Kinetic insights into the role of the reductant in H₂O₂-driven degradation of chitin by a bacterial lytic polysaccharide monooxygenase

Abstract

Lytic polysaccharide monooxygenases (LPMOs) are monocopper enzymes that catalyze oxidative cleavage of glycosidic bonds in polysaccharides in the presence of an external electron donor (reductant). In the classical O₂-driven monooxygenase reaction, the reductant is needed in stoichiometric amounts. In a recently discovered, more efficient H₂O₂-driven reaction, the reductant would be needed only for the initial reduction (priming) of the LPMO to its catalytically active Cu(I) form. However, the influence of the reductant on reducing the LPMO or on H₂O₂ production in the reaction remains undefined. Here, we conducted a detailed kinetic characterization to investigate how the reductant affects H₂O₂-driven degradation of ¹⁴C-labeled chitin by a bacterial LPMO, SmLPMO10A (formerly CBP21). Sensitive detection of ¹⁴C-labeled products and careful experimental set-ups enabled discrimination between the effects of the reductant on LPMO priming and other effects, in particular enzyme-independent production of H₂O₂ through reactions with O₂. When supplied with H₂O₂, SmLPMO10A catalyzed 18 oxidative cleavages per molecule of ascorbic acid, suggesting a “priming reduction” reaction. The dependence of initial rates of chitin degradation on reductant concentration followed hyperbolic saturation kinetics, and differences between the reductants were manifested in large variations in their half-saturating concentrations (K_mR^app). Theoretical analyses revealed that K_mR^app decreases with a decreasing rate of polysaccharide-independent LPMO reoxidation (by either O₂ or H₂O₂). We conclude that the efficiency of LPMO priming depends on the relative contributions of reductant reactivity, on the LPMO's polysaccharide monooxygenase/peroxygenase and reductant oxidase/peroxidase activities, and on reaction conditions, such as O₂, H₂O₂, and polysaccharide concentrations.

Introduction

Lytic polysaccharide monooxygenases (LPMOs)² are oxidative enzymes that cleave glycosidic bonds in poly- and oligosaccharides (1). Their oxidative mechanism was reported in 2010 when CBP21, a “chitin-binding protein” from Serratia marcescens known to boost chitinase activity (2), was shown to cleave β-1,4-glycosidic bonds in chitin with concomitant formation of C1 oxidized chito-oligosaccharides (3). It was shown that the oxygen atom introduced at C1 originated from molecular oxygen and that the reaction depended on an external electron donor that delivers two electrons per catalytic cycle (R-H + 2e⁻ + 2H⁺ → R-OH + H₂O) to complete the four-electron reduction of O₂ (3). The LPMO activity of CBP21, hereafter called SmLPMO10A, can be driven by ascorbic acid (AscA) (3, 4), which today is a common reductant used in LPMO reactions. However, other reductants may be used, such as reduced GSH (3), Fe(II)SO₄ (3), gallic acid (GA) (5), and fungal cellobiose dehydrogenase (4). Studies on a wide variety of cellulose-active LPMOs have shown that many other reductants will work (6), including reductants that are intrinsic to lignocellulosic feedstocks (7). These reductants include GA (8–15), various phenolics (9, 10, 12, 13, 16–18), and lignin fragments (9, 12, 19–21).

Recent studies have shown that LPMOs can use H₂O₂ as co-substrate (22, 23). For a review of the (potential) involvement of H₂O₂ and LPMOs in degradation of lignocellulose, see Bissaro et al. (24). Although the possibility of a (O₂-using) monooxygenase reaction is still being debated, the fact is that the (H₂O₂-using) peroxygenase reaction is several orders of magnitude more efficient (23, 25, 26). Computational studies also support the plausibility of H₂O₂ as co-substrate for LPMOs (27–29). The peroxygenase-like reaction also depends on an external electron donor; however, in principle, the reductant is only needed for initial activation (priming) of the Cu(II) resting state to the catalytically active Cu(I) state (23). Once activated, the LPMO would in principle be capable of catalyzing multiple cleavage reactions, if fueled with H₂O₂. Of note, LPMO priming by a one-electron reduction is part of some of the currently proposed monooxygenase mechanisms (30), but these mechanisms do not alleviate the need for two electrons being delivered per catalytic cycle.

Provided with H₂O₂ as cosubstrate and a low amount (0.1 mm) of AscA, SmLPMO10A shows high efficiency in chitin oxidation with the k_cat/K_m for H₂O₂ being on the order of 10⁶ m⁻¹ s⁻¹ (k_cat = 6.7 s⁻¹ and K_m for H₂O₂ = 2.8 μm) (25). Recent work on a fungal LPMO also indicated high k_cat values for the peroxygenase activity on cello-oligosaccharides (26). Such a high catalytic efficiency and the low K_m for H₂O₂ call for revisiting existing kinetic data for LPMO reactions with different reductants. Next to priming the LPMO, reductants will be involved in the generation of H₂O₂, by enzyme-independent reduction of O₂ and/or by fueling LPMO-catalyzed production of H₂O₂ (16, 31). The futile reaction of reduced mediators with O₂ is a well-known challenge in the application of monooxygenases and is referred to as the oxygen dilemma (32). H₂O₂ is a product of these oxidations, as it can be formed directly by two-electron reduction of O₂ or indirectly from a superoxide radical intermediate generated by one-electron reduction of O₂, as shown, for example, for the oxidation of AscA (33). Notably, the incorporation of the ¹⁸O label into the oxidized chitooligosaccharide products generated by SmLPMO10A in reactions with ¹⁸O₂, seemingly supporting a monooxygenase mechanism (3), could also result from a reaction with H₂¹⁸O₂, originating from in situ reduction of ¹⁸O₂ either by the reductant or by the LPMO (23). The interpretation of data are further complicated by irreversible inactivation of the LPMO by a surplus of H₂O₂ in the presence of reductant (5, 23, 25, 34). Consequently, it is difficult to judge whether and to what extent the differences between different reductants reported so far reflect the efficiency of the reductant in reducing the LPMO or the effect of the reductant on the production of H₂O₂ in the reaction mixtures.

Here, we set out to obtain insights into the effects of the reductant on LPMO activity using the chitin-active SmLPMO10A as a model enzyme. Sensitive detection of ¹⁴C-labeled degradation products of chitin enabled accurate initial rate measurements and allowed discrimination between the effects of the reductant on priming of the LPMO and side effects involving the production of H₂O₂. We show that the different tested reductants, namely AscA, GA, and methylhydroquinone (MHQ), differ in their ability to support SmLPMO10A activity, and we provide a kinetic and theoretical framework for understanding reductant function.

Results

Experimental set-up

The kinetics of H₂O₂-driven oxidation of chitin (¹⁴C-labeled crystalline α-chitin nanowhiskers (CNWs)) by SmLPMO10A with 0.1 mm AscA as reductant has been characterized in detail before (25). The k_cat value for oxidative cleavage of CNWs was 6.7 s⁻¹, and K_m values were 2.8 μm and 0.58 mg ml⁻¹ for H₂O₂ and CNWs, respectively (25). The quantification of the radioactivity in the supernatants of SmLPMO10A reactions enables sensitive detection of soluble products, the concentration of which is expressed in GlcNAc equivalents (NAG_eq). Under the conditions used in previous studies, the turnover of one molecule of H₂O₂ results in one oxidative cleavage, which is manifested in the release of four soluble NAG_eq (25). Because about 50% of the oxidized groups remain associated with the insoluble substrate, the average degree of polymerization of soluble products is 8 NAG_eq (25). Importantly, under these conditions (typically 1 mg ml⁻¹ CNWs, 42 nm SmLPMO10A, and reaction time < 10 min), and as long as enzyme inactivation and H₂O₂ depletion are avoided, release of NAG_eq is linear over time up to levels of 190 μm released soluble NAG_eq (25). Therefore, in the present study, we used 1 mg ml⁻¹ CNWs as initial substrate concentration, and the maximum levels of released NAG_eq remained well below the 190 μm threshold. If not stated otherwise, the concentration of SmLPMO10A was 42 nm. Time courses of the release of NAG_eq were measured at two initial H₂O₂ concentrations, 20 and 100 μm, and the nature and the concentration of the reductant were varied.

Kinetic studies of LPMOs with H₂O₂ as cosubstrate in an aerobic environment are complicated by the simultaneous presence of O₂ and a reducing agent. Because of the low K_m value of SmLPMO10A for H₂O₂ (25), low H₂O₂ concentrations must be used in kinetic studies, and care must be taken to account for the formation/depletion of the H₂O₂ in enzyme-independent and enzyme-dependent background reactions. For example, the formation of H₂O₂ in the reaction between AscA and molecular oxygen is well-known (35). Furthermore, the oxidation of AscA by O₂ is catalyzed by metal ions, like Cu(II) (36, 37), so that their presence even in trace amounts can significantly affect the formation of H₂O₂ and interpretation of the results of LPMO kinetics. The problem is further amplified by the occurrence of metal ions in polymeric substrates of LPMOs, such as cellulose and chitin (38–40).

When setting up our experimental conditions, we first assessed the activity of SmLPMO10A in experiments without added H₂O₂, hereinafter referred to as “background activity.” Considerable variation in background activity was observed between different CNW batches in reactions containing AscA (0.1 mm) as reductant (Fig. 1A). This indicates that the quality of the water and/or buffer components used in the preparation of CNW batches (see “Experimental procedures”) affects background activity. The addition of peroxidase completely removed the background activity, indicating that the formation of H₂O₂ drives catalysis under these conditions (Fig. 1B). The addition of EDTA (5 mm) also completely removed the background activity, suggesting that divalent metal ions are involved (41). The addition of catalase caused more than 10-fold reduction of the background activity (Fig. 1B). Collectively, these results suggest that, under the conditions used here, the background activity is related to the formation of H₂O₂ by divalent metal ion-catalyzed oxidation of AscA by O₂.

Figure 1.

Control experiments for setting of the reaction conditions. All experiments were carried out in NaAc buffer (50 mm, pH 6.1) at 25 °C and contained CNWs (1.0 mg ml⁻¹) and SmLPMO10A (42 nm). A, formation of soluble products in the experiments with different batches of CNWs. Experiments were set up with 20 μm initial H₂O₂ load (filled symbols) or with no added H₂O₂ (open symbols). The experiments without added H₂O₂ show what in this study is referred to as “background activity.” Two different CNW batches (see “Experimental procedures”) designated as high-background (black symbols) and low-background (blue and red symbols) were used as substrates. Black square, 100 μm AscA; black diamond, 100 μm AscA + 5 mm EDTA; red circle, 100 μm AscA; blue triangle, 10 μm AscA; blue diamond, 10 μm AscA + 5 mm EDTA. Due to a shortage of the “high-background” batch, some of the series were done without duplicates. B, effect of catalase (200 units ml⁻¹) and peroxidase (2.0 units ml⁻¹ + 150 μm Amplex Red) on the formation of soluble products in background experiments (no added H₂O₂). Here, the low-background CNWs were used, and the time scale was extended to obtain higher signals. Filled symbols, reactions without catalase; open symbols, reactions with catalase. The concentration of AscA was 1 mm (diamond) or 100 μm (square). The solid line shows linear regression of the data. The crosses show products generated in a reaction containing 1 mm AscA, horseradish peroxidase, and Amplex Red. Error bars, S.D. derived from at least two independent experiments.

It is worth noting that, whereas EDTA removed the background activity, the LPMO remained active when supplied with H₂O₂. The ability of LPMOs to withstand EDTA has been reported before (8) and coincides with very high reported copper affinities (8, 42). Apparently, the strong binding of copper and slow off rates (43) ensure that there is no significant transfer of copper from the LPMO to EDTA, provided that the contact time is kept short, as is the case here. Accordingly, in experiments with 0.1 mm AscA as reductant and 20 μm load of H₂O₂, EDTA had no effect on the kinetics of SmLPMO10A. However, at lower AscA concentration (10 μm), EDTA became inhibitory for the LPMO reaction (Fig. 1A). One may speculate that EDTA can sequester only the Cu(II) from LPMO and becomes inhibitory when the reduction of LPMO becomes less efficient. Considering these complications, the use of EDTA was judged not applicable in further studies of reductant efficiency.

To minimize complexity, in further work, we only used “low-background” CNW in the experiments with AscA as reductant. Unlike reactions with AscA, within the time scales used, there was no significant background signal when GA and methyl hydroquinone (MHQ) were used as reductants (Fig. S1), regardless of the CNW batch used. This is corroborated by a reported much higher stability of the latter reductants against oxidation (i.e. H₂O₂ production) compared with AscA (10). We also attempted to use 1,4-dihydroxy-2,6-dimethoxybenzene, an efficient reductant in supporting AA9's activity (10), but the high background activities (data not shown) did not permit accurate measurement of SmLPMO10A kinetics, and this reductant was omitted from further studies. As found for AscA before (25), the order of the addition of reductant and H₂O₂ had no effect on the outcome of the experiment (Fig. S1). The experimental set-up where the reductant was added to premixed CNW/SmLPMO10A 30 s before the reaction was started by the addition of H₂O₂ (zero time point) was used throughout this study.

Importantly, considering the above, and as a result of precisely tailoring the reaction conditions, in all of the experiments described below, the rates of SmLPMO10A without added H₂O₂ were always insignificant compared with reactions containing a 20 or 100 μm initial load of H₂O₂. Thus, in the experiments described below, the effects of various possible O₂-, reductant-, and/or SmLPMO10A-dependent H₂O₂ producing pathways are negligible.

Time curves with different reductants

We studied the effect of reductants on chitin oxidation using either 20 or 100 μm H₂O₂ as initial load. We have previously shown (25), using 0.1 mm AscA as reductant, that at the lower H₂O₂ initial load (represented here by experiments with 20 μm H₂O₂), chitin oxidation by SmLPMO10A is fast enough compared with enzyme inactivation (i.e. most H₂O₂ is incorporated into oxidized products), whereas at the higher H₂O₂ initial load (represented here by experiments with 100 μm H₂O₂), the kinetics is governed by enzyme inactivation. A few experiments with the higher load of H₂O₂ were included in this study to illustrate the effect of reductant on H₂O₂-driven enzyme inactivation.

First, we followed the formation of soluble NAG_eq in time using a 20 μm initial H₂O₂ load. AscA, GA, and MHQ at different concentrations were used as reductants. Based on the stoichiometry of 4.0 ± 0.3 NAG_eq released per H₂O₂ consumed (25), one can calculate that a maximum of 80 ± 6 μm NAG_eq can be released in the experiments with 20 μm H₂O₂. Data presented in Fig. 2 show that the rate of NAG_eq formation depends on the nature and concentration of the reductant. Although there is some uncertainty associated with the precise measurement of the plateau values for reactions with largely different rates, all time curves tend to reach a plateau value of NAG_eq ([NAG_eq]_max) close to 80 μm (Fig. 2). Of note, the consistency between the measured and expected [NAG_eq]_max values shows that side reactions leading to production or consumption of H₂O₂ are not significant under our experimental conditions.

Figure 2.

Time curves of the formation of soluble products (in NAG_eq) in the presence of different reductants. All experiments were carried out in NaAc buffer (50 mm, pH 6.1) at 25 °C and contained CNWs (1.0 mg ml⁻¹), H₂O₂ (20 μm), and SmLPMO10A (42 nm). The reductant was AscA (A), GA (B), or MHQ (C). The concentration of the reductant is shown in the plot. The horizontal line indicates (shadowing shows S.D.) the maximum amount of soluble NAG_eq that can be produced in the experiments with 20 μm H₂O₂ (80 ± 6 μm NAG_eq). The arrows in A indicate the addition of fresh AscA to a final concentration of 100 μm. The solid lines show nonlinear regression analysis (up to the 1,200-s time point) according to Equation 1. Error bars, S.D. derived from at least two independent experiments.

As a clear exception, the experiments with AscA at low concentrations (0.2 and 0.5 μm) yielded lower values of [NAG_eq]_max (Fig. 2A). The addition of a fresh portion of AscA (up to 100 μm) at the point in the time curve where the formation of NAG_eq had decayed caused a new burst in activity and subsequent leveling off at a [NAG_eq]_max around 80 μm (Fig. 2A). This indicates that at low AscA concentrations, depletion of the reductant limits the reaction. [NAG_eq] values from experiments with low [AscA] were fitted to Equation 1 (25) using nonlinear regression analysis (Fig. 2A).

$$\left[{\text{NAG}}_{\text{eq}}\right]={\left[{\text{NAG}}_{\text{eq}}\right]}_{\max }\left(1-e^{-k_{\text{obs}}t}\right)$$ (Eq. 1)

Plotting the [NAG_eq]_max values obtained for the three lowest AscA concentrations against the concentration of AscA results in a linear correlation, showing that 73 ± 20 NAG_eq are released per molecule of AscA (Fig. 3). Considering that 4.0 ± 0.3 NAG_eq are released per H₂O₂ consumed and that one H₂O₂ supports one oxidative cleavage (23, 25), we can thus estimate that an average of 18 ± 5 oxidative cleavages are performed by SmLPMO10A per reduction (priming) event under our experimental conditions.

Figure 3.

Correlation between the maximum concentration of soluble products ([NAGeq]_max) and the concentration of ascorbic acid. The [NAG_eq]_max values were found by nonlinear regression analysis of the data in Fig. 2A (for the series with [AscA] of 0.2, 0.5, and 1.0 μm and for the time points up to 1,200 s) according to Equation 1. The solid line shows linear regression of the data. Error bars, S.D. derived from at least independent experiments.

Previous studies using 0.1 mm AscA as reductant showed that, under otherwise identical conditions, the initial addition of 100 μm H₂O₂ yields reaction kinetics that are dominated by enzyme inactivation. Under these conditions, a maximum of only 25 μm NAG_eq was reached (i.e. 6.25% of the theoretical maximum) (25). Dose–response experiments with GA as reductant (Fig. 4) show that increasing the GA concentration leads to faster product formation and faster enzyme inactivation. For instance, fitting of the data to Equation 1 shows half-lives of 220 and 90 s for the reactions with 50 and 500 μm GA, respectively. At GA concentrations above 100 μm, the maximum levels of NAG_eq reach 25 μm (Fig. 4), which is similar to what was observed for AscA with a 100 μm H₂O₂ load (25). The [NAG_eq]_max value is a result of the competition between (i) H₂O₂-driven catalysis of chitin oxidation by SmLPMO10A and (ii) irreversible inactivation of reduced SmLPMO10A by H₂O₂. The fact that different [GA] yield different enzyme half-lives but similar [NAG_eq]_max values suggests that less efficient priming of SmLPMO10A slows down both catalysis and inactivation.

Figure 4.

Time curves for the formation of soluble products (in NAG_eq) at high H₂O₂ load. All experiments were carried out in NaAc buffer (50 mm, pH 6.1) at 25 °C and contained CNWs (1.0 mg ml⁻¹), H₂O₂ (100 μm), and SmLPMO10A (42 nm). Gallic acid was used as reductant, and its concentration is shown in the plot. The solid lines show nonlinear regression analysis according to Equation 1. Error bars, S.D. derived from at least two independent experiments.

Dependence of initial rates on the nature and concentration of the reductant

The linear regions of the time curves of NAG_eq formation were used to calculate the initial rates of reactions carried out with AscA, GA, and MHQ and an initial H₂O₂ load of 20 μm. Fig. 5A shows the results obtained for GA, and Fig. S2 shows the results obtained for the other reductants. Fig. 5B shows the dependences of the initial rates of NAG_eq formation on the concentration of reductants. The initial rates versus [reductant] curves were in general accordance with hyperbolic saturation kinetics (Equation 2).

$$v_i=\frac{V_{\max }^{\text{app}}\left[\text{R}\right]}{K_{m\text{R}}^{\text{app}}+\left[\text{R}\right]}$$ (Eq. 2)

In Equation 2, the v_i is the initial rate of the formation of NAG_eq, [R] is the concentration of the reductant, and K_mR^app and V_max^app represent the apparent half-saturating concentration of the reductant and apparent V_max of the formation of NAG_eq. The V_max^app is related to the apparent catalytic constant for oxidative cleavage of CNWs (k_cat^app) according to V_max^app = nk_cat^app E_tot where n is an average number of NAG_eq released per oxidative cleavage (n = 4) and E_tot is the total concentration of enzyme. The parameters of Equation 2 are referred to as apparent parameters because the reaction was studied at one set of substrate concentrations only (20 μm H₂O₂ and 1 mg ml⁻¹ CNWs). Table 1 lists the values of K_mR^app and V_max^app and shows that differences between the reductants are primarily manifested by largely different K_mR^app values amounting to 2, 86, and 700 μm, for AscA, GA, and MHQ, respectively. The differences between V_max^app values obtained with different reductants were less prominent (Table 1), which is as expected because close to full reduction of the LPMO should be reached in all cases, regardless of the reductant.

Figure 5.

Initial rates of the formation of soluble products (in NAG_eq) and their dependence on the nature and concentration of the reductant. All experiments were carried out in NaAc buffer (50 mm, pH 6.1) at 25 °C and contained CNWs (1.0 mg ml⁻¹) and SmLPMO10A (42 nm). A, linear regions of the progress curves (full progress curves are shown in Fig. 2B) obtained with 20 μm H₂O₂ and GA as reductant. Solid lines, linear regression of the data. The slopes of the lines correspond to the initial rates (v_i). For the linear regions of the time curves obtained with other reductants and with GA and 100 μm H₂O₂, see Fig. S2. B, dependence of initial rates on the nature and concentration of the reductant. The type of reductant and the concentration of H₂O₂ are indicated in the plot. Solid lines, nonlinear regression of the data according to Equation 2. Error bars, S.D. derived from at least two independent experiments.

Table 1.

Kinetic parameters for different reductants in oxidation of chitin by SmLPMO10A

V_max^app and K_mR^app were determined by fitting the experimental data depicted in Fig. 5B to Equation 2. k_cat^app represents the rate constant for oxidative cleavages and is calculated from V_max^app, E_tot, and a stoichiometry of 4 NAG_eq/oxidative cleavage.

Reductant	Vmaxapp	kcatapp	KmRapp	kcatapp/KmRapp
	μm NAGeq s−1	oxidative cleavages s−1	μm	m−1 s−1
Ascorbic acid	0.53 ± 0.05	3.2 ± 0.3	2.0 ± 0.9	1,600,000 ± 700,000
Gallic acid	0.30 ± 0.01	1.8 ± 0.03	86 ± 7	21,000 ± 1,500
Methylhydroquinone	0.38 ± 0.02	2.3 ± 0.1	700 ± 100	3,300 ± 400
Gallic acida	0.20 ± 0.01	1.2 ± 0.1	180 ± 30	6,700 ± 1,000

^a Results obtained using 100 μm H₂O₂; in all other cases, the initial H₂O₂ load was 20 μm.

With GA as reductant, the K_mR^app and V_max^app were determined for two different initial H₂O₂ loads, 20 and 100 μm. At the higher H₂O₂ load, the V_max^app decreased (indicating inhibition by H₂O₂), whereas the K_mR^app increased (Table 1). The reasons for this apparent decrease in reductant efficiency (k_cat^app/K_mR^app in Table 1) with increasing H₂O₂ concentration are addressed in the theoretical analysis, described below.

Theoretical analysis of the reduction/oxidation of the LPMO in the absence of chitin

The reduced LPMO will be subject to nonproductive reoxidation primarily when it is not bound to substrate, because that exposes the reduced copper site to both O₂ and H₂O₂. To gain further insight into the dependence of K_mR^app and V_max^app on the nature of the reductant and on the concentration of H₂O₂, we thus first considered oxidation/reduction of the LPMO in the absence of chitin, as outlined in Fig. 6A (the effects of including chitin into the reaction scheme are discussed below). In Fig. 6A, the chemically inert complexes ECu(I)–R, ECu(II)–O₂, and ECu(II)–H₂O₂ are omitted for clarity. The scheme of Fig. 6A assumes that both reduction and oxidation of the LPMO are irreversible. This assumption is plausible when analyzing initial rates, as product concentrations and, thus, rates of possible reverse reactions are negligible.

Figure 6.

Theoretical analysis of the reduction/oxidation of LPMO in the absence of polysaccharide substrate. A, in the scheme, E represents LPMO, and Cu(I) and Cu(II) denote the oxidation state of the active-site copper. The reductant (R) and oxidants (O₂ and H₂O₂) are allowed to bind only to the Cu(II) and Cu(I) forms of LPMO, respectively. The routes shown in gray are not included in the analysis. These include (i) formation of H₂O₂ that is subsequent to reduction of O₂ by the LPMO and (ii) oxidation of LPMO by H₂O₂ that leads to the irreversible inactivation (i.e. p_ik_oxH2O2, where p_i is the probability that LPMO is inactivated in the reaction with H₂O₂). The equations show the dependence of the Cu(I) form of the LPMO and the parameters of Equation 2 (i.e. the apparent catalytic constant for chitin oxidation (k_cat^app) and the apparent half-saturating concentration of the reductant (K_mR^app)) on the concentration of O₂ and H₂O₂. The steady-state equations were derived using the assumption that only the Cu(I) form of LPMO is catalytically active. The binding of R, O₂, and H₂O₂ is represented by corresponding dissociation constants (K_R, K_O2, and K_H2O2), as defined in the figure. Note that the true reducing capability (efficiency) of the reductant is reflected in K_R and k_red, whereas the parameters of Equation 2 refer to the LPMO reaction. B–E show the dependence of the fraction of the Cu(I) form of the LPMO on the concentration of the reductant (B) and the dependence of the parameters of Equation 2, k_cat^app (C), K_mR^app (D), and k_cat^app/K_mR^app (E), on the efficiency of reductant (k_red/K_R). The data shown in B–E were calculated using the equations shown in the figure. In all calculations, the values of second-order rate constants (k₁, k₃, and k₅) were arbitrarily set to 10³ mm⁻¹ s⁻¹, whereas the values of first-order dissociation rate constants were arbitrarily set to 100 s⁻¹ for k₋₃ and k₋₅ and to 10³ s⁻¹ for k₋₁. This results in the values of 1, 0.1, and 0.1 mm for K_R, K_O2, and K_H2O2, respectively. The concentrations of O₂ and H₂O₂ were set to 0.2 and 0.02 mm, respectively. In B, the values of the catalytic rate constant for reduction (k_red) were varied as shown in the plot, and the catalytic rate constants for oxidation by O₂ (k_oxO2) and H₂O₂ (k_oxH2O2) were set to 0.01 and 0.1 s⁻¹, respectively. The latter implies that the rate of reoxidation of the reduced LPMO [ECu(I)] by O₂ and H₂O₂ (terms k_oxO2[O₂]/K_O2 and k_oxH2O2[H₂O₂]/K_H2O2 in Equation 3) are equal (0.02 s⁻¹). In C–E, a higher value of k_oxH2O2 (1.0 s⁻¹) is also considered (the color code is given in the plot), implying that the rate of reoxidation of the LPMO by H₂O₂ (0.2 s⁻¹) is 10-fold faster compared with reoxidation by O₂. In calculations for C–E, the K_R was kept at 1 mm, and k_red was varied between 0.01 and 10 s⁻¹. k_cat(max)^app represents the maximum value of k_cat^app (i.e. it represents the hypothetical situation when all enzyme is in its active form, [ECu(I)] = E_tot). For the calculations in C and E, the value of k_cat(max)^app was set to 10 s⁻¹.

The catalytically active form of SmLPMO10A is the ECu(I) form, which has been proposed to bind to chitin followed by the binding of H₂O₂ and oxidative cleavage of the glycosidic bond (25, 44). The equation for steady-state [ECu(I)] is given in Fig. 6. The dependence between the initial rates of NAG_eq formation (v_i) and [ECu(I)] is given by v_i = nk_cat(max)^app[ECu(I)], where n is the number of NAG_eq released per oxidative cleavage and k_cat(max)^app is the k_cat^app for oxidative cleavages in the conditions where all enzyme is in ECu(I) form. Replacing [ECu(I)] with E_tot, one can arrive to Equation 2 and derive equations that show the dependence of K_mR^app and k_cat^app on the concentration of O₂ and H₂O₂ (Fig. 6). Note that in doing so, we assume that concentration of CNWs has no effect on the steady-state [ECu(I)]. Using these equations, one can predict how the apparent catalytic properties of the LPMO depend on the efficiency of the reductant, k_red/K_R, as shown in Fig. 6 (B–E).

Fig. 6B shows that in the presence of O₂ and H₂O₂, the concentration of the catalytically active ECu(I) form is always lower than E_tot. The equation for k_cat^app shows that this parameter is always reduced compared with its maximum value, when all enzyme is expected to be in its Cu(I) form (k_cat(max)^app), at least by a factor of 1 + [O₂]/K_O2 + [H₂O₂]/K_H2O2. This is evident by inspection of the expression of k_cat^app in Fig. 6, assuming conditions of very efficient reduction (i.e. when the value of the rate constant for the reduction (k_red) is much higher than the rate constants for oxidation of LPMO by O₂ (k_oxO2) and H₂O₂ (k_oxH2O2)). The similar V_max^app values for different reductants (Table 1) indicate that their k_red values are always higher than k_oxH2O2 and k_oxO2. At very low concentration of both oxidants, O₂ and H₂O₂, the k_cat^app approaches k_cat(max)^app (regardless of k_red), whereas the value of K_mR^app approaches zero (Fig. 6). Fig. 6C shows that, as expected, the k_cat^app increases with an increasing k_red value of the reductant (Fig. 6C), whereas the K_mR^app decreases (Fig. 6D). The dependence of k_cat^app/K_mR^app on the efficiency of the reductant (k_red/K_R) was found to be linear (Fig. 6E).

Kinetics at low levels of O₂

Because the steady-state concentration of catalytically active Cu(I) form of LPMO is expected to depend in part on [O₂] (Fig. 6), we also measured the kinetics of the degradation of CNWs by SmLPMO10A under reduced [O₂] (i.e. under N₂ atmosphere). AscA, GA, and MHQ were used as reductants. The concentration of added H₂O₂ was 20 and 100 μm. In general, the kinetics (time curves and initial rates) measured in N₂ atmosphere were similar to those measured in normal air conditions (Fig. 7 and Fig. S3). This result is in line with earlier observations that, when provided with H₂O₂, the presence of O₂ has no effect on the activity of an LPMO (23). Of note, the binding of LPMO to cellulose has also been shown to be independent of the presence of O₂ (45). Collectively, these results suggest that in the experimental set-ups with initial H₂O₂ supply, the O₂-involving pathways in Fig. 6 are not significant.

Figure 7.

Kinetics of the degradation of CNWs by SmLPMO10A at different levels of O₂. All experiments were carried out in NaAc buffer (50 mm, pH 6.1) at 25 °C and contained CNWs (1.0 mg ml⁻¹) and SmLPMO10A (42 nm). A, initial rates (for the time curves see Fig. S3) of the formation of soluble products (in NAG_eq) in the experiments done under N₂ atmosphere (anaerobic) or under atmospheric O₂, as indicated. The nature and the concentration of the reductant along with the concentration of H₂O₂ are given in the plot. B, time curves of the formation of soluble products in the experiments with 100 μm H₂O₂ and GA as reductant (concentrations are shown in the plot). Experiments were done under N₂ atmosphere (anaerobic) or under atmospheric O₂, as indicated. Error bars, S.D. derived from at least two independent experiments.

Binding to CNWs in the absence of reductant

The binding of SmLPMO10A to CNWs in the presence of 0.1 mm AscA has been measured before (25) and resulted in the half-saturating concentration for CNWs of 0.68 ± 0.01 mg ml⁻¹. In the absence of reductant, the binding of LPMO to CNWs results in the formation of a nonproductive ECu(II)–CNW complex. In the case of nonefficient priming, this nonproductive complex can exist also in the presence of reductant but may be converted to a productive complex if reduction of LPMO within the complex with CNW is possible (e.g. through long-range electron transfer). To assess possible nonproductive binding, we measured the binding of SmLPMO10A to CNWs in the absence of reductant. SmLPMO10A was incubated with nonlabeled CNWs, and after separation of bound enzyme by centrifugation, the concentration of chitin-free enzyme was measured by measuring its activity in the supernatant using ¹⁴C-labeled CNWs (1.0 mg ml⁻¹) and H₂O₂ (20 μm) as substrates and 0.1 mm AscA as reductant. As a control experiment, we measured the time curves of the degradation of CNWs at different concentrations of SmLPMO10A (Fig. 8A).

Figure 8.

Binding of SmLPMO10A to chitin in the absence of reductant. All experiments were carried out in NaAc buffer (50 mm, pH 6.1) at 25 °C. Error bars, S.D. from two independent experiments. A, control experiment showing time curves for the formation of soluble products (in NAG_eq) upon incubation of CNWs (1.0 mg ml⁻¹) with SmLPMO10A at different concentrations (as indicated in the plot). The concentration of H₂O₂ and AscA was 20 and 100 μm, respectively. B, control experiment showing the dependence of initial rates of NAG_eq formation (derived from the data in A) on the concentration of SmLPMO10A. C, binding of SmLPMO10A to unlabeled CNWs in the absence of H₂O₂ and AscA was assessed by measuring SmLPMO10A activity in supernatants of the binding experiment using ¹⁴C-labeled CNWs supplemented with AscA and H₂O₂, as described under “Experimental procedures.” The plot shows the ratio of initial rates of SmLPMO10A activity in the supernatants from binding experiments in the presence (v_CNW) and absence (v_{CNW = 0}) of CNWs. The concentration of CNWs in the binding experiment is shown on the x axis. The values of v_CNW and v_{CNW = 0} were derived from the initial slopes of the time curves for the formation of soluble NAG_eq upon degradation of ¹⁴C-labeled CNWs by unbound SmLPMO10A in the supernatants of the binding reactions (for the time curves, see Fig. S4). The solid line is the best-fit line according to Equation 4.

The time curves for reactions with lower concentration of SmLPMO10A (below 25 nm) leveled off below the theoretical [NAG_eq]_max value (80 ± 6 μm) that can be released in experiments with 20 μm H₂O₂. As pointed out before, the [NAG_eq]_max value is a product of two competing processes, catalysis and inactivation. The inactivation is first-order with enzyme, and changing the concentration of SmLPMO10A has no influence on its half-life. At the same time, the rate of chitin oxidation is linear with the concentration of SmLPMO10A. Therefore, it is likely that at low concentration of SmLPMO10A, the enzyme is inactivated before it can perform enough oxidative cleavages to reach the theoretical [NAG_eq]_max value.

Fig. 8B shows a linear correlation between the initial rate of the LPMO reaction and the total concentration of SmLPMO10A. Of note, such a linear relationship has never been reported for O₂-driven reactions, in line with the notion that in such reactions, other factors, such as H₂O₂ generation, are rate-determining. Besides providing the control for the linearity between the initial rate and the enzyme concentration, this result indicates that the binding sites on CNWs were in excess at all enzyme concentrations used (up to 84 nm). The latter condition is a prerequisite for the applicability of steady-state kinetic models, but it is not often easy to achieve in the case of heterogeneous interfacial catalysis (46).

The time curves for the formation of NAG_eq by nonbound SmLPMO10A present in the supernatants of binding experiments are shown in Fig. S4. Analysis of the dependence of the ratio of initial rates measured in the presence (v_CNW) and absence (v_{CNW = 0}) of chitin in the binding experiment on the concentration of CNWs in binding experiments (Fig. 8C) according to Equation 4 results in a half-saturating concentration for CNWs of 3.0 ± 0.7 mg ml⁻¹. This concentration is 4.5-fold higher than the corresponding concentration measured in the presence of 0.1 mm AscA (25), which is well in line with the recent results by Kracher et al. (45), who demonstrated that the binding of the Cu(I) form of Neurospora crassa LPMO9C to cellulose is stronger compared with binding of the Cu(II) form. This latter study showed that reduction of N. crassa LPMO9C increased both the binding strength and binding capacity but that the effect was strongest (4.7-fold increase) at the level of the partition coefficient (initial slope of the Langmuir binding isotherm) (45). Because here we measured LPMO binding in conditions satisfying an excess of binding sites, the effects of reductant on binding (4.5-fold increase) reflect mainly effects on the partition coefficient of the Langmuir binding isotherm. The fact that similar results were obtained with Neurospora crassa LPMO9C and SmLPMO10A may be taken to suggest that the effects of the redox state of active site copper on binding of LMPOs to their polysaccharide substrate may be similar for LPMOs belonging to different families, but may also be coincidental because the two LPMOs have different substrates.

Discussion

Recent findings showing that H₂O₂ acts as a cosubstrate of LPMOs (22, 23, 25, 26) call for detailed kinetic studies of H₂O₂-driven degradation of polysaccharides. Such studies are complicated by the low K_m of LPMOs for H₂O₂, which necessitates the use of H₂O₂ at low micromolar concentrations. The problem is further amplified by the requirement for the priming reduction of the LPMO, because reducing agents are sensitive toward oxidation by O₂, especially in the presence of metal ions. Importantly, H₂O₂ is often a product of these oxidations. Oxidation of reducing agents by H₂O₂ is also plausible. Finally, enzyme inactivation is commonly observed for LPMOs, in both O₂- and H₂O₂-driven reactions (4, 23, 25).

Here, we occasionally encountered significant activity of SmLPMO10A in experiments without added H₂O₂. This background activity depended on the nature of the reductant and showed strong variation between different working batches of chitin (Fig. 1A). Numerous control experiments (Fig. 1B) suggested that the background activity is related to generation of H₂O₂ in the divalent metal ion-catalyzed oxidation of the reductant. Therefore, the addition of copper salts to reaction mixtures likely has effects beyond supplementing the LPMO with catalytically essential copper, and great care must be taken when doing so (47).

Regarding our studies on background activities, the effect of catalase is worth a short discussion. In some studies, it has been noticed that the addition of catalase improves the efficiency of LPMO-containing enzyme mixtures, which was ascribed to the potentially damaging effect of radical formation when H₂O₂ reacts with transition metals in the reaction mixture (48). For a similar reason, catalase is added in experimental set-ups where LPMO reactions are supplied with excess copper (49). Considering the now well-established beneficial effect of H₂O₂ on LPMO activity (22, 23, 25, 26), one might expect that catalase inhibits LPMO activity in such enzyme mixtures, which is not compatible with the observed overall beneficial effect of catalase (48, 50). Moreover, neither Bissaro et al. (23) nor Møllers et al. (51) observed an effect of catalase on LPMO activity, except in situations with very high H₂O₂ concentrations (22). Bearing in mind the high affinity of LPMOs for H₂O₂ (23), being as low as 2.8 μm for SmLPMO10A (25), and the very low affinity of catalase to H₂O₂ (52, 53), special care must be taken in dosing catalase and interpreting potential effects. The K_m for H₂O₂ of the bovine liver catalase used in this study is around 100 mm (52). Based on the rate of NAG_eq formation in the background reaction (0.36 μm NAG_eq min⁻¹; Fig. 1B) and the values of kinetic parameters of SmLPMO10A published before (25), one can estimate a steady-state concentration of 0.025 μm for H₂O₂ in the experiment with 0.1 mm AscA. As defined by the commercial suppliers, the unit of catalase activity (μmol of H₂O₂ min⁻¹) is measured at an H₂O₂ concentration around 10 mm. Therefore, the activity of catalase in our experiment is expected to be 10,000 μm/0.025 μm = 4 × 10⁵-fold lower than the activity that could be estimated from the information provided by the supplier. This may be the reason why an apparently huge dose of catalase (200 units ml⁻¹) did not completely inhibit the background activity of SmLPMO10A (Fig. 1B).

As pointed out before, the optimal harnessing of the catalytic potential of LPMO takes an experimental set-up with controlled in situ production (e.g. using oxidases) or continuous external supply of H₂O₂ (23, 25, 54). Provided that enzyme inactivation is avoided, the effect of reductant in these set-ups is expected to be similar to the effects of reductant on initial rates measured here (Fig. 5 and Table 1). Note that whereas the necessary reductants under some conditions may contribute to H₂O₂ generation, in the present study, conditions were such that the observed efficiency of reductants reflects the efficiency in priming of the LPMO. The stoichiometry of 18 ± 5 oxidative cleavages made per molecule of reductant, as measured in reactions with low concentrations of AscA (Fig. 3), supports the use of reductant in priming reduction and not as substrate in our experiments. This result is in accordance with conclusions drawn from reactions in which cellulose was degraded by an LPMO-containing commercial cellulase mixture, which showed that, when controlling the supply of H₂O₂, ∼200 μm AscA was needed to produce ∼3000 μm oxidized products, leading to an approximate stoichiometry of 15 (23).

In accordance with previous studies (10, 12), the present data show that different reductants have very different efficiencies (k_cat^app/K_mR^app) in supporting SmLPMO10A catalysis. Although different reductants have somewhat different k_cat^app values, we show that the differences in efficiency primarily relate to largely different apparent half-saturating concentration (K_mR^app) (Table 1), which again is strongly correlated with the rate constant for LPMO reduction (k_red) (Fig. 6D; K_mR^app decreases sharply as k_red increases). Of note, measuring the true electron transfer efficiency of the reductant (i.e. the values of k_red and K_mR) implies measurements of electron transfer in the pre-steady state regime. Stopped-flow experiments with cellobiose dehydrogenase indicate a second-order rate constant (cf. k_red/K_mR) of 6.4 10⁵ m⁻¹ s⁻¹ for electron transfer between SmLPMO10A and the cellobiose dehydrogenase of Myriococcum thermophilum (4).

The different efficiencies of the reductants are likely related to their varying midpoint potentials, as demonstrated in an earlier study by Kracher et al. (10). Kracher et al. (10) showed that MHQ is a moderately effective reductant, which coincides with its high K_mR^app (and, thus, low k_red) value reported here. MHQ undergoes two successive one-electron and one-proton transfers, leading to the semiquinone and quinone form, with redox potentials of 23 and 460 mV versus NHE, respectively. The standard potential for the direct transfer of two electrons and two protons to reduce the quinone to MHQ is 230 mV versus NHE (55). SmLPMO10A has a midpoint potential of 275 mV versus NHE (42). Ascorbic acid and gallic acid, two of the most commonly used reductants in LPMO research, are oxidized irreversibly (56, 57), which prevents the determination of standard midpoint potentials, and which, as such, may affect their efficiency. Comparative studies of oxygen reduction by Kracher et al. (10) indicated that ascorbic acid is a much better reductant than gallic acid and MHQ, which is compatible with the differences observed here.

From the analysis of the reaction scheme in Fig. 6, it follows that there is a linear relationship between the value of k_cat^app/K_mR^app for chitin degradation and the value of the efficiency constant for LPMO priming by the reductant (k_red/K_mR) according to Equation 3.

$$\frac{k_{\text{cat}}^{\text{app}}}{K_{m\text{R}}^{\text{app}}}=\frac{k_{\text{cat}\left(\max \right)}^{\text{app}}}{k_{{\text{oxO}}_2}\frac{\left[{\text{O}}_2\right]}{K_{{\text{O}}_2}}+k_{{\text{oxH}}_2{\text{O}}_2}\frac{\left[{\text{H}}_2{\text{O}}_2\right]}{K_{{\text{H}}_2{\text{O}}_2}}}\left(\frac{k_{\text{red}}}{K_{\text{R}}}\right)$$ (Eq. 3)

The rate constants and binding constants of Equation 3 are defined in the legend to Fig. 6. To what extent the value of k_cat^app/K_mR^app (Table 1) reflects the value of (k_red/K_mR) depends on the rates of LPMO reoxidation (i.e. the terms in the denominator of Equation 3; note that these are all terms referring to a situation without bound substrate, which is the most common, if not the only, situation in which reduction occurs). To decide which route, O₂- or H₂O₂-driven, governs LPMO reoxidation, one needs numerical estimates for the corresponding terms in the denominator of Equation 3. The K_H2O2 in this denominator reflects the binding of H₂O₂ to free reduced LPMO (Fig. 6A). This binding mode has been proposed to be responsible for the inactivation of SmLPMO10A and has low affinity (K_H2O2 > 100 μm as estimated from the data in Ref. 25). The k_oxH2O2 is the catalytic constant for the oxidation of ECu(I) by H₂O₂. This H₂O₂-dependent oxidative route can lead to irreversible inactivation of LPMO but also to reoxidation of LPMO to its ECu(II) form without inactivation (as in Fig. 6A). The second-order rate constant for inactivation of SmLPMO10A is on the order of 10³ m⁻¹ s⁻¹ (25). Using this number, one can estimate the rate of reoxidation (k_obs(oxH2O2)) of SmLPMO10A to be around 0.02 s⁻¹ (H₂O₂ at 20 μm). Note that the true reoxidation rate by H₂O₂ is expected to be 1/p_i-fold higher, where p_i is the probability that enzyme is inactivated upon reacting with H₂O₂ in the absence of chitin (Fig. 6A). The rate of O₂-mediated reoxidation of SmLPMO10A is not known, but similar initial rates measured under normal air and anaerobic conditions (Fig. 7) suggest that under the conditions used in our study, reoxidation is governed by H₂O₂. Regarding other LPMOs, the data for the rate of O₂-mediated reoxidation of nonsubstrate-bound LPMOs vary a lot. Some studies support rate constants below 1 min⁻¹ (10, 31, 34, 58), whereas others support values above 10 min⁻¹ (59, 60). Recently, Breslmayr et al. (61) showed that the oxidation of 2,6-dimethoxy phenol by N. crassa NcLPMO9C is driven by H₂O₂ and does not depend on the presence of O₂, suggesting that in the absence of a polysaccharide substrate, O₂-driven reoxidation of the LPMO is much slower than H₂O₂-driven reoxidation, although this of course would depend on the concentrations of the two co-substrates. More studies are needed to make general conclusions about the contribution of possible oxidase/peroxidase reactions in reoxidation of substrate-free LPMOs.

The simplest mechanism of H₂O₂-driven degradation of CNWs by SmLPMO10A that can account for the observations and considerations described here is depicted in Fig. 9A. For simplicity, binding of the Cu(II) form of the LPMO to chitin was omitted. This simplification is, at least to some extent, justified by the experimentally observed weaker binding of the Cu(II) form of SmLPMO10A compared with the Cu(I) form (Fig. 8C and data in Ref. 25). Weaker binding of the Cu(II) form is also supported by computational studies of SmLPMO10A (44). It has been proposed that H₂O₂-driven catalysis by SmLPMO10A follows a compulsory order ternary complex mechanism, with chitin being the first substrate to bind to the reduced enzyme (25, 44). Simulations suggest that when SmLPMO10A is bound to chitin, a channel connecting the bulk solvent to the active site would regulate access of reagents to the latter (44). These considerations imply that the binding modes of H₂O₂ to free SmLPMO10A and to the complex of SmLPMO10A with CNWs are different. Therefore, binding of the LPMOCu(I)–H₂O₂ complex to CNWs is omitted in the scheme of Fig. 9A. Because the O₂-driven mechanism has been shown to follow a random-order mechanism (60) we do not make a similar assumption for binding of O₂ (Fig. 9A). Because in the experimental conditions used to determine the initial rates, the oxidation of chitin without external H₂O₂ supply was not significant, the possible O₂-driven route of chitin oxidation is omitted in the scheme of Fig. 9A.

Figure 9.

Theoretical analysis of the effects of reductant on H₂O₂–driven degradation of polysaccharides by an LPMO. A, the reaction scheme is constructed by adding the chitin (CNW) to the scheme shown in Fig. 6A. The routes shown with gray are not included in the analyses. Numerical solutions for the steady-state are shown in B–G. If not stated otherwise, the values of k₁, k₃, k₅, k₉, and k₁₀ were set to 10³ mm⁻¹ s⁻¹. The values of k₋₃, k₋₅, and k₋₉ were set to 100 s⁻¹, whereas the value of k₋₁ was 10³ s⁻¹. Therefore, the K_R, K_O2, and K_H2O2 (for definitions, see the legend to Fig. 6) had values of 1, 0.1, and 0.1 mm, respectively. The binding constants for CNWs were adjusted to 0.5 mg ml⁻¹ by setting k₇ = k₈ = 100 ml mg⁻¹ s⁻¹ and k₋₇ = k₋₈ = 50 s⁻¹. For the dissociation of H₂O₂ from its ECu(I)–CNW complex, the value of k₋₁₀ was set to 2 s⁻¹. The value of k_cat(max)^app was set to 10 s⁻¹. Concentrations of O₂, H₂O₂, and CNWs were set to 0.2 mm, 0.02 mm, and 1.0 mg ml⁻¹, respectively. B, dependence of the initial rate of NAG_eq formation (v_i) on the concentration of reductant at different k_red values (as shown in the plot). The concentration of enzyme was set to 0.05 μm, and four NAG_eq are released per molecule of H₂O₂. The values of k_oxO2 and k_oxH2O2 were set to 0.01 and 0.1 s⁻¹, respectively (i.e. the rates of reoxidation of LPMO by O₂ and H₂O₂ are both 0.02 s⁻¹). The v_i based on the scheme with inclusion of chitin (CNW+) was found from numerical solution of the scheme in A, whereas the v_i based on the scheme without explicit inclusion of chitin (CNW−) was calculated from the dependence of the concentration of the Cu(I) form of LPMO on the concentration of the reductant, as shown in Fig. 6B. C, dependence of k_cat^app/K_mR^app for chitin oxidation on the efficiency of reductant (k_red/K_R) based on reaction schemes with (CNW+; Fig. 9A) or without (CNW−; Fig. 6A and Equation 3) the inclusion of chitin. Obviously, there is no LPMO activity without chitin, but the calculations without chitin were included as they represent a prediction of what will happen in the situation where chitin does not influence the steady-state [ECu(I)]. In the calculations, the K_R was kept at 1 mm, and k_red was varied between 0.01 and 10 s⁻¹. As in B, the rates of LPMO reoxidation by O₂ and H₂O₂ were both set to 0.02 s⁻¹. D, dependence of initial rates of NAG_eq formation (v_i) on the concentration of reductant at different concentrations of H₂O₂ (as shown in the plot). The concentration of enzyme was set to 0.05 μm, and four NAG_eq are released per molecule of H₂O₂. The values of k_red and k_oxH2O2 were set to 0.01 and 0.1 s⁻¹, respectively. The binding strength of O₂ (K_O2) was set to 1.0 or 10 mm as indicated in the plot. E–G, dependence of k_cat^app (E), K_mR^app (F), and k_cat^app/K_mR^app (G) on the rate of LPMO reoxidation (k_obs(ox)). k_obs(ox) is the sum of the rate constants of oxidation by O₂ (k_obs(oxO2)) and H₂O₂ (k_obs(oxH2O2)), which were calculated according to k_obs(oxO2) = k_oxO2[O₂]/K_O2 and k_obs(oxH2O2) = k_oxH2O2[H₂O₂]/K_H2O2. The value of k_red was set to 1.0 s⁻¹. The values of k_obs(oxO2) and k_obs(oxH2O2) were varied by varying the values of k_oxO2 (between 0.0025 and 5.0 s⁻¹) and k_oxH2O2 (between 0.025 and 50 s⁻¹). The effect of k_obs(ox) on apparent parameters of the reductant was assessed using three different scenarios: (i) k_obs(oxO2) = k_obs(oxH2O2), (ii) k_obs(oxH2O2) = 0 (mimicked by setting k₅ = 0), and (iii) k_obs(oxO2) = 0 (mimicked by setting [O₂] = 0). All scenarios were analyzed at two concentrations of CNWs, 1.0 mg ml⁻¹ (solid lines) and 10 mg ml⁻¹ (dashed lines). Note that the results of the first two scenarios (blue and green lines) overlap. Note also that the scale for k_cat^app/K_mR^app is logarithmic.

The steady-state solution for the mechanism in Fig. 9A was found using the King–Altman procedure (62) and analyzed numerically. In general, the numerical solutions for the dependence of chitin degradation on the efficiency of reductant (Fig. 9) are in qualitative accordance with the analytical solutions found using the scheme that does not consider chitin (Fig. 6). The dependence of initial rates on the concentration of reductant followed Michaelis–Menten saturation kinetics regardless of the k_red (Fig. 9B), similar to the dependence of the concentration of the ECu(I) form of LPMO on the concentration of reductant found by analytical solutions to the scheme without CNWs (Fig. 6B). The k_cat^app increased and the K_mR^app decreased with increasing efficiency of reduction (data not shown), similarly to predictions derived from the model without explicit inclusion of chitin (Fig. 6, C and D). Both reaction schemes yielded a linear dependence of k_cat^app/K_mR^app on the true efficiency of reductant (k_red/K_mR), but the slope found in the scheme with the explicit presence of chitin (Fig. 9A) was higher than that predicted in the absence of chitin (Fig. 9C). This can be accounted for by the reduced rate of LPMO reoxidation in the presence of chitin, as only the reoxidation of chitin-free enzyme is considered in Fig. 9A. Numerical solutions to the scheme in Fig. 9A also predicted inhibition by H₂O₂ (Fig. 9D), as indeed observed when using GA as reductant (Fig. 5B). Of note, this hitherto nonobserved phenomenon refers to true inhibition, reflected in reduced initial rates, and should not be confused with irreversible inactivation caused by H₂O₂.

Regarding the application of LPMOs, it is important to note that the requirements for the reductant (e.g. its half-saturating concentration) in supporting polysaccharide degradation depend not only on the catalytic efficiency of the reductant (k_red/K_mR) (Figs. 6 (C–E) and 9C) but also on the rate of LPMO reoxidation (Fig. 9, E–G). The increase of K_mR^app (Fig. 9F) and decrease of k_cat^app (Fig. 9E) with increasing rate of LPMO reoxidation is caused by the competition between the possible polysaccharide monooxygenase/peroxygenase and reductant oxidase/peroxidase activities of the LPMO. Increasing the polysaccharide concentration favors the monooxygenase/peroxygenase reactions, at the expense of the oxidase/peroxidase reactions that take place with free LPMOs, and this is reflected in decreasing K_mR^app (Fig. 9F) and increasing k_cat^app (Fig. 9E). Because the irreversible inactivation of LPMO is also related to the reoxidation of free LPMO by H₂O₂ (23, 25, 26), the high dry matter conditions used in industrial scale degradation of lignocellulosic biomass (63) should reduce the needed amounts of the reductant for the priming of LPMO as well as reduce the inactivation of LPMO. Obviously, careful control of the amount of available H₂O₂ is also of importance, because this will avoid unproductive oxidation of the reductant as well as LPMO inactivation (54).

Regarding the natural environment of LPMOs and potential sources of H₂O₂, a plethora of H₂O₂ suppliers as potential candidate partner enzymes have been identified for fungal LPMOs involved in lignocellulose conversion, although most connections remain to be clearly established (24). In the case of bacterial LPMOs, even less is known because no obvious, conserved redox partner, like cellobiose dehydrogenase in fungi, has been identified yet. As to the matter of the availability of H₂O₂, reported concentrations in natural ecosystems span several orders of magnitude, but existing data need to be interpreted with caution because H₂O₂ concentrations measured in vivo may reflect the result of multiple, and sometimes unknown, production and consumption fluxes (24). Of note, the very low K_m values for H₂O₂ (2.8 μm for SmLPMO10A) indicate that H₂O₂-driven LPMO action should be compatible with most forms of microbial life.

In summary, the optimal concentration and the performance of the reductant depend on the relative activities of the different reactions catalyzed by a particular LPMO, but also on process conditions like the concentrations of O₂, H₂O₂, and the polysaccharide substrate.

Experimental procedures

Substrates and enzymes

¹⁴C-Labeled CNWs were prepared from α-chitin of crab shells (Sigma C7170) using N-acetylation of free amino groups with 1-¹⁴C acetic anhydride, as described by Kuusk et al. (64). The specific radioactivity of the CNW preparation was 4.18 × 10⁶ dpm mg⁻¹. A mother stock suspension of CNWs (8.5 mg ml⁻¹) was kept at 4 °C in 50 mm NaAc buffer pH 6.1 (0.01% NaN₃). The batches of CNWs were prepared from the mother stock by washing out the NaN₃ through repetitive centrifugation (5 min, 2,900 × g) and resuspension (in 50 mm NaAc buffer, pH 6.1) steps. Although the water was Milli-Q ultrapure water (>18.2 megaohms cm⁻¹), we encountered variation in the background activity between working batches of CNWs when AscA was used as reductant, as described under “Results.” Apart from the experiments shown in Fig. 1, only the working batches of CNWs with low background activity were used in the experiments with AscA as reductant. There were no differences in background activities between working batches of CNWs when GA or MHQ were used as reductants, and the results with these reductants represent results obtained using different working batches of CNWs. Nonlabeled CNWs used in the binding experiment were prepared from α-chitin of crab shells (Sigma C7170) exactly as described by Kuusk et al. (64).

Ascorbic acid (Sigma, A7506), gallic acid (Sigma, G7384), methylhydroquinone (Sigma-Aldrich, 89600), and 1,4-dihydroxy-2,6-dimethoxybenzene (Aldrich, 565032) stock solutions were prepared in water less than 10 min before use. EDTA and AmplexRed were from Sigma. Dilutions of a commercial H₂O₂ stock solution (Honeywell, lot SZBG2070) with known concentration (30 weight %, 9.8 m) were prepared in water less than 10 min before use. The water was Milli-Q ultrapure water (>18.2 megaohms cm⁻¹) throughout.

SmLPMO10A was produced and purified as described previously (65). Purified SmLPMO10A was saturated with copper by incubating with CuSO₄ and following removal of unbound copper by ultracentrifugation exactly as described by Kuusk et al. (25). The concentration of SmLPMO10A was determined by absorbance at 280 nm using a theoretical molar extinction coefficient of 35,200 m⁻¹ cm⁻¹. Catalase from bovine liver (Sigma, C9322) and peroxidase from horseradish (Sigma, P8375) were used as purchased.

Degradation of CNWs by SmLPMO10A

Reactions were prepared essentially as described by Kuusk et al. (25). All reactions were performed in 1.5-ml polypropylene microcentrifuge tubes in NaAc buffer (50 mm, pH 6.1) at 25 °C without stirring, although the suspension was mixed by pipetting before withdrawing samples, at each sampling time point. If not stated otherwise, the reaction mixture contained ¹⁴C-labeled CNW (1.0 mg ml⁻¹), SmLPMO10A (42 nm), H₂O₂, and reductant. The total volume of the reactions was 0.8 ml. SmLPMO10A was added to CNWs, and after 5–10 min of incubation, the reductant was added. 30 s after the addition of reductant, the reactions were initiated (zero time point) by the addition of H₂O₂ to a desired concentration. At selected time points, 0.1-ml aliquots were withdrawn and mixed with 25 μl of 1.0 m NaOH to stop the reaction. Nonlabeled CNWs (to 3 mg ml⁻¹) in 0.2 m NaOH were added to improve the sedimentation of the CNWs during centrifugation (25). After centrifugation (5 min, 10⁴ × g), 50 μl of supernatant was withdrawn, and the radioactivity in the supernatant was measured using a scintillation counter (PerkinElmer Life Sciences). The sample for the zero-time point was withdrawn before the addition of H₂O₂ and was treated as the other samples. The reading of the zero-time point was subtracted from the readings of all time points. The reactions without the addition of H₂O₂ were performed exactly as described above with the addition of an equal amount of NaAc buffer (50 mm, pH 6.1) instead of H₂O₂. Concentrations of soluble products were expressed in NAG_eq, which were calculated from the radioactivity readings exactly as described by Kuusk et al. (25). At least two independent replicates were carried out for each experiment (S.D. values are derived from at least two experiments).

Control reactions for testing the effect of the order of the addition of reductant and H₂O₂ (Fig. S1) were set up with 1 mm GA or MHQ as reductants. The reaction mixtures contained SmLPMO10A (42 nm), CNWs (1.0 mg ml⁻¹), and H₂O₂ (20 μm) as described above but with an opposite order of addition for the reductant and H₂O₂. In one set of experiments, the reaction was started by the addition of H₂O₂, but the preincubation time of CNW/SmLPMO10A mixture with reductant was extended from the usual 30 s to 5 min.

For the degradation of CNWs in anaerobic conditions, the reactions were performed in a glovebox under N₂ atmosphere. Before use, the buffer (50 mm NaAc, pH 6.1) was N₂-saturated by five cycles of 5 min of vacuum degassing and 5 min of bubbling with N₂ gas. The CNW working batch (8 ml in a 15-ml tube) and stock solution of SmLPMO10A (0.5 ml in a 1.5-ml microcentrifuge tube) were treated by repeated flowing of N₂ gas into the headspace of the tube. The stock solutions of H₂O₂ and reductant were prepared in a glovebox using N₂-saturated buffer. The N₂-saturated buffer constituted 85% of the total volume of the SmLPMO10A reactions set up in the glovebox.

Binding of SmLPMO10A to CNWs in the absence of reductant

In the binding experiment, SmLPMO10A (84 nm) was incubated with nonlabeled CNWs (at different concentrations), and after 5 min, the CNWs were pelleted by centrifugation (1 min, 10⁴ × g). The concentration of unbound SmLPMO10A was estimated by measuring LPMO activity in the supernatant using ¹⁴C-labeled CNWs as substrate. To do this, labeled CNWs and AscA were added to the supernatant. 30 s after the addition of AscA, the reaction was initiated by the addition of H₂O₂, and the release of ¹⁴C-labeled products over time (in NAG_eq) was measured. The final concentrations of ¹⁴C-labeled CNWs, AscA, and H₂O₂ were 1.0 mg ml⁻¹, 0.1 mm, and 20 μm, respectively. In this procedure, the supernatant of the binding reaction was diluted 2-fold, meaning that the maximal (i.e. no binding of SmLPMO10A to nonlabeled CNWs) total concentration of SmLPMO10A in the activity measurement was 42 nm. The control experiment ([CNW] = 0 mg ml⁻¹) was undertaken exactly as described above but without the nonlabeled CNWs in the binding experiment. The results of the binding experiments were fitted to Equation 4.

$$\frac{v_{\text{CNW}}}{v_{\text{CNW}=0}}=\frac{1}{1+\frac{\left[\text{CNW}\right]}{K_{i\left(\text{CNW}\right)}}}$$ (Eq. 4)

In Equation 4, v_CNW and v_{CNW = 0} are the initial rates of CNW degradation by the supernatants from the binding experiment made in the presence and absence of CNWs, respectively. [CNW] is the concentration of CNWs in the binding experiment, and K_i(CNW) is the half-saturating concentration of CNWs at which 50% of the SmLPMO10A molecules are bound to CNWs and 50% are free in the solution.

Author contributions

P. V. conceived and coordinated the study. S. K., R. K., and P. V. designed, performed, and analyzed the experiments, interpreted data, and wrote the paper. A. H. assisted with the experiments. P. K. derived the rate equations. B. B., M. S., and V. G. H. E. interpreted data and wrote the paper. All authors reviewed the results and approved the final version of the manuscript.