The α-l-arabinofuranosidases (α-l-ABFs) are exoenzymes involved in the hydrolysis of α-l-arabinosyl linkages in plant cell wall polysaccharides. They play a crucial role in the degradation of arabinoxylan and arabinan and they are used in many biotechnological applications. Analysis of the genome of R. cellulolyticum showed that putative cellulosomal α-l-ABFs are exclusively encoded by the xyl-doc gene cluster, a large 32-kb gene cluster. Indeed, among the 14 Xyl-Doc enzymes encoded by this gene cluster, 6 are predicted to be α-l-ABFs belonging to the CAZyme families GH43 and GH62.
The biochemical characterization of these six Xyl-Doc enzymes revealed that four of them are α-l-ABFs. GH4316-1229 (RcAbf43A) which belongs to the subfamily 16 of the GH43, encoded by the gene at locus Ccel_1229, has a low specific activity on natural substrates and can cleave off arabinose decorations located at arabinoxylan chain extremities. GH4310-1233 (RcAbf43Ad2,3), the product of the gene at locus Ccel_1233, belonging to subfamily 10 of the GH43, can convert the double arabinose decorations present on arabinoxylan into single O2- or O3-linked decorations with high velocity (kcat = 16.6 ± 0.6 s−1). This enzyme acts in synergy with GH62-1234 (RcAbf62Am2,3), the product of the gene at locus Ccel_1234, a GH62 α-l-ABF which hydrolyzes α-(1 → 3) or α-(1 → 2)-arabinosyl linkages present on polysaccharides and arabinoxylooligosaccharides monodecorated. Finally, a bifunctional enzyme, GH62-CE6-1240 (RcAbf62Bm2,3Axe6), encoded by the gene at locus Ccel_1240, which contains a GH62-α-l-ABF module and a carbohydrate esterase (CE6) module, catalyzes deacylation of plant cell wall polymers and cleavage of arabinosyl mono-substitutions. These enzymes are also active on arabinan, a component of the type I rhamnogalacturonan, showing their involvement in pectin degradation.
Arabinofuranosyl decorations on arabinoxylan and pectin strongly inhibit the action of xylan-degrading enzymes and pectinases. α-l-ABFs encoded by the xyl-doc gene cluster of R. cellulolyticum can remove all the decorations present in the backbone of arabinoxylan and arabinan, act synergistically, and, thus, play a crucial role in the degradation of plant cell wall polysaccharides.
Arabinose-containing polysaccharides are found mainly in hemicelluloses and pectins. Arabinoxylan, the most abundant hemicellulosic component, is found in a wide range of plant species, hard wood, and annual plants . Arabinoxylan is constituted of a linear backbone of β-(1 → 4)-linked d-xylopyranosyl units (Xylp) decorated by α-l-arabinofuranosyl (Araf) substituents attached through O-2 and/or O-3. In arabinoxylan, O-acetyl substitutions of xylose residues are common, but 4-O-methyl-d-glucuronic acid is also found . The ratio Araf/Xylp in plant cell wall varies considerably, from 0.6 to 0.07 . Some of the arabinose residues are linked to ferulic acids by ester bonds and the formation of ferulate dimers creates arabinoxylan–arabinoxylan cross-links. Pectins, a highly complex and heterogeneous group of polysaccharides, are composed of two types of regions: homogalacturonan and type I rhamnogalacturonan. Arabinan and arabinogalactan I are found in type I rhamnogalacturonan side chains. Arabinan backbone consists of α-(1 → 5)-linked-Araf units branched at O-3 or O-2 with Araf units and AGI consists of a β-(1 → 4)-d-galactopyranosyl backbone with O-3 substitutions of α-linked-Araf residues .
α-l-Arabinofuranosidases (α-l-ABFs), also called arabinoxylan-arabinofuranohydrolases (AXHs), are exoenzymes involved in the cleavage of α-(1 → 2), α-(1 → 3) or α-(1 → 5)-linked arabinosyl decorations in arabinoxylan, arabinan, arabinogalactan, and arabinoxylooligosaccharides (AXOS) [5,6,7,8,9]. α-l-ABFs are found in glycoside hydrolase (GH) families 2, 3, 43, 51, 54, and 62. They are divided into three types depending on their mode of action on arabinoxylan. AXHs-m cleave off arabinose from mono-substituted (α-(1 → 2) or α-(1 → 3)-linked) xylose residues. AXHs-d act on double-substituted xylose residues and remove either α-(1 → 2)-Araf linkages or α-(1 → 3)-Araf linkages. Finally, AXHs-m,d display a dual activity on mono-substituted and double-substituted xylose residues located at chain extremities and/or internal [6, 10]. α-l-ABFs able to cleave α-(1 → 5)-Araf linkages catalyze the cleavage of terminal α-l-arabinofuranosyl residues on decorated or linear arabinan. Arabinosyl substitutions in hemicelluloses and pectins may hinder the catalytic interaction of the enzymes with the substrate backbone, and negatively affect the hydrolysis. Moreover, arabinosyl substitutions participate in the cross-linkage of polysaccharides within the plant cell wall. α-l-ABFs are, thus, required for the complete degradation of hemicelluloses and pectins, and exhibit a cooperative effect with other lignocellulose-degrading enzymes .
Ruminiclostridium cellulolyticum H10 (ATCC 35319), a non-ruminal, mesophilic cellulolytic bacterium , secretes carbohydrate-active enzymes (CAZymes) involved in the degradation of plant cell wall polysaccharides (http://www.cazy.org). Some of them, known as cellulosomal enzymes, bear a dockerin module and can interact with a non-catalytic scaffolding protein, CipC, carrying eight complementary cohesin modules, to form cellulosomes. Many of the well-characterized cellulosomal enzymes are encoded by the cip-cel operon (loci Ccel_0728 to Ccel_0739) or belong to the GH9 family [13,14,15]. The 26-kb cip-cel operon encodes essential subunits for the formation of cellulosomes and cellulose degradation [16, 17]. Another large gene cluster of 32 kb, called xyl-doc (loci Ccel_1229 to Ccel_1242), was identified and encodes secreted dockerin-bearing proteins which all exhibit a carbohydrate-binding module (CBM) predicted to target hemicelluloses (mainly CBM6) and a catalytic module putatively involved in the degradation of hemicelluloses (glycoside hydrolases GH2, 10, 27, 30, 43, 59, 62, and 95 and carbohydrate esterases CE1 and CE6) . The expression of xyl-doc genes is activated by XydR, a response regulator, in the presence of straw, arabinoxylan, xylose, or arabinose . The R. cellulolyticum genome analysis showed that the 6 cellulosomal putative α-l-ABFs are encoded exclusively by the xyl-doc gene cluster . Four genes (loci Ccel_1229, Ccel_1231, Ccel_1233, and Ccel_1235) encode GH43 enzymes. Two genes (loci Ccel_1234 and Ccel_1240) encode GH62-containing proteins. The gene at locus Ccel_1240 encodes a putative bifunctional enzyme with one GH62 module and one CE6 module putatively involved in the hydrolysis of acetyl groups presents in xylan backbone. All these enzymes were detected in the cellulosomes [18, 19] and could drive the removal of arabinose decorations in arabinoxylans, an indispensable stage for complete hydrolysis. All of these enzymes were characterized biochemically and a mode of action was proposed for each of them based on generated three-dimensional models. In summary, these enzymes can remove all the Araf decorations and acetyl substitutions that arabinoxylan can carry and are also active on arabinan. They play a crucial role in the degradation of these compounds.
Bacterial strains and plasmids
Genomic DNA from R. cellulolyticum ATCC 35319 (NCBI Reference Sequence: NC_011898.1) served as a template for amplification by PCR of the genes at loci Ccel_1229, Ccel_1231, Ccel_1233, Ccel_1234, Ccel_1235, and Ccel_1240 encoding the mature forms of the putative α-l-arabinofuranosidases. The list of primers used in this study is provided in Additional file 1: Table S1. The amplicons were cloned in pET22b(+) (Novagen) at NdeI/XhoI sites or in pET28b(+) (Novagen) at NcoI/XhoI, to introduce six histidine codons at the 3′ extremity of the coding sequence, except, for the gene encoding, the putative α-l-arabinofuranosidases GH4329-1231 which is cloned in the pGEX-5X-2 vector (Sigma-Aldrich) at BamHI/XhoI, to add a GST-tag (glutathion S-transferase) at the N-terminal extremity of the recombinant protein. In this case, the DNA sequence encoding the hexa-histidine tail was introduced in the primer 1231pGexrev. Positive clones were verified by DNA sequencing (Genewiz). The BL21(DE3) Escherichia coli strain was used as production strain for all the recombinant proteins, except for GH4329-1231 for which the DH5α (New England Biolabs) strain was used.
Protein production and purification
Recombinant proteins were purified from 700 mL cultures in lysogenic broth medium supplemented with glycerol (0.85%) and the appropriate antibiotic (at the final concentration of 200 µg/mL for ampicillin and 100 µg/mL for kanamycin). After growth at 37 °C until A600nm = 1.5–2, the cultures were cooled, and induction of the gene expression was performed overnight at 18 °C with 150 μM of IPTG (isopropyl β-d-1-thiogalactopyranoside). After 16 h of induction, the cells were harvested by centrifugation (3000×g, 15 min, 4 °C), resuspended in 30 mM Tris–HCl, pH 8.0, 1 mM CaCl2, supplemented with few milligrams of DNase I (Roche Applied Science), and broken in a French press. The crude extract was centrifuged 15 min at 10,000g at 4 °C and loaded on 2–5 mL of nickel–nitrilotriacetic acid resin (Qiagen) equilibrated in the same buffer. The proteins of interest were eluted with 100 mM imidazole in 30 mM Tris–HCl, pH 8.0, 1 mM CaCl2. Except for GH4329-1231, the purification was achieved on Q-Sepharose fast flow (GE Healthcare) equilibrated in 30 mM Tris–HCl, pH 8.0, 1 mM CaCl2. The protein GH4329-1231 was mainly produced under insoluble form, despite the presence of the GST-Tag, and a low quantity of protein was purified on nickel-affinity column. For this particular recombinant protein, only one chromatography was done. The purified proteins were dialyzed and concentrated by ultrafiltration against 20 mM Tris‐maleate, pH 6.0, 1 mM CaCl2, and stored at − 80 °C. The concentration of the proteins was estimated by measurement of the absorbance at 280 nm and the use of the molar extinction coefficient calculated by ProtParam tool (https://web.expasy.org/protparam: GH4316-1229 101,565 M−1 cm−1, GH4329-1231 154,310 M−1 cm−1, GH4310-1233 144,550 M−1 cm−1, GH62-1234 127,130 M−1 cm−1, GH4329-1235 114,305 M−1 cm−1, GH62-CE6-1240 172,300 M−1 cm−1). Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was performed using precast gels 4–15% acrylamine (Bio-Rad). Gels were stained with coomassie blue.
WAXY-I, WAXY-RS, arabinan (from sugar beet), and linear arabinan were purchased from Megazyme and OSX was purchased from Sigma-Aldrich. WAXY-I and WAXY-RS have almost the same ratio Araf/Xylp (36/51 and 38/62, respectively). Insoluble substrates WAXY-I and OSX were washed with milliQ water onto the Stericup® vacuum filtration system (0.22 µm pore size, polyethersulfone membrane, Sigma-Aldrich). Activities were determined at 37 °C, under mild shaking (70 rpm) by mixing 4 mL of substrate at 17.5 g/L in 20 mM Tris‐maleate, pH 6.0, 1 mM CaCl2, 0.01% (w/v) azide with 40 μl of an appropriate concentration of enzyme (between 10 nM to 1 µM). At specific intervals, aliquots of 500 μL were cooled on ice. For the insoluble substrates, aliquots were centrifuged at 4 °C for 5 min at 15,000g. 0.2 mL of soluble products released were mixed with 50 μL of 0.5 M sodium hydroxide and analyzed by high‐pressure anion exchange chromatography coupled with pulsed amperometric detector (HPAEC‐PAD) (Dionex CarboPac PA1 column) as previously described . Arabinoxylooligosaccharides (AXOS) A3X, A2XX, XA3XX, A2+3XX, and XA2+3XX were purchased from Megazyme. Activities were determined at 37 °C, under 750 rpm shaking (Eppendorf Thermomixer), by mixing 90 μL of substrate at 1.11 mM in 20 mM Tris-maleate, pH 6.0, 1 mM CaCl2 with 10 μL of enzyme (ranging from 10 nM to 1 µM). Samples (0.2 mL) were collected at specific time intervals, cooled on ice, and mixed with 50 μL of 0.5 M sodium hydroxide before to be analyzed by HPAEC-PAD.
Kinetic parameters were determined on washed WAXY-I by incubating at 37 °C, under 750 rpm shaking (Eppendorf Thermomixer), 1.5 mL of substrate at various concentrations (17.5 g/L–10 g/L–5 g/L–3.33 g/L–2 g/L–1.33 g/L–1 g/L) in 20 mM Tris‐maleate, pH 6.0, 1 mM CaCl2, 0.01% (w/v) sodium azide with GH4310-1233 at 100 nM (final concentration) or GH62-1234 and GH62-CE6-1240 at 1 µM (final concentration). At specific intervals, 500 μL aliquots were cooled on ice and centrifuged at 4 °C for 5 min at 15,000g. Samples (0.2 mL) were mixed with 50 μL of 0.5 M sodium hydroxide and analyzed by HPAEC-PAD. Injection of samples containing arabinose, xylose, and AXOS at known concentrations was used to identify and quantify the released sugars. Specific activities are given in IU/μmol (1 IU/µmol = 1 µmol of arabinose released per minute by 1 µmol of enzyme under the experimental conditions used). These are calculated by determination of the quantity of arabinose released after 10 min of reaction by HPAEC-PAD. kcat in s−1 and Km in g/L. The Km and Vmax values were obtained by Lineweaver–Burk plots and the kcat calculated from the Vmax values. The program OriginPro was also used to fit the Michaelis–Menten representation.
The acetyl-xylan esterase activity was measured using washed WAXY-I at the concentration of 17.5 g/L. GH62-CE6-1240 at the final concentration of 1 µM was incubated with 4 mL of substrate, at 37 °C under mild shaking (70 rpm). At specific intervals, aliquots of 500 µL were cooled on ice and centrifuged at 4 °C for 5 min at 15,000g. 0.2 mL of soluble products released were mixed with 50 μL of 25 mM H2SO4 and analyzed using an IOTA 2 Differential Refractometer and an HPX-87H HPLC column (Bio-Rad) preceded by the corresponding guard column (Micro-Guard HPLC column protection system). The acid acetic released was eluted with H2SO4 5 mM at 55 °C, and a constant flow rate of 0.6 mL/min. Acetic acid at known concentrations (ranging from 0.5 to 10 mM) was used as the standard. The specific activity (IU/μmol) is the amount (in µmol) of released acetic acid per minute and per µmol of enzyme.
Synergistic action between GH4310-1233 and GH62-1234 was measured as follows: 4 mL of washed WAXY-I at the concentration of 17.5 g/L was incubated with 40 µL of GH4310-1233 (final concentration 0.1 µM) and 40 µL of GH62-1234 (final concentration 1 µM) 24 h at 37 °C under mild shaking (70 rpm). Aliquots of 500 µL were collected, cooled on ice, and centrifuged at 4 °C for 10 min at 10,000g. 0.2 mL of soluble products were mixed with 50 μL of 0.5 M sodium hydroxide and analyzed by HPAEC-PAD.
Three-dimensional structure modeling
We used various servers for the construction of 3D models. Each α-l-ABF characterized was modeled using each program (SWISS-MODEL, Phyre2, and RaptorX). The identification of structural template(s) comes from the initial protein primary structures from the KEGG SSDB Database [cce:Ccel_1229, cce:Ccel_1233, cce:Ccel_1234, and cce:Ccel_1240]. The SWISS-MODEL server was used to build protein models, using a template structure and target-template sequences alignments. For each selected template, a 3D protein model was automatically generated and the assessment of the model’s quality: the GMQE score and QMEAN score were calculated. GMQE (Global Model Quality Estimation) score and QMEAN score provided a quality estimation of the structural features observed for the model. GMQE score is expressed as a number between 0 and 1. QMEAN scores around zero indicate good agreement between the model structure and experimental structures.
The Phyre2 (Protein Homology/AnalogY Recognition Engine) server was also used to generate 3D models after searching for homologous sequences in the HHblits database (HMM–HMM-based lightning-fast iterative sequence search). For each model constructed, the ProQ2 quality assessment was checked [20, 21].
RaptorX was generally used to generate models for protein sequences without close homology (less than 30% sequence identity in PDB). It assigns confidence scores to evaluate the quality of the predicted 3D model: P value for the relative global quality (the smaller the P value, the higher quality the model), GDT (global distance test), and uGDT (un-normalized GDT) for the absolute global quality. For a protein with > 100 residues, uGDT > 50 and GDT > 50 are good indicators.
PyMOL (http://www.pymol.org) was used to visualize, analyze the 3D model, and to construct the figures shown in the present article.
Modular organization and putative function of Xyl-Doc enzymes
The 14 putative cellulosomal proteins (called Xyl-Doc enzymes) encoded by the xyl-doc gene cluster (Ccel_1229 to Ccel_1242) all display a modular organization (Table 1) and are predicted to be involved in hemicellulose degradation. The xyl-doc gene cluster was specifically expressed when arabinoxylan, arabinose, or xylose was used as substrates, and Xyl-Doc enzymes were consequently detected by proteomic analysis in the cellulosomes produced by R. cellulolyticum grown in these specific culture conditions [18, 19]. Some of them, which have a catalytic module belonging to families GH43 and GH62, are annotated as α-l-arabinofuranosidases (Table 1). In the CAZy database, GH43 family contain β-xylosidases, α-l-arabinofuranosidases, arabinanases, xylanases, galactan 1,3-β-galactosidases, exo-α-1,5-l-arabinofuranosidases, exo-α-1,5-l-arabinanases, and β-1,3-xylosidases. They are organized in 37 subfamilies . The xyl-doc gene cluster encodes four GH43 containing proteins: GH4316-1229 belongs to subfamily 16, GH4329-1231 and GH4329-1235 belong to subfamily 29, and GH4310-1233 belongs to subfamily 10. Their predictive functions according to the subfamily membership are reported in the Table 1. Most characterized enzymes, belonging to the subfamilies 16, 29 and 10, subfamilies found in the Xyl-Doc GH43, show some polyspecificity: β-xylosidase, xylanase, and α-l-ABF activities . Only a thorough enzymatic characterization using natural substrates can allow discrimination between these activities. The catalytic modules of the two proteins belonging to the subfamily 29 (Xyl-Doc enzymes: GH4329-1231 and GH4329-1235) share 67% of identity, whereas these modules display approximatively 30% of identity with those found in other Xyl-Doc GH43 enzymes.
An X19 module is found associated with the GH43 subfamily 10 module of the protein GH4310-1233. The function of the X19 modules, which have a β-propeller fold and are found in subfamilies 9 to 14 and 36, is still unclear. They might be involved in the stabilization on the three-dimensional structure of the associated catalytic modules . The whole GH4310-1233 protein shows 89% and 91% of identity, respectively, with uncharacterized proteins from Ruminiclostridium papyrosolvens DSM 2782 (EPR10800) and Clostridium sp BNL1100 (AEY65170), two other Clostridium displaying similar xyl-doc gene cluster organization and surrounding genes.
GH62 family only contains enzymes described as α-l-arabinofuranosidases. Two Xyl-Doc GH62-enzymes (GH62-1234 and GH62-CE6-1240) could be involved in the removal of Araf decoration in arabinoxylan. The product of the gene at locus Ccel_1240 encodes an enzyme with two distinct catalytic modules (GH62 and CE6). This enzyme is predicted to have an α-arabinofuranosidase activity and an acetyl-xylan esterase activity. The GH62 catalytic module of the two GH62-Xyl-Doc enzymes are strictly identical, except for one amino-acid at position 326 (T in GH62-1234 replaced by E in GH62-CE6-1240). It is interesting to note that the adjacent CBM6 of these two catalytic modules are also strictly identical. In terms of nucleotidic sequences, the 5′-parts of these two genes show more than 97% of identity. A bifunctional cellulosomal enzyme, RjAbf62A-Axe6A, from Ruminiclostridium josui JCM 1788, with exactly the same modular organization and a high sequence identity (91%) was recently characterized and its role in the arabinoxylan degradation was demonstrated . This enzyme is the first GH62 characterized that has an endoxylanase activity in addition to its α-l-ABF function.
GH43 and GH62 enzymes display a five-bladed β-propeller fold (clan F). The active site is located in a deep cavity in the center of the β-propeller. They are inverting enzymes. Two conserved carboxylate residues, an aspartate and a glutamate, that act as catalytic acid and base, respectively, are required to catalyze the hydrolysis of the glycosidic bond. An additional conserved aspartate is necessary for the catalysis, though its role is not fully established. It could help to maintain the correct alignment of the general acid residue with the substrate . The three conserved acidic amino-acids were found in the primary structure of the six Xyl-Doc GH43 and GH62 modules, as determined by sequence alignment. In the structural models generated with SWISS-MODEL, Phyre2, and RaptorX, these residues occupy the center of the cavity as expected. The additional X19 module of the protein GH4310-1233 (encompassing amino-acids 344 to 533) displays a canonical β-sandwich fold (Fig. 1). The structure of a GH43 β-xylosidase from Geobacillus stearothermophilus showed that the cleft of the active site is partly blocked by a loop originating from the X19 module . This loop closes the cleft on one side to form the catalytic pocket. The crystal structure of HiAXHd3 (GH43 subfamily 36) from Humicola insolens also showed an overlap between the GH43 module and the X19 module, and the interface between these two modules bears the substrate-binding site and contributes to the overall configuration of the substrate-binding pocket . It has been demonstrated that the stable folding of each module is dependent upon the other showing conformational and catalytic importance of the overlap between these two modules . A loop coming from the X19 module is also found at the proximity of the catalytic cavity of GH4310-1233 (Fig. 1).
Putative Xyl-Doc α-l-ABFs production and purification
To investigate the function of these six putative α-l-ABFs, the mature form of each protein (encompassing the catalytic module, the CBM6, and the dockerin module) was produced in Escherichia coli. Recombinant proteins were tagged with a hexa-histidine sequence at their C-terminus. Proteins were purified using an immobilized metal affinity chromatography (IMAC) followed by an anion exchange chromatography step to achieve a significantly high yield and high purity of proteins. Production of recombinant GH4329-1231 in E. coli resulted in aggregation into inclusion bodies regardless of the growth temperature and inducer (IPTG) concentration. To obtain a soluble form of the protein, a fusion with glutathione S-transferase was done. Recombinant proteins containing a GST-tag at their N-terminal extremity were often reported to have greater solubility . However, the recombinant GH4329-1231 was still predominantly insoluble and only a very low amount of purified protein could be obtained. The purity of the proteins was checked by SDS-PAGE (Additional file 2: Figure S1) and their migration was in agreement with the molecular masses predicted from the polypeptide sequences of the recombinant forms (GH4316-1229: 56,281 Da; GH4329-1231: 81,774 Da; GH4310-1233: 79,932 Da; GH62-1234: 57,165 Da; GH4329-1235: 54,169 Da and GH62-CE6-1240, 86,290 Da).
Putative Xyl-Doc α-l-ABFs activities on polysaccharides
To determine the substrate specificities and the functions of these six enzymes, each was tested for its capacity to degrade arabinoxylan and the released products were analyzed by HPAEC-PAD (Table 2). Wheat-flour arabinoxylan for reducing-sugar assays (WAXY-RS), a highly decorated soluble substrate (ratio Araf/Xylp = 38/62), was first used to analyze the nature of the released sugars and discriminate α-l-ABFs from other enzymatic functions (Table 2). The recombinant protein GH4329-1231 showed a β-xylosidase activity, since only xylose is released from arabinoxylan, with a relatively low specific activity at around 0.9 IU/µmol. This enzyme is the first member of this subfamily described as a β-xylosidase.
The Xyl-Doc enzyme GH4329-1235 appears polyspecific, having both β-xylosidase and α-l-ABF activities with mainly a β-xylosidase activity on arabinoxylan. GH43 polyspecific enzymes have previously been characterized . Recently, the arabinoxylan degradation profile of a GH4329 subfamily enzyme, AxB8 from Clostridium thermocellum, was shown to release xylose as the main hydrolysis product . Only small amounts of arabinose are released by AxB8 such as GH4329-1235.
Four enzymes, GH4316-1229, GH4310-1233, GH62-1234, and GH62-CE6-1240, released only arabinose from WAXY-RS and can, thus, be considered as α-l-ABFs (Table 2). The two GH62 α-l-ABFs have the same level of specific activities from 3 to 4 IU/µmol, GH4310-1233 exhibits a high enzymatic activity at around 450 IU/µmol and GH4316-1229 is the least efficient (specific activity around 0.2 IU/µmol). Further characterization of the activities of these four α-l-ABFs was performed.
Αctivities of the four α-l-ABFs on various polysaccharides and kinetic parameters
To further characterize the four Xyl-Doc α-l-ABFs, various substrates with different compositions were used (Table 3). Insoluble wheat-flour arabinoxylan insoluble (WAXY-I) and wheat-flour arabinoxylan for reducing-sugar assays (WAXY-RS) have almost the same Araf/Xylp ratios, 36/51 and 38/62, respectively. The procedure of extraction and purification of WAXY-I allow maintenance of ferulic acid cross-links and acetylation of xylose residues. WAXY-RS contains no ferulic acid cross-links or acetylated xylose residues. A poorly decorated xylan from oat spelt (OSX) with a ratio Araf/Xylp of 10/70 was also tested.
Whatever the arabinoxylan used, the two GH62-α-l-ABFs display the same specific activities (Table 3). GH4310-1233 has the same specific activity against WAXY-RS and WAXY-I, but is around fivefold less active on OSX. GH4316-1229 is around ten times less active on insoluble substrates OSX and WAXY-I compared with WAXY-RS. Kinetic parameters on WAXY-I were determined, except for GH4316-1229, which is not active enough (Table 3). The turnover number, kcat, and the Km, found for the two cellulosomal GH62-α-l-ABFs, are in the same range as those reported for other characterized GH62-α-l-ABFs . The kcat of GH4310-1233 is around 150 times higher than the kcat of the two cellulosomal GH62-α-l-ABFs and GH4310-1233 has a high catalytic efficiency (kcat/Km) despite an elevated Km value.
Some α-l-ABFs are able to remove the Araf decorations from arabinan. When sugar-beet arabinan was used as substrate, the GH4310-1233 and the two GH62-α-l-ABFs released arabinose, showing that they could recognize and hydrolyse Araf decoration, whereas GH4316-1229 was found to be completely inactive (Table 3). Linear arabinan was degraded by none of these enzymes, indicating that none of them have an α-l-arabinanase activity or an exo-α-(1 → 5)-degradative function.
The bifunctional enzyme GH62-CE6-1240 can release acetate from WAXI-I (specific activity: 2.83 ± 0.21 IU/µmol), indicating that this enzyme is able to remove two distinct decorations of the main chain: Araf and acetylation.
α-l-ABFs activities on arabinoxylooligosaccharides
The results given above showed that the xyl-doc gene cluster encodes 4 α-l-ABFs: GH4316-1229, GH4310-1233, GH62-1234, and GH62-CE6-1240. Activities were also explored using arabinoxylooligosaccharides (AXOS) containing single Araf decorations or double Araf decorations (Table 4). According to the one-letter code system already published , we used the following nomenclature: the letter X corresponds to an unsubstituted xylose residue and Ax with a superscript number (x) describes a xylose residue decorated by an Araf residue. If Araf is α-1,2-linked, the decorated Xylp residue is noted A2. If Araf is α-1,3-linked, the decorated Xylp residue is noted A3. If the xylodextrin contains a double decoration, the substituted xylose residue substituted is noted: A2+3.
GH4316-1229 is totally inactive on the tested AXOS (Table 4), despite the high enzyme concentration used (5 µM). The Araf decorations located on the non-reducing Xylp residue or penultimate Xylp residue from the non-reducing end in these AXOS, are not cleaved by GH4316-1229. We can hypothesize that this enzyme is an α-l-ABF specific of the decorations located at the reducing end of the main chain. However, in the absence of appropriate substrates commercially available, it is not possible to confirm this hypothesis. Consequently, we are also unable to provide further information with respect to its specificity for mono- or double Araf decorations. Anyway, the singularity of the sugar-motif recognized by this enzyme could explain its very low specific activity on arabinoxylan.
GH4310-1233 cleaves Araf decorations from internal di-substituted xylodextrins (Table 4) and is inactive on mono-substituted substrates. This enzyme can cleave α-(1 → 2) and α-(1 → 3). HPAEC-PAD analysis of the products, from A2+3XX and XA2+3XX hydrolysis showed that the main products, apart from Araf, are, respectively, A3XX and XA2XX indicating that, depending upon the position of the double substitution in the AXOS, the cleaved linkage varies (Fig. 2a, b).
The two GH62-α-l-ABFs encoded by the xyl-doc gene cluster were found to be active on polysaccharides and mono-substituted AXOS. For both enzymes, a lower specific activity was found on arabinoxylobiose A3X and on arabinoxylotetraose XA3XX, maybe because the preferential cleaved glycosidic bond is α-(1 → 2). However, these enzymes seem to prefer polysaccharides. Indeed, the specific activity is almost 100-fold higher on WAXY-I, WAXY-RS, and OSX than on AXOS. The same specific activity was observed on WAXY and OSX despite a lower amount of Araf decorations in OSX. This observation can probably be explained by the fact that, in OSX, the xylose backbone predominantly harbors Araf mono-substitutions. This hypothesis might also explain the lower specific activity of GH4310-1233 on OSX, which specifically recognizes double Araf decorations.
Synergistic action between GH4310-1233 and GH62-1234
According to their modes of action, GH4310-1233 and GH62-1234 should act in synergy. Indeed, the single Araf substitutions (O2- or O3-α-linked) generated by GH4310-1233 on polysaccharides could be recognized and cleaved by GH62-1234. This hypothesis was confirmed, as shown in Fig. 3. A synergy factor of 1.53 (ratio between the experimental quantity of arabinose released by the mixture of GH4310-1233 and GH62-1240 and the calculated sum of arabinose released by each enzyme alone) was found between these two enzymes on WAXY-I.
Three-dimensional structure modeling
Computational methods for protein structure modeling (SWISS-MODEL, Phyre2 and RaptorX) were used to shed some light on our biochemical results. The 3D structure of the catalytic module of the protein GH4316-1229 (residues 33–314 based on KEGG SSDB Database, cce:Ccel_1229) was modeled to refine the understanding of its mode of action. Structural models were built using the 3D structure of Ct43Araf (PDB ID: 5a8d.1.A) (subfamily 16) from Clostridium thermocellum, which shares 64% sequence identity with the GH43 module of GH4316-1229. GH4316-1229 displays a five-bladed β-propeller structure (each propeller formed by four antiparallel β-strands) and its molecular surface reveals a narrow and deep pocket for the accommodation of only one Araf residue. The three putative conserved catalytic amino-acids were identified by multiple alignment (D40, D159, and E208) (Fig. 4a). Similarly to GH4316-1229, Ct43Araf has the ability to catalyze the hydrolysis of terminal arabinosyl residues in arabinoxylans [33, 34]. Ct43Araf hydrolyses the Araf decorations (α-1,2 or α-1,3) located at the non-reducing extremity of arabinoxylans, whereas our biochemical results suggest, for GH4316-1229, an action at the reducing extremity. According to the narrowness of the catalytic pocket, GH4316-1229 could be specific of Araf mono-substitutions. GH4316-1229 and Ct43Araf act only on arabinoxylans and are unable to cleave off arabinose decorations from arabinan. These biochemical results are consistent with the structure of Ct43Araf which reveals the presence of a substrate-binding cleft designed to interact with arabinoxylan but unable to accommodate the nonlinear backbone of arabinan [33, 34].
The N-terminal part of GH4310-1233, composed of two modules (GH43: residues 35 to 300 and X19: residues 344 to 533), was modeled using the 3D structure of HiAXHd3 (PDB ID: 3zxl.A) (subfamily 36) from Humicola insolens which shares 37% identity with the GH43-X19 part of GH4310-1233 (Figs. 1 and 4b). HiAXHd3 is, like GH4310-1233, highly specific for substrates containing double Araf substitutions . The structure of the catalytic cavity of GH4310-1233 shows a shallow but extended pocket able to accommodate two Araf residues. The three putative catalytic amino-acids (D46, D156, and E207) are localized, according to our model, around the groove inside which Araf decorations might interact (Fig. 4b). Like other GH43 proteins carrying an additional X19 module, the 3D model shows structural connections between these two modules.
The primary structures of the two GH62 catalytic modules from GH62-1234 (residues 35 to 295 based on KEGG SSDB Database, cce:Ccel_1234) and GH62-CE6-1240 (residues 35 to 294 based on KEGG SSDB Database, cce:Ccel_1240) are strictly identical except for one residue located at position 236. For both GH62 catalytic modules, the template used was the GH62-α-l-ABFs from Coprinopsis cinerea (PDB ID: 5b6s.1.A) . The three-dimensional modeling revealed, as expected for an enzyme belonging to the clan F, a structure in five-bladed β-propeller with a narrow pocket, carrying the three putative catalytic amino-acids, identified by multiple alignment. GH62-α-l-ABFs from C. cinerea accommodates a single Araf residue . The narrowness of the catalytic cavities of both GH62-1234 and GH62-CE6-1240 also suggests an interaction with a single Araf substitution, in agreement with the enzymatic characteristics described above (Fig. 4c). The serine–histidine–glycine (SHG) motif conserved within the GH62 family [8, 36, 37] and in some GH43 members was also found in the four R. cellulolyticum α-l-ABFs described in this paper. In the arabinase, BsArb43B from Bacillus subtilis, the histidine, at the bottom of the catalytic cavity, coordinates a calcium ion and plays a structural role .
Cellulosomes of Ruminiclostridium cellulolyticum are extracellular multi-enzyme machineries, which efficiently degrade plant cell wall polysaccharides. 62 annotated CAZymes contain a type I-dockerin module and are integrated into the cellulosomes through interactions with the scaffolding protein CipC. Many cellulosomal enzymes display activity against cellulose, hemicelluloses (arabinoxylans and xyloglucans), and pectins, and it was shown, by transcriptomic and proteomic analyses, that the composition of R. cellulolyticum cellulosomes is modulated according to the growth substrates [13, 19, 38, 39]. Genes encoding cellulases, especially those gathered in the cip-cel cluster, are regulated by carbon catabolite repression, while the expression of most of the genes encoding other CAZymes and accessory enzymes is activated via two-component regulation systems as it was shown for the xyl-doc gene cluster [17,18,19, 39]. This 32-kb gene cluster encodes modular CAZymes putatively involved in the degradation of hemicelluloses components such as arabinoxylan which can constitute up to 40% of the total plant cell wall dry mass depending on the plant species, especially in the cell walls of wheat grain . xyl-doc gene expression is induced by the presence of straw, and arabinoxylan, and XydR, a response regulator acts as an activator of the transcription of this operon. The complete degradation of arabinoxylans requires endo-β-1,4-xylanases, β-xylosidases, and several accessory enzymes such as α-l-ABFs. The role of α-l-ABFs in the plant cell wall degradation is crucial, because these enzymes enhance the hydrolytic rate of endo-β-1,4-xylanases, particularly in substrates from agricultural residues such as wheat straw, corn fiber, and rice straw. α-l-ABFs are, therefore, promising tools in various processes like, for instance, the production of bioethanol . R. cellulolyticum cellulosomal α-l-ABFs are exclusively encoded by the xyl-doc genes at loci Ccel_1229, Ccel_1233, Ccel_1234, and Ccel_1240 . These enzymes are members of family GH43 and family GH62, and each carries a CBM6.
CBM6s are found in various CAZymes: arabinofuranosidases, xylanases, acetyl-xylan esterases, endoglucanases, agarases, mannanases, and glucanases [23, 41]. CBM6s interact with diverse carbohydrate targets. Some of them are able to interact with internal residues within a polysaccharide chain and have multiple binding subsites, other recognize a single sugar residue. The biochemical properties and structural analysis of the CBM6 of the Xyl-Doc enzyme GH59-1238, a putative α-galactosidase, demonstrated an interaction with a single xylose residue located at the extremity of a xylan chain or Xylp decorations present on xyloglucans . This module is unable to interact with glucose, mannose, arabinose, and galactose, and the same substrate specificity has been found for the CBM6 of the Xyl-Doc putative α-galactosidase GH27-1237 . The CBM6 modules of these two enzymes, probably involved in the degradation of galacto(gluco)mannan, share 93% of sequence identity. The primary structure of the CBM6s of the α-l-ABFs studied in this paper diverge from the two CBM6s formerly characterized (only 73% of identity) and it has been suggested by Abbott et al., that some of these modules might exhibit different binding specificities . CBM6s are often linked to GH43 catalytic modules belonging to subfamilies 15, 16, 29 and 2. The case of GH4310-1233 whose catalytic module belongs to subfamily 10, but exhibits a CBM6 is, thus, unusual. CBM6s are also not universal in GH62 enzymes, since only 2% of these enzymes are appended with a CBM6 .
The biochemical characterization of the putative Xyl-Doc α-l-ABFs showed that four are true α-l-ABFs and belong to the type-B. Indeed, type-B α-l-ABFs show activity from side chains of arabinan or arabinoxylan polysaccharides and decorated oligosaccharides . According to the biochemical data obtained and the three-dimensional structure models, we proposed a specific mode of action for each of them.
GH4316-1229, which belongs to the subfamily 16 of the GH43, encoded by the gene at locus Ccel_1229, has a very low specific activity on arabinoxylan and oat spelt xylan, and could be involved in the degradation of some particular Araf motifs located at the reducing-end extremity of the xylan backbone. We have renamed this enzyme RcAbf43A. Like Ct43Araf from Clostridium thermocellum, RcAbf43A has a strict selectivity for arabinoxylans and cannot degrade Araf decorations in arabinans . The previous proteomic and transcriptomic analyses have shown that the gene at locus Ccel_1229, the first gene of the operon, is highly expressed in corn fiber and that RcAbf43A is abundantly found in cellulosomes when natural complex plant cell wall materials are used as growth substrates [18, 19, 39]. This enzyme, despite its low activity, might play a crucial role in arabinoxylan degradation due to its particular mode of action. Its involvement may take place in the first stage of the hemicelluloses degradative process. RcAbf43A acts on Araf decorations located on chain extremities, probably more easily available than the internal decorations due to the presence of ester cross-linkage and the other surrounding polysaccharides, and may release few molecules of arabinose which in turn could act as inducers of the expression of the xyl-doc genes. Thus, XydS, the histidine kinase of the two-component system (XydS/XydR), could sense this inducer and activate the regulator XydR. The xyl-doc operon is, thus, highly expressed allowing the massive production of the Xyl-Doc enzymes for a complete degradation of arabinoxylans.
GH4310-1233 renamed RcAbf43Ad2,3 is an α-l-ABF able to convert the double Araf substitutions present on arabinoxylan and arabinan into single O2- or O3-linked decorations. The possibility to accommodate two Araf residues is consistent with the 3D model showing a shallow-binding cleft adjacent to the active site pocket. Up to now, three GH43 α-l-ABFs specific of the double Araf substitutions have been characterized, one from Bifidobacterium adolescentis (BadAbf43A, subfamily 10), one from Humicola insolens (HiAXHd3 subfamily 36), and one from Chrysosporium lucknowense C1(Abn7 subfamily 36) [27, 42,43,44]. They exclusively cleave α‐(1 → 3)‐Araf residues linked to bisubstituted Xylp. In this respect RcAbf43Ad2,3 of R. cellulolyticum has an unusual behavior, since the cleaved linkage can be either α‐(1 → 3) or α‐(1 → 2) depending upon the position of the double substitution along the xylan backbone. To our knowledge, RcAbf43Ad2,3 is the first enzyme reported to have such an activity pattern. Its specific activity on arabinoxylan and on arabinan is rather high, and despite a Km relatively high 24.8 g/L, this enzyme was found to be the most efficient R. cellulolyticum cellulosomal α-l-ABF, with a turnover value of 16.6 s−1 on arabinoxylan. Furthermore, it is active on highly decorated substrates such as WAXY-I or WAXY-RS, OSX, a less substituted substrate, arabinan, and arabinoxylooligosaccharides. This protein is massively found in cellulosomes purified from growth culture on straw . Its substrate specificity and mode of action make this enzyme essential for the degradation of both arabinoxylan and arabinan, two substrates displaying a high content in bisubstituted xylosyls.
The two cellulosomal GH62-enzymes, products of genes at loci Ccel_1234 and Ccel_1240, found in the cellulosomes of R. cellulolyticum release single Araf decorations, α-(1 → 2) or α-(1 → 3) linked, present on arabinoxylan, sugar-beet arabinan and arabinoxylooligosaccharides. Moreover, the bifunctional enzyme, the product of gene at locus Ccel_1240, contains an additional CE6 module able to cleave off acetyl substitutions present on WAXY-I, and, thus, harbors an acetyl-xylan esterase activity. Several plant cell wall polysaccharides in hardwood, softwood, or annual plants are esterified with acetic acid and carbohydrate esterases (CEs) catalyze the deacylation of these substrates and thereby facilitate the action of GHs. Presently, the CAZy database contains 16 CE families. The family CE6 is composed of only few characterized enzymes, all of them having an acetyl-xylan esterase activity. According to its mode of action, GH62-1234 was renamed RcAbf62Am2,3 and GH62-CE6-1240 was renamed RcAbf62Bm2,3Axe6. The recently characterized bifunctional enzyme from R. josui RjAbf62A-Axe6A harbors an acetyl-xylan esterase activity with a specific activity of 0.0170 ± 0.0002 IU/µmol on WAXY-I . Although the two enzymes show 91% of identity, this activity is relatively low compared to the acetyl-xylan esterase activity of the homologous R. cellulolyticum RcAbf62Bm2,3Axe6 determined to be 2.83 ± 0.21 IU/µmol. Nevertheless, the GH62 module of RjAbf62A-Axe6A is around 50-fold more active on insoluble arabinoxylan and this enzyme is, in addition, polyspecific. Indeed, the GH62 module of RjAbf62A-Axe6 also has an endoxylanase activity in addition to the α-l-ABF activity typically found in all other GH62 enzymes.
The topology of the active site of the GH62 module, of both RcAbf62Am2,3 and RcAbf62Bm2,3Axe6, modeled in this study is in good agreement with our biochemical data. The deep and narrow catalytic cavity in the center of the β-propeller is likely to accommodate only a single sugar. These two modular enzymes RcAbf62Am2,3 and RcAbf62Bm2,3Axe6 have identical N-terminal moiety (GH62-CBM6) primary structure except for one amino-acid located far away from the catalytic cavity at the end of the GH62 module. The DNA sequence of the corresponding genes also shows a very high identity (97%) suggesting a gene duplication. To utilize the broad range of vegetal substrates, cellulolytic bacteria have to adapt and diversify their enzymatic arsenal by gene duplications and/or horizontal gene transfers, thereby generating novel catalytic properties.
A synergistic action was demonstrated when RcAbf62Am2,3 and RcAbf43Ad2,3 are mixed together with arabinoxylan, leading to a 1.5-improvement. Similarly, it would be interesting to assay the synergy of the characterized α-l-ABFs with other GHs, especially endo-β-1,4-xylanases. Indeed, it was already known that substrates containing large amounts of arabinoxylan are not easily cleaved by endoxylanases without prior or simultaneous incubation with arabinofuranosidases [45, 46]. The potentially synergistic effect of these enzymes towards arabinan should also be explored.
The results of the present study have allowed us to analyze the role of the α-l-ABFs encoded by the xyl-doc cluster. They are involved in hemicellulose degradation and more particularly in arabinoxylan degradation, but also in the degradation of arabinan a component of the type I rhamnogalacturonan. Complete enzymatic degradation of arabinoxylan requires the action of debranching enzymes, also called accessory enzymes, α-l-arabinofuranosidases, feruloyl esterases, acetyl-xylan esterases, and α-l-glucuronidases. Four Xyl-Doc enzymes act as α-l-ABFs, with complementary modes of actions. Together, they can remove all the Araf decorations, both at chain extremities and within the xylan backbone (Fig. 5). Moreover, the bifunctional enzyme RcAbf62Bm2,3Axe6 removes the acetyl substituent, thereby impeding the enzymatic degradation of plant cell wall polysaccharides by GHs targeting the main chain. Arabinosyl decorations in hemicelluloses and pectins participate in the cross-linking within the plant cell wall polysaccharides and, thus, α-l-ABFs can also favor the activity of feruloyl esterase by their capacity to unpack the plant cell wall polymers . Indeed, synergistic action between α-l-ABFs and ferulyol esterases, another category of accessory enzymes, has previously been shown . A putative feruloyl esterase is encoded by a gene present in the xyl-doc cluster (locus Ccel_1232). This enzyme may also play a pivotal role in the degradation of arabinoxylans and pectins (Fig. 5). Oncoming studies of the yet uncharacterized Xyl-Doc enzymes will certainly increase our understanding of the degradative efficiency of R. cellulolyticum towards plant cell wall polysaccharides.
Availability of data and materials
All data generated or analyzed during this study are included in this published article and its additional file.
- Araf :
oat spelt xylan
insoluble wheat-flour arabinoxylan
wheat-flour arabinoxylan for reducing-sugar assays
- Xylp :
Chakdar H, Kumar M, Pandiyan K, Singh A, Nanjappan K, Kashyap PL, Srivastava AK. Bacterial xylanases: biology to biotechnology. 3 Biotech. 2016;6:150.
Dervilly-Pinel G, Thibault J-F, Saulnier L. Experimental evidence for a semi-flexible conformation for arabinoxylans. Carbohydr Res. 2001;330:365–72.
Kurakake M, Kisaka W, Ouchi K, Komaki T. Pretreatment with ammonia water for enzymatic hydrolysis of corn husk, bagasse, and switchgrass. Appl Biochem Biotechnol Part A Enzym Eng Biotechnol. 2001;90:251–9.
Huisman MMH, Brüll LP, Thomas-Oates JE, Haverkamp J, Schols HA, Voragen AGJ. The occurrence of internal (1 → 5)-linked arabinofuranose and arabinopyranose residues in arabinogalactan side chains from soybean pectic substances. Carbohydr Res. 2001;330:103–14.
Biely P, Singh S, Puchart V. Towards enzymatic breakdown of complex plant xylan structures: state of the art. Biotechnol Adv. 2016;34:1260–74.
Mechelke M, Koeck DE, Broeker J, Roessler B, Krabichler F, Schwarz WH, Zverlov VV, Liebl W. Characterization of the arabinoxylan-degrading machinery of the thermophilic bacterium Herbinix hemicellulosilytica—six new xylanases, three arabinofuranosidases and one xylosidase. J Biotechnol. 2017;257:122–30.
Linares-Pastén JA, Falck P, Albasri K, Kjellström S, Adlercreutz P, Logan DT, Karlsson EN. Three-dimensional structures and functional studies of two GH43 arabinofuranosidases from Weissella sp. strain 142 and Lactobacillus brevis. FEBS J. 2017;284:2019–36.
Wilkens C, Andersen S, Dumon C, Berrin JG, Svensson B. GH62 arabinofuranosidases: structure, function and applications. Biotechnol Adv. 2017;35:792–804.
Borsenberger V, Dornez E, Desrousseaux ML, Massou S, Tenkanen M, Courtin CM, Dumon C, O’Donohue MJ, Fauré R. A1H NMR study of the specificity of α-l-arabinofuranosidases on natural and unnatural substrates. Biochim Biophys Acta. 2014;1840:3106–14.
Koutaniemi S, Tenkanen M. Action of three GH51 and one GH54 α-arabinofuranosidases on internally and terminally located arabinofuranosyl branches. J Biotechnol. 2016;229:22–30.
Numan MT, Bhosle NB. α-l-arabinofuranosidases: the potential applications in biotechnology. J Ind Microbiol Biotechnol. 2006;33:247–60.
Petitdemange E, Caillet F, Giallo J, Gaudin C. Clostridium cellulolyticum sp. nov., a cellulolytic, mesophilic: species from decayed grass. Int J Syst Bacteriol. 1984;34:155–9.
Ravachol J, Borne R, Tardif C, de Philip P, Fierobe H-P. Characterization of all family-9 glycoside hydrolases synthesized by the cellulosome-producing bacterium Clostridium cellulolyticum. J Biol Chem. 2014;289:7335–48.
Reverbel-Leroy C, Pagès S, Bélaïch A, Bélaïch JP, Tardif C. The processive endocellulase CelF, a major component of the Clostridium cellulolyticum cellulosome: purification and characterization of the recombinant form. J Bacteriol. 1997;179:46–52.
Gal L, Pagès S, Gaudin C, Bélaïch A, Reverbel-Leroy C, Tardif C, Bélaïch JP. Characterization of the cellulolytic complex (cellulosome) produced by Clostridium cellulolyticum. Appl Environ Microbiol. 1997;63:903–9.
Maamar H, Valette O, Fierobe H-P, Bélaïch A, Bélaïch J-P, Tardif C. Cellulolysis is severely affected in Clostridium cellulolyticum strain cipCMut1. Mol Microbiol. 2004;51:589–98.
Abdou L, Boileau C, de Philip P, Pagès S, Fierobe H-P, Tardif C. Transcriptional regulation of the Clostridium cellulolyticum cip-cel operon: a complex mechanism involving a catabolite-responsive element. J Bacteriol. 2008;190:1499–506.
Celik H, Blouzard J-C, Voigt B, Becher D, Trotter V, Fierobe H-P, Tardif C, Pagès S, de Philip P. A two-component system (XydS/R) controls the expression of genes encoding CBM6-containing proteins in response to straw in Clostridium cellulolyticum. PLoS ONE. 2013;8:e56063.
Blouzard J-C, Coutinho PM, Fierobe H-P, Henrissat B, Lignon S, Tardif C, Pagès S, de Philip P. Modulation of cellulosome composition in Clostridium cellulolyticum: adaptation to the polysaccharide environment revealed by proteomic and carbohydrate-active enzyme analyses. Proteomics. 2010;10:541–54.
Ray A, Lindahl E, Wallner B. Improved model quality assessment using ProQ2. BMC Bioinf. 2012;13:224.
Kelley LA, Mezulis S, Yates CM, Wass MN, Sternberg MJE. The Phyre2 web portal for protein modeling, prediction and analysis. Nat Protoc. 2015;10:845–58.
Lombard V, Golaconda Ramulu H, Drula E, Coutinho PM, Henrissat B. The carbohydrate-active enzymes database (CAZy) in 2013. Nucleic Acids Res. 2014;42:D490–5.
Mewis K, Lenfant N, Lombard V, Henrissat B. Dividing the large glycoside hydrolase family 43 into subfamilies: a motivation for detailed enzyme characterization. Appl Environ Microbiol. 2016;82:1686–92.
Wang Y, Sakka M, Yagi H, Kaneko S, Katsuzaki H, Kunitake E, Kimura T, Sakka K. Ruminiclostridium josui Abf62A-Axe6A: a tri-functional xylanolytic enzyme exhibiting α-l-arabinofuranosidase, endoxylanase, and acetylxylan esterase activities. Enzyme Microb Technol. 2018;117:1–8.
de Sanctis D, Inácio JM, Lindley PF, de Sá-Nogueira I, Bento I. New evidence for the role of calcium in the glycosidase reaction of GH43 arabinanases. FEBS J. 2010;277:4562–74.
Brüx C, Niefind K, Ben-David A, Leon M, Shoham G, Shoham Y, Schomburg D. Crystallization and preliminary crystallographic analysis of a family 43 beta-d-xylosidase from Geobacillus stearothermophilus T-6. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2005;61(Pt 12):1054–7.
McKee LS, Peña MJ, Rogowski A, Jackson A, Lewis RJ, York WS, Krogh KB, Viksø-Nielsen A, Skjøt M, Gilbert HJ, Marles-Wright J. Introducing endo-xylanase activity into an exo-acting arabinofuranosidase that targets side chains. Proc Natl Acad Sci. 2012;109:6537–42.
Costa S, Almeida A, Castro A, Domingues L. Fusion tags for protein solubility, purification and immunogenicity in Escherichia coli: the novel Fh8 system. Front Microbiol. 2014;5:63.
de Camargo BR, Claassens NJ, Noronha EF, Kengen SWM, Quirino BF, Noronha EF, Kengen SWM. Heterologous expression and characterization of a putative glycoside hydrolase family 43 arabinofuranosidase from Clostridium thermocellum B8. Enzyme Microb Technol. 2018;109:74–83.
Till M, Goldstone D, Card G, Attwood GT, Moon CD, Arcus VL. Structural analysis of the GH43 enzyme Xsa43E from Butyrivibrio proteoclasticus. Acta Crystallogr Sect F Struct Biol Commun. 2014;70(Pt 9):1193–8.
Cartmell A, McKee LS, Peña MJ, Larsbrink J, Brumer H, Kaneko S, Ichinose H, Lewis RJ, Viksø-Nielsen A, Gilbert HJ, Marles-Wright J. The structure and function of an Arabinan-specific alpha-1,2-arabinofuranosidase identified from screening the activities of bacterial GH43 glycoside hydrolases. J Biol Chem. 2011;286:15483–95.
Fauré R, Courtin CM, Delcour JA, Dumon C, Faulds CB, Fincher GB, Fort FS, Fry GSC, Halila S, Kabel GMA, Pouvreau KL, Quemener IB, Rivet A, Saulnier GL, Schols JHA, Driguez IH. A brief and informationally rich naming system for oligosaccharide motifs of heteroxylans found in plant cell walls. Aust J Chem. 2009;62:533.
Goyal A, Ahmed S, Sharma K, Gupta V, Bule P, Alves VD, Fontes CM, Najmudin S. Molecular determinants of substrate specificity revealed by the structure of Clostridium thermocellum arabinofuranosidase 43A from glycosyl hydrolase family 43 subfamily 16. Acta Crystallogr Sect D Struct Biol. 2016;72:1281–9.
Ahmed S, Luis AS, Bras JLA, Ghosh A, Gautam S, Gupta MN, Fontes CM, Goyal A. A Novel α-l-arabinofuranosidase of family 43 glycoside hydrolase (Ct43Araf) from Clostridium thermocellum. PLoS ONE. 2013;8:e73575.
Tonozuka T, Tanaka Y, Okuyama S, Miyazaki T, Nishikawa A, Yoshida M. Structure of the catalytic domain of α-l-arabinofuranosidase from Coprinopsis cinerea, CcAbf62A, provides insights into structure–function relationships in glycoside hydrolase family 62. Appl Biochem Biotechnol. 2017;181:511–25.
Siguier B, Haon M, Nahoum V, Marcellin M, Burlet-Schiltz O, Coutinho PM, Henrissat B, Mourey L, O’Donohue MJ, Berrin JG, Tranier S, Dumon C. First structural insights into α-l-arabinofuranosidases from the two GH62 glycoside hydrolase subfamilies. J Biol Chem. 2014;289:5261–73.
Wang W, Mai-Gisondi G, Stogios PJ, Kaur A, Xu X, Cui H, Turunen O, Savchenko A, Master ER. Elucidation of the molecular basis for arabinoxylan-debranching activity of a thermostable family GH62 α-l-arabinofuranosidase from Streptomyces thermoviolaceus. Appl Environ Microbiol. 2014;80:5317–29.
Badalato N, Guillot A, Sabarly V, Dubois M, Pourette N, Pontoire B, Robert P, Bridier A, Monnet V, Sousa DZ, Durand S, Mazéas L, Buléon A, Bouchez T, Mortha G, Bize A. Whole proteome analyses on Ruminiclostridium cellulolyticum show a modulation of the cellulolysis machinery in response to cellulosic materials with subtle differences in chemical and structural properties. PLoS ONE. 2017;12:e0170524.
Xu C, Huang R, Teng L, Wang D, Hemme CL, Borovok I, He Q, Lamed R, Bayer EA, Zhou J, Xu J. Structure and regulation of the cellulose degradome in Clostridium cellulolyticum. Biotechnol Biofuels. 2013;6:73.
Freeman J, Ward JL, Kosik O, Lovegrove A, Wilkinson MD, Shewry PR, Mitchell RAC. Feruloylation and structure of arabinoxylan in wheat endosperm cell walls from RNAi lines with suppression of genes responsible for backbone synthesis and decoration. Plant Biotechnol J. 2017;15:1429–38.
Abbott DW, Ficko-Blean E, van Bueren AL, Rogowski A, Cartmell A, Coutinho PM, Henrissat B, Gilbert HJ, Boraston AB. Analysis of the structural and functional diversity of plant cell wall specific family 6 carbohydrate binding modules. Biochemistry. 2009;48:10395–404.
van den Broek LAM, Lloyd RM, Beldman G, Verdoes JC, McCleary BV, Voragen AGJ. Cloning and characterization of arabinoxylan arabinofuranohydrolase-D3 (AXHd3) from Bifidobacterium adolescentis DSM20083. Appl Microbiol Biotechnol. 2005;67:641–7.
Wang W, Andric N, Sarch C, Silva BT, Tenkanen M, Master ER. Constructing arabinofuranosidases for dual arabinoxylan debranching activity. Biotechnol Bioeng. 2018;115:41–9.
Pouvreau L, Joosten R, Hinz SWA, Gruppen H, Schols HA. Chrysosporium lucknowense C1 arabinofuranosidases are selective in releasing arabinose from either single or double substituted xylose residues in arabinoxylans. Enzyme Microb Technol. 2011;48:397–403.
Polizeli MLTM, Rizzatti ACS, Monti R, Terenzi HF, Jorge JA, Amorim DS. Xylanases from fungi: properties and industrial applications. Appl Microbiol Biotechnol. 2005;67:577–91.
de Vries RP, Kester HCM, Poulsen CH, Benen JAE, Visser J. Synergy between enzymes from Aspergillus involved in the degradation of plant cell wall polysaccharides. Carbohydr Res. 2000;327:401–10.
Koseki T, Okuda M, Sudoh S, Kizaki Y, Iwano K, Aramaki I, Matsuzawa H. Role of two α-l-arabinofuranosidases in arabinoxylan degradation and characteristics of the encoding genes from shochu koji molds, Aspergillus kawachii and Aspergillus awamori. J Biosci Bioeng. 2003;96:232–41.
The authors would like to thank Goetz Parsiegla from the laboratoire Bioénergétique et Ingénierie des Protéines (BIP) Research group: Enzymology of Supramoleculaires Systems (Aix‐Marseille Université) for helpful discussions concerning the 3D modeling. The authors would also like to thank Athel Cornish-Bowden from the laboratoire Bioénergétique et Ingénierie des Protéines (BIP) a native English speaker for the correction of the manuscript.
This research was supported by a fellowship from the Ministère de l’Enseignement Supérieur et de la Recherche to MM and by the CNRS, France.
Ethics approval and consent to participate
Consent for publication
All authors have approved biotechnology for biofuels for publication.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Mroueh, M., Aruanno, M., Borne, R. et al. The xyl-doc gene cluster of Ruminiclostridium cellulolyticum encodes GH43- and GH62-α-l-arabinofuranosidases with complementary modes of action. Biotechnol Biofuels 12, 144 (2019). https://doi.org/10.1186/s13068-019-1483-y
- Biomass degradation
- Characterization of enzymes
- Substrate specificity
- Ruminiclostridium cellulolyticum