Following the Fate of Lytic Polysaccharide Monooxygenases under Oxidative Conditions by NMR Spectroscopy

Lytic polysaccharide monooxygenases (LPMOs) are copper-dependent enzymes that catalyze oxidative cleavage of polysaccharides, such as cellulose and chitin. LPMO catalysis requires a reductant, such as ascorbic acid, and hydrogen peroxide, which can be generated in situ in the presence of molecular oxygen and various electron donors. While it is known that reduced LPMOs are prone to autocatalytic oxidative damage due to off-pathway reactions with the oxygen co-substrate, little is known about the structural consequences of such damage. Here, we present atomic-level insights into how the structure of the chitin-active SmLPMO10A is affected by oxidative damage using NMR and circular dichroism spectroscopy. Incubation with ascorbic acid could lead to rearrangements of aromatic residues, followed by more profound structural changes near the copper-active site and loss of activity. Longer incubation times induced changes in larger parts of the structure, indicative of progressing oxidative damage. Incubation with ascorbic acid in the presence of chitin led to similar changes in the observable (i.e., not substrate-bound) fraction of the enzyme. Upon subsequent addition of H2O2, which drastically speeds up chitin hydrolysis, NMR signals corresponding to seemingly intact SmLPMO10A reappeared, indicating dissociation of catalytically competent LPMO. Activity assays confirmed that SmLPMO10A retained catalytic activity when pre-incubated with chitin before being subjected to conditions that induce oxidative damage. Overall, this study provides structural insights into the process of oxidative damage of SmLPMO10A and demonstrates the protective effect of the substrate.


■ INTRODUCTION
The discovery of lytic polysaccharide monooxygenases (LPMOs) more than a decade ago has changed our understanding of enzymatic biomass degradation. 1−5 LPMOs are monocopper enzymes that degrade recalcitrant polysaccharides like cellulose and chitin 4,6−10 by disrupting the crystalline structure of these polysaccharides. 4,6−12 The reaction mechanism used by LPMOs is not fully uncovered 13,14 but involves either O 2 or H 2 O 2 as a co-substrate and an external reductant. 4,14−18 The copper-active site is part of a solvent-exposed substrate-binding surface. It comprises two conserved histidines (one of which is the N-terminal residue) that coordinate a single copper ion. 6,8−10 LPMOs share similar three-dimensional structures, consisting of a β-sandwich core with β-strands linked through loops containing several short helices, believed to influence substrate specificity and oxidative regioselectivity. 19−21 Chitin-and cellulose-active LPMOs cleave the β(1,4)glycosidic bonds in their substrates in a regioselective manner, acting on the C1-and/or C4-carbon. 21,22 The LPMO reaction ( Figure 1) starts with the reduction of the copper atom in the active site from Cu(II) to Cu(I) by an external reductant. The solvent-exposed active site 6,9 enables copper reduction by both small-molecule reductants such as ascorbic acid and gallic acid 4,6,23 and redox enzyme partners such as cellobiose dehydrogenase, 10,24 oligosaccharide oxidases, 25,26 and pyrroloquinoline-quinone-dependent dehydrogenase. 27 The activated LPMO then reacts with molecular O 2 or H 2 O 2 to create a reactive oxygen species capable of abstracting a hydrogen atom from C1 or C4 of the scissile glycosidic bond. Hydrogen abstraction is followed by hydroxylation through a rebound mechanism, which destabilizes the glycosidic bond and leads to its cleavage. 10,[13][14][15]17,28 The peroxygenase reaction with H 2 O 2 is much faster than the monooxygenase reaction with O 2 , 18,29−31 and there is some debate in the field regarding the kinetic relevance of the latter. 17,18 Although copper reduction has been shown to increase substrate affinity, 9,32 it is still unclear if LPMOs bind the oxygen co-substrate before or after they bind to the substrate. LPMO activity significantly boosts the performance of hydrolytic enzymes involved in the depolymerization of crystalline substrates. 2,4,5,33 However, to efficiently harness the power of LPMOs in commercial enzyme cocktails for polysaccharide saccharification, challenges related to LPMO inactivation must be resolved. 34 In solution, reduced LPMOs are susceptible to inactivation caused by reactions between the activated LPMO and its cosubstrate in a process referred to as oxidative selfinactivation. 15,35,36 Oxidative self-inactivation occurs when H 2 O 2 reacts with the non-substrate-bound reduced LPMO, leading to off-pathway reactions that may cause oxidative damage to catalytically important residues. 15,35,37 Even in LPMO reactions without added H 2 O 2 , this oxidant will be available due to several side reactions. LPMOs will produce H 2 O 2 when incubated with an appropriate reductant (like ascorbic acid) and O 2 in the absence of their substrate. 15,38,39 In addition, H 2 O 2 may be generated through reactions between the reductant and O 2 . 23,40 The oxidation of certain reductants (including the much-used ascorbic acid) is strongly affected by the presence of transition metals in solution. A recent study showed that even low micromolar concentrations of free copper (i.e., concentrations similar to typically used LPMO concentrations) promote oxidation of ascorbic acid and concomitant production of H 2 O 2 , which may lead to inactivation of LPMOs. 16 As LPMOs are efficient peroxygenases, 15,29 LPMO reactions can be fueled by extraneously supplied H 2 O 2 . When doing so, the supply of H 2 O 2 must be carefully controlled to prevent LPMO inactivation. 15,30,34 The protective effect of the substrate against oxidative selfinactivation has been shown by carrying out LPMO reactions with varying substrate concentrations. 15,39,41 The presence of carbohydrate binding modules, which are believed to keep the LPMO in close proximity to its substrate, also reduces LPMO inactivation. 41 Substrate binding will position the reactive oxygen species produced during catalysis in a way that promotes bond cleavage rather than damaging off-pathway reactions. 28,32,42 Interestingly, Bissaro et al. identified a tunnel connecting the active site of the chitin-bound SmLPMO10A to the bulk solvent. This would make it possible for the reduced LPMO to activate O 2 or H 2 O 2 while substrate-bound, 42 creating a controlled reaction environment.
Currently, limited information is available regarding the structural effects of oxidative off-pathway reactions in LPMOs. Mass spectrometry-based studies of oxidative damage in cellulose-active ScLPMO10C have shown that oxidative damage primarily occurs close to the copper-active site, with the two catalytic histidines and nearby aromatic residues being most exposed to oxidation. 15 Other parts of the structure seemed largely unaffected. 15 A more detailed molecular understanding of the oxidative self-inactivation of LPMOs is needed as it may eventually help optimize these enzymes for utilization in industrial biomass saccharification. In addition, one may speculate about the existence of possible protective hole hopping pathways 43 in LPMOs, which would be expected to largely involve aromatic residues. The propagation of oxidative damage through the LPMO molecule likely relates to such pathways.
Here, we have studied structural changes in an LPMO that is particularly rich in aromatic residues near the catalytic center under oxidative conditions. NMR 15 N-HSQC (heteronuclear single-quantum coherence) spectra and circular dichroism (CD) spectra were used to monitor structural changes occurring in chitin-active SmLPMO10A (also known as CBP21) incubated with ascorbic acid under aerobic conditions and/or H 2 O 2 in the absence of the substrate. Similar treatments in the presence of chitin were used to investigate (1) A copper atom must be bound to the LPMO-active site for the enzyme to be catalytically active. LPMO activity is also dependent on (2) a priming reaction where Cu(II) is reduced to Cu(I) by a reductant such as ascorbic acid. (3) In the presence of the substrate, the reduced LPMO (provided with an oxygen containing co-substrate) catalyzes oxidative cleavage of glycosidic bonds in the substrate. (4) Alternatively, an off-pathway reaction occurs, which may lead to oxidative damage by uncontrolled reactive oxygen species. (5) The ratio between on-pathway and off-pathway reactions depends on the concentrations of both substrate and H 2 O 2 . (6) The off-pathway reactions may lead to copper oxidation without enzyme damage, meaning that re-reduction of the LPMO is needed to become active. Of note, damage to the enzyme may lead to the release of free copper into the solution.
the protective effect of the substrate against oxidative damage and self-inactivation.

■ MATERIALS AND METHODS
The methods and experimental design used to monitor structural changes in SmLPMO10A under oxidative conditions in either the presence or absence of chitin are summarized in Figures 2 and S1.
Sample Preparation of apo-SmLPMO10A for NMR. Cloning, production, purification, and NMR sample preparation of both natural isotope abundance, 15 N-and 13 C, 15 Nisotopically enriched SmLPMO10A samples (UniProt entry Figure 2. Flow diagram of the incubations and NMR experiments done to investigate copper-active site reduction and oxidative self-inactivation of SmLPMO10A. The rectangular dark green text boxes represent the time points when the NMR spectra of SmLPMO10A were recorded. The experiments were done using three treatment series. In all the treatment series, copper was added to a sample of apo-SmLPMO10A in a 3:1 molar ratio to obtain holo-SmLPMO10A. Note that excess (i.e., unbound) Cu(II) ions were removed from the samples by desalting, and the samples were then concentrated to samples of ∼150 μM protein. Ascorbic acid (AscA) was added as indicated, without (series 1) or with (series 2 and 3) prior to the addition of chitin. After approximately 2 days, ascorbic acid and hydrogen peroxide were added, as indicated, and time points from this second phase of the incubation experiment are referred to as t′ rather than t. Time points t = 0 and t′ = 0 refer to approximately 5 min after addition. The recording time for each 15 N-HSQC spectrum was 40 min. O83009, residues 28−197) were carried out as previously described. 44 The protein concentration was determined by measuring the A 280nm value of the samples using a NanoDrop ND-1000 spectrophotometer (NanoDrop). The A 280nm value was then used to calculate the protein concentration based on the theoretical extinction coefficient (ε = 35,200 M −1 cm −1 ) obtained using the ProtParam tool (https://web.expasy.org/ protparam/).
To obtain the apo form of SmLPMO10A, the purified enzyme samples were incubated at 4°C overnight in acetate buffer (25 mM sodium acetate and 10 mM NaCl, pH 5.0) with 2 mM EDTA. A buffer exchange with acetate buffer (25 mM sodium acetate and 10 mM NaCl, pH 5.0) was performed to remove EDTA. A sample was then concentrated to ∼150 μM with a volume of ∼500 μL using first a VivaSpin centrifugation filter (10 kDa cut-off, Sartorius), followed by an Amicon Ultra centrifugation filter (3 kDa cut-off, Merck). The sample was transferred to a 5 mm NMR tube (LabScape Essence), and 10% D 2 O (99.9% D, Sigma-Aldrich) was added.
To obtain the samples of holo-SmLPMO10A, CuCl 2 was added to apo-SmLPM10A in a 3:1 molar ratio in acetate buffer (25 mM sodium acetate and 10 mM NaCl, pH 5.0). The sample was then incubated for 30 min at room temperature, and excess copper was removed by gel filtration using a PD MidiTrap G-25 desalting column (GE Healthcare; Uppsala, Sweden) equilibrated with acetate buffer (25 mM sodium acetate, 10 mM NaCl, pH 5.0). 45 The resulting sample of holo-SmLPMO10A was concentrated to ∼150 μM using an Amicon Ultra centrifugation filter (3 kDa cut-off, Merck), and the sample was split into three equal volumes of ∼500 μL. One of the resulting samples was transferred to a 5 mm NMR tube (LabScape Essence), and 10% D 2 O (99.9%, D, Sigma-Aldrich) was added.
Insoluble β-chitin fibers (Mahtani Chitosan) were mechanically treated by milling and sieving to a particle size of approximately 0.5 mm. 10 mg of these β-chitin particles were transferred into two separate 5 mm NMR tubes (LabScape Essence). The two remaining holo-SmLPMO10A samples were then added to each of the NMR tubes, and 10% D 2 O (99.9%, D, Sigma-Aldrich) was added to the samples. NMR Spectroscopy. All NMR spectra were recorded at 25°C on a Bruker AVANCE III HD 800 MHz spectrometer equipped with a 5 mm Z-gradient CP-TCI (H/C/N) cryogenic probe at the NV-NMR-Centre/Norwegian NMR Platform (NNP) at the Norwegian University of Science and Technology (NTNU). 1 H chemical shifts were internally referenced to the water signal at 25°C, while 15 N and 13 C chemical shifts were indirectly referenced to the water signal based on absolute frequency ratios (Zhang et al. 2003). 46 The NMR data were processed using TopSpin version 4.0.7. The processed NMR spectra were analyzed using CARA version 1.5.5. The previously published NMR assignment of SmLPMO10A (Aachmann et al., 2012) with BMRB entry number 17160 was used.
SmLPMO10A under Oxidative Conditions by NMR Spectroscopy. The effects of oxidative conditions on the structure of SmLPMO10A were investigated by NMR using the experimental design outlined in Figure 2. Holo-SmLPMO10A was subjected to conditions designed to promote oxidative self-inactivation in three treatment series. In treatment series 1, holo-SmLPMO10A was incubated in oxidative conditions without the substrate, while in treatment series 2 and 3, the enzyme was pre-incubated with chitin for 1 and 24 h, respectively, before being subjected to oxidative conditions (with the substrate still present).
Copper Binding. Cu(II) binding to the active site of SmLPMO10A was evaluated by recording a 15  . At this point, a 1 H spectrum was recorded to verify that no more ascorbic acid could be observed. Additional ascorbic acid (5 mM) was added together with H 2 O 2 (2 mM) to potentially induce extensive oxidative damage, and 1 H, 15 N-HSQC, HNCA, and HNCACB spectra were recorded over the course of 1 day (Figure 2: 15 N-HSQC 12−13). Finally, a 1 H spectrum was recorded to verify that no more ascorbic acid could be observed.
In treatment series 3 (24 h pre-incubation with chitin), 1 H and 15 N-HSQC spectra were recorded every 24 h for a total of 3 days (Figure 2: 15 N-HSCQ 15−18), and the 1 H spectrum was recorded to verify that no more ascorbic acid could be observed. Additional ascorbic acid (5 mM) was added together with H 2 O 2 (2 mM) to potentially induce extensive oxidative damage. Following this step, 15  where ΔδH = δH obs − δH free and ΔδN = δN obs − δN free are the absolute changes in the chemical shifts in parts per million (ppm) of the amide proton and amide nitrogen, respectively. 48 CSPs were calculated for pairs of consecutively recorded 15 N-HSQC spectra in each of the treatment series and between each recorded 15 N-HSQC spectrum and the 15 N-HSQC reference spectrum of apo-SmLPMO10A ( 15 N-HSQC 1). Line Width from Signal Intensity Calculations. The line widths of 1 H− 15 N signals in all the recorded 15 N-HSQC spectra of both apo-and holo-SmLPMO10A were measured by integration of the signal using the CARA software. 49 The relative change in line widths was then estimated by calculating the intensity ratios between 15  Changes in Mobility in Response to Oxidative Conditions. Heteronuclear { 1 H}-15 N NOEs, T 1 -and T 2relaxation time experiments were recorded for apo-SmLPMO10A and for holo-SmLPMO10A after a 96 h incubation time with 10 mM ascorbic acid. A 96 h incubation period was chosen to ensure that structural changes in SmLPMO10A happening in response to ascorbic acid were completed before recording the spectra. After 96 h, a 1 H NMR spectrum was recorded to verify that the initially added amount of ascorbic acid was exhausted. Additional ascorbic acid (5 mM) was added together with H 2 O 2 (2 mM) to induce extensive oxidative damage, and new NOEs, T 1 -and T 2relaxation time experiments, were acquired after 72 h of incubation. The { 1 H}-15 N heteronuclear NOEs as well as T 1 and T 2 relaxation data were derived from the spectra using Dynamics Center software version 2.3.1 (Bruker BioSpin).
Spectroscopy. CD spectroscopy was used to monitor structural changes for treatment series 1 ( Figure 2) using samples prepared correspondingly to NMR experiments. Aliquots were collected from samples of 100 μM apo-SmLPMO10A alone, 100 μM apo-SmLPMO10A after the addition of 10 mM ascorbic acid (at t = 4 min, 24 h, and 48 h), and holo-SmLPM10A after the addition of 10 mM ascorbic acid (at t = 4 min, 30 min, 4 h, 20 h, 24 h, 45 h, and 48 h). All samples were in acetate buffer (25 mM sodium acetate and 10 mM NaCl, pH 5.0). The collected aliquots were diluted to a protein concentration of 2.5 μM using acetate buffer, and far-UV CD spectra and UV absorption spectra were recorded using a Chirascan qCD spectropolarimeter (Applied Photophysics Ltd. Leatherhead, Great Britain). A 0.1 cm path length quartz cuvette (Sigma-Aldrich) was employed. The reported CD spectra are based on measuring the ellipticity at 185−260 nm and absorbance in the range of 185−340 nm (resolution of 2 nm, step size of 0.5 nm, and 1.3 s per time point) and represent the average of eight scans expressed as mean residue molar ellipticity [θ]MRE. All data was recorded at 23°C. The content of secondary structural elements was analyzed using the BeStSel (Beta Structure selection) server (http://bestsel. elte.hu/index.php). 50,51 Functional Studies of SmLPMO10A under Oxidative Conditions. Activity tests were performed to assess the catalytic capacity of holo-SmLPMO10A after being subjected to treatments like those in treatment series 1 and 3 shown in Figure 2. Two samples of ∼100 μM holo-SmLPMO10A in acetate buffer (25 mM sodium acetate, 10 mM NaCl, pH 5.0), with a volume of 500 μL, were prepared as described above. The samples were each transferred to 5 mm NMR tubes (LabScape Essence). To study the protective effect of chitin, 10 mg of mechanically milled β-chitin particles with a size of ∼0.5 mm (Mahtani Chitosan) were added to one of the two NMR tubes while no substrate was added to the other NMR tube. The samples were then pre-incubated for 24 h at room temperature, after which ascorbic acid (10 mM) was added to both samples. 10 μL aliquots were collected from the NMR tubes at t = 0 h (i.e., 30 s after the addition of ascorbic acid) and at t = 24 h and diluted with 990 μL of acetate buffer (25 mM sodium acetate and 10 mM NaCl, pH 5.0) containing 10 mg chitin from shrimp shells (Sigma-Aldrich, practical grade powder). Ascorbic acid was then added to the reactions (1 mM final concentration) followed by incubation for 24 h at 40°C in an Eppendorf thermomixer set to 800 RPM. The reactions were stopped by separating the enzyme and soluble products from the insoluble chitin substrate by filtration using 96-well filter plates (Merck). The concentration of oxidized LPMO products was also determined in aliquots taken from the preincubation reaction with β-chitin to allow accounting for carryover of products potentially formed in the pre-incubation phase.
To measure the maximum (i.e., 100%) enzyme activity of holo-SmLPMO10A, a reaction with 1 μM LPMO and 10 mg of shrimp chitin was carried out in 500 μL of 25 mM sodium acetate buffer, pH 5.0, at 40°C and 800 RPM using an Eppendorf thermomixer. The enzyme was pre-incubated with the substrate for 24 h at room temperature. The reaction was initiated by adding ascorbic acid to a 1 mM final concentration and quenched after 24 h by filtration as described above. A control reaction lacking enzyme was prepared by incubating 10 mg of chitin particles from shrimp shells in 500 μL of 25 mM sodium acetate buffer, pH 5.0, at 40°C and 800 RPM for 24 h using an Eppendorf thermomixer.
Quantification of Soluble Oxidized LPMO Products by HPAEC-PAD. Soluble oxidized products generated upon incubating chitin with the LPMO were analyzed by highperformance anion−exchange chromatography with pulsed amperometric detection (HPAEC-PAD) using a Dionex ICS5000 system (Thermo Scientific, San Jose, CA, USA) equipped with a CarboPac PA200 analytical column. Prior to analysis, the reaction samples were diluted two times with 50 mM sodium phosphate buffer, pH 6.0, and treated with 1 μM SmGH20 chitobiase (30°C, overnight) 45 to convert native and oxidized chito-oligosaccharide products to a mixture of Nacetylglucosamine (GlcNAc) and chitobionic acid (GlcNAcGlcNAC1A). A stepwise gradient with an increasing amount of eluent B (eluent B: 0.1 M NaOH and 1 M NaOAc; eluent A: 0.1 M NaOH) was applied to the column at 0.5 mL/ min flow rate according to the following program: 0−10% B over 5 min, 10−100% B over 4.5 min, 100−0% B over 6 s, and 0% B over 13.5 min. Chromeleon 7.0 software was used for data analysis. Standard solutions of chitobionic acid were prepared in-house as previously described. 45 Biochemistry pubs.acs.org/biochemistry Article ■ RESULTS

Exposure to Ascorbic Acid in the Absence of Chitin Affects Residues Near the Copper-Active
Site. The addition of Cu(II) to apo-SmLPMO10A ( 15 N-HSQC 1 and 2) was accompanied by rapid relaxation of 1 H− 15 N signals belonging to residues located within ∼10 Å from the copperactive site ( Figure S2). The increased relaxation is due to PRE 52 (paramagnetic relaxation enhancement) caused by Cu(II) binding to the histidine brace formed by H28 and H114 and is in agreement with previous results obtained by Aachmann et al. 9 To follow the effect of ascorbic acid on the structure of holo-SmLPMO10A (treatment series 1), a total of three 15 N-HSQC spectra were recorded at time points t = 0 h (i.e., just after addition), 20, and 40 h. The immediate (t = 0 h) effect of ascorbic acid ( Figure S3, 15 N-HSQC 3) was the loss of PRE on 1 H− 15 N signals belonging to residues ∼10 Å from the active site, indicating that the copper was reduced. As PRE influenced the NMR spectra of Cu(II) holo-SmLPMO10A, all the 15 N-HSQC spectra of holo-SmLPMO10A recorded after the addition of ascorbic acid were compared with the reference spectrum of apo-SmLPMO10A ( 15 N-HSQC 1), which is not affected by PRE.
Compared with apo-SmLPMO10A, copper binding and addition of ascorbic acid (i.e., binding of reduced copper) produced changes in both the chemical shifts and line widths for several of the 1 H− 15 N signals in the HSQC spectrum ( Figure S3, 15 N-HSQC 3). CSPs indicate changes in the chemical environment of the nuclei giving rise to the signal, 53 whereas changes in line width (seen by signal intensity) reflect changes in residue dynamics. 54 Relative to the apo enzyme, the reduced, copper-containing holo enzyme showed large CSPs (>50 Hz) for residues near the active site ( Figure S4) immediately (t = 0 h) after the addition of ascorbic acid ( Figure 3A,C). Among these were residues known to be involved in substrate binding and residues in or near the copper site, such as T111, H114, D182, and F187. 3,9 The binding of reduced copper and surplus concentrations of ascorbic acid also affected larger parts of the protein, including residues such as A63, A72, G150, and A71 whose signals were The arrows indicate the direction of the change in chemical shifts upon copper binding and reduction. Note that there are no (red) signals for residues within ∼10 Å due to the PRE caused by Cu(II). (B) Change in the line width, seen as an increase in signal intensity (found by peak integration) for H114 upon addition of ascorbic acid (at t = 0 h) compared to apo-SmLPM10A. (C) CSPs in response to the addition of ascorbic acid (a t = 0 h) and copper reduction compared to apo-SmLPMO10A highlighted on the X-ray crystal structure of SmLPMO10A (PDB ID: 2BEM). The magnitude of the change in Hz is indicated by the color scheme. Non-detectable residues are shown in red. No CSPs >50 Hz were observed for the gray-colored areas of the structure. (D) Changes in line widths in response to the addition of ascorbic acid (at t = 0 h) and copper reduction compared with apo-SmLPMO10A highlighted on the X-ray crystal structure of SmLPMO10A. An increase in signal intensity is shown in orange, while a decrease is shown in blue. Non-detectable residues are shown in red. Biochemistry pubs.acs.org/biochemistry Article not affected by PRE ( Figure 3A), in accordance with previous studies. 9 The addition of ascorbic acid also led to narrower line widths (relative to the apo-enzyme), seen by increased signal Figure 4. Residual activity of SmLPMO10A after various periods of pre-incubation with 10 mM ascorbic acid in the absence or presence of ≈2% (w/w) β-chitin. The pre-incubation reaction contained 100 μM LPMO in 25 mM sodium acetate buffer, pH 5.0, with 10 mM NaCl, and the total volume was 500 μL. Residual LPMO activity was determined by taking 10 μL aliquots from pre-incubated samples (filtrated first in the case of preincubation with β-chitin) and setting up 24 h reactions with ≈1% (w/w) β-chitin in 25 mM sodium acetate buffer, pH 5.0, supplied with 10 mM NaCl and 1 mM ascorbic acid with a total volume of 1 mL at 40°C. The bars represent relative amounts of soluble oxidized products released by the LPMO in these reactions, compared to a reaction with a control enzyme sample that was not pre-incubated. Soluble oxidized products were subjected to treatment with chitobiase to produce chitobionic acid, which was quantified. Carryover of LPMO products generated during the preincubation phase of the experiment was detectible after 24 h of pre-incubation with β-chitin and amounted to ≈12% percent of the total observed signal. This background signal was subtracted from the product amount used to produce this figure. Error bars indicate standard deviations between triplicates. Typical chromatograms showing soluble LPMO products generated in these experiments are shown in Figure S5.   Figure S4). These increased intensities were accompanied by a relative decrease in signal intensity for the remaining parts of the structure, as shown in Figure 3D. Line width analysis can provide information about protein dynamics. 47,54 The increased 1 H− 15 N signal intensities near the copper-active site are intriguing as they could indicate slow exchange between different conformations of holo-SmLPMO10A 47 or, interestingly, increased local mobility, following the addition of ascorbic acid. Taken together, the CSP and the changes in signal intensity show that exposure to ascorbic acid has a major effect on the copper-binding region of the LPMO. These changes exceed the structural effect of copper binding, which is not expected to increase structural flexibility based on previous results. 9 The most likely explanation for these changes is oxidative damage of the catalytic center. Indeed, LPMO activity measurements showed reduced activity at t = 0 (measured at approximately 30 s after adding ascorbic acid), whereas LPMO activity was completely abolished at later time points, demonstrating that the LPMO is rapidly damaged when exposed to a high concentration of reductant in the absence of the substrate (Figures 4 and S5).
Longer incubation times (t > 0 h) with ascorbic acid did not result in additional CSPs compared to those observed at t = 0 h. The number of amino acid residues with narrow line widths relative to apo-SmLPMO10A first grew ( Figure S4) at t = 20 h after adding ascorbic acid ( Figure S6, 15 N-HSQC 4), potentially meaning that a larger part of the structure displayed local increases in mobility. Further evidence of structural  (Figure 5B,C), which is indicative of structural changes in the intermediate exchange regime 54 and may reflect a decline in protein's structural stability. 55 As a result, the number of residues with narrow line widths/high signal intensities decreased and was lower than the intensities found at t = 0 h sample ( Figure 5A). It is conceivable that extended exposure to ascorbic acid in the presence of free copper ions released from damaged LPMOs leads to abiotic formation of reactive oxygen species that can damage proteins. 56 Subsequent incubation with fresh ascorbic acid and H 2 O 2 resulted in the signals of several residues disappearing from the 15 N-HSQC spectra (Figures S8−S10, 15 N-HSQC 6−8). This indicates that the protein gets damaged and is unfolded/ denatured when exposed to a combination of free copper (leaking from damaged LPMOs), ascorbic acid, and H 2 O 2 . Immediately after adding H 2 O 2 ( Figure S8, 15 N-HSQC 6), signals of residues primarily located near the copper-active site became undetectable, while new signals appeared in the region between 8 and 9 ppm in the spectrum (Figure 6), indicating the partial loss of SmLPMO10A's structural integrity. 57 More pronounced changes, no longer limited to the copper-binding region, were observed in the 15 N-HSQC spectrum recorded at t′ = 44 h after the addition of H 2 O 2 (Figure S10, 15 N-HSQC 8). First, around 25% of the signals for the protein could no longer be observed in the spectrum. Second, the line width of many of the remaining detectable signals became broader ( Figure S11), and there was an overall decrease in signal dispersion with signals accumulating in the range between 8 and 9 ppm in the spectrum (Figures 6 and S12). Third, narrow line widths and CSPs >50 Hz were observed for still detectable residues near the active site ( Figure S4). These spectral changes, taken together, strongly suggest a loss of 3D structure and protein unfolding.

CD Spectroscopy Reveals Conformational Rearrangement of Aromatic Residues near the Active Site in
Response to Ascorbic Acid. CD spectroscopy in the far UV range (185−260 nm) was used to investigate changes in the overall structure of holo-SmLPMO10A in response to incubation with 10 mM ascorbic acid in the absence of chitin (the same conditions as in treatment series 1), while UV absorbance at 250 nm was used to monitor ascorbic acid consumption. 58 The CD spectrum of apo-SmLPMO10A was in agreement with previous findings 55 (Figure S13), showing a negative maximum at 219 nm and a zero-ellipticity crossover at 214 nm indicative of a structure with a high degree of β-strands. 60 The spectrum also featured a negative band between 185 and 187 nm indicating some random coils 60 and a positive maximum at 232 nm. The positive maximum signal at 232 nm is considered to arise from π−π* excitation coupling between the side chains of tryptophan, phenylalanine, and tyrosine residues that are less than 1 nm apart 61 and has been shown to disappear when SmLPMO10A is unfolded. 59 Indeed, SmLPMO10A contains a cluster of aromatic residues just beneath the copper site ( Figure S13). The spectrum for the holo-enzyme showed only minor changes relative to the apo-enzyme, which can be due to the expected rigidification at the copper site ( Figure S14). 9  The addition of 10 mM ascorbic acid resulted in increased intensity of the positive maximum at 233 nm and appearance of a second positive band at 240−255 nm for both holo-SmLPMO10A ( Figure 7) and catalytically inactive apo-SmLPMO10A ( Figure S15). Control spectra of ascorbic acid alone (Figures 7 and S15) showed that these spectral changes observed immediately after adding ascorbic acid are to a large extent due to absorbance of ascorbic acid.
Interestingly, only in the case of holo-SmLPMO10A, the positive maximum at ∼233 nm remained increased (Figure 7), even after all ascorbic acid had been consumed ( Figure S16). In contrast, the spectrum of apo-SmLPMO10A at t = 48 h after the addition of ascorbic acid was very similar to the spectrum recorded after 4 min ( Figure S15). These findings could indicate that the addition of ascorbic acid leads to changes in the structure of holo-SmLPMO10A that change the orientation of aromatic side chains with respect to each other. In addition, the negative minimum showed a blue shift from 219 to 218 nm at t = 48 h after the addition of ascorbic acid.
Backbone Dynamics in the Fast Exchange Regime. After adding ascorbic acid, no consistent change in the mobility of holo-SmLPMO10A on the pico-and nanosecond timescales was detected across the recorded heteronuclear { 1 H}− 15 N NOEs, T 1 -and T 2 -relaxation time experiments, as  Figure S17. Clear changes in the structural flexibility of holo-SmLPMO10A would be observed in the same area of the structure in all the three experiments. As a result, no definitive conclusion about mobility at the pico-and nanosecond timescales could be reached. Chitin Protects SmLPMO10A against Oxidative Damage. The impact of the substrate was investigated by pre-incubating holo-SmLPMO10A with β-chitin fibers for 1 h before adding ascorbic acid and, subsequently, fresh ascorbic acid and H 2 O 2 (treatment series 2). As in treatment series 1, 15 N-HSQC spectra were recorded at varying time points (see Figure 2 for an overview). In these experiments, the binding of the LPMO to chitin is expected to result in reduced signal intensities, as the substrate-bound LPMO is undetectable by solution-state NMR. Consequently, the observations described below concern the non-substrate-bound fraction of the LPMO.
Changes indicative of large structural changes and increased structural flexibility were observed in the recorded 15 N-HSQC spectrum immediately (t = 0 h) after adding ascorbic acid in the presence of chitin ( Figure S18, 15 N-HSQC 10). Compared to apo-SmLPMO10A, many residues showed narrower line widths, observed as an increase in the intensity of their 1 H− 15 N signals ( Figure 8B). The affected residues were similar to those shown in Figure 5A (i.e., effects of ascorbic acid in the absence of chitin), but the total number of affected residues was slightly higher (Figures S19 and S20). The affected Structural changes were also reflected by the CSP (>50 Hz) observed for residues near the copper-active site and residues G29, K63, and A181 becoming non-detectable ( Figure S19). As a whole, these combined changes seem to point at a high degree of structural changes in the ascorbic acid exposed, copper-containing enzyme, as also seen in treatment series 1.
In contrast to what was observed in treatment series 1 ( Figure 5A), only two residues (Y30 and H114) kept displaying narrower line widths compared with the reference spectrum of apo-SmLPMO10A ( Figure S21) at t = 20 h after the addition of ascorbic acid ( Figure S22, 15 N-HSQC 11). The remaining residues showed a sharp decrease in signal intensity or became undetectable ( Figures S19 and S21). The decrease in signal intensities can in part be explained by line broadening ( Figure S23) resulting from loss of structural integrity, as could be expected based on the results of treatment series 1. However, comparing the average signal intensities of the recorded spectra (which differ from the local variations in signal intensities discussed above), the overall reduction is larger than what was seen in treatment series 1 (Figure 9). The most likely explanation for these findings is that ascorbic acid, that is, reduction of the LPMO combined with the availability of in situ generated H 2 O 2 , 32 promotes binding of the LPMO to chitin, making the enzyme invisible by NMR.
Underpinning the very different situation in the presence of the substrate, activity measurements showed that, in the presence of the substrate, the LPMO was still active after 20 h of incubation with ascorbic acid and chitin (Figures 4 and S5). In contrast, in treatment series 1, the addition of ascorbic acid led to almost immediate enzyme inactivation. Of note, the activity assay does not allow a full quantitative description of a gradual decrease in catalytically competent LPMOs since catalysis in the activity assay is to a large extent limited by access to in situ generated H 2 O 2 and not only related to the amount of catalytically competent enzyme. It is thus possible to reconcile the observations that, on the one hand, a fraction of the LPMOs is being damaged, as suggested by the changes observed immediately after adding ascorbic acid, while, on the other hand, a fraction of the LPMOs remains active.
It is important to note that the apparent inactivation of SmLPMO10A after only 30 s of pre-incubation with the reductant and β-chitin ( Figure 4) is at least in part due to the fact that, prior to setting up the assay for the determination of residual activity, the β-chitin particles used in the preincubation were removed via filtration. Thus, chitin-bound SmLPMO10A was removed, leading to underestimation of the residual activity. The key point illustrated by Figure 4 is that after 24 h of pre-incubation with β-chitin, the enzyme is still active, in contrast to the enzyme pre-incubated in the absence of β-chitin.
The addition of fresh ascorbic acid together with H 2 O 2 was expected to boost the ongoing chitin degradation reaction but also to considerable enzyme inactivation. Previous work has shown that high concentrations of ascorbic acid and H 2 O 2 make the LPMO reaction very fast 31 while increasing the risk of oxidative damage. 34 Initially, the addition of ascorbic acid and H 2 O 2 ( Figure S24, 15 N-HSQC 12) led to a further reduction of the average signal intensity ( Figure 9B) and left nine residues undetectable ( Figure S19). This is likely due to a combination of increased substrate binding and loss of signal due to structural deterioration of soluble enzyme. Upon addition of ascorbic acid and H 2 O 2 , the number of residues with large CSPs (>50 Hz) also grew ( Figure S19), indicating that, indeed, structural changes were taking place.
Interestingly, longer incubation times (t′ = 20 h) with ascorbic acid and H 2 O 2 ( Figure S25, 15 N-HSQC 13) restored the average signal intensity ( Figure 9B). Furthermore, residues that were undetectable in the previously recorded spectrum ( Figure S24, 15 N-HSQC 12) reappeared, and the overall   Figure S19), again pointing at the damage near the catalytic copper. In addition, there were multiple CSPs >50 Hz, particularly in the environment of the copper site ( Figure S19). Nevertheless, the protective effect of the substrate against oxidative self-inactivation is evident when comparing the sample exposed to ascorbic acid and H 2 O 2 in treatment series 1 and 2 ( 15 N-HSQCs 7 and 13). In treatment series 2, no evidence of structural unfolding could be observed, with 1 H− 15 N signals in the spectrum remaining dispersed and narrow ( Figure 10) and with residues staying detectable. Figure 11 illustrates the large difference in the number of residues ending up as being undetectable between treatment series 1 and 2.
To verify the results of treatment series 2, treatment series 3 was performed with minor differences; the pre-incubation period with chitin prior to the addition of ascorbic acid was extended to 24 h. 15 N-HSQC spectra were recorded over a longer period, compared to treatment series 2 (0−78 h with ascorbic acid alone, followed by 0−48 h after the addition of fresh reductant and H 2 O 2 : Figure S26−S33, 15 N-HSQCs 14− 21). When monitoring the effect of adding ascorbic acid alone over this long time, indications of protein degradation became increasingly evident ( Figure S34). Furthermore, the average signal intensity decreased with time, and by t = 78 h, the signals of many residues were undetectable ( Figure S34). The addition of fresh ascorbic acid and H 2 O 2 (t′ = 0 h) to treatment series 3 resulted in signals corresponding to the active form of holo-SmLPMO10A to seemingly reappear ( Figure S31, 15 N-HSQCs 19), as evidenced by missing residues reappearing and restoration of the signal dispersion in the spectrum ( Figure S35) and an overall increase in the average signal intensities ( Figure S36), although it was less pronounced than in treatment series 2. The latter is compatible with the prolonged preceding incubation with ascorbic acid only, which would lead to a larger degree of enzyme damage, as was indeed observed ( Figure S36). Longer incubation (t′ > 0 h) resulted in a drastic decrease in the average signal intensity due to the enzyme suffering further oxidative damage. ■ DISCUSSION LPMOs are prone to oxidative self-inactivation caused by offpathway reactions involving the reduced LPMO and oxygen or H 2 O 2 . 15,37 Reactions with the oxygen co-substrate generate reactive oxygen species, including hydroxyl radicals, that may emerge upon homolytic cleavage of H 2 O 2 . 15,28,62 As the substrate helps confining the reactive oxygen species, offpathway reactions are exacerbated in the absence of the substrate. 15,31,35,63 The chemical details of the destructive redox chemistry that happens in LPMO-active sites remain partly unknown. It has been shown that the copper-binding histidine's and nearby aromatic residues are particularly vulnerable. 15 In experiments with ascorbic acid only (as in the present study), H 2 O 2 is generated through the reductant oxidase activity of the LPMO 38 and through abiotic oxidation of the reductant. 16 In reactions with bacterial family AA10 LPMOs, such as SmLPMO10A, such in situ production of H 2 O 2 will be rather slow, 64 meaning that there is a large difference between the incubations with and without H 2 O 2 added. Biochemistry pubs.acs.org/biochemistry Article NMR enables real-time observation of changes in the overall structure of proteins and of changes in individual residues, while at the same time providing dynamic information. NMR thus provides the possibility to gain insights into the structural changes that accompany oxidative self-inactivation of LPMOs. In the present study, we have used 15 N-HSQC "fingerprint" spectra as reporters of structural changes during treatment series (summarized in Figure 2) that were designed to induce oxidative damage. Importantly, the setup of this study was such that we observed both autocatalytic damages, as one would expect during the initial phase of incubating with ascorbic acid, and more general damage by reactive oxygen species, which happens when the LPMO is exposed to high concentrations of hydrogen peroxide in the presence of transition metals. Changes in the structure and dynamics of SmLPMO10A were monitored by CSPs and 1 H− 15 N signal line widths, respectively, while overall structural integrity was monitored by the overall chemical shift dispersion. It is important to note that the 15 N-HSQC spectra only reveal changes that affect the backbone H N and N atoms of SmLPMO10A.
In agreement with previous findings, 9 1 H− 15 N signals belonging to residues near (∼10 Å) the copper-active site became undetectable due to PRE caused by binding of Cu(II) to apo-SmLPMO10A. The addition of ascorbic acid resulted in the reappearance of signals near the copper site due to reduction of the copper to diamagnetic Cu(I). Importantly, oxidative damage to the copper-coordinating histidines, H28 and H114, likely leads to release of the copper atom from the active site. 56 Thus, in all treatment series, the continuous persistence of signals sensitive to PRE may mean that the copper stays reduced but also that the copper dissociates from the enzyme.
In this study, large amounts of ascorbic acid were added, which, in addition to reducing the LPMO and relieving PRE, would result in in situ generation of H 2 O 2 , causing autocatalytic inactivation of the reduced enzyme. Data from treatment series 1 showed that the addition of ascorbic acid to Cu(II) holo-SmLPMO10A caused residues near the copperactive site to display CSPs >50 Hz and narrower line widths, when comparing with apo-SmLPMO10A ( Figure 3). Both effects indicate changes in the backbone structure of SmLPMO10A. Minor conformational rearrangements and reduced conformational flexibility brought about by copper binding have been reported in the literature. 9,65 However, while copper binding is expected to result in broadening due to reduced conformational flexibility, the presence of ascorbic acid led to the opposite, namely, narrowing of line widths for residues near the copper center. This narrowing could result from a slow exchange regime between multiple structural states of holo-SmLPMO10A or result from increased local flexibility near the active site. This indicates structural changes in the catalytic center that are likely due to oxidative damage caused by an autocatalytic process, aligning well with an observed rapid decrease in catalytic activity. Longer incubations with ascorbic acid led to general line broadening, suggesting that oxidative damage propagated through the enzyme and that overall structural integrity was being lost. In accordance with these structural signs of enzyme damage, activity measurements showed that the enzyme was completely inactive after 24 h of incubation with ascorbic acid.
Conformational rearrangements in response to ascorbic acid were also observed by CD spectroscopy. The CD spectrum of both apo-and holo-SmLPMO10A showed an unusual positive maximum at 233 nm, which can be explained by π−π* excitation coupling between aromatic side chains, 61 which causes a strong negative peak at 213 nm and a positive peak around 230 nm. 66−68 The increased intensity of the positive maximum at 233 nm could indicate stronger contributions from π−π* excitation couplings between aromatic side chains following incubation with ascorbic acid. 67,68 SmLPMO10A contains at least three pairs of aromatic residues able to exhibit π−π* excitation coupling: Y30−Y39, W108−W119, and W178−F187. In addition, excitation coupling between W119 and W178 or F187 is possible ( Figure S12). Based on the NMR data, conformational changes in these aromatic pairs are most likely to occur for W119−F178 since both involved residues showed CSPs >50 Hz in response to the addition of ascorbic acid (Figure 3). The pair is located directly below the copper-active site ( Figure S13). Interestingly, it has been speculated that conserved tyrosine and tryptophan residues near the active site are protecting LPMOs from inactivation through a "hole hopping" pathway, 30,43,69 and the changes observed here through CD measurements and NMR could relate to such pathways. In this respect, it is worth noting that CSPs >50 Hz were also observed for residues Y30 and W108, meaning that conformational rearrangements of these residues may also contribute to increased π−π* excitation couplings. The residue Y39 was not assigned, meaning that no structural information about this residue was obtained, while W119 did not show a significant CSP.
Of note, it cannot be excluded that absorbance from ascorbic acid oxidation products also influences the peak maximum at ∼233 nm. Ascorbic acid is initially oxidized to dehydroascorbic acid, 70,71 which is rapidly degraded into a plethora of oxidation products, 23,71 all of which could potentially influence the CD spectra to a variable degree.
After the incubation with ascorbic acid and enzyme inactivation, subsequent addition of fresh ascorbic acid, H 2 O 2 , and possible free copper in solution could lead to a highly damaging environment allowing, for example, Fentonlike chemistry catalyzed by reduced free copper. Since free copper promotes oxidation of ascorbic acid, hydrogen peroxide levels will be relatively high. This situation likely leads to massive enzyme damage, which is no longer autocatalytic. Indeed, the 15 N-HSQC spectra (Figures S8−S10, 15 N-HSQC 6−8) showed features commonly associated with protein unfolding, 55 which became increasingly pronounced as time progressed (Figure 6). Looking at which residues became undetectable, it would seem that β-strands 1 and 5 were particularly prone to damage and (partial) unfolding. Structural elements that seemed to remain intact include parts of the β-sandwich core and the distal parts of the L2-loop (∼13 Å from the active site), as shown in Figure 6.
In treatment series 2, the addition of ascorbic acid after preincubation with chitin ( Figure 2, 15 N-HSQC 10) was again immediately (t = 0 h) accompanied by CSPs and indications of increased structural flexibility. Longer incubation with ascorbic acid led to a notable reduction in the average signal intensity, which was mainly due to binding of the LPMO to the substrate and not to structural disintegration since signal intensities were restored later on in the treatment series, upon addition of fresh ascorbic acid and H 2 O 2 ( Figure 9). Adding β-chitin particles to the samples creates a biphasic system where the substratebound SmLPMO10A becomes invisible. 9 Together with activity data, the NMR data show that reduction and the presence of in situ generated H 2 O 2 promote substrate binding Biochemistry pubs.acs.org/biochemistry Article and that such binding protects the LPMO from damage and inactivation. The signs of protein degradation, a process that eventually will also contribute to a decrease in signal intensities, observed in response to ascorbic acid originate from unbound SmLPMO10A, which suffers from oxidative damage. The results of treatment series 3 led to a similar conclusion; in this case, damage during the ascorbic acid-only phase was more extensive due to the much longer incubation time.
The addition of fresh ascorbic acid and H 2 O 2 in treatment series 2 and 3 (i.e., with the presence of chitin) led to an increase in the signal intensities of nearly all observable residues. Signals that were unobservable before the addition of H 2 O 2 also became visible again. Moreover, the overall appearance of the spectra in terms of signal dispersion became more similar to the spectra collected before the initial addition of ascorbic acid. The reappearance of signals for the intact, reduced LPMO is compatible with the notion that the addition of H 2 O 2 drastically speeds up the rate of chitin cleavage, which could lead to release of catalytically competent LPMO into the solution, which is visible to NMR again.
Based on these results, we propose the chain of events outlined in Figure 12. When the copper-active site of SmLPMO10A is reduced, the substrate affinity of the enzyme increases, and the equilibrium shifts toward more protein being substrate-bound, as previously demonstrated by Kracher and colleagues. 32,39,72 The addition of externally supplied H 2 O 2 increases the catalytic activity, resulting in substrate oxidation and the release of intact LPMO back into the solution. This chain of events implies that the presence of chitin, that is, a good substrate, protects the LPMO against self-oxidative damage, as was confirmed by activity measurements, showing that loss of enzyme activity was much reduced in the reactions with chitin. It has been shown that binding of the LPMO to the substrate increases the enzyme's reactivity toward H 2 O 2 . 28,37 Thus, the presence of the substrate has multiple effects. On the one hand, reduced LPMOs are removed from the solution, which limits the risk of futile turnover of H 2 O 2 , which could lead to enzyme damage. On the other hand, the increased consumption of H 2 O 2 in the enzyme substrate complex removes available H 2 O 2 from the solution, further reducing the chance of futile, potentially damaging turnover by non-substrate-bound LPMOs.

■ CONCLUSIONS
Overall, by using NMR spectroscopy, CD spectroscopy, and activity assays, our results shed light on the process of oxidative self-inactivation of SmLPMO10A over time. Whereas CSPs place the process of oxidative damage in a structural context, by providing direct insights into the chemical environment of the observed nuclei, changes in signal intensities indicate structural flexibility and/or that the native tertiary structure of SmLPMO10A is partially lost. Using these tools, we show how oxidative damage first happens near the copper site and then propagates through the protein, and we show that chitin protects SmLPMO10A from oxidative self-inactivation. Our CD studies suggest that aromatic residues in the core of SmLPMO10A play a role in determining the fate of redoxactive species generated at the copper site, and further studies of the potential role of these residues in protecting the enzyme from autocatalytic inactivation would be of interest.
Experimental design overview, NMR spectra, structural changes indicated in the primary sequence, plots of change in 1 H− 15 N signal intensity, signal dispersion, line widths, aromatic pairs, CD and absorbance spectra, protein dynamics, structural changes, and chromatograms (PDF)