Characterization and interpretation of the reaction mechanism of a novel GH115 α-Glucuronidase from Flavobacteriaceae

1 Characterization and interpretation of the reaction mechanism of a novel GH115 α-Glucuronidase from Flavobacteriaceae. Savvina Leontakianakou*, Anders Sundin†, Eva Nordberg Karlsson Division of Biotechnology, Department of Chemistry, Lund University, PO Box 124, Lund SE 22100, Sweden Institution †Centre of Analysis and Synthesis, Department of Chemistry Lund University, PO Box 124, Lund SE-22100, Sweden Institution *Corresponding author: [email protected] KEYWORDS: GH115, Glucuronidase, SN2 catalytic mechanism, CBM, recognition site ABSTRACT: Xylan is a major component in the complex network of the plant cell wall. Xylan consists of xylose units connected by β-(1,4) glycosidic bonds, which can be heavily substituted, increasing its degradation recalcitrancy. Such substituents include α-D-glucuronic acid (glucopyranosyluronic acid, GlcAp) or 4-O-methyl GlcAp (MeGlcAp) at the O-2 position. Removal of these substituents can be catalyzed by α-glucuronidases. Here we report a novel two-domain α-glucuronidase from GH family 115, termed FAgu115A, that acts on both polymeric and oligomeric xylan. A homology model was constructed, revealing the significance of residues Asp303 and Asp177 for catalysis. These residues are conserved in all characterized proteins belonging to this family. The catalytic activity of these residues was verified by site-specific mutagenesis corroborating the hypothesis that Asp303 serves as the proton donor, while Asp177 increases the nucleophilicity of the reactive water through a hydrogen bond. Additionally, a potential CBM located at the C terminus is of importance for dimerization. Its significance was confirmed by producing a truncated variant followed by determination of the molecular mass by analytical sizeexclusion chromatography. Kinetic and thermal inactivation analysis were performed resulting in a Km of 3.7 g/l on beechwood xylan and an activation energy (Eα) of 67 kJ/mol. Introduction Glucuronoxylan (GX) and (methyl)-glucuronoxylan (M)GX) constitute major components of hardwood hemicellulose, and are also found in the xylan of conifers [1]. Both are key types of xylans in various wood lignocellulosic feedstocks, including poplar [2]. The backbone is comprised of β-1,4 linked Dxylopyranose (Xylp) residues, which are frequently acetylated [3]. The main substituents are α-Dglucuronic acid (glucopyranosyluronic acid, GlcAp) or 4-O-methyl-GlcAp (MeGlcAp) at the O-2 position [4] [4b]. The motif and quantity of different substituents vary among species, with GlcAp and MeGlcAp being the primary branch type in hardwood xylans [5]. Complete conversion (debranching) of such branched xylans can be achieved through enzymatic modification, facilitating the development of xylan-derived https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 2 polymers with specific characteristics applicable in biofuels, cosmetics, and the feed and food industries [6]. Such advancements can lead to more efficient biorefinery processes. (M)GX debranching is typically achieved with α-glucuronidases (EC 3.2.1.131/139), which hydrolyze the glycosidic bond between the (Me)GlcAp and the xylose units [7]. These enzymes are present in both fungi and bacteria and are categorized under three glycoside hydrolases (GH) families: GH4, GH67, and GH115, according to the Carbohydrate-Active Enzyme database (CAZy, www.CAZy.org). GH4 and GH67 are only active on the reducing end, while GH115 is active on any GlcA branch on the xylan backbone [8].GH115 glucuronidases feature a β/α barrel structure and have previously been suggested to employ a single displacement inverting mechanism [9], although at this stage the catalytic residues have not been uniformly identified. As of now, the CAZy database contains over 3000 sequences related to these enzymes, of which only 8 have been characterized, with 5 having their crystal structure elucidated [10] Characterized GH115 glucuronidases have been reported to possess multiple domains, including potential carbohydrate-binding modules (CBMs) [11] [10]. CBMs are described as non-catalytic domains that fold separately and are connected via a linker to the catalytic domain [12]. CBMs are typically comprising 100-150 residues, are primarily composed of β-sheets and play a crucial functional role in associating the substrate with the catalytic site [13]. Similar to carbohydrate-active enzymes, CBMs are classified into different CAZy families, however, the potential CBMs associated with GH115 enzymes have not yet been assigned to any specific family. In this study, we selected a gene encoding a GH115 candidate deposited in Genbank, from the genome of the Flavobacteriaceae bacterium strain 3519-10. After cloning, the enzyme was produced in Escherichia coli, and a homology model was built. Point mutations were introduced, both in silico and in vitro, replacing the suggested catalytic residues, and the effect of the exchange, combined with substrate docking and theozyme modelling allowed us to identify the residues responsible for catalysis via an SN2 reaction mechanism. The importance of the potential CBM domain for the activity on polymeric and oligomeric substrates was evaluated by producing a truncated version consisting only of the catalytic domain (CD). Supporting Information Sequence analysis, and multiple sequence alignment: The gene encoding for WT FAgu115A (GenBank Accession No. ACU08180.1) was retrieved from the Genbank. The molecular mass, pI and the theoretical molar extinction coefficient were calculated using ProtParam1 : the pI was 6.41, the molecular mass was 95 kDa and the extinction coefficient was 177 050 M-1 cm-1. Sequences were aligned using ClustalW2. ESPript3 was used to generate the figure with the annotated areas3. Expression of the genes encoding FAgu115A and mutant variants: The genes encoding the FAgu115A, FAgu115A _D177A, FAgu115A _D303A, FAgu115A _D177/303A and FAgu115A _CD were chemically synthesized using native codons and cloned into the expression vector pET21b (+) with a His-tag inserted at the N-terminus. The resulting plasmids (pET21b+_ FAgu115A) were transformed into E. coli BL21(DE3) competent cells (Sigma) via heat-shock transformation. A 1% pre-inoculum of the respective recombinant strain, pre-cultivated in LB medium overnight at 37 °C with 100 μg/ml of ampicillin as a selection marker, was added to 100 ml LB medium. The cultures were incubated under shaking at 37 °C until an optical density OD600nm of 0.6-0.8 was reached. Then, they were induced with 1 mM isopropyl βD-1-thiogalactopyranoside (IPTG), and the incubation was continued for 4 hours at 26 °C. The cells were harvested, lysed by sonication, and centrifuged at 26,000 g to separate soluble proteins from the cell debris. FAgu115 and variants were purified by immobilized metal ion affinity chromatography using an ÄKTA start protein purification system (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) with a HisTrap™ High-Performance column (1 ml, Cytiva). The system was equilibrated, and unbound proteins were washed out using a binding buffer consisting of 50 mM HEPES, 500 mM NaCl, 50 mM Imidazole https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 3 and 5% v/v glycerol pH 7.4. The His-tagged, target protein was eluted isocratically with a buffer containing 50 mM HEPES, 500 mM NaCl, 500 mM Imidazole pH 7.4. The proteins concentration was determined spectrophotometrically by measuring the absorbance at 280 nm, and the purity was estimated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). PD10 buffer exchange: For biochemical characterization and activity assays, all enzymes underwent buffer exchange using PD10 gravity columns following the procedure recommended by the manufacturer. In brief, the column was equilibrated with 25 ml of the exchange buffer, followed by the addition of the enzyme in elution buffer. Subsequently, the eluate was further processed by adding an equal volume of the exchange buffer for a final elution in the new environment. Thermal Stability (Tm) under different pH values and additive screen: The Prometheus NT 48 nanoDSF (NanoTemper Technologies, GmbH, Munich Germany)4 was used to determine the unfolding of purified FbGlu115A and mutants under different pH conditions. Briefly, enzyme samples at 0.2 g/L were prepared in pH 4.4, 5.6, 6.4, 7.0, 7.8, 8 (McIlvaine buffer system). Each sample (10 μl) was loaded to the capillaries of the instrument. The intrinsic fluorescence was then monitored by measuring the emission at wavelengths of 330 nm and 350 nm while subjecting the samples to a temperature gradient ranging from 20°C to 90°C, with a temperature increase of 1°C per minute. All data analysis was performed using the PR control software (Version 2.0, Munich Germany), and the Tm was extracted from the first derivative of the 350/330 absorbance ratio. Following the DSF results for the optimal pH, other parameters that can affect protein stability were tested similarly using the Rubic Additive Screen (Molecular Dimensions) as suggested by the manufacturer. In brief, 2 µl of the protein in the optimal buffer were added to each well of a 96-well PCR plate containing SYPRO Orange solution and the selective additive. Each well was then loaded into a capillary and the melting temperature was measured as described above. Irreversible Deactivation and Arrhenius constant: Thermal inactivation studies were conducted by incubating WT FAgu115A (2.8 mg/ml) in McIlvaine buffer at pH 6.0, at temperatures of 44.5°C, 45.5°C, and 48°C without substrate. Aliquots were taken at predetermined intervals and immediately cooled on ice. Each aliquot was then combined with 1% w/v beechwood xylan in the same buffer, reaching a final concentration of 0.01 g/l, and incubated at 38°C for 24 hours. Activity was measured using HPAEC-PAC, as described below, and all samples were assessed in duplicate. For this study on thermodynamic denaturation5, the thermal inactivation rate constant (k) of the enzyme was determined at various temperatures using Equation (1), where E0 represents the initial enzyme activity at time zero, and E(t) denotes the residual activity at each time point: 𝐸(𝑡)= 𝐸0 × 𝑒−𝑘𝑡 (1) By computing the thermal inactivation rate constant, the natural logarithm of each value was obtained and plotted against the inverse fraction of temperature to ascertain the Arrhenius constant (A) and calculate the activation energy, Eα, utilizing the Arrhenius equation (2). Here, R stands for the universal gas constant 8.314 J mol-1 K-1. 𝑘 = 𝐴 × 𝑒−𝐸𝑎 𝑅𝑇 (2) The times required for the enzyme to lose half (t1/2) of its activity and 90% (D value) of it was calculated using the equation (3) and (4) respectively. https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 4 𝑡12 ⁄=ln (2) 𝑘 ⁄ (3) 𝐷 = ln (10) 𝑘 ⁄ (4) For the determination of the thermal inactivation rate constant (k) the least square method was utilized, excel was used for all functions, equations, fittings, and graphs. Size Exclusion Chromatography (SEC): Analytical gel filtration experiment was set for the determination of the molecular weight of WT FAgu115A and truncated. Samples were run on Superdex 200 Increase 10/300 GL (Cytiva) with McIlvaine buffer pH 6 at room temperature. Protein concentrations varied between runs and were typically in the micromolar range. Calibration of the column was enabled using standard markers, ferritin, aldolase, conalbumin, albumin and blue dextran for the calculation of void volume. The apparent partition coefficient (Kav) was calculated for all peaks. Molecular Modellin:, Homology modelling and molecular mechanics minimizations were performed with Schrödinger release 2023 -3, using OPLS4 force field (Schrödinger, LLC, New York, NY). A homology model of FAgu115 was constructed from the X-ray crystal structure of GH115 α-1,2-glucuronidase 7PUG in complex with xylopentaose6. The protein bound xylopentaose could only be branched with a not sterically hindered glucuronic acid side chain at the third xylose. This yielded a glucuronic acid branch in a cavity with significant protein contact. Therefore, the glucuronic acid side chain was added manually to the third xylose and was energy optimized in the catalytic cavity. Theozyme model of SN2 Glucuronic acid hydrolysis: To explore the possibility of an SN2 enzymatic mechanism a theozyme model was constructed. A theozyme is a hypothetical enzyme with only the ligand and the catalytic residues present in the reaction. With QM calculations based on such a model, the geometries of the starting material, the product as well as the transition state can be found. Residues Asp303 and Asp 177 were included in the model however the backbone except for C-α was not included. The ligand was simplified to the disaccharide Xyl 2,1-GlcA. A water molecule from the 7PUG crystal was maintained to function as nucleophile on C1 atom of GlcA. In the model we assume that the backbone does not move, therefore the C-α and C-β atoms were locked. We also assume that the ligand does not move therefore the O-1 and O-4 of the xylose were also locked. Finally, in the GlcA the carbon of the carboxyl group has multiple hydrogen bonds and was additionally locked. This provided a model for the starting material. The product was constructed by a QM calculation (B3LYP-D3) on the starting material with a 1.4 Å constraint between the C1 of GlcA and the oxygen of the nucleophilic water. The transition state was found with a Linear Synchronous Transit (LST) search between the starting model and product model. Activity assays: The activity was assessed through spectrophotometric measurements, while the quantification of uronic acid release was determined using the K-URONIC kit (Megazyme), with small modifications, through a coupled enzymatic assay. The amount of NADH formed in this reaction, measured at λ=340 nm, was used as an indicator directly proportional to the concentration of glucuronic acid. The substrates used in the assay were Beechwood, 2,2-(4-O-Methyl-α-D-Glucuronyl)-xylobiose (OUX), and 2,3-(4-O-Methyl-α-D-Glucuronyl)-xylotetraose (O-XUXX) (Megazyme). Additionally, the activity was tested with HPAEC-PAD, ICS 6000, Dionex (Thermo Scientific). A CarboPac PA200 analytical column (250 mm × 3 mm, 5.5 µm) equipped with a respective guard column (50 mm × 3 mm) was used for the analysis of samples before and after the enzymatic treatment. The mobile phase consisted of 100 mM NaOH and a gradient of sodium acetate of 0-100 mM during the first 10 min after which it was kept constant at 100 mM until the end of the run at 23 min for the Beechwood analysis, while the gradient was elongated for the first 30 min when oligosaccharides were analyzed[23] (manuscript). Methylated https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 5 glucuronic acid of 4.8mM-0.6 mM was used as standard for identification and quantification. For the calculation of the maximum velocity (Vmax) and the Michaelis constant (Km) the least square method was utilized, excel was used for all functions, equations, fittings, and graphs. Results and Discussion Sequence analysis, production, and molecular mass of FAgu115A and variants. To the best of our knowledge FAgu115A is the first characterized glycoside hydrolase from the unclassified Flavobacteriaceae bacterium strain 3519-10. The genome of the microorganism encodes a large number of glycoside hydrolases, including a polysaccharide utilization locus (PUL) for xylan degradation (supplementary Fig. S1), highlighting the potential of the bacterium to utilize this substrate. The xylan degradation PUL include genes coding for putative glycoside hydrolase family 43 enzymes (a xylosidase/arabinosidase (GH43_1) and a β-xylosidase with a CBM91 (EC 3.2.1.37, GH43_12)), a gene coding for a GH10 β-1,4-D-xylanase (EC 3.2.1.8), a gene encoding a GH67 potential α-glucuronidase (EC 3.2.1.139), and a GH115 (noted as hypothetical protein in the database). The gene encoding the GH115 candidate was selected for cloning, production (in E. coli) and bioinformatics analysis, revealing a deduced amino acid sequence for wild-type (WT) FAgu115A of 836 amino acid residues, composed of one catalytic domain (CD) and one potential, but unclassified, carbohydrate binding module (CBM). A truncated form (FAgu115A _CD), lacking the C-terminal CBM was also produced, and consisted of 629 amino acids. The theoretical molecular masses of FAgu115A (including the catalytic residue variants FAgu115A_D177A, FAgu115A_D303A, FAgu115A_D177/303A) and the truncated form FAgu115A _CD were determined to be 95 and 72 kDa, respectively. These values corresponded well with the molecular mass calculated from the SDS-polyacrylamide gel electrophoresis (SDS-PAGE) of the produced proteins. Based on the information in CAZy, sequences of all structure determined GH115 candidates (five crystal structures, four deposited models, and FAgu115A from this work), were aligned using ClustalW. The multiple sequence alignment (MSA) revealed amino acids conserved across all GH115 sequences (Fig. 1B), that based on current literature data include amino acids crucial for catalysis [7b, 10] , along with their surrounding residues. The structure determined wtsAgu115A (PDB 7PUG) isolated from a metagenome in an anerobic digester [10], displayed a two-domain structure corresponding to that of FAgu115A, with 41.9 % sequence similarity, and was used to generate a homology model using the Schrödinger suite. The monomeric units of the GH115 proteins (Fig. 1A) were superimposed, showing great similarity of the catalytic domains (CDs), with the root mean square deviation (RMSD) of the helical domains being between 0.2-1.6Å. All structures exhibited a common fold with well-preserved secondary structure near the ligand binding site. Their main differences were observed in the loop regions, and in the presence and location of C-terminal accessory domains. A C-terminal β sheet CBM-like domain is superimposed at a matching location in all the enzymes except for BtGH115A (PDB 5BY3) [14] shown in copper in Fig. 1A, that differs from FAgu115 both in substrate specificity (acting on arabinogalactan) and oligomeric state (being a monomer in solution) [14]. Interestingly, this second differently located C-terminal domain (compared with FAgu115A) is present in some of the other GH115s, like SdeAgu115A (PDB 4ZMH) and AxyAgu115A (PDB 6NPS) as well as in BtGH115A. https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 6 Figure 1. The 3D-model of Fagu115 with docked ligand, superimposed on deposited GH115 structures (from PDB) or deposited models: 7PUG, 6NPS, 4C90, 5BY3 and 4ZMH, whereas B5H8Y8, A3LY17, and F5BAP8 originated from α-fold models. FAgu115A (gray, this work), 7PUG (green), https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 7 6NPS (pink), 4C90 (cyan), 5BY3 (copper), 4ZMH (lime), B5H8Y8 (magenta), A3LY17(yellow), F5BAP8 (blue).4C90, 4ZMH and 5BY3 are also shown separately to highlight all the different location scenarios of C-terminal domains (circled) present in family GH115 B. Part of the Multiple Sequence Alignment of all characterised or structure determined GH115 proteins present in CAZy. The catalytic amino acids are denoted with a star, on top of a black frame, with all conserved residues in the motifs framed. The sequence similarity matrix of the GH115s can be found in the supplementary (Fig. S2) along with the full sequence alignment. Thermal Stability (Tm) at different pHs and using additive screening The thermostability of both the WT enzyme and its mutated variants was assessed by determining the melting temperature (Tm) at various pH conditions. The WT and single or double residue variants demonstrated enhanced stability at the lower pH in the pH range from pH 6 to 8, as indicated in Table 1. Notably, FAgu115A_D303A exhibited a slightly elevated Tm at pH 6.4; however, the standard deviation was more pronounced in this sample. One-way ANOVA was conducted, and the resulted p-value was 0.443 suggesting no significant difference between FAgu115A_D303A, other point mutation variants and WT. The truncated enzyme, FAgu115A_CD, displayed a significantly lower melting temperature, which dependent on pH showed a ΔTm decrease ranging from 5°C to 14°C, implying a crucial role of the domain in the C terminus in preserving the structural stability of the protein. The thermostability was, however, relatively high considering the isolation site of the bacterium, which was a deep Antarctic ice core [15]. Table 1. Melting Temperature (Tm) of the enzymes at different pH values. All measurements were performed in duplicates. pH FAgu115A FAgu115A _CD FAgu115A _D177/303A FAgu115A _D177A FAgu115A _D303A 5.6 N.D. 34.95±0.35 50.55±0.05 50.9±0.5 47.8±0.5 6 50.7±0.5 36.7±0.1 50.35±0.15 50.65±0.15 50.4±0 6.4 49.3±0.3 38±0.1 49.6±0.1 50.35±0.15 51.05±1.45 7 46.85±0.05 38.9±0 47.45±0.15 47.8±0.1 46.95±0.05 7.8 43.8±0 39±0 44.85±0.05 45.35±0.05 44.05±0.05 8 43.55±0.05 38.85±0.05 44.5±0 44.9±0 43.6±0 Further investigations of the thermostability of the WT were made using additive screening, where the effects of small molecules were monitored, including salts, monovalent ions, multivalent ions, reducing agents, cofactors, nucleotides, and imidazole. Our aim was to observe how these additives influenced the Tm, by e.g. affecting aggregation and solubility of the protein at otherwise maintained buffer conditions (Table 2). Table 2. The melting temperature of WT FAgu115A at pH 6 influenced by the compounds included in the additive screen. The experiments were made in McIlvaine buffer pH 6. Compound Tm (°C) Compound Tm (°C) Compound Tm (°C) Compound Tm (°C) Ultapure water 49.9 CHAPS 1mM ND EDTA 5mM 38.9 Deoxyribonucleic acid 20μM ND Sodium acetate trihydrate 100mM 49.9 CHAPSO 1mM ND EGTA 5mM 38.8 ATP/ 1mM MgCl2 90.2 Calcium acetate hydrate 100mM 44.7 n-Octyl-β-D-Glycopyranoside 1mM ND Urea 0.1M ND ATPyS/ 1mM MgCl2 ND Potassium acetate 100mM 50.1 n-Decyl-β-D-maltopyranoside 1mM ND Urea 0.5M 48.4 cAMP/ 1mM MgCl2 ND Ammonium acetate 100mM 49.9 n-Dodecyl-β-D-Maltopyranoside 1mM 81.3 Urea 1M 44.8 GTP/ 1mM MgCl2 57 Sodium sulfate 100mM 50 monosaccharides mix 25mM ND Urea 2M 39.4 GTPγS/ 1mM MgCl2 ND Magnesium sulfate heptahydrate 100mM 54.8 D-Glucose 25mM ND Urea 4M 28.9 cGMP/ 1mM MgCl2 ND Potassium sulfate 100mM 50.8 Sucrose 25mM 51 Guanidine hydrochloride 150mM 49.4 NADH/ 1mM MgCl2 ND https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 8 Ammonium sulfate 100mM 49.9 Maltose 25mMCarboxylic acids mix 50mM ND Guanidine hydrochloride 500mM 42.5 NADPH/ 1mM MgCl2 ND Sodium phosphate monobasic monohydrate 100mM ND Carboxylic acids mix 50mM 49.7 NDSB 195 1mM 50.3 Polyethyleneimine 800 5mM 46.2 Sodium phosphate dibasic 100mM 46.2 L-Proline 50mM 49.4 NDSB 201 1mM 50 Imidazole 200mM 40.1 Potassium phosphate monobasic 100mM ND Glycine 50mM 48.8 Fos-Choline-12 1mM 47.5 Imidazole 400mM 34.8 Potassium phosphate dibasic 100mM 46.3 L-Glutamic acid monosodium salt hydrate 50mM 50.1 Iron (III) chloride hexahydrate 1mM 49.5 Spermidine 1mM 49.7 Sodium tartrate dibasic dihydrate 100mM ND L-Glutamic acid monosodium salt hydrate 500mM 56.6 Zinc chloride 1mM 48.3 Spermine tetrahydrochloride 1mM 49.3 Sodium citrate tribasic dihydrate 100mM 49.3 L-Arginine 50mM 41.4 Cobalt (II) chloride hexahydrate 1mM 50.2 Sarcosine 1mM ND Sodium malonate dibasic monohydrate 100mM 49.4 L-Arginine 500mM 34.7 Nickel (II) chloride hexahydrate 1mM ND Nicotinic acid 5mM 49.3 Sodium nitrate 100mM 48 L-Glutamic acid monosodium salt hydrate 50mM/L-Arginine 50mM 43 Potassium Chloride 100mM ND PEG 400 5% v/v ND Sodium formate 100mM ND L-Glutamic acid monosodium salt hydrate 500mM/L-Arginine 500mM ND Ammonium Chloride 100mM 49.3 PEG 1000 5% v/v 50 Potassium formate 100mM ND Gly-Gly-Gly 50mM ND Sodium iodide 100mM 47.5 PEG 3350 5% v/v 50.2 Sodium fluoride 100mM 49.8 Oxaloacetic acid 5mM 51 Potassium iodide 100mM 47.5 DTT 5mM 48.6 Potassium fluoride 100mM 51.1 Dimethylsulfoxide 5% v/v DMSO 44.7 Sodium bromide 100mM ND TCEP 5 mM ND Ammonium fluoride 100mM 50.6 Ethylene Glycol 5 % v/v ND Magnesium chloride hexyhydrate 1mM ND Biotin 5mM 49.6 Lithium Chloride 100mM 50.2 Glycerol 5 % v/v 50.3 Calcium chloride dihydrate 1mM ND Betaine hydrocloride 5mM 85.6 Sodium Chloride 100mM ND Glycerol 20 % v/v 49.6 Manganese (II) chloride tetrahydrate 1mM 50 Co enzyme A 5mM 50.9 N.D. = not determined The majority of the additives did neither exhibit a stabilizing effect nor a destabilizing effect on the protein, with the melting temperature (Tm) remaining stable around 50°C. Notably, the chelating agents, EDTA and EGTA, had destabilizing effects, reducing the Tm to 38.9°C and 38.8°C, respectively, indicating removal of interacting metal ions. This hypothesis is strengthened by the fact that sodium and calcium ions were present in the crystal of BoAgu115A and wtsAgu115A, respectively [7b, 10]. These two proteins are the ones with closest structural similarity to FAgu115A. A similar but smaller effect was also observed for 0.1 M phosphate groups. In addition, urea and imidazole had a negative impact on stability, with the effect increasing proportionally to the concentration of each compound, possibly leading to protein unfolding. For that reason, buffer exchange was performed after protein purification. Conversely, stabilizing effects were observed with the addition of the following compounds: magnesium sulfate-hepta-hydrate at 100mM (Tm 54.8°C), L-glutamic acid monosodium salt hydrate at 500mM (Tm 56.6°C), n-Dodecyl-β-D-Maltopyranoside at 1mM (Tm 81.3°C), betaine hydrochloride at 5mM (Tm 85.6°C), and ATP at 1mM with magnesium chloride (Tm 90.2°C). This stabilization could be due to a ligand-like binding of these compounds to the protein’s active site. Irreversible Deactivation and Arrhenius constant The thermal inactivation profile of FAgu115A, when incubated at temperatures just below the Tm, is illustrated in Figure 2 below. Table 3 reports the thermal inactivation rate constant (k), the half-life (T1/2), and the decimal reduction time (D-value). As expected, k increases while T1/2 and the D-value decrease with increasing temperature. Equations 1, 3, and 4 from the Method section were respectively employed for these calculations. There is a high temperature dependence on the enzyme's activity, with complete loss of activity after 5 min incubation at 48°C. Notably, a threshold is observed between 44.5°C and 45.5°C. For the former complete inactivation occurs after approximately 20 min of incubation, whereas https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 9 the enzyme appears to stabilize with 40% of its initial activity at 44.5°C until the end of the 30-min experiment, which may be due to refolding at the two lower temperatures. Figure 2. Thermal inactivation of WT FAgu115A over time. E/E0 is the residual enzyme activity, where E is the enzymatic activity at each time point and E0 the activity at time 0. The data are mean values of duplicates, and the bars denote the standard deviation. The natural logarithm of the calculated k values was used to construct the Arrhenius plot (Fig. 3). The resulting equation of this plot enabled the calculation of the activation energy (Eα) for inactivation by fitting the data to equation 2. The activation energy Eα was determined to be 377.2 kJ/mol for beechwood xylan. To our knowledge, similar calculations have not been previously conducted within this enzyme family, making it difficult to compare FAgu115A with other enzymes of the same family. It is important to highlight that these results agree with the theozyme calculations for the activation energy of the chemical process to reach the transition state (Fig. 8B), in which the activation energy is approximately 5 times lower compared to the Eα for inactivation. https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 16 Figure 9. A. HPAEC-PAD chromatograms of FAgu115 enzymes after 24h reaction with 0.5% Beechwood xylan. The peak between 9-10 minutes corresponds to 4-O-methyl glucuronic acid and the area is proportional to concentration with WT FAgu115 (black) having the larger peak, followed by FAgu115_D177A (blue) while there is no product formation in the other three enzymes. B. The chromatograms demonstrate the product formation over time for WT FAgu115 and FAgu115_D177A at 0h, 1h and 24h. It stands out that the product formation at 1 h is much slower for FAgu115_D177A (green) compared to the wildtype (pink). Xylooligosachharides standards from X2-X6 and 4-O-methyl glucuronic acid are showing in grey. To further explore the extent of its enzymatic capacity, kinetic studies were conducted on both FAgu115 and FAgu115_D177A (selected time points shown in Fig. 9B). During these studies, the turnover rate (kcat) and catalytic efficiency (kcat /Km) were calculated (Fig. 10), taking into account each catalytic site of the enzyme. Considering that the studied enzymes are dimers, two catalytic sites per enzyme unit were included in the calculation. Figure 10. Kinetic parameters on Beechwood xylan of WT FAgu115A (A) and FAgu115A_D177A (B). https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 17 The activity of WT FAgu115A was further examined using two distinct oligomeric substrates, one with 4-O-methyl-GlcA substitution at the reducing end and one with substitution on the non-reducing end (Table 4). This investigation aimed to assess the enzyme's activity in comparison to the other glucuronidase-families, which selectively cleave glucuronic acid from the reducing end. Additionally, the enzyme's affinity for oligomeric chains of varying sizes was evaluated. Table 4. Specific activity of FAgu115A on xylotetraose with 4-O-Methyl GlcA substitution on the 2nd xylose unit on the reducing end (XUXX) and on xylobiose with 4-O-Methyl GlcA substitution on the non-reducing end xylose (UX) Specific Activity (U/mg enzyme) Substrate (g/l) XUXX UX 1 14.4 ± 0.72 1.7 ± 0.21 3 27.0 ±0.23 10.0 ± 0.06 5 - 11.7 ± 0.54 10 21.3 ± 2.57 21.9 ± 4.12 15 18.1 ±0.49 - The homology model indicates conservation of the binding site of the C terminal domain as well as of the catalytic residues in the CD. Manual addition of glucuronic acid to the xylooligosaccharide ligand aided in proposing the residues involved in catalysis, D177 and D303 (Fig. 11A), which was also verified in the lab through point mutations of the amino acids of interest. As expected, the mutations D303A and D177/303A resulted in complete loss of activity (Fig. 8A). Surprisingly, the single mutation D177A only slowed the reaction (Fig. 9B). The remaining activity could, however, be explained after further investigation of the structure. It can be seen that the placement of an additional residue (D407) allows an extended proton shuttle from water to D407 (instead of only to D177). The negative charge of D177, that is facilitating a hydrogen bond to the nucleophilic water, arises from its interaction with D407. Mutating D177 to a much smaller Ala-residue created a space that can be filled by water molecules (Fig. 11B). It is probable that these water molecules, along with D407, form a hydrogen bonding network, once again resulting in nucleophilic attack, although at a slower pace. This was also evident from the catalytic efficiency (kcat /Km) in beechwood of 4.35 s-1 (g/l)-1 for the wild type, which is 4.5 times higher than that of D177A (Fig. 10). The specific activity of FAgu115A on XUXX and UX proves that the enzyme is active in both external and internal MeGlcA substitutions. Furthermore, the specific activity was 8.4 and 2.7 times higher in XUXX compared to UX at 1 and 3 g/l, respectively. Based on the ligand recognition in the binding site it is evident both +3R and +2NR (Fig. 5) of FAgu115 participate in the binding, of XUXX. BoAgu115A [7b] similarly exhibited higher activity on XUXX in comparison to UXX, fairly assuming that more subsites participate in the xylan binding. From our superimposition analysis and binding site interpretation, residue Trp633 (responsible for the interactions with +2NR) is conserved in BoAgu115A (PDB 4C90) and identified as Trp650, explaining the increased affinity towards longer substrates. That suggests a higher affinity and stronger binding from the backbone (Table 4). https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 18 Figure 11. (A) Catalytic residues and nucleophile water surrounding the ligand. (B) Proton "shuttle" from Asp 407 to nucleophile water after the mutation D177A. Conclusion Here, we present a novel GH115 glucuronidase from Flavobacteriaceae bacterium strain 3519-10, designated as FAgu115A. The dimerization state of FAgu115A is defined by a butterfly arrangement of each protomer. The CBM-like domain in the C-terminus is critical for maintaining both the dimerization state and the functionality of the protein. For the first time, the catalytic residues have been defined and explained through both site-specific mutation and QM calculations. Additionally, a bioinformatics analysis of all characterized GH115 proteins revealed a well-conserved structural geometry and sequence similarity around the catalytic site. Author Contributions SL: Validation, Investigation, Data Curation, WritingOriginal Draft, SL and ENK: Conceptualization, Project administration, SL and AS: Methodology, Formal analysis, https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 19 Visualization, ENK and AS: Resources, SL, AS and ENK: WritingReview & Editing, ENK: Supervision, Funding acquisition, All authors have given approval to the final version of the manuscript. Funding Sources This work is supported by the EU Horizon 2020 project EnXylaScope, Grant number 101000831. ACKNOWLEDGMENT We would like to thank Carl Grey and Javier Linares Pasten for their fruitful discussion about enzyme kinetics. REFERENCES [1] E. Sjostrom, Wood chemistry: fundamentals and applications, Elsevier, 2013. [2] P. Sannigrahi, A. J. Ragauskas, G. A. Tuskan, Biofuels, Bioproducts and Biorefining 2010, 4, 209-226. [3] a H. O. Bouveng, P. Garegg, B. Lindberg, Acta Chem Scand 1960, 14, 742-748; b A. Teleman, M. Tenkanen, A. Jacobs, O. Dahlman, Carbohydrate research 2002, 337, 373-377. [4] aA. Martínez-Abad, J. Berglund, G. Toriz, P. Gatenholm, G. Henriksson, M. Lindström, J. Wohlert, F. Vilaplana, Plant physiology 2017, 175, 1579-1592; bM. S. Izydorczyk, C. G. Biliaderis, Carbohydrate Polymers 1992, 17, 237-247. [5] J. Rao, Z. Lv, G. Chen, F. Peng, Progress in Polymer Science 2023, 140, 101675. [6] C. Amorim, S. C. Silvério, K. L. Prather, L. R. Rodrigues, Biotechnology Advances 2019, 37, 107397. [7] aT. Nagy, D. Nurizzo, G. J. Davies, P. Biely, J. H. Lakey, D. N. Bolam, H. J. Gilbert, Journal of Biological Chemistry 2003, 278, 20286-20292; bA. Rogowski, A. Baslé, C. S. Farinas, A. Solovyova, J. C. Mortimer, P. Dupree, H. J. Gilbert, D. N. Bolam, Journal of Biological Chemistry 2014, 289, 53-64. [8] O. Ryabova, M. Vršanská, S. Kaneko, W. H. van Zyl, P. Biely, FEBS letters 2009, 583, 1457-1462. [9] K. Kolenová, O. Ryabova, M. Vršanská, P. Biely, FEBS letters 2010, 584, 4063-4068. https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0 20 [10] C. Wilkens, M. Vuillemin, B. Pilgaard, I. Polikarpov, J. P. Morth, Acta Crystallographica Section D: Structural Biology 2022, 78, 658-668. [11] W. Wang, R. Yan, B. P. Nocek, T. V. Vuong, R. Di Leo, X. Xu, H. Cui, P. Gatenholm, G. Toriz, M. Tenkanen, Journal of Biological Chemistry 2016, 291, 14120-14133. [12] O. Shoseyov, Z. Shani, I. Levy, Microbiology and molecular biology reviews 2006, 70, 283-295. [13] A. B. Boraston, D. N. Bolam, H. J. Gilbert, G. J. Davies, Biochemical journal 2004, 382, 769-781. [14] F. Aalbers, J. P. Turkenburg, G. J. Davies, L. Dijkhuizen, A. L. Van Bueren, Journal of molecular biology 2015, 427, 3935-3946. [15] J. A. Raymond, B. C. Christner, S. C. Schuster, Extremophiles 2008, 12, 713-717. [16] K. L. Hudson, G. J. Bartlett, R. C. Diehl, J. Agirre, T. Gallagher, L. L. Kiessling, D. N. Woolfson, Journal of the American Chemical Society 2015, 137, 15152-15160. [17] L. S. McKee, M. J. Peña, A. Rogowski, A. Jackson, R. J. Lewis, W. S. York, K. B. Krogh, A. Viksø-Nielsen, M. Skjøt, H. J. Gilbert, Proceedings of the National Academy of Sciences 2012, 109, 6537-6542. [18] E. Gasteiger, C. Hoogland, A. Gattiker, S. e. Duvaud, M. R. Wilkins, R. D. Appel, A. Bairoch, Protein identification and analysis tools on the ExPASy server, Springer, 2005. [19] J. D. Thompson, D. G. Higgins, T. J. Gibson, Nucleic acids research 1994, 22, 46734680. [20] X. Robert, P. Gouet, Nucleic acids research 2014, 42, W320-W324. [21] F. H. Niesen, H. Berglund, M. Vedadi, Nature protocols 2007, 2, 2212-2221. [22] Y. Ando, S. Hagiwara, H. Nabetani, Journal of Food Engineering 2017, 199, 9-18. https://doi.org/10.26434/chemrxiv-2024-wmxtb ORCID: https://orcid.org/0000-0003-3799-112X Content not peer-reviewed by ChemRxiv. License: CC BY-NC-ND 4.0