Modularity impacts cellulose surface oxidation by a lytic polysaccharide monooxygenase from Streptomyces coelicolor

Two variants of LPMO10C from Streptomyces coelicolor (UniProt ID: Q9RJY2), full-length ScLPMO10C (AA10-linker-CBM2, residues 35-364) and its catalytic domain only, ScAA10 (AA10, residues 35-230), were recombinantly expressed in Escherichia coli. Both variants were purified by a two-step protocol, starting with anion exchange chromatography followed by size exclusion chromatography, as previously described (Forsberg et al. 2014). Prior to use, the two LPMO variants were incubated with a three-fold molar surplus of copper sulphate in room temperature at pH 6.0 for 30 min. Subsequently, excess copper was removed by gel filtration using PD MidiTrap G-25 desalting columns (GE Healthcare, Chicago, USA) and the buffer was simultaneously exchanged to 50 mM sodium phosphate (pH 6.0).

The GH6 endoglucanase from Thermobifida fusca (TfCel6A, residue 32-441; UniProt ID: Q47R05 (Irwin et al. 1993) was cloned into the pNIC-CH (Addgene) expression vector by ligation-independent cloning. This cloning procedure adds a poly-histidine tag (His₆ tag) to the C-terminus of the protein. After verification of the sequence by Sanger sequencing, the expression vector was transformed into chemically competent E. coli One Shot® BL21 StarTM (DE3) cells (Invitrogen). Cells harboring the plasmid were pre-cultured for 8 h in 5 mL of LB medium (Lysogenic Broth) containing kanamycin (50 µg/mL) and subsequently used to inoculate 500 mL of TB medium (Terrific Broth) containing kanamycin that was then incubated for 16 h at 20 °C in an LEX-24 Bioreactor (Harbinger Biotechnology, Canada) using compressed air for aeration and mixing. Expression was induced by adding isopropyl β-D-thiogalactopyranoside (IPTG) to a final concentration of 0.1 mM at an absorbance at 600 nm (A₆₀₀) of 0.8, followed by incubation for 24 h at 20 °C. Cells were harvested by centrifugation (5500×g, 10 min) and suspended in lysis buffer (50 mM Tris/HCl, pH 8.0, 500 mM NaCl, and 5 mM imidazole). Cells were disrupted by two 2-min cycles of pulsed sonication on ice (3 s on and 2 s off) and cell debris was removed by centrifugation (75,000×g, 30 min). The supernatant was sterilized by filtration using a 0.22-µm syringe filter and directly loaded onto a 5-ml HisTrap HP nickel-Sepharose column (GE Healthcare, Chicago, USA) equilibrated with 50 mM Tris/HCl, pH 8.0, 500 mM NaCl, 5 mM imidazole (Buffer A). Protein was eluted by applying a 25-column volume linear gradient to 100% Buffer B (50 mM Tris/HCl, pH 8.0, 500 mM NaCl and 500 mM imidazole) at a flow rate of 2.5 ml/min. Protein-containing fractions were analyzed by SDS-PAGE and subsequently concentrated, with concomitant buffer exchange to 20 mM Tris/HCl, pH 8.0, using an Amicon® ultracentrifugal filter (Millipore, Burlington, MA, USA) with a 10-kDa cutoff.

Protein concentrations were measured using A₂₈₀ and the protein’s calculated molar extinction coefficients (ε^ScLPMO10C = 75,775 M⁻¹ cm⁻¹, ε^ScAA10 = and 52,160 M⁻¹ cm⁻¹ and ε^TfCel6A = 81,275 M⁻¹ cm⁻¹).

Activity was assessed on bacterial microcrystalline cellulose (BMCC), a gift from the late Prof. David Wilson at Cornell University, filter paper (Whatman no.1), Avicel® PH-101 purchased from Sigma Aldrich (St. Louis, USA), and phosphoric acid swollen cellulose (PASC), which was prepared from Avicel essentially as described by Wood (Wood 1988). Suphanilic acid treated Avicel (SA- Avicel) was produced by treating Avicel with sulphanilic acid to block pre-existing carboxylic acid groups. Avicel was suspended in dimethylformamide (DMF) (Bioshop Ontario, Canada) at a concentration of 50 g/L before adding the coupling reagent (12.5 mM benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate (PyBOP) (Sigma Aldrich Co. Missouri, USA) and 200 mM N,N-di-isopropylethylamine (DIPEA) (Sigma Aldrich Co. Missouri, USA). After gentle mixing for 30 min at room temperature, 100 mM sulphanilic acid was added and the mixing continued for an additional 60 min. The reaction mixture was then centrifuged and the solid fraction was washed several times with DMF to remove unbound sulphanilic acid from the labeled-Avicel (SA-Avicel). The SA-Avicel product was suspended in MilliQ water to a final concentration of 20 g/L.

Unless stated otherwise, reactions were performed with 0.5 µM LPMO (ScLPMO10C or ScAA10) in 50 mM sodium phosphate buffer (pH 6.0) in the presence of 1 mM reductant (ascorbic acid dissolved in trace select water, or gallic acid dissolved in 100% dimethyl sulfoxide (DMSO)) at 40 °C and 800 rpm in an Eppendorf Thermomixer (Eppendorf, Hamburg, Germany). All reactions were performed in triplicates.

For qualitative analysis of cellulolytic activity, ScLPMO10C and ScAA10 were incubated with 10 g/L filter paper or Avicel, or 1 g/L PASC or BMCC. After separation of the solid material by vacuum filtering through a 96-well filter plate (Millipore, Burlington, MA, USA), supernatants were analyzed by HPAEC-PAD (see below) for detection of oxidized cello-oligosaccharides.

To determine the concentration of solubilized oxidized products, purified TfCel6A was added to the supernatants to a final concentration of 1 µM, followed by static incubation overnight at 37 °C. To determine the total concentration of oxidized products (i.e., soluble and insoluble), the LPMOs were inactivated by boiling the complete reaction mixture for 15 min at 100 °C at the various time points. Subsequently, the reaction mixtures were diluted with 50 mM sodium phosphate buffer (pH 6.0) to a final concentration of 2 g/L LPMO-treated Avicel or 0.1 g/L of LPMO-treated BMCC after which TfCel6A was added (5 µM final concentration), followed by incubation for another 48 h at 50 °C with shaking at 800 rpm in an Eppendorf Thermomixer (Eppendorf, Hamburg, Germany). As a result of this procedure cellulose is almost completely solubilized, and oxidized products appear as dimers and trimers that are quantified by HPAEC-PAD (see below).

Native and oxidized products were analyzed with a high-performance anion-exchange chromatography method (Westereng et al. 2013) using a Dionex™ ICS-5000 system (Thermo Scientific, Sunnyvale, CA) set up with a disposable electrochemical gold electrode. Samples of 5 µL were injected on a CarboPac PA200 (3 × 250 mm) column operated with 0.1 M NaOH (eluent A) at a flow rate of 0.5 mL/min and a column temperature of 30 °C. Elution was achieved using a stepwise gradient with increasing amounts of eluent B (0.1 M NaOH + 1 M NaOAc), as follows: 0–5.5% B over 3 min; 5.5–15% B over 6 min; 15–100% B over 11 min; 100–0% B over 0.1 min; and 0% B (reconditioning) for 5.9 min. Chromatograms were recorded using Chromeleon 7.0 software.

C1-oxidised cellobiose and cellotriose standards were prepared in-house according to a previously published protocol (Zamocky et al. 2006; Bissaro et al. 2016).

BMCC (3 g/L) or SA-Avicel (10 g/L) was incubated with 1 µM enzyme (ScLPMO10C or ScAA10) and 1 mM gallic acid in 50 mM sodium phosphate buffer (pH 6.0) at 40 ˚C with 800 rpm orbital shaking for 24 h. After incubation, the supernatant was separated from the remaining insoluble cellulose fiber by centrifugation at 20,000×g for 10 min and the fiber was suspended in DMF. The cellulose fiber was fluorescently labeled through conventional amide bond formation, where the enzymatically introduced carboxylic acid group reacts with the amine group of the Rhodamine chloride dye (Fig. 1). Specifically, ≤ 0.9 mg of cellulose fiber in 1 mL of Dimethylformamide (DMF) containing 200 µM of N,N-Diisopropylethylamine (DIPEA) and 100 µM of benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate (PyBop) was gently mixed for 5 min at room temperature. Subsequently, 12.5 µL of Rhodamine 110 chloride dye (final concentration 100 µM; excitation max = 498 nm; emission max = 520 nm) was added to the reaction mixture, followed by gentle mixing for 5 min at room temperature. The reaction mixtures were then centrifuged at 20,000×g for 10 min, followed by repeated washing of the solid fraction with DMF to remove unbound dye from the labeled-cellulose fiber. The washing liquids were fluorescently monitored until all unbound dye was removed. The labeled-cellulose fiber (~ 1–10 g/L) was finally suspended in 500 µL MilliQ water and visualized under a confocal microscope. Because different sample quantities were prepared for microscopic analysis, measures of total fluorescent intensity cannot be correlated to extent of LPMO activity.

The labeled cellulose fiber was visualized using a confocal microscope (Leica TCS SP5 Wetzlar, Germany) equipped with an argon laser used at 20% power. Images were collected using a 100 × oil-immersion lens and a fluorescein isothiocyanate (FITC) filter. Four random spots were imaged for each sample. The Leica Application Suite (LAS) was used for data processing and annotation of captured images.

To study if the impact of the CBM depends on the type of cellulose, the activity of ScLPMO10C and ScAA10 was tested towards four types of cellulosic materials with different properties such as crystallinity, degree of polymerization (DP), available surface area and particle size (Park et al. 2010). These reactions, and all other activity assays described in this study, were done using so called “monooxygenase conditions”, which means that the reaction is limited by the in-situ generation of H₂O₂ through a combination of abiotic and enzyme-catalyzed oxidation of the reductant. We used both ascorbic acid and gallic acid as reductants and we used enzyme preparations that were devoid of free copper, which would affect abiotic oxidation of the reductant in the case of ascorbic acid (Stepnov et al. 2021; Golten et al. 2023).

The wildtype enzyme, possessing a CBM, is more effective in oxidizing and solubilizing products compared to the CBM-deficient enzyme on all four substrates (Fig. 2). This is an expected result since the CBM-deficient enzyme will be rapidly inactivated under the conditions used, especially at low substrate concentrations (Courtade et al. 2018; Stepnov et al. 2022). Notably, both enzyme variants generated most oxidized products in reactions containing BMCC and Avicel, which are substrates with expected high crystallinity and low DP (~ 250 compared to a DP of > 1000 for filter paper) (Hall et al. 2010; de Oliveira et al. 2011). It would thus seem that this particular LPMO more readily engages in productive interactions with highly crystalline substrates. BMCC seems to be a particularly good substrate, since reactions with BMCC gave high product levels while this substrate was used at 1 g/L concentration, as opposed to 10 g/L for Avicel.

Comparison of the amount of oxidized products in the soluble fraction to the total amount of generated oxidized products showed that the fraction of solubilized products was lower for the substrates that were used at higher concentrations (Avicel & filter paper) (Fig. 2), which is in line with previous observations on the concentration dependency of the fraction of soluble oxidized products made with Avicel (Courtade et al. 2018). For all four substrates, the fraction of solubilized products was higher for the full-length enzyme, compared to the CBM-deficient enzyme. Courtade et al., have suggested that this is due, at least in part, to CBM-mediated anchoring of the LPMO on the substrate. Such anchoring would lead to more localized oxidation of the substrate, meaning that there is a higher chance of multiple chain cleavages in the same polysaccharide chain, which is what is needed to generate short, soluble products.

Intriguingly, the impact of the CBM on the fraction of solubilized products varied between the cellulose substrates (Fig. 2). In the reactions with Avicel and filter paper, the wildtype enzyme released approximately 57–58% of all oxidized products, while ScAA10 solubilized only 43% and 17%, respectively. For PASC and BMCC, these fractions were 87% and 100%, and 45% and 87%, respectively. BMCC stands out in that a high fraction of the products are soluble at this low substrate concentration. One possible explanation for these notable differences may be that the various substrates have different surface morphologies that affect how LPMOs and CBMs preferably bind. In this respect, it is worth noting that the Avicel and filter paper used in this study are particles with an average size of 50–100 µm, while BMCC and PASC are celluloses likely exhibiting a higher surface-to-volume ratio. The substrates may substantially differ in terms of the amount and distribution of accessible productive LPMO binding sites, and it has been observed before that the product profile generated by LPMOs depends on the type of cellulose (Sun et al. 2021). We can only speculate why BMCC is so special; perhaps it has a limited number of cellulose chains that are particularly prone to LPMO action; this would mean that the same chains are cleaved many times, which would give a high fraction of soluble products. In this respect, it is interesting to note an early study by Bothwell et al. (1997) showing that cellulases from T. fusca showed 9–30 times higher adsorption levels when incubated with BMCC than when incubated with Avicel (Bothwell et al. 1997). While this observation does not explain the substrate-dependent variations observed here, it does show that there is huge variation between various cellulose types in terms of how enzymes bind.

To elucidate the potential effect of substrate loading on LPMO efficacy, reactions with different concentrations of BMCC (0.1–3.0 g/L) were set up with both enzyme variants, using copper-insensitive gallic acid as reductant. ScLPMO10C performs well at low substrate concentrations (0.1–1 g/L) (Fig. 3A). This good performance results from the combination of steady slow H₂O₂ production, which is primarily due to abiotic oxidation of the reductant, gallic acid, with efficient substrate binding enabled by the CBM (Courtade et al. 2018; Stepnov et al. 2022). Unexpectedly, at higher substrate loads (2–3 g/L) the full-length enzyme became less efficient. While there is no obvious explanation for this observation, it is conceivable that the substrate has a limited number of high-affinity non-productive binding sites next to a high, perhaps close to saturating, number of productive binding sites; in such a situation, the relative importance of non-productive binding will increase at increasing substrate concentration. Another explanation, which may be related to non-productive binding, entails that H₂O₂ availability decreases at higher substrate concentrations because substrate-binding reduces the oxidase activity of the LPMO (Kittl et al. 2012), although it has been claimed that for enzymes such as ScLPMO10C, under these experimental conditions, H₂O₂ production primarily results from abiotic oxidation of the reductant (Stepnov et al. 2021).

Conversely, the CBM-deficient variant showed an increase in product formation with increased substrate concentration (Fig. 3B). This can be explained by the increased number of binding sites that comes with higher substrate concentration, which promotes binding of this weak-binding CBM-free enzyme variant, which again prevents off-pathway reactions and enzyme inactivation (Courtade et al. 2018; Stepnov et al. 2022). While the efficiency of ScLPMO10C showed a maximum at 1 g/L BMCC (Fig. 3A,C), 3 g/L (i.e., the highest used BMCC concentration) was not enough to observe saturation for ScAA10 (Fig. 3B, C). A similar finding was described by Courtade et al. (2018) using Avicel as substrate, albeit at higher substrate concentrations: while ScLPMO10C showed maximum efficiency at 5 g/L substrate concentration, ScAA10 performed better with increasing substrate concentration, all the way up to 40 g/L.

For labeling reactions, we used 3 g/L BMCC, as both enzymes generated a similar amount of oxidized sites at this substrate loading (Fig. 3C). We also assessed activity on 10 g/L Avicel, since this represents a commonly used substrate loading for LPMO reactions.

In addition to investigating the effect of LPMO modularity on the release of soluble products from cellulose fibers, the pattern of surface oxidation of the residual cellulose fiber was investigated using fluorescence microscopy. BMCC and Avicel were prioritized for this study given the preferred activity of ScLPMO10C on these cellulosic substrates. Notably, Avicel is prepared using bleached cellulose and so retains carboxylic acids in the cellulose substrate before LPMO treatment (Lewin 1997). To eliminate the corresponding background fluorescence from untreated Avicel, pre-existing carboxylic acids were blocked using sulphanilic acid, yielding SA-Avicel (Fig. S1). Importantly, ScLPMO10C activity on Avicel and SA-Avicel were equivalent (Fig. S2). The fluorescent labeling method applied herein builds from previous fluorescent labeling approaches to study LPMOs on insoluble substrates (Vuong et al. 2017; Wang et al. 2018) and quantify soluble oxidized products (Budischowsky et al. 2022). Whereas the method reported in (Vuong et al. 2017) uses a water-soluble fluorophore (ANDA) and coupling reagent (EDAC) to covalently label cellulose after LPMO treatment, a polar aprotic solvent (DMF) was used here along with DIPEA and PyBOP as the activating and coupling reagents, respectively, to permit labeling with the Rhodamine chloride fluorophore. This change to the more photostable fluorophore (Yatzeck et al. 2008) was necessary for image analysis by confocal microscopy. Whereas Wang et al. (2018) previously used DIPEA and PyBOP reagents to label LPMO-treated alpha-chitin with fluorescein-5-thiosemicarbazide (FTSC), the corresponding confocal microscopy images were not sufficiently resolved to evaluate oxidation pattern (Wang et al. 2018). Besides fluorescent labeling, earlier approaches to study LPMO action on cellulose fibers have applied atomic force microscopy to distinguish preferential activity at crystalline versus amorphous regions of the cellulose substrate (Eibinger et al. 2014, 2017; Song et al. 2018; Karnaouri et al. 2020; Uchiyama et al. 2022), and radioisotope labeling (Wang et al. 2018) and X-ray photoelectron spectroscopy (Karnaouri et al. 2020) to quantify enzymatically introduced oxidation sites. The current fluorescent labeling technique extends previously reported methods by permitting qualitative assessment of oxidized positions on cellulose fiber surfaces and how the surface oxidation pattern might be impacted by the LPMO domain structure.

Full-length ScLPMO10C and the CBM-truncated ScAA10 comprising only the catalytic domain differed in the pattern of oxidation left on residual cellulose fiber surfaces. Whereas treating BMCC and SA-Avicel with ScAA10 led to a dispersed pattern of surface oxidation, treatment of the same substrates with ScLPMO10C led to a punctate pattern of surface oxidation overlying the dispersed oxidation pattern (Fig. 4). The observed punctate pattern of oxidation left by ScLPMO10C agrees with the higher amount of short soluble products (2–4 DP) released from Avicel by the full-length over truncated enzyme, first reported by (Courtade et al. 2018). The punctate pattern of surface oxidation observed for ScLPMO10C also supports the formation of oxidation “hot spots” suggested by Courtade et al. based on classical soluble product analyses (Courtade et al. 2018). Unfortunately, however, the resolution of the microscopy method was insufficient to quantify the fluorescent signal from the hot spots versus underlying dispersed oxidation sites, which would reveal whether the presence of the CBM generally leads to local oxidations or else targets ScLPMO10C to minor structural features on the substrate surface.

Applying the fluorescent labeling method to a larger set of LPMOs with and without CBMs is needed to establish the broader potential of CBMs to control the pattern of cellulose surface oxidation. Considering the varying oxidative regioselectivities (C1, C4, C1/C4) of natural LPMOs, such broader studies would require development of specific labeling methods that allow visualization of C4-oxidized sites on the fiber. Such methods are challenging and not readily available, one reason being that C4-oxidized chain ends are chemically unstable at alkaline pH (Westereng et al. 2016). The subsequent key question concerns the possible biological and industrial impact of the oxidation “hot spots” visualized in Fig. 4. In an AFM study from 2018, Song et al. visualized LPMO movement for fungal TrAA9A, an enzyme with mixed C1/C4-oxidation pattern from Trichoderma reesei possessing a family 1 cellulose-binding domain (CBM1). The movement was shown to be along, across, and intriguingly also penetrating the cellulose fibers leading to release of fibrils with shorter diameters. Notably, the LPMO was observed to stay in the same position longer than it was moving on the cellulose and was barely visible for periods as it was penetrating the cellulose. Upon LPMO treatment, the width of the fibers decreased, and the cellulose surface transitioned from being relatively smooth to become rougher and with an increase in surface area which led to ~ 10% higher degree of conversion by a cellobiohydrolase (Song et al. 2018). Song et al. also found that longer incubation times led to the accumulation of multiple LPMO molecules on particular cellulose fibers, an observation that seems compatible with the observation of LPMO oxidation hot spots in the present study. Relatedly, Koskela et al. showed that single domain C1-oxidizing NcLPMO9F was more efficient in introducing oxidations on the cellulose surface compared to CBM1-containing C1-oxidizing NcLPMO9E, while the latter CBM-containing enzyme solubilized the fiber more efficiently than the CBM-free LPMO. Hence it was suggested that the absence of a CBM led to enhanced movement of the enzyme, allowing for more disperse oxidation over the cellulose surface (Koskela et al. 2019).

Accumulating evidence shows that the impact of LPMOs on enzymatic degradation of cellulose relates in part to the ability to disintegrate cellulose fibers, generating microfibrils that are accessible to cellulases (Song et al. 2018). The other major effect concerns the introduction of accession points for cellobiohydrolases in crystalline regions (Eibinger et al. 2014, 2017; Uchiyama et al. 2022). Considering the above, one may envisage that the presence or absence of CBMs in LPMOs affects the synergistic effects with cellulases. Besides enhancing cellulose saccharification, LPMOs may find applications in production of materials, including nanocellulose fibers (Hu et al. 2018; Koskela et al. 2019; Moreau et al. 2019; Valls et al. 2019; Karnaouri et al. 2020, 2022; Marjamaa et al. 2023), functionalized chitin fibers (Wang et al. 2018), and production of densified wood films (Koskela et al. 2023). In these applications, one aims at uniformly oxidizing fiber surfaces while minimizing the release of soluble products. In this context, LPMOs lacking a CBM might be preferred.

In conclusion, we show here that, at least for some LPMOs and certain cellulose substrates, carbohydrate-binding modules are a critical determinant of the pattern of oxidation on cellulose fibers. We envisage that the visualization method employed herein, with acceptable background levels, using a more precise coupling reagent, benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate (PyBop), and a fluoro-stable dye, Rhodamine chloride 110, is a valuable analytical tool that can be utilized for obtaining deeper insight into LPMO functionality. This includes better assessment of the potential of LPMOs to modify fiber surface properties for material applications.