The “life-span” of lytic polysaccharide monooxygenases (LPMOs) correlates to the number of turnovers in the reductant peroxidase reaction

Lytic polysaccharide monooxygenases (LPMOs) are monocopper enzymes that degrade the insoluble crystalline polysaccharides cellulose and chitin. Besides the H2O2 cosubstrate, the cleavage of glycosidic bonds by LPMOs depends on the presence of a reductant needed to bring the enzyme into its reduced, catalytically active Cu(I) state. Reduced LPMOs that are not bound to substrate catalyze reductant peroxidase reactions, which may lead to oxidative damage and irreversible inactivation of the enzyme. However, the kinetics of this reaction remain largely unknown, as do possible variations between LPMOs belonging to different families. Here, we describe the kinetic characterization of two fungal family AA9 LPMOs, TrAA9A of Trichoderma reesei and NcAA9C of Neurospora crassa, and two bacterial AA10 LPMOs, ScAA10C of Streptomyces coelicolor and SmAA10A of Serratia marcescens. We found peroxidation of ascorbic acid and methyl-hydroquinone resulted in the same probability of LPMO inactivation (pi), suggesting that inactivation is independent of the nature of the reductant. We showed the fungal enzymes were clearly more resistant toward inactivation, having pi values of less than 0.01, whereas the pi for SmAA10A was an order of magnitude higher. However, the fungal enzymes also showed higher catalytic efficiencies (kcat/KM(H2O2)) for the reductant peroxidase reaction. This inverse linear correlation between the kcat/KM(H2O2) and pi suggests that, although having different life spans in terms of the number of turnovers in the reductant peroxidase reaction, LPMOs that are not bound to substrates have similar half-lives. These findings have not only potential biological but also industrial implications.

Lytic polysaccharide monooxygenases (LPMOs) are monocopper enzymes that catalyze the cleavage of glycosidic bonds in various polysaccharides and oligosaccharides. The most noteworthy property of LPMOs is their ability to break glycosidic bonds in recalcitrant, highly crystalline regions of insoluble substrates-cellulose and chitin. This can be achieved because of the flat and open active site architecture of LPMOs that is suited to interact with multiple polysaccharide chains in an ordered crystalline lattice (1)(2)(3)(4). The catalytically essential copper atom is held in a solvent exposed histidinebrace like structure that is part of a flat substrate-binding surface (5)(6)(7). This enables LPMOs to catalyze breakage of glycosidic bonds in polysaccharides while being in a regular crystal lattice. Thus, LPMO action does not depend on the energetically unfavorable lifting of the polysaccharide chain out of this lattice, which contrasts with canonical glycoside hydrolases, that act on single polysaccharide chains employing acid-base catalysis (8). LPMOs work synergistically with glycoside hydrolases and boost the rate of the degradation of recalcitrant polysaccharides (9)(10)(11)(12)(13)(14)(15)(16)(17).
Although initially described as monooxygenases, in 2017 Bissaro et. al., showed that LPMOs use H 2 O 2 as a cosubstrate (18). Since then, several studies have confirmed the peroxygenase nature of LPMO catalysis (19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29) while the existence of a true monooxygenase activity is debated. Confusion regarding the nature of the cosubstrate stems from the fact that both peroxygenase and monooxygenase reactions rely on the Cu(I) form of the enzyme (30). Therefore, LPMOs need the presence of reductant that, for the peroxygenase activity, is used only for the initial priming of the Cu(II) resting state to the catalytically active Cu(I) form (18,27,31). Besides the initial priming reduction, the monooxygenase reaction requires stoichiometric delivery of two electrons per one glycosidic bond cleavage (9). Ascorbate (AscA) is the reductant most often used in LPMO research. Unfortunately, AscA is amenable to enzyme-independent oxidation by O 2 and the product of such oxidation is H 2 O 2 , the true cosubstrate of LPMOs. Furthermore, the oxidation of AscA by O 2 is catalyzed by copper (32)(33)(34), which is a plausible contaminant in LPMO reactions-it may be attached to sugar substrates or be present in LPMO preparations (22,23,35,36). The situation is further complicated by the reductant oxidase activity of LPMOs. When not protected by the bound substrate, the Cu(I) active site of LPMOs can be reoxidized by O 2 leading, again, to the formation of H 2 O 2 (37).
Today, it is becoming widely accepted that the apparent monooxygenase activity in many LPMO reactions is a consequence of the H 2 O 2 -producing side reactions. The absence, or at least the lack of kinetic relevance, of the monooxygenase reaction is reflected in the low activity of LPMOs under typical "monooxygenase" experiment setups (38) and the strong stimulation of LPMO activity by factors stimulating the rate of H 2 O 2 -producing side reactions, like irradiation of lightsensitive redox-active compounds with visible light (39)(40)(41)(42)(43)(44)(45). Regarding interpretation of kinetic data, a serious drawback of LPMO studies performed with "monooxygenase" experimental setups is that the catalytic rates are limited by LPMO independent H 2 O 2 -producing side reactions (23,(46)(47)(48)(49) and, thus, do not reveal the true catalytic ability of the LPMO of interest. At best, catalytic rates obtained with these setups may reflect the reductant oxidase activity of the substrate-free LPMO under the given substrate load.
Similar to the oxidase activity described above, the Cu(I) active site of LPMOs can also be reoxidized by H 2 O 2 in a reductant peroxidase reaction (24,50,51). The results of single-turnover measurements with AA9 and AA10 LPMOs have shown that the re-oxidation of Cu(I) by H 2 O 2 is several orders of magnitude faster than re-oxidation by O 2 (52,53). An unwanted side reaction of reductant peroxidase activity is the irreversible inactivation of the enzyme. As proposed by Bissaro et al., in 2017, the polysaccharide peroxygenase activity involves Fenton-type chemistry (18), that is, homolytic cleavage of H 2 O 2 (53), which generates a hydroxyl radical. Within the enzyme-substrate complex, the highly reactive hydroxyl radical intermediate is optimally positioned for productive chemistry, leading to hydrogen atom abstraction from the C1 and/or C4 carbon of the substrate (18,54). However, in the absence of substrate the hydroxyl radical will engage in nonproductive reactions, such as oxidation of the enzyme, which may lead to the loss of catalytic activity. It has been shown that amino acids close to the catalytic copper are primary targets of oxidative damage (18,55,56).
Although structurally and biochemically well characterized, kinetic studies of peroxygenase catalysis by LPMOs are still scarce. To date only two in-depth kinetic studies of LPMOs acting on insoluble substrates are available, for the bacterial chitin-active family AA10 LPMO of Serratia marcescens (SmAA10A) (19), and for the fungal family AA9 LPMO of Trichoderma reesei (TrAA9A) (24). Kinetic characterization with soluble oligosaccharides is available for AA9 LPMOs of Neurospora crassa (NcAA9C) (26) and Lentinus similis (LsAA9A) (26,28), and an AA11 of Aspergillus fumigatus (AfAA11B) (25). The kinetics of the reductant peroxidase reaction of LPMOs is also poorly characterized, as is the rate of enzyme inactivation associated with this reaction. In-depth kinetic characterization of the inactivation of LPMO through peroxidase reactions has only been described for one LPMO, TrAA9A (24). The stability of LPMOs is of utmost importance for their application in biotechnological valorization of lignocellulosic biomass (57). However, the scarcity of kinetic data does not allow to conclude about possible activity-stability trade-offs.
To fill these knowledge gaps regarding the kinetic properties and stability of LPMOs, here, we provide the first kinetic characterization of the cellulose peroxygenase activity of a bacterial LPMO, ScAA10C of Streptomyces coelicolor. We also provide in-depth kinetic characterization of the AscA peroxidase activity of this enzyme and three additional wellstudied LPMOs, fungal cellulose-active TrAA9A and NcAA9C, and bacterial chitin-active SmAA10A.

Cellulose peroxygenase reaction
To date an in-depth kinetic characterization of the polysaccharide peroxygenase reaction is available only for two LPMOs. Using 14 C-labeled polymeric substrates we have characterized the degradation of chitin by SmAA10A (19) and bacterial microcrystalline cellulose (BMCC) by TrAA9A (24). Here we carried out similar studies for NcAA9C and ScAA10C. We chose pH 5.0 for kinetic characterization of LPMOs since this is the optimal pH for glycoside hydrolases that operate in synergy with LPMOs in degradation of lignocellulose. NcAA9C is special because it is active on several soluble glycans and even cello-oligomers (58)(59)(60)(61). Although NcAA9C was able to release soluble products from 14 C-labeled BMCC, product levels were low and enzyme activity decayed rapidly (Fig. 1A). Differently from TrAA9A (24) and ScAA10C (see below), NcAA9C showed significant activity also in the experiments without added H 2 O 2 . This apparent "monooxygenase" activity also decayed rapidly and using 1.0 mM AscA as reductant the reactions with and without added H 2 O 2 reached the same plateau value of the released soluble products (Fig. 1A). These results suggest that BMCC is not a good substrate for NcAA9C because of inefficient binding.
ScAA10C had high activity on BMCC. Using 1.0 mM AscA as reductant, the release of soluble products in the experiments without added H 2 O 2 was insignificant ( Fig. 1B; note that these are 10 min reactions; reported LPMO activity in reductantdriven reactions is typically based on multi-hour incubation times). The decay of the release of 14 C-labeled soluble products (expressed in glucose equivalents, Glc eq ) in time was too fast to capture the linear-range of the progress curves (Figs. 1B and S1A). Therefore, the progress curves were fitted to a single exponential function (Equation 1) and initial rates were calculated as v i = [Glc eq ] max k obs (time derivative of Equation 1 in the limiting conditions of time approaching to zero).
The dependency of the initial rates of the release of soluble products on the concentration of [H 2 O 2 ] is shown in Figure 1C. Unfortunately, the K M for H 2 O 2 appeared to be too low for determining its value. Within the error limits the activity was saturated with H 2 O 2 already at the lowest concentration of H 2 O 2 applicable (5.0 μM). The rates measured using 1.0 and 1.5 g L −1 BMCC were also the same within error limits, suggesting that the concentration of cellulose was saturating. Using the average values of initial rates measured using 5.0 to 100 μM H 2 O 2 and 1.0 and 1.5 g L −1 BMCC we found V max to be 0.18 ± 0.03 μM Glc eq s −1 . In order to convert the rates measured in Glc eq s −1 to the turnover number for glycosidic bond cleavage, a stoichiometry coefficient (n) showing the number of soluble Glc eq released per one glycosidic bond cleavage must be known. The value of n is measured under experimental conditions that favor the cellulose peroxygenase reaction, that is, at low H 2 O 2 and high cellulose concentrations (24). Under such conditions, stoichiometric conversion of added H 2 O 2 to cleaved glycosidic bonds can be assumed. A value of n = 3.7 ± 0.8 μM Glc eq /H 2 O 2 was found using the average [Glc eq ] max (Fig. S1B) values obtained with 1.0 and 1.5 g L −1 BMCC, and 5.0 and 10 μM H 2 O 2 . We note that n is an empirical parameter that depends on the average degree of polymerization of the soluble products as well as on the probability of LPMO products being in the soluble fraction (19). However, n is useful for the purpose of calculating the k cat of the cellulose peroxygenase reaction as it relies only on the assumption that the cleavage of a glycosidic bond depends on H 2 O 2 and the stoichiometry is one glycosidic bond cleavage per one H 2 O 2 molecule. Using an n of 3.7, the k cat for the ScAA10C-catalyzed cellulose peroxygenase reactions was calculated to be 2.4 ± 0.5 s −1 . Table 1 shows an overview of currently available kinetic parameters for the polysaccharide peroxygenase reaction catalyzed by LPMOs. The data show similar k cat values for all three LPMOs (two AA10s and 1 AA9) but indicate that the AA10s have lower K M values for H 2 O 2 .

Ascorbate peroxidase reaction
AscA is the most often used reductant in LPMO research. However, to date the kinetic characterization of the AscA peroxidase reaction is available only for TrAA9A (24). Here, we extended the studies of the AscA peroxidase reaction to three model LPMOs, NcAA9C, ScAA10C, and SmAA10A. Characteristic progress curves for AscA oxidation are shown in Figure In Equation 2 [S] stands for the concentration of the substrate, the concentration of which was varied within the series (H 2 O 2 or AscA), and k cat app and K M(S) app are apparent catalytic and Michaelis constants, respectively.  Enzyme catalyzed reactions involving two substrates obey either the ternary complex or the ping-pong kinetic mechanism. The kinetic signature of the ping-pong mechanism is that the apparent k cat /K M is always the true value independent of the concentration of the other substrate (62). This was shown to be the case for the AscA peroxidase reaction of TrAA9A (24  Table 2. Fungal AA9 enzymes showed higher k cat and k cat /K M(H2O2) values compared to their bacterial AA10 counterparts. On the other hand, AA10 enzymes tend to have higher apparent affinity for AscA (lower K M(AscA) ). The differences between fungal and bacterial LPMOs in the terms of K M(H2O2) and k cat /K M(AscA) were less obvious ( Table 2).

Inactivation of LPMOs
A drawback in LPMO catalysis is the irreversible inactivation of the enzyme in the reductant peroxidase reaction. For quantitative analysis of inactivation the time curves of AscA oxidation (Figs. 2A and S4) were analyzed according to the Equation 4 (24). In The term p i in Equation 5 stands for the probability of LPMO inactivation in the reductant peroxidase reaction. Using the data depicted in Figure 3A (with underlying progress curves shown in Fig. S7), we determined Δ[AscA] max values and the p i for the different LPMOs and the data are summarized in Table 3. Fungal enzymes had much higher stability, turning over more than 100 AscA molecules before inactivation, while ScAA10A and SmAA10A were inactivated after 38 and 10 turnovers, respectively.
To verify these differences and exclude possible reductantspecific effects on LPMO inactivation, we also tested the peroxidation of a phenolic reductant, methyl hydroquinone (MHQ) by LPMOs. At high MHQ and H 2 O 2 concentrations, the formation of the oxidized product (methyl quinone, MQ) decayed because of inactivation of LPMO, as shown for TrAA9A in Fig. S8. We were not able to obtain stock solutions of MHQ without small amounts of H 2 O 2 , which precluded proper correction for background (O 2 -driven) MHQ turnover and Michaelis-Menten analysis. Therefore, the data from experiments without LPMO were used as the background signal (Fig. S8). Based on progress curves for the oxidation of MHQ by LPMOs (Fig. S9)  To verify that enzyme inactivation is general and not related to the peroxidase reaction only, we did preincubation experiments with TrAA9A and ScAA10C. After preincubation with H 2 O 2 and the reductant (AscA and MHQ; peroxidase conditions) for selected times, we measured residual BMCC peroxygenase activity. In all cases, the loss of the activity in the reductant peroxidase reaction was reflected in the loss of activity in the cellulose peroxygenase reaction (Fig. S10), suggesting that both reactions are similarly affected by inactivation.
Comparison of the kinetic parameters for the AscA peroxidase reaction (Table 2) and the maximum turnover numbers (Table 3) revealed a positive correlation (increasing n max with increasing  (Fig. 2E). c Derived from the dependency of k cat app on [H 2 O 2 ] (Fig. 2D). ). C, correlation between n max for AscA and k cat /K M(H2O2) for the ascorbate peroxidase reaction. The data for TrAA9A are from Ref (24). For the correlation between n max for AscA and other parameters of the ascorbate peroxidase reaction Fig. S11. Shown are average values ± SD (n = 2, independent experiments). AscA, ascorbate; LPMOs, lytic polysaccharide monooxygenases.

Discussion
Recent studies have shown that LPMOs are efficient polysaccharide peroxygenases (19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29). For the catalysis of this unique reaction, these monocopper enzymes rely on a single Cu(I) in their active site (30). Since the resting state of copper in aerobic environments is Cu(II), LPMOs need the presence of a reductant for their activation. To date there is no evidence for the reoxidation of the Cu(I) LPMO in the polysaccharide peroxygenase reaction. On the contrary, multiple studies have shown that, once reduced, an LPMO can perform multiple peroxygenase reactions (18,27,31). Although it has been shown that the Cu(I) form of LPMO binds to the substrate with higher affinity compared to the Cu(II) form (31,63), in real systems there is always a population of substrate-free LPMO-Cu(I). This population is amenable to reoxidation by O 2 and H 2 O 2 . Since LPMOs acting on crystalline surfaces have evolved a flat, solvent-exposed active site architecture, off-pathway reoxidation of the active site copper may be an unavoidable side reaction. Reoxidation by O 2 to generate H 2 O 2 has been studied quite well and it has been speculated that the reductant oxidase activity of LPMO may serve as a source of H 2 O 2 to be used by LPMOs (25,49) or by other H 2 O 2 consuming enzymes like lignin peroxidases (64,65). On the other hand, less is known about reoxidation by H 2 O 2 , whereas this reaction actually is crucial because it may lead to the irreversible enzyme inactivation of LPMO. Of note, it is hard to see any biological rationale for the reductant peroxidase activity of LPMOs especially in the light of enzyme inactivation.
It is important to minimize the flux of H 2 O 2 through the reductant peroxidase reaction, to maximize LPMO stability and minimize futile turnover of oxidant. Assuming the experimentally supported ping-pong mechanism, the rate of H 2 O 2 turnover in the reductant peroxidase reaction (v ox R ) is given by Equation 6.
In Equation 6, R denotes the reductant, and the kinetic parameters of the reductant peroxidase reaction ( Equation 7 shows that, under these assumptions, the rate of the reductant peroxidase reaction depends linearly on the H 2 O 2 concentration. k cat R and K M(H2O2) R may vary between reductants and will, as shown in this study, vary between LPMOs ( Table 2).
The LPMO peroxygenase reaction has been shown to follow the ternary complex mechanism (19,24), the kinetics of which is given by Equation 8.
In Equation 8, S denotes the sugar substrate, and the kinetic parameters for the peroxygenase reaction (Table 1) (Fig. 1C) precluded determination of the k cat /K M(H2O2) value of the cellulose peroxygenase reaction catalyzed by ScAA10C, it seems, that also this bacterial enzyme has a strong preference, by a factor 50 or higher, in favor of the peroxygenase reaction (Tables 1 and 2). As shown by Equation 9, the H 2 O 2 fluxes depend on the concentrations of substrate-bound and free LPMO, which again will depend on the substrate concentration. Indeed, it has been shown in several studies that substrate affinity is an important contributor to LPMO stability (66). Stronger binding increases the [LPMO] bound /[LPMO] free ratio and drives the flux of H 2 O 2 through the peroxygenase reaction (Equation 9). The results obtained when characterizing the cellulose peroxygenase kinetics of NcAA9C (Fig. 1A) may serve as an example of the effects of inefficient binding. Although the initial activity seems to be high, the enzyme is rapidly inactivated in the experiments with added H 2 O 2 . NcAA9C had also relatively high initial activity (compared to ScAA10C (Fig. 1B) and TrAA9A (24)) in the experiments without added H 2 O 2 but also in this case the enzyme was rapidly inactivated (Fig. 1A). All in all, these results indicate that the reaction with BMCC is substrate-limited and, thus, that BMCC is not a good substrate for NcAA9C. Although NcAA9C has a carbohydrate binding module, inefficient binding of the catalytic domain would leave the active site free for "self-production" of H 2 O 2 in the AscA oxidase reaction but also for inactivation in the AscA peroxidase reaction. Collectively these data suggest that crystalline BMCC is not a good substrate for NcAA9C. It is conceivable that the enzyme only acted on a minor fraction of more amorphous material in the BMCC, and the effective substrate concentration thus was very low. Of note, activity of this enzyme on crystalline cellulose (Avicel) (58,67) has been demonstrated, but only under rather extreme conditions (high enzyme loads, long incubations, and sensitive detection without quantitative reporting), that are very different from the conditions used here. The lack of activity on crystalline substrates is intriguing, especially considering that NcAA9C clearly is a competent LPMO when acting on other substrates. For example, a k cat value of 124 ± 27 s −1 (at 4 C) has been reported for H 2 O 2 driven cleavage of soluble cellopentaose (26).
Although the k cat /K M(H2O2) values of the fungal enzymes (Tables 1 and 2) were less supportive for the flux of H 2 O 2 through the peroxygenase reaction, the fungal LPMOs were more resistant toward inactivation in the reductant peroxidase reaction. The probability of inactivation of SmAA10A in the reductant peroxidase reaction is about an order of magnitude higher compared to TrAA9A and NcAA9C (Table 3). Thus, the less pronounced preference for the peroxygenase reaction in fungal LPMOs seems to be, at least to some extent, counterbalanced by a higher resistance toward oxidative inactivation in the peroxidase reaction. The strong positive correlation between the k cat /K M(H2O2) of the reductant peroxidase reaction and the number of turnovers made before the inactivation (Fig. 3C) suggests that the higher stability of fungal LPMOs has coevolved with the catalytic efficiency in the peroxidase reaction. One may further speculate that the latter is an unavoidable "coproduct" of evolution toward higher reductant oxidase efficiency needed for being "self-supporting" with H 2 O 2 cosubstrate.
A characteristic structural feature of natural AA9 enzymes expressed in fungi is the N-methylation of the Cu coordinating N-terminal histidine (35). Studies of H 2 O 2 -fueled LPMO reactions have led to the suggestion that this posttranslational modification helps protect against oxidative damage to this vital residue (68). Importantly, of the two AA9s used in this study, only one, TrAA9A, carried the methylation, whereas the two enzymes showed almost identical susceptibilities to inactivation through the peroxidase reaction (Table 3). LPMOs show large functional differences also within the same family, (69) and the similar stability of NcAA9C and TrAA9A does not rule out an important role of the methylation. However, the present results show that other structural features also play important roles. Another difference between fungal and bacterial LPMOs is that the fungal enzymes have a Tyr in the second coordination sphere located close to what could be called the proximal axial coordination position of the copper, whereas the corresponding position in about 90 % of bacterial AA10 enzymes, including the two studied here, is occupied by Phe (3,70). Tyr and Trp have been proposed to protect redox enzymes against oxidative damage by providing hole hopping pathways for reactive radical intermediates (71). The presence of Trp and Tyr radical intermediates in the reoxidation of Cu(I) has indeed been demonstrated for fungal LPMOs, including TrAA9A (53). Thus, the Tyr (instead of Phe) in the second coordination sphere may contribute to the higher stability of fungal LPMOs compared to their bacterial AA10 counterparts.
The results presented here suggest that the probability of the inactivation of LPMO in the reductant peroxidase reaction is independent on the nature of the reductant (Fig. 3, A and B and Table 3). This is expected for the ping-pong mechanism, where there is no ternary complex comprised of the LPMO, the reductant, and H 2 O 2 . Given that the inactivation takes place in the reaction of reduced Cu(I) LPMO with H 2 O 2 , and not in the reaction with Cu(II) LPMO, it is not surprising that it does not matter which specific reductant was responsible for generating the Cu(I) LPMO.
The linear correlation between the k cat /K M(H2O2) for the reductant peroxidase reaction and the stability of the LPMO (Fig. 3C)  ) but considering the ping-pong mechanism is expected to be independent of the concentration of the reductant (i.e., the apparent k cat R / K M(H2O2) R is independent on [R], Fig. 2F). Although not the focus of this study, it is worth contemplating on the possible impact of the findings and considerations described above on the processing of cellulosic biomass with LPMO-containing enzyme cocktails (72). As degradation reactions proceed, the effective substrate concentration decreases, whereas the degradation of the remaining substrate, which likely is the most recalcitrant fraction of the starting material, would benefit from LPMO action. Instead, as a result of the lower substrate concentration, a larger fraction of the LPMOs will be in a substrate-free form, which leads to increased nonproductive use of H 2 O 2 and increased enzyme inactivation. It is conceivable that process optimization could be achieved by nonconventional dosing of the LPMOs, rather than adding the LPMOs all at the start of the reaction. Further understanding and optimization of LPMO performance would benefit from more in-depth knowledge of substrate binding kinetics and of possible protective mechanisms (27,55,71) that may be affected by substrate binding.
Back to biology, one may wonder whether the different kinetic signatures of fungal and bacterial LPMOs reported here may reflect adaptation to different steady-state levels of H 2 O 2 and/or the nature of the reductants present in their native environments. As recently discussed by Hemsworth (73), more detailed information about the conditions in the natural environments of LPMOs is needed to understand LPMO functionality in natural ecosystems and to reveal the biological relevance of functional differences described above.

Experimental procedures
Materials MHQ (lot # BCBH9920V) and L-ascorbic acid (AscA, lot # SLBM0850V) were from Sigma-Aldrich. Chelex 100 resin (50-100 mesh, sodium form) was from Bio-Rad. The H 2 O 2 stock solution (lot # SZBG2070) was from Honeywell. A 0.5 M stock solution of the sodium acetate buffer, pH 5.0, was kept overnight with beads of Chelex 100 resin after preparation. Dilutions of the commercial H 2 O 2 stock solution (30 wt %, 9.8 M) were prepared in Chelex-treated sodium acetate buffer directly before use. AscA (50 mM in water) was kept as frozen aliquots at −18 C and the aliquots were melted directly before use. The water was Milli-Q ultrapure water that had been passed through a column with Chelex 100 resin.
TrAA9A, NcAA9C, ScAA10C, and SmAA10A were produced and purified as described in Kont et al. , respectively. The purified LPMOs were saturated with copper by overnight incubation with excess (3:1 M ratio) CuSO 4 . The unbound copper was removed using a Toyopearl HW-40 desalting column. The concentration of the LPMOs was determined by measuring the absorbance at 280 nm using theoretical extinction coefficients of 54,360, 46,910, 75,775, and 29,450 M −1 cm −1 for TrAA9A, NcAA9C, ScAA10C and SmAA10A, respectively. 14 C-BMCC (specific radioactivity 2.0 × 10 6 dpm mg −1 ) was prepared as described earlier (24). To remove possible cellulose bound metal ions the 14 C-BMCC was incubated with 10 mM EDTA in 10 mM Tris-HCl, pH 8.0 overnight. Finally, EDTA was removed by washing with 50 mM sodium acetate (pH 5.0) using repetitive centrifugation and resuspension steps. The stock solutions of 14 C-BMCC and LPMOs were kept in 50 mM sodium acetate (pH 5.0) at 4 C.

Reductant peroxidase reaction
LPMO was added to the reductant (AscA or MHQ) and the reaction was started by the addition of H 2 O 2 . The oxidation of AscA was followed by the decrease in absorbance at 265 nm using appropriate calibration curves. The oxidation of MHQ was followed by the increase in absorbance at 251 nm using the extinction coefficient of 21,450 M −1 cm −1 (67). The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C, without stirring, in a spectrophotometer cuvette.

Cellulose peroxygenase reaction
LPMO and the reductant (AscA or MHQ) were added to 14 C-BMCC, and 30 s after the addition of the reductant the reaction was started by the addition of H 2 O 2 . At selected times 0.18 ml aliquots were withdrawn (from a total reaction volume of 1.35 ml) 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. For zero time points aliquots were withdrawn before the addition of the reductant and H 2 O 2 . The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C without stirring.
Measuring concentration of LPMO active in cellulose peroxygenase reaction LPMO (250 nM) was preincubated with reductant (50 μM AscA or 1.0 mM MHQ) and 100 μM H 2 O 2 at 25 C. At defined times 72 μl aliquots (from a total reaction volume of 0.5 ml) were withdrawn and added to 108 μl of a mixture containing 14 C-BMCC (1.67 g L −1 ), AscA (1.67 mM), and H 2 O 2 (0.83 mM), followed by incubation for 15 min. The insoluble substrate was removed by centrifugation (2 min, 10 4 × g) and the soluble products were quantified by measuring the radioactivity in the supernatant. Under these conditions the kinetics is governed by the inactivation of LPMO and the amount of released products scales linearly with the concentration of active LPMO. Calibration curves were made using different LPMO concentrations but in the absence of reductant and H 2 O 2 in preincubation. The reactions were made in 50 mM sodium acetate (pH 5.0) at 25 C without stirring.

Data availability
All data are available within the article and its Supporting Information File and from the corresponding author upon reasonable request.