Replacement of carbohydrate binding modules improves acetyl xylan esterase activity and its synergistic hydrolysis of different substrates with xylanase

Background Acetylation of the xylan backbone was a major obstacle to enzymatic decomposition. Removal of acetyl groups by acetyl xylan esterases (AXEs) is essential for completely enzymatic hydrolysis of xylan. Appended carbohydrate binding modules (CBMs) can promote the enzymatic deconstruction of plant cell walls by targeting and proximity effects. Fungal acetyl xylan esterases are strictly appended to cellulose-specific CBM1. It is still unclear whether xylan-specific CBMs have a greater advantage than CBM1 in potentiating the activity of fungal deacetylating enzymes and its synergistic hydrolysis of different substrates with xylanase. Results Three recombinant AXE1s fused with different xylan-specific CBMs, together with wild-type AXE1 with CBM1 and CBM1-deleted mutant AXE1dC, were constructed in this study. The optimal temperature and pH of recombinant AXE1s was 50 °C and 8.0 (except AXE1dC-CBM6), respectively. Cellulose-specific CBM1 in AXE1 obviously contributed to its catalytic action against substrates compared with AXE1dC. However, replacement of CBM1 with xylan-specific CBM4-2 significantly enhanced AXE1 thermostability and catalytic activity against soluble substrate 4-methylumbelliferyl acetate. Whereas replacements with xylan-specific CBM6 and CBM22-2 were more effective in enzymatic release of acetic acid from destarched wheat bran, NaClO2-treated wheat straw, and water-insoluble wheat arabinoxylan compared to AXE1. Moreover, replacement with CBM6 and CBM22-2 also resulted in higher degree releases of reducing sugar and acetic acid from different substrates when simultaneous hydrolysis with xylanase. A good linear relationship exists between the acetic acid and reducing sugar release. Conclusions Our findings suggested that the replacement with CBM6 and CBM22-2 not only significantly improved the catalysis efficiency of AXE1, but also increased its synergistic hydrolysis of different substrates with xylanase, indicating the significance of targeting effect in AXE1 catalysis mediated by xylan-specific CBMs. Electronic supplementary material The online version of this article (doi:10.1186/s12896-016-0305-6) contains supplementary material, which is available to authorized users.

Synergistic actions between xylanases and acetyl xylan esterases have been reported. An acetyl xylan esterase (BnaA) from Neocallimastix patriciarum showed a significant synergistic effect in combination with a recombinant xylanase (XynA) on the degradation of delignined spear grass. More reducing sugars were released by adding BnaA and XynA simultaneously than by XynA alone [8]. Tong et al. [9] found that the overall xylose yield from wheat arabinoxylan hydrolysis was 8 % with xylanase treatment and increased to 34 % when xylanase was combined with AXE from Chaetomium thermophilum. Furthermore, such enzymatic release of acetyl groups by AXE not only enhanced the solubilization of xylan to some extent and, but also increased the subsequent hydrolysis of cellulose by cellulases [10].
The bacterial cellulolytic and hemicellulolytic enzymes contain various CBMs with diverse polysaccharide recognition specificity, paralleling the considerable diversity in the target substrate of the catalytic module of the enzymes. Thus, the catalytic domains from cellulases [21,22], xylanases [23] and mannanases [24] are generally appended to cellulose, xylan and mannan-binding CBMs, respectively. Therefore, different types of CBMs can assist the appended catalytic domain targeting to a particular substrate and potentiate the catalytic rates of enzymes [25][26][27][28][29]. In contrast, the catalytic domains in fungal cellulolytic and hemicellulolytic enzymes are strictly appended to CBM1. That is, numerous enzymes that do not cleave cellulose contain CBM1 that recognize the celluolose [30,31]. CBM1 can potentiate enzyme activity of the catalytic module against hemicellulose due to its proximity effects, because in intact plant cell wall, hemicelluloses are in close association with cellulose [32].
However, this potentiation by targeting and proximity effect of CBMs might rely on the types of enzymes, and probably substrates as well [33,34]. For example, CBM15, a xylan-binding module, can potentiate the activity of Xyl11A from Neocallimastix patricarium but not that of Xyl10B from Cellvibrio mixtus [32]. The enhancements of enzyme activity by additional CBM were extensively studied in recent years, but most of them were related to the main-chain degrading enzymes such as cellulases and xylanases [35][36][37][38]. It is still unclear whether xylan-specific CBMs are more effective than CBM1 in potentiating the activity of deacetylating enzymes by targeting and proximity effects when against substrates with different structures and cellulose contents. With this end, an acetyl xylan esterase (AXE1) from Volvariella volvacea, which consisted of a family 1 catalytic domain and a CBM1 linked with serine and threonine-rich peptide, was investigated in this study [39]. Three CBMs from family 4, 6, and 22 [40][41][42], respectively, which display different specificity for xylans, were selected to replace CBM1 in AXE1. The influences of CBMs replacement on its enzyme activity and its synergistic hydrolysis of different substrates with xylanase were comparatively studied. The replacement of CBM1 in AXE1 with xylan-specific binding modules resulted in a better catalysis performance than wild-type enzyme did, which provided a rational method to design better AXEs via engineered fusion to CBMs with different binding specificities.

Recombinant enzymes
To evaluate the influences of CBMs replacement on enzyme activity and its synergistic hydrolysis for different substrates with xylanase, five recombinant AXE1s with and without CBMs were designed, that is, AXE1, a wild-type enzyme; AXE1dC, a CBM1-deleted form but containing the linker; and AXE1dC-CBM4-2/6/22-2, a fusion of AXE1dC and CBM4-2/6/22-2, respectively. Schematic representations of five rAXE1s were shown in Fig. 1.

Purification and molecular weight analysis
The recombinant proteins were purified in a one-step procedure by affinity chromatography using Ni-NTA agarose gel. The theoretical molecular weights for purified rAXE1s were estimated to be 32.3 kDa (AXE1dC), 36.4 kDa (AXE1), 46.5 kDa (AXE1dC-CBM6), 50.5 kDa (AXE1dC-CBM4-2) and 50.6 kDa (AXE1dC-CBM22-2), respectively. However, SDS-PAGE analysis revealed that the purified enzymes migrated as two bands. The two forms were attributed to different degrees of glycosylation at the two N-glycosylation sites present in the protein. After endoglycosidase H (endo H) treatment, the proteins became a single band except AXE1dC-CBM4-2. But the molecular masses of rAXE1s were still slightly higher than the estimated ones (Fig. 2). The incomplete deglycosylation of rAXE1s may be due to O-glycosylation sites located on rAXE1s, which were found by O-β-GlcNAc attachment sites during the eukaryotic protein sequence determination using the YinOYang 1.2 tool (http://www.cbs.dtu.dk/services/ YinOYang/).

Specific activity and biochemical properties of rAXE1s towards soluble substrate
The biochemical properties of rAXE1s were compared by using 4-methylumbelliferyl acetate as the substrate. Temperature optima for all rAXE1s were 50°C and pH optima for rAXE1s were pH 8.0 except AXE1dC-CBM6, which was pH 8.5 ( Fig. 3a and b). AXEdC-CBM4-2 retained almost 40 % of activity after incubation at 55°C for 90 min, in contrast, AXE1 retained only 3 %, indicating that replacement of CBM1 with CBM4-2 resulted in more than 7-fold increase of the thermostability compared to wild-type enzyme. AXE1dC-CBM22-2 also showed slightly higher thermostabilty than AXE1 ( Fig. 3c). No significant differences were found in the pH stability of these enzymes. All of the enzymes had a broad range of pH tolerance (Fig. 3d).
Thermostability improvement by fusion with CBM4-2 and CBM22-2 might benefit from their thermophilic properties. CBM4-2 from the thermophilic bacterium Rhodothermus marinus Xyn10A had an estimated denaturation temperature of 87.4°C, indicating it is a thermostable CBM [43]. CBM22-2 from Clostridium thermocellum Xyn10B had previously been described as a thermostabilizing domain on account of their apparent ability to confer thermostability to the catalytic module [44]. Thermostability improvement by fusion with thermostable CBM was previously reported for some cellulases and hemicellulases. For example, the chimeric enzyme comprising feruloyl esterase and CBM42 retained 40 % of its activity after incubation at 55°C for 30 min, whereas, the native enzyme almost lost its activity [45]. Voutilainen et al. [46] reported that the adding of CBM1, CBM2 or CBM3 to the single-module cellobiohydrolase (Cel7A) from a thermophilic fungus Talaromyces emersonii increased the unfolding temperatures from 65 to 74°C, 72 and 75°C, respectively. Thermostable enzymes offer potential benefits in the hydrolysis of substrates, leading to improvement of the overall economy of the process [47].
Furthermore, fusion of CBM4-2 significantly enhanced the specific activity, the activity of AXE1dC-CBM4-2 towards 4-methylumbelliferyl acetate, was twice as high as AXE1. Kinetic parameters analysis also showed that the V max and K cat /K m value of AXE1dC-CBM4-2 was nearly two-fold higher compared to AXE1 (Table 1). But no significant difference of the specific activity was observed between AXE1 and other mutant enzymes (Table 1). It might be due to the intrinsic property of  CBM4-2 or/and a favorable structural arrangement between the binding module and the catalytic domain [48].

Binding specificity of rAXE1s
The ligand-binding properties of CBM1, 4-2, 6 and 22-2 had been characterized previously. CBM1 bound specifically to cellulose surfaces but not to mannan and xylan [49]. CBM4-2 showed high affinity for different soluble xylan, including birch xylan, wheat arabinoxylan, oat-spelt xylan and larchwood xylan, and phosphoric-acid swollen cellulose, but to Avicel very slightly [43,50]. CBM6 had five sugar-binding sites and could accommodate highly decorated xylans but interacted weakly with soluble forms of cellulose [41,51]. CBM22-2 had a cleft harboring a binding site for xylan. This protein could bind specifically to xylooligosaccharides and various substituted xylans, especially oat spelt xylan, wheat and rye arabinoxylan [42].
However, these substrates were restricted to purified polysaccharides that had an invariant chemical structure of β-1,4-linked xylose moieties. Here, three more complex substrates, including destarched wheat bran, NaClO 2 -treated wheat straw and water-insoluble wheat arabinoxylan (inAX), together with Avicel, were used in binding experiments to examine the specificity of CBMs. Table 2 gave the composition of three lignocellulose substrates.
Compared with earlier studies with pure xylans, AXE1dC-CBM4-2, AXE1dC-CBM6 and AXE1dC-CBM22-2 displayed a less capability to bind to substrates. As shown in Fig. 4, the proteins were mainly left in unbound fraction when rAXE1s were incubated with substrates at 0°C for 1 h, except AXE1, which bound   effectively to Avicel (Fig. 4a). AXE1dC was totally unable to bind to Avicel and other substrates. AXE1dC-CBM6 and AXE1dC-CBM22-2 bound slightly more to lignocellulose substrates than Avicel. Avicel is not pure cellulose but contains also xylan, which explained the binding of xylan-specific CBM variants with Avicel. However, AXE1dC-CBM4-2 could just bind to these four insoluble substrates very slightly. AXE1 also bound weakly to destarched wheat bran, NaClO 2 -treated wheat straw and inAX, probably due to simultaneous existence of cellulose in these substrates (Fig. 4b, c, d).

Influence of different CBMs on the release of acetic acid from arabinoxylans
During the first hour, the release of acetic acid was very poor (below 3 %), but as time increased, it raised dramatically (Fig. 5). After a hydrolysis for 24 h, the cumulative release of acetic acid from inAX was the highest (over 15 %), followed by destarched wheat bran (about 14 %) and NaClO 2 -treated wheat straw (less than 10 %), although NaClO 2 -treated wheat straw was richest in acetyl groups (Table 3). Thus might be due to the low accessibility of acetyl groups in NaClO 2 -treated wheat straw. Overall, the release of acetic acid by rAXE1s in present study was much higher than that of AXE from T. reesei against hydrothermally pretreated wheat straw [10]. To some extent, the differences of pretreatment method of substrates and enzymes origins might cause the distinction of enzymatic hydrolysis yield.
As expected, all rAXE1s with CBM displayed a higher capability in release of acetic acid from destarched wheat bran, NaClO 2 -treated wheat straw and inAX than AXE1dC, indicating the potentiation effect of CBM on rAXE1s activity. Furthermore, compared to AXE1, the replacement of CBM1 with CBM6 and CBM22-2, which bind to a broad range of xylans, resulted in 10.64, 13 and 13.45 %, and 8.47, 6.11 and 5.57 % more increases in release of acetyl groups from destarched wheat bran, NaClO 2 -treated wheat straw and inAX, respectively. AXE1dC-CBM4-2 displayed similar capacity in release of acetic acid as AXE1 (Table 3). CBMs can potentiate  the appended carbohydrate-active enzyme action on lignocellulosic substrates via targeting and proximity effects depended on the sources of lignocellulosic substrates [13]. Mai-Gisondi et al. [52] investigated the ability of a family 3 cellulose-binding module (CBM3) fusion to enhance the activity of an acetylxylan esterase from Aspergillus nidulans (AnAXE, AN6093.2). It was found that the CBM3 fusion did not affect AnAXE catalytic efficiency on soluble substrates such as pnitrophenyl acetate and 4-methylumbelliferyl acetate, or corncob xylan, but increased the activity by two to four times on cellulose acetate. The elevated deacetylating capacity of AXE1 was mainly attributed to the proximity effect of CBM1 because of its cellulose-specific binding property. Thus, CBM1 increased the concentration of AXE1 in the vicinity of the target substrate.
On the other hand, the hydrolysis with AXE1dC-CBM4-2, AXE1dC-CBM6 and AXE1dC-CBM22-2 might be affected by targeting effect via assisting the enzyme attached directly to xylan. These results indicated the significance of targeting effect in AXE1 acetic acid release by xylan-specific CBMs, especially with CBM6 and CBM22-2.

Influence of different CBMs on the synergistic actions between rAXE1s and XynII
It could be seen that the combination of rAXE1s and XynII was more effective in each case than XynII alone (Fig. 6), implying the synergistic actions between rAXE1s and XynII against destarched wheat bran, NaClO2treated wheat straw and water-insoluble wheat arabinoxylan (inAX) [53]. Overall, the release of total reducing sugar from inAX was the highest, followed by that of destarched wheat bran and NaClO 2 -treated wheat straw. In general, the combination of rAXE1s with CBMs and XynII displayed a more obvious synergy in comparison to the combination of AXE1dC and XynII, implying the positive effect of CBMs on synergistic actions between rAXE1s and XynII against substrates. Moreover, AXE1dC-CBM6 and AXE1dC-CBM22-2 had a more positive   [51]. Since the xylan-specific CBMs, especially for CBM6 and CBM22-2, can target enzymes directly on the particular substrate xylan and potentiate the catalytic rates of AXE1dC-CBM6 and AXE1dC-CBM22-2, therefore increasing the efficiency of synergistic hydrolysis in comparison to AXE1 and AXE1dC. The released acetic acid was also quantified in the synergism experiments. With the same rAXE1s enzyme loading, the adding of XynII improved the release of acetic acid as a whole (Fig. 7). As for destarched wheat bran and inAX, the releases of acetic acid were increased by about 15-17 % in each interval in synergism experiments than rAXE1 alone. However, a distinction of acetic acid release pattern existed between NaClO 2treated wheat straw with destarched wheat bran and inAX. The adding of XynII dramatically accelerated the release rate of acetic acid particularly in the first 12 h when NaClO 2 -treated wheat straw was used. Specially, the simultaneous action of XynII and AXE1dC-CBM4-2 increased the release of acetic acid by 90 and 38 % as compared to AXE1dC-CBM4-2 alone at the first and sixth hour. For synergism of XynII and AXE1dC-CBM6 or AXE1dC-CBM22-2, the corresponding values were 86, 46 and 100 %, 61 %, respectively. However, after 24 h, the improvement was not distinct between AXE1dC-CBM4-2 or AXE1dC-CBM22-2 alone with corresponding synergism with XynII. Perhaps, xylanspecific CBM4-2 and CBM22-2 targeted to the xylan in earlier stages when xylan in NaClO 2 -treated wheat straw was available, leading to accelerate the release of acetic acid during the first 12 h (Fig. 7c and e). But the targeting effect faded with the removal of xylan during the last 12 h [32].
The hydrolysis products of synergistic actions between rAXE1s and XynII were determined by HPAEC-PAD. The main products were X1 and X2 with tiny of X3. X2 was much higher than X1, whereas no X4 or above was detected. As shown in Table 4, rAXE1s with CBM6 and CBM22-2 caused the highest increases in the total amount of X1, X2 and X3 than other rAXE1s when simultaneous hydrolysis with XynII. This result was consistent well with the reducing sugar determined by Somogyi method (Fig. 6). It is interesting to note the ratios of X1 and X2 released from different substrates are different. Although the hemicellulose fraction of these three lignocellulose substrates is arabinoxylan, the structure and composition of arabinoxylan are still different, which may cause the different enzymatic hydrolysis efficiency and then the ratios of X1 and X2 from different substrates.  The results are expressed as % of xylan (as xylose) in destarched wheat bran, NaClO 2 -treated wheat straw or water-insoluble wheat arabinoxylan (inAX). dC AXE1dC, 1 AXE1; 4-2 AXE1dC-CBM4-2, 6 AXE1dC-CBM6 22-2 AXE1dC-CBM22-2; X XynII A high linear correlation exists between the release of acetic acid and reducing sugar (Fig. 8). The R 2 of destarched wheat bran, NaClO 2 -treated wheat straw and inAX was 0.8493, 0.9296 and 0.9512, respectively. This suggested that the removal of acetyl groups contributed to the release of xylose from xylans. For the synergistic actions between rAXE1s and XynII, the highest enhancement was obtained in inAX due to the highest release of acetic acid as compared with destarched wheat bran and NaClO 2 -treated wheat straw. This suggested that the synergy was more pronounced in the materials containing a high accessibility of acetyl groups. These results were in good agreement with reports on the combination of endoxylanase and AXE in increasing the xylan conversion, and consequently the cellulose conversion to glucose by cellulolytic enzymes [3,54].

Conclusion
Cellulose-specific CBM1 in AXE1 obviously contributes to its catalytic action against lignocellulosic substrates compared with AXE1dC. However, the replacement of CBM1 with xylan-specific CBM4-2 significantly enhanced AXE1 thermostability and catalytic activity against soluble substrate 4-methylumbelliferyl acetate. Whereas replacement with xylan-specific CBM6 and CBM22-2 were more effective in enzymatic release of acetic acid from destarched wheat bran, NaClO 2 -treated wheat straw, and water-insoluble wheat arabinoxylan compared to AXE1. Moreover, replacement with CBM6 and CBM22-2 also resulted in higher degree releases of xylose and acetic acid from different lignocellulosic substrates when simultaneous hydrolysis with xylanase. A good linear relationship exists between acetic acid and the reducing sugar release. Our findings suggested the significance of targeting effect in AXE1 catalysis against lignocellulosic substrates mediated by xylan-specific CBMs.

Gene, strains and chemicals
The plasmid pPICZαA-axe1, which has been previously constructed to express mature AXE1 [39], was used as a template for recombinant acetyl xylan esterases construction. The gene sequences encoding xylanspecific CBM4-2, CBM6, and CBM22-2 modules were originally from Rhodothermus marinus Xyn10A, Clostridium thermocellum Xyn11A, and Clostridium thermocellum Xyn10B, respectively (Additional file 1) [40][41][42]. The corresponding gene fragments were synthesized by Springen Biotech Co. (Jiangsu, China) and carried into plasmids pUC57 with modified codon according to the codon preference of Pichia pastoris. Escherichia coli DH5α was used for plasmid construction and propagation. P. pastoris KM71H and the plasmid pPIC-ZαA from Invitrogen (Carlsbad, CA, USA) were used for heterologous expression.
The substrates 4-methylumbelliferyl acetate and Avicel were purchased from Sigma-Aldrich (St. Louis, MO). Wheat bran was obtained from Yongfang (Shandong, China) and wheat straw was obtained from Lian Yungang (Jiangsu, China). Water-insoluble wheat arabinoxylan (inAX) was purchased from Megazyme (P-WAXYI, Wicklow Ireland). TransStart FastPfu DNA polymerase and T4 DNA ligase were purchased from TransGen Biotech (Beijing, China). The restriction enzymes and endoglycosidase H were purchased from New England Biolabs (Beverly, MA, USA). All other chemicals used were analytical grade.

Expression, purification and molecular weight analysis
The plasmids were linearized using SacI and transformed into Pichia pastoris KM71H competent cells by electroporation according to the Pichia expression system manual from Invitrogen. All rAXE1s, including AXE1, were produced by methanol induction at a final concentration of 2.5 % according to the method described previously [55]. Supernatants from 25 mL cultures were collected by centrifugation (5000 g for 15 min) and then the crude enzymes were purified by affinity chromatography using Ni-NTA Agarose gel (Qiagen, Valencia, CA, USA) according to the manufacturer's manual. The enzyme homogeneity and molecular weights of purified enzymes were estimated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) [12 % (w/v)]. Purified rAXE1s were deglycosylated with endoglycosidase H and analyzed by SDS-PAGE. XynII was produced and purified as described previously [53].

Specific activity and biochemical properties of rAXE1s towards soluble substrate
The rAXE1s activity were determined spectrophotometrically at 50°C by measuring the increasing in A 354 nm during the initial 1 min of the assay resulting from the release of 4-methylumbelliferone from 4-methy lumbelliferyl acetate [56]. Reaction mixtures consisted of: 1398 μL 1 M potassium phosphate buffer (pH 8.0), 100 μL 10 mM 4-methylumbelliferyl acetate and 2 μL purified enzyme. One unit of enzyme activity (IU) was defined as the quantity of enzyme required to release 1 μmol of 4-methylumbelliferone per minute.
Optimal pH and temperature values were determined over the ranges pH 6.0-9.0 (universal buffer: 50 mM H 3 PO 4 , 50 mM CH 3 COOH, 50 mM H 3 BO 3 , pH adjusted by 0.2 M NaOH at 25°C) and 40-60°C, respectively. To determine pH stability, rAXE1s were incubated at different buffer (pH 3.0-11.0) at 4°C for 12 h. Thermostability of rAXE1s were determined by being incubated at 55°C for 0-90 min. Residual activities towards 4-methylumbelliferyl acetate were compared with control samples. The kinetic constants for rAXE1s were assayed by measuring the rate of 4-methylumbelliferyl acetate hydrolysis under the standard assay conditions using a substrate concentration range of 0.33-1.33 mM.

Substrate composition analysis
The preparation of destarched wheat bran and NaClO 2treated wheat straw was described by Wang et al. [57] and Saarnio et al. [58,59], respectively. The carbohydrate and lignin content of destarched wheat bran, NaClO 2 -treated wheat straw and inAX was determined by the standard procedure of NREL [60]. The acetic acid content was determined after alkaline hydrolysis (2 M NaOH, 4 h, 70°C) [61] and quantified by the Acetic Acid Assay Kit (Megazyme) according to the manufacturer's instructions.

Binding specificity of rAXE1s
Since the optimal pH of xylanase used in synergistic hydrolysis was 7.0, the substrate binding and hydrolysis experiment on insoluble substrates were both carried out at pH 7.0. Reaction mixtures (2 mL) containing 50 mg substrate and 250 μg purified rAXE1s in pre-cooled 100 mM potassium phosphate buffer (pH 7.0) were incubated on a rotating shaker (200 r.p.m.) at 0°C for 60 min. After incubation, samples were centrifuged (10,000 g, 4°C) for 3 min. The unbound proteins in the supernatant were concentrated by PEG6000. The precipitations were washed three times with pre-cooled 100 mM potassium phosphate buffer (pH 7.0). The unbound and bound proteins in the supernatants and precipitates, respectively, were boiled with SDS-sample buffer for 5 min and analyzed by SDS-PAGE analysis.
Influence of different CBMs on the enzymatic release of acetic acid from arabinoxylans Reaction mixtures consisted of: 0.1 g destarched wheat bran, NaClO 2 -treated wheat straw or inAX [2 % (w/v) suspension in 100 mM sodium citrate buffer, pH 7.0] with 0.005 μmol of purified enzyme in a total volume of 5 mL. Mixtures were incubated at 50°C for 1, 6, 12, 24 h with orbital shaking (150 r.p.m.) and then boiled at 99°C for 10 min. All hydrolysis experiments were carried out in duplicates. After centrifugation, the acetic acid in the supernatant was quantified using the Acetic Acid Assay Kit (Megazyme) according to the manufacturer's instructions.