with the Michaelis-Menten equation (Figs. 3E and S10A). The apparent k cat and K mðH 2 O 2 Þ values increased with increasing [AscA] (Fig. S10, B and C), with k cat =K mðH 2 O 2 Þ being independent of [AscA] (Fig. 3F). The same kinetic signatures were observed also in the experiments, where the reduction of ABTS cation radical was used to measure the concentration of AscA (Fig. S7B) (22) instead of measuring the absorbance at 265 nm (Fig. S10). Combining the results of these two different experiment setups resulted in an average k cat =K mðH 2 O 2 Þ value of 26,900 ± 3000 M −1 s −1 for the ascorbate peroxidase reaction of TrAA9A (Table 1). The dependency of initial rates of AscA oxidation on [AscA] is shown in Figure 4A. The apparent k cat (Fig. 4B) and K m(AscA) (Fig. 4C) (Fig. 4D). Analysis of the dependency of apparent k cat and K m(AscA) on [H 2 O 2 ] (Fig. 4, B and C) suggested the true values of k cat and K m(AscA) of 2.1 s −1 , and 140 μM, respectively. The best estimate for the true K mðH 2 O 2 Þ was calculated as a ratio of k cat = 2.1 ± 0.2 s −1 (Fig. 4B) and k cat =K mðH 2 O 2 Þ = 0.027 ± 0.003 μM −1 s −1 (Fig. 3F). This resulted in a K mðH 2 O 2 Þ value of 78 ± 8.6 μM. However, these figures must be treated with caution because the highest [AscA] applicable in the experiments (100 μM) was not sufficiently saturating (Fig. 4A).  peroxidase reactions) by LPMO. Random order ternary complex and ping-pong mechanisms were used for modeling the cellulose peroxygenase and the reductant peroxidase reaction, respectively. For simplicity, the weak binding of Cu(II) enzyme forms to cellulose is omitted. Because the reactivity with O 2 (cellulose oxygenase and reductant oxidase reaction) was negligible (Fig. S1), we have also omitted possible complexes with O 2 . The reductant peroxidase reaction was assumed to lead to the irreversible inactivation of LPMO (with a probability of P i ). Chemical reactions were considered to be irreversible, and the mechanism in Figure 5A was solved using an equilibrium assumption for all complexes. For derivation of the rate equations, see Supporting information (Supplementary results along with Supplementary equations Equations S1-S24).

Theoretical analysis of the H 2 O 2 -driven catalysis
The initial rates of both, the reductant peroxidase (Equation S3) and the cellulose peroxygenase (Equation S13), reactions were consistent with the Michaelis-Menten equation with apparent parameters for H 2 O 2 shown in Figure 5B. In the absence of the cellulose substrate (S), the apparent parameters of the reductant peroxidase reaction reduce to those expected for the ping-pong mechanism   . The values of the parameters used in calculations were the following: (C) k R cat ¼ 2:1 s −1 , P i = 0.0072, K R mðH2O2Þ = 78 μM, K H2O2ðSÞ = 30 μM (K mðH2O2Þ for the cellulose peroxygenase reaction in Table 1), K R mðRÞ = 140 μM (from Fig. 4C), and K s = 0.25 g l −1 . Because the results of the binding experiments in the presence of AscA came with high uncertainty (Fig. S5), the estimate of K s was taken as the half-saturating concentration of BMCC for the apparent k cat of the cellulose peroxygenase reaction in Fig. S4B. The concentration of the reductant ([R]) was set to 1000 μM in calculations. D, the same as in panel C, but the calculations were made using 4.5fold higher P i (P i = 0.0324) and 2-fold higher K s (K s = 0.5 g l −1 ). AscA, ascorbic acid/ascorbate; BMCC, bacterial microcrystalline cellulose; LPMO, lytic polysaccharide monooxygenase.
(Equations S12-S14), with k cat /K m being independent of the concentration of the reagents as also observed in the experiments (Figs. 3F and 4D). Cellulose acts as a mixed-type inhibitor for the reductant peroxidase reaction, with apparent k cat =K mðH 2 O 2 Þ reduced by the competitive (K s ) component of inhibition (Fig. 5B).
The apparent parameters of the cellulose peroxygenase reaction depend on the concentration of cellulose (Fig. 5B) as expected for the ternary complex mechanism. The apparent k cat and K mðH 2 O 2 Þ of the cellulose peroxygenase reaction increase with the increasing concentration of the reductant and the rate constant of the active site copper reduction (k red ) but decrease with the increasing rate constant of the active site copper reoxidation (k ox ). However, the apparent k cat =K mðH 2 O 2 Þ of the cellulolytic peroxygenase reaction is independent on the kinetic parameters of the reduction and reoxidation of the active site copper as well as on the concentration of the reductant (Fig. 5B).
A drawback in the application of LPMOs is the irreversible inactivation of the enzyme in the reductant peroxidase reaction. The rate of enzyme inactivation is a product (Equation S20) of the rate of the reductant peroxidase reaction and the probability of enzyme inactivation in the reductant peroxidase reaction (P i ). Therefore, the apparent kinetic parameters for inactivation are the same as for the reductant peroxidase reaction, but the k cat of the reductant peroxidase reaction ðk R cat Þ must be replaced with k R cat P i (in Fig. 5B and in Equations S4, S6, S7, S9, S12, and S14). In the presence of the reductant (R), H 2 O 2 , and cellulose (S), the half-life of LPMO (t (0.5) ) is given by the following equation: For the definition of constants, see Figure 5. The half-life of LPMO is the lowest in the absence of cellulose, but, starting from this minimum, the half-life is expected to increase linearly with an increasing cellulose concentration (Fig. 5, C  and D).

Discussion
Despite a wealth of structural and biochemical data, the quantitative kinetic studies of LPMOs are scarce. Here, we used a 14 C-labeled cellulose substrate for the kinetic characterization of the LPMO from the cellulose-degrading model fungus T. reesei. TrAA9A revealed high catalytic efficiency of the cellulolytic peroxygenase reaction. The k cat =K mðH 2 O 2 Þ value of 0.29 μM −1 s −1 measured on BMCC (Table 1) is well in line with the k cat =K mðH2O2Þ value of 0.27 μM −1 s −1 measured for the same enzyme on wood-derived cellulose (Avicel) using competition experiments with other H 2 O 2 -consuming enzymes (22). Unlike the competition experiments, the analyses made here also provide the values of individual k cat and K mðH 2 O 2 Þ (Table 1). To date, the corresponding values are available only for the chitinolytic peroxygenase reaction of SmAA10A (20). TrAA9A and SmAA10A have similar k cat values for the polysaccharide peroxygenase reaction, 8.5 s −1 and 6.7 s −1 , respectively. However, SmAA10A had about an order of magnitude lower K mðH 2 O 2 Þ (2.8 μM and 30 μM for SmAA10A and TrAA9A, respectively) and, hence, higher k cat =K mðH 2 O 2 Þ . Another major difference is in the apparent half-saturating concentration of AscA, which is much lower for SmAA10A (around 2 μM (11) and 0.1 mM for SmAA10A and TrAA9A, respectively). The values of second-order rate constants of 4 × 10 4 (34) and 6.9 × 10 3 M −1 s −1 (33) have been reported for the reoxidation of the active site copper by H 2 O 2 under single-turnover conditions for TrAA9A and SmAA10A, respectively. Thus, the requirement for higher AscA concentration in the polysaccharide peroxygenase reaction may arise from the high reductant peroxidase activity of TrAA9A. Recent kinetic study of the peroxygenase activity of the AfAA11B with the chitotetraose substrate (k cat = 4.0 s −1 and K mðH 2 O 2 Þ = 8.9 μM) also revealed a high half-saturating concentration of the AscA (around 0.5 mM) (21). Because AfAA11B has unusually high AscA oxidase activity, these results also indicate to a link between the requirement for the higher AscA concentration in the peroxygenase reaction and the rate of the reoxidation of the active site copper (21).
The rate constant of 4 × 10 4 M −1 s −1 (pH 6, 4 C) measured for the reoxidation of TrAA9A-Cu(I) by H 2 O 2 under singleturnover conditions (34) is somewhat higher (considering the temperature difference) than the k cat =K mðH 2 O 2 Þ of 2.7 × 10 4 M −1 s −1 found here for the AscA peroxidase reaction (Table 1). If so, the rate-limiting step of reoxidation is after the electron transfer resulting in the formation of TrAA9A-Cu(II) (Jones et al. (34), measured the rate by following the growth of the Cu(II) signal). Because the reductant peroxidase reaction can lead to the irreversible inactivation of LPMO, an in-depth understanding of the kinetics and mechanism of this reaction is of utmost importance regarding the application of LPMOs. An experiment setup presented here (Fig. 3, C and D) allows the determination of the probability of inactivation of LPMO in the AscA peroxidase reaction (P i ), which was 0.0072 (Table 1). Using the estimate of k cat for the AscA peroxidase reaction of 2.1 s −1 (Table 1) and P i , one can estimate the k cat for inactivation of TrAA9A of 0.015 s −1 . In the conditions of low cellulose and high H 2 O 2 concentrations, the decay of the rate of the cellulose peroxygenase reaction (k obs in Equation 1) is determined by the inactivation of LPMO (20). In this regard, we note that the k obs values measured in the cellulose peroxygenase reaction (Fig. 2C) are somewhat higher than expected from the values of P i and k cat of the AscA peroxidase reaction (see above). The inactivation of LPMO is caused by hydroxyl radicals that form on the homolytic cleavage of O-O bond in H 2 O 2 (16, 43). In the presence of the substrate, the hydroxyl radical is "caged" and hydrogen atom abstraction is directed toward the formation of Cu(II)-oxyl intermediate, which eventually leads to the hydroxylation of the substrate (44) and breakage of the glycosidic bond. Thus, inactivation of the enzyme in the cellulose peroxygenase reaction is unlikely. In the absence of the substrate, there is more freedom for the reactivity of hydroxyl radical and hydrogen atoms can be abstracted not only from the reductant (45) (as in Fig. 5A) but also from the enzyme. The latter can lead to the inactivation of enzyme. The primary targets for the oxidative damage are the copper-coordinating histidine residues, but modifications of other amino acids have also been observed (16, 43,46). The AscA peroxidase reaction apparently assumes less contacts between the enzyme and substrate (36) than the polysaccharide peroxygenase reaction, which requires multipoint precision binding to the polysaccharide substrate (47)(48)(49)(50)(51). Therefore, it is possible that some oxidative damages that are deleterious for the polysaccharide peroxygenase reaction can be tolerated by the AscA peroxidase reaction. If so, the inactivation measured using the AscA peroxidase reaction may overestimate the performance of LPMO in the cellulose peroxygenase reaction.
The measured half-life of TrAA9A increased with an increasing cellulose concentration, but the "protective effect" of cellulose was less prominent than that predicted using Equation 3 and the best estimates of the kinetic parameters (Fig. 5, C and D). Besides the P i and k cat of the AscA peroxidase reaction, the protective effect of cellulose is determined by the affinity of LPMO to cellulose (K s in Equation 3), which was high for the TrAA9A/BMCC system (Fig. S5). TrAA9A consists of a catalytic domain and a carbohydrate-binding module, which has been shown to increase the affinity to cellulose (40). The existence of nonproductive complexes, where an enzyme is bound to cellulose only by the carbohydrate-binding module, is well-known for glycoside hydrolases (52) and has been proposed also for LPMOs (53). Because the copper active site of LPMOs in such nonproductive complexes is susceptible for inactivation, their existence reduces the protective effect of the polysaccharide. Further studies will reveal the relationships between the stability of LPMOs and possible different binding modes to the polysaccharide substrate.
For economic reasons, the enzymatic degradation of lignocellulose is performed under high dry-matter consistency (54). To maximize the stability of LPMOs, the concentration of H 2 O 2 must be kept low (far below the K mðH 2 O 2 Þ ). Under these conditions, the rate of the LPMO reaction is governed by the k cat =K mðH 2 O 2 Þ values. For TrAA9A, the k cat =K mðH 2 O 2 Þ of the peroxygenase reaction is an order of magnitude higher than that for the AscA peroxidase reaction, which, in turn, was about two orders of magnitude higher than the corresponding figure for inactivation (Table 1). Furthermore, the apparent k cat =K mðH 2 O 2 Þ of the AscA peroxidase reaction and inactivation decrease with the increasing cellulose concentration (Fig. 5B). Thus, the values of the kinetic parameters and their dependency on cellulose concentration suggest that the flow of H 2 O 2 through the cellulolytic peroxygenase reaction is strongly favored over side reactions. However, because the contact times required to achieve target conversion are usually in the range of 72 to 96 h (55), the stability of LPMOs is a major issue in the application of these enzymes in lignocellulose conversion (24,56). Our results suggest that LPMOs and their variants with low efficiency of the reductant peroxidase reaction and low probability of enzyme inactivation should be more stable. Further studies are needed to reveal structural determinants of the probability of inactivation and possible trade-offs between the efficiency of the cellulose peroxygenase and the reductant peroxidase reaction.
TrAA9A was produced and purified as described in Kont et al., 2019 (26). The purified TrAA9A was saturated with copper by overnight incubation with excess CuSO 4 . The unbound copper was removed using a Toyopearl HW-40 desalting column. The concentration of TrAA9A was determined by the absorbance at 280 nm using a theoretical extinction coefficient of 54,360 M −1 cm −1 . 14 C-Labeled bacterial cellulose (2.0 × 10 6 dpm mg −1 ) was prepared by the laboratory fermentation of Gluconobacter xylinum (ATCC 53582) as described in (58), but the cultivation medium was supplied with uniformly 14 C-labeled glucose (1.25 mCi g −1 glucose). 14 C-Labeled BMCC was prepared using the treatment of 14 C-labeled bacterial cellulose with HCl as described (26). The stock solutions of cellulose and TrAA9A were kept in 50 mM sodium acetate (pH 5.0) at 4 C. The water was Milli-Q ultrapure water that had been passed through a column with Chelex 100 resin. Cellulose substrates were incubated with 10 mM EDTA in 10 mM Tris, pH 8.0, overnight, followed by the removal of EDTA by thorough washing with 50 mM sodium acetate (pH 5.0) using repetitive centrifugation and resuspension steps.

Cellulose peroxygenase reaction
TrAA9A and AscA were added to 14 C-labeled BMCC, and 30 s after the addition of AscA, the reaction was started by the addition of H 2 O 2 . At selected times, 0.18-ml aliquots were withdrawn and added to 20 μl of 1.0 M NaOH to stop the reaction. Cellulose was separated by centrifugation (3 min, 10 4 g), and the soluble products were quantified by measuring the radioactivity in the supernatant. The reading of the zero data point (for that, the aliquot was withdrawn before the addition of H 2 O 2 ) was subtracted from the readings of the time points. The concentration of soluble products was calculated based on the radioactivity in the supernatant and the total radioactivity of cellulose in the experiment. Initial rates were calculated as the activity at 30 s. The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C without stirring (BMCC forms a stable suspension).
AscA peroxidase reaction H 2 O 2 was added to AscA, and the reaction was started by the addition of TrAA9A. The oxidation of AscA was followed by the decrease in absorbance at 265 nm. The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C without stirring in a spectrophotometer cuvette.
Binding of TrAA9A to cellulose TrAA9A (100 nM) was incubated with a nonlabeled BMCC (0-1.5 g l −1 ) for 2 min. In one series, the binding experiments were also supplied with 1 mM AscA. Cellulose was separated by centrifugation (1 min, 10 4 g), and the concentration of free TrAA9A in the supernatant was measured by measuring its cellulose peroxygenase activity. For that, the supernatant was reacted with 14 C-labeled BMCC (0.5 g l −1 ) and H 2 O 2 (20 μM) for 30 s. Reactions with the supernatants from the binding experiments made without AscA were also supplied with 1 mM AscA before the measurement of cellulose peroxygenase activity. Reference reactions for 100% free TrAA9A were made exactly as described above, but BMCC was omitted from the binding experiments. All reactions were made in sodium acetate (50 mM, pH 5.0) at 25 C.
Measurement of the stoichiometry (n) of cellulose peroxygenase reaction of TrAA9A using in situ generation of H 2 O 2 by glucose/GO reaction For the calibration of the rate of H 2 O 2 formation, ABTS (1 mM) was mixed with glucose (10 mM) and HRP (0.1 μM) in a spectrophotometer cuvette. The reaction was started by the addition of GO (0.05 g l −1 ), and the [H 2 O 2 ] was measured by measuring the absorbance at 420 nm using the ε 420 of 32,000 M −1 cm −1 and stoichiometry of 2ABTS + /H 2 O 2 . For the cellulose peroxygenase reaction, 14 C-labeled BMCC was mixed with glucose (10 mM), TrAA9A, and AscA (1 mM). The reactions were started by the addition of GO (0.05 g l −1 ). At selected times, 0.18-ml aliquots were withdrawn and added to a 20 μl of 1.0 M NaOH to stop the reaction. Cellulose was separated by centrifugation (3 min, 10 4 g), and the soluble products were quantified by the radioactivity in the supernatant. The concentration of soluble products was calculated based on the radioactivity in the supernatant and the total radioactivity of cellulose in the experiment and was expressed in Glc eq . The reading of the zero data point (for that, the aliquot was withdrawn before the addition of GO) was subtracted from the readings of the time points. The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C without stirring.
Measurement of the AscA peroxidase activity of TrAA9A using ABTS + for the measurement of the concentration of AscA AscA (25, 50, or 75 μM) was mixed with TrAA9A (0.125 μM), and the reaction was started by the addition of H 2 O 2 . After 120 s, an aliquot of the reaction mixture was added to the appropriately diluted mixture of ABTS/ABTS + , and the concentration of AscA was determined by the decrease in absorbance at 420 nm. The volume of the aliquot of the reaction mixture was selected so that the maximum concentration of AscA in the cuvette was 10 μM. The dilution of ABTS/ABTS + was selected so that the maximum concentration of ABTS + after the addition of the reaction mixture was 25 μM. A small amount of AscA consumed in the 120-s reactions without added H 2 O 2 was subtracted from the corresponding values measured in the presence of H 2 O 2 before the calculation of the initial rates (always less than 20% of AscA was consumed in these experiments). The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C. The ABTS/ ABTS + stock solution was made by incubating ABTS (2.0 mM) with potassium persulfate (0.5 mM) as described (22).

Data availability
All data are available within the article and its Supporting Information file and from the corresponding author upon reasonable request.
Supporting information-This article contains supporting information.