Characterizing a Halo-Tolerant GH10 Xylanase from Roseithermus sacchariphilus Strain RA and Its CBM-Truncated Variant

A halo-thermophilic bacterium, Roseithermus sacchariphilus strain RA (previously known as Rhodothermaceae bacterium RA), was isolated from a hot spring in Langkawi, Malaysia. A complete genome analysis showed that the bacterium harbors 57 glycoside hydrolases (GHs), including a multi-domain xylanase (XynRA2). The full-length XynRA2 of 813 amino acids comprises a family 4_9 carbohydrate-binding module (CBM4_9), a family 10 glycoside hydrolase catalytic domain (GH10), and a C-terminal domain (CTD) for type IX secretion system (T9SS). This study aims to describe the biochemical properties of XynRA2 and the effects of CBM truncation on this xylanase. XynRA2 and its CBM-truncated variant (XynRA2ΔCBM) was expressed, purified, and characterized. The purified XynRA2 and XynRA2ΔCBM had an identical optimum temperature at 70 °C, but different optimum pHs of 8.5 and 6.0 respectively. Furthermore, XynRA2 retained 94% and 71% of activity at 4.0 M and 5.0 M NaCl respectively, whereas XynRA2ΔCBM showed a lower activity (79% and 54%). XynRA2 exhibited a turnover rate (kcat) of 24.8 s−1, but this was reduced by 40% for XynRA2ΔCBM. Both the xylanases hydrolyzed beechwood xylan predominantly into xylobiose, and oat-spelt xylan into a mixture of xylo-oligosaccharides (XOs). Collectively, this work suggested CBM4_9 of XynRA2 has a role in enzyme performance.

It is likely that in the wild-type R. sacchariphilus RA, XynRA2 is exported across the cytoplasmic membrane by the Sec pathway due to the presence of a signal peptide (Met 1 to Ala 33 ). In addition, XynRA2 has a CTD that enables the protein to be secreted across the outer membrane by T9SS. The T9SS protein secretion pathway is also known as Por secretion system (PorSS) [33,34] which was discovered in Porphyromonas gingivalis for secreting a potent protease gingipains [34]. Besides being identified in R. sacchariphilus RA, we noticed that in another genome sequencing project, some other annotated cellulases and hemicellulases incorporated a CTD; however, there has been little research describing the actual function of T9SS to GH enzymes. The two closest homologs of XynRA2, the Xyn10A from R. marinus and Xyl2091 from M. roseus also possessed a similar CTD [28,35]. The CTD possesses five short motifs, in which Motif B, Motif D and Motif E are important for the extensive modification by T9SS [36,37]. By aligning the CTD region of XynRA2 with other xylanases, the well-conserved Gly residues were identified in Motif B and Motif D, whereas Arg substituted the almost-conserved Lys in Motif E [37] (Figure 2c). Proteins that possessed the CTD were found to be cell-anchoring or rely on CTD for secretion, such as Xyn10A from R. marinus DSM 4252 as well as SprB, RemA, and ChiA from Flavobacterium johnsoniae [35,36]. Collectively, this suggests that XynRA2 could be a cell-anchoring enzyme. However, further experimental validation is required. It is likely that in the wild-type R. sacchariphilus RA, XynRA2 is exported across the cytoplasmic membrane by the Sec pathway due to the presence of a signal peptide (Met 1 to Ala 33 ). In addition, XynRA2 has a CTD that enables the protein to be secreted across the outer membrane by T9SS. The T9SS protein secretion pathway is also known as Por secretion system (PorSS) [33,34] which was discovered in Porphyromonas gingivalis for secreting a potent protease gingipains [34]. Besides being identified in R. sacchariphilus RA, we noticed that in another genome sequencing project, some other annotated cellulases and hemicellulases incorporated a CTD; however, there has been little research describing the actual function of T9SS to GH enzymes. The two closest homologs of XynRA2, the Xyn10A from R. marinus and Xyl2091 from M. roseus also possessed a similar CTD [28,35]. The CTD possesses five short motifs, in which Motif B, Motif D and Motif E are important for the extensive modification by T9SS [36,37]. By aligning the CTD region of XynRA2 with other xylanases, the well-conserved Gly residues were identified in Motif B and Motif D, whereas Arg substituted the almost-conserved Lys in Motif E [37] (Figure 2c). Proteins that possessed the CTD were found to be cell-anchoring or rely on CTD for secretion, such as Xyn10A from R. marinus DSM 4252 as well as SprB, RemA, and ChiA from Flavobacterium johnsoniae [35,36]. Collectively, this suggests that XynRA2 could be a cell-anchoring enzyme. However, further experimental validation is required. Based on an InterPro analysis, the CBM of XynRA2 was annotated as CBM4_9. The closest biochemically characterized xylanase (Xyn10A) from R. marinus DSM 4252 has two dissimilar CBM4_9s arranged in tandem (Figure 1b), which were denoted as "CBM4-1" and "CBM4-2" in the original article [38]. Another close homolog, a characterized xylanase (Xyl2091) from M. roseus, also possessed a CBM4_9 [28]. Interestingly, the amino acid stretch of the CBM4_9 from R. sacchariphilus RA is only 70% and 51% identical to R. marinus and M. roseus counterparts respectively, suggesting that the affinity of the three enzymes against hemicellulose might be different. Different families of CBMs such as CBM6_36 for XynG1-1 [39], CBM13 for XynAS27 [40], and dual CBM9-CBM22 for XynSL3 [24] were often reported in GH10 xylanases. According to the CAZy, other CBMs associated with xylanases are from families 1, 2, 3, 10, 15, 35, and 37. The CBM4 family from xylanases usually binds to xylan β-glucan, and/or amorphous cellulose [41,42]. We anticipated the substrate specificity of CBM4_9 in XynRA2 to be similar. Several reports have shown that the removal of the CBMs affected the biochemical properties of their partnering xylanases [39,40,43]. Therefore, we Putative structure of (a) CBM4_9 and (b) GH10 catalytic domain of XynRA2. The models were colored with the rainbow scheme (blue N-terminus, follows by green, yellow, and red C-terminus); (c) multiple sequence alignment of XynRA2 CTD with the counterpart of Xyn10A from R. marinus, Xyl2019 from M. roseus, ChiA from F. johnsoniae, as well as CTD proteins from P. gingivalis and Parabacteroides distasonis. Amino acid stretch for Motif B, D, and E are indicated by red, blue, and yellow boxes, respectively. Asterisks (*) indicate fully conserved amino acids while colon (:) indicates amino acid groups of similar properties.
Based on an InterPro analysis, the CBM of XynRA2 was annotated as CBM4_9. The closest biochemically characterized xylanase (Xyn10A) from R. marinus DSM 4252 has two dissimilar CBM4_9s arranged in tandem (Figure 1b), which were denoted as "CBM4-1" and "CBM4-2" in the original article [38]. Another close homolog, a characterized xylanase (Xyl2091) from M. roseus, also possessed a CBM4_9 [28]. Interestingly, the amino acid stretch of the CBM4_9 from R. sacchariphilus RA is only 70% and 51% identical to R. marinus and M. roseus counterparts respectively, suggesting that the affinity of the three enzymes against hemicellulose might be different. Different families of CBMs such as CBM6_36 for XynG1-1 [39], CBM13 for XynAS27 [40], and dual CBM9-CBM22 for XynSL3 [24] were often reported in GH10 xylanases. According to the CAZy, other CBMs associated with xylanases are from families 1, 2, 3, 10, 15, 35, and 37. The CBM4 family from xylanases usually binds to xylan β-glucan, and/or amorphous cellulose [41,42]. We anticipated the substrate specificity of CBM4_9 in XynRA2 to be similar. Several reports have shown that the removal of the CBMs affected the biochemical properties of their partnering xylanases [39,40,43]. Therefore, we constructed a mutant enzyme (designated as XynRA2∆CBM) by deleting the CBM4_9 but retaining the linker connecting the CBM to the catalytic domain to evaluate the effect of its truncation on the xylanase.

Expression of Recombinant XynRA2 and XynRA2∆CBM
The gene fragments encoding for mature XynRA2 (2349 bp) and XynRA2∆CBM (1857 bp) were cloned in pET28a(+), expressed in E. coli BL21 (DE3) and purified using Ni-NTA columns. The purified enzymes migrated as two distinct bands around 90 kDa and 70 kDa on SDS-PAGE, which were consistent with the theoretical molecular weight of XynRA2 (89.5 kDa) and XynRA2∆CBM (68.5 kDa) respectively (Figure 3a). constructed a mutant enzyme (designated as XynRA2ΔCBM) by deleting the CBM4_9 but retaining the linker connecting the CBM to the catalytic domain to evaluate the effect of its truncation on the xylanase.

Expression of Recombinant XynRA2 and XynRA2ΔCBM
The gene fragments encoding for mature XynRA2 (2349 bp) and XynRA2ΔCBM (1857 bp) were cloned in pET28a(+), expressed in E. coli BL21 (DE3) and purified using Ni-NTA columns. The purified enzymes migrated as two distinct bands around 90 kDa and 70 kDa on SDS-PAGE, which were consistent with the theoretical molecular weight of XynRA2 (89.5 kDa) and XynRA2ΔCBM (68.5 kDa) respectively (Figure 3a).

Effect of pH and Temperature
Using beechwood xylan as the substrate, the purified XynRA2 had maximum activity at pH 8.5 and retained a relatively high activity between pH 7−9. Truncation of the CBM broadened the pH profile (pH 5−9) with the optimum pH shifted to 6.0 (Figure 3b). Similarly, CBM4_9 truncation changed the optimum pH from 7.5 to 7.0 for a xylanase PX3 from Paenibacillus terrae HPL-003 [44]. The working pH for the mutant PX3 also narrowed to pH 5−10, while the native PX3 had an active pH ranging from 3−12. The optimum pHs of xylanase Xyn10A from R. marinus and Xyl2091 from M. roseus were 7.5 and 6.5 respectively [28,45], while that of truncated counterparts was not reported.
The optimum temperature for the activity of native XynRA2 and XynRA2ΔCBM was 70 °C. Overall, the temperature profiles for both enzymes were identical (Figure 3c). To evaluate the thermostability, XynRA2 and XynRA2ΔCBM were incubated at 70 °C without substrate for a specific interval prior to measuring residual activity. The half-life of both XynRA2 and XynRA2ΔCBM at 70 °C was approximately 45 min; however, XynRA2ΔCBM was more sensitive to prolonged temperature treatment (Figure 3d). The optimum temperatures of Xyn10A and Xyl2091 were 80 °C and 65 °C respectively and their half-lives were about 90 min (80 °C) and 160 min (60 °C), respectively. Truncation of the CBM in Xyn10A from R. marinus also resulted in a decrease in thermostability, indicating that the CBM with this xylanase also contributed to enzyme stability [46]. Truncation of the CBM from xylanases from Streptomyce rochei L10904 (Srxyn10) [43], Paenibacillus campinasensis G1-1 (XynG1-1) [39], and Streptomyces sp. S27 (XynAS27) [40] showed that removal of

Effect of pH and Temperature
Using beechwood xylan as the substrate, the purified XynRA2 had maximum activity at pH 8.5 and retained a relatively high activity between pH 7-9. Truncation of the CBM broadened the pH profile (pH 5-9) with the optimum pH shifted to 6.0 (Figure 3b). Similarly, CBM4_9 truncation changed the optimum pH from 7.5 to 7.0 for a xylanase PX3 from Paenibacillus terrae HPL-003 [44]. The working pH for the mutant PX3 also narrowed to pH 5-10, while the native PX3 had an active pH ranging from 3-12. The optimum pHs of xylanase Xyn10A from R. marinus and Xyl2091 from M. roseus were 7.5 and 6.5 respectively [28,45], while that of truncated counterparts was not reported.
The optimum temperature for the activity of native XynRA2 and XynRA2∆CBM was 70 • C. Overall, the temperature profiles for both enzymes were identical (Figure 3c). To evaluate the thermostability, XynRA2 and XynRA2∆CBM were incubated at 70 • C without substrate for a specific interval prior to measuring residual activity. The half-life of both XynRA2 and XynRA2∆CBM at 70 • C was approximately 45 min; however, XynRA2∆CBM was more sensitive to prolonged temperature treatment (Figure 3d). The optimum temperatures of Xyn10A and Xyl2091 were 80 • C and 65 • C respectively and their half-lives were about 90 min (80 • C) and 160 min (60 • C), respectively. Truncation of the CBM in Xyn10A from R. marinus also resulted in a decrease in thermostability, indicating that the CBM with this xylanase also contributed to enzyme stability [46]. Truncation of the CBM from xylanases from Streptomyce rochei L10904 (Srxyn10) [43], Paenibacillus campinasensis G1-1 (XynG1-1) [39], and Streptomyces sp. S27 (XynAS27) [40] showed that removal of the CBM did not affect the optimum temperature of xylanases. However, the truncated versions of XynG1-1 and XynAS27 displayed a significant decrease in thermostability [39,40]. In contrast, the CBM-truncated variant of Srxyn10 from S. rochei L10904 exhibited a substantial increase in thermostability at 60-70 • C despite sharing similar optimum temperature with its native counterpart [43].

NaCl Tolerance
The R. sacchariphilus RA was capable of growing in media containing a high concentration of NaCl [20]. Since XynRA2 is probably expressed as an extracellular cell-bound enzyme, we decided to investigate the effect of NaCl on xylanase activity. Multiple xylanases are known to exhibited moderate halo-tolerance, but only limited reports have demonstrated extreme halo-tolerance ability as displayed by XynRA2 (Table 1). The relative activity of XynRA2 and XynRA2∆CBM was slightly enhanced when the catalytic reactions were supplemented with 1.0 M NaCl. Notably, XynRA2 retained 94% of initial activity at 4.0 M, and 71% at 5.0 M NaCl. Although the mutant XynRA2∆CBM was more salt-sensitive, the enzyme retained the relative activity of 79% at 4.0 M and 54% at 5.0 M (Figure 3e). The reason for the lower halo-tolerance is unknown. In addition, there is a lack of literature elucidating the relationship between CBM and halo-xylanase activity.
A homology model of XynRA2 catalytic domain demonstrated a high distribution of acidic amino acids on the protein surface resulting in an overall negative electrostatic potential (Figure 4), which might explain the excellent protein stability in high NaCl concentration. Theoretically, halo-tolerant enzymes contain more acidic residues (Asp and Glu) than non-polar residues (Val, Ile, Leu, Met, and Phe). Halo-tolerant enzymes are also enriched with small residues (Ala, Val, Ser, and Thr) but lack Lys residue [47]. It has been proposed that excess acidic residues could facilitate the weakening of hydrophobicity or strengthening of hydrophilic forces on the enzyme surface, which increases water-binding capacity and prevent proteins aggregation at high salt concentration [48,49].

Enzyme Kinetics
The specific activities and the turnover rate (k cat ) of the purified XynRA2 and XynRA2∆CBM were determined by reacting the enzymes with soluble beechwood xylan. The specific activities of XynRA2 and XynRA2∆CBM were 300 U/mg and 160 U/mg respectively. The k cat of the native and mutant enzymes were 24.8 s −1 and 15.7 s −1 , respectively. We found that the truncation of CBM significantly affected the performance of the enzymes. This finding was in consistent with XynG1-1 from P. campinasensis that CBM truncation reduced the k cat by 20% [39]. Removal of CBM alone did not affect the k cat of XynAS27 from Streptomyces sp. S27. However, truncating CBM together with the linker reduced k cat value by 25% [40]. On the other hand, the xylanase variant of Srxyn10 with a CBM truncation had a three-fold higher specific activity on beechwood xylan than its native counterpart [43].

Substrate and Product Specificities
The purified XynRA2 and XynRA2∆CBM were incubated with various substrates before analyzing them using HPLC. Generally, XynRA2 and XynRA2∆CBM showed similar substrate specificities. Both enzymes were active on beechwood xylan, oat-spelt xylan, and xylo-oligosaccharides (XOs) such as X 6 , X 5 , X 4 but not on X 3 and X 2 . Except for xylose-based carbohydrates, the enzymes were unable to hydrolyze glucose-, maltose-, and arabinose-derived polymers such as carboxymethylcellulose (CMC), Avicel™, starch, pullulan, d-cellobiose, and arabinan. The results indicated that the enzymes did not possess either a cellulase or arabinase activity, suggesting that XynRA2 is a specific GH10 xylanase. This is in agreement with a recent statistical study that showed most of the characterized GH10 xylanases were mono-specific (96.8%, n = 350) towards xylanosic substrates [4].
We compared the product formation pattern of XynRA2 and XynRA2∆CBM by reacting the purified enzymes with beechwood xylan and oat-spelt xylan ( Figure 5). Upon reacting XynRA2 with beechwood xylan, the products constituted a mixture of XOs ranging X 6 , X 5 , X 4 , X 3 , and X 2 at the beginning of the reaction (15 min). After a prolonged hydrolysis (24 h), xylobiose (X 2 ) was accumulated as the primary product together with detectable X 3 and X 1 (Figure 5a), and the product formation pattern for XynRA2∆CBM against beechwood xylan was shown in Figure 5b. For reactions of 15 min and 24 h, the product profile for XynRA2∆CBM was similar to that of XynRA2. Yet, the ratio of X 3 and X 2 was slightly different in the 1 h and 3 h reactions. Previous reports on xylanase rXTMA from Thermotoga maritima and xylanase A from Caldibacillus cellulovorans also showed a variation in the profiles of XOs produced by native and CBM-depletion xylanases [55,56].
Although the same reaction conditions were used with oat-spelt xylan, we obtained lower sugar yields, probably due to the physical structure of oat-spelt xylan which consisted of both insoluble and soluble fractions. Furthermore, the product profiles were also different for beechwood xylan and oat-spelt xylan. After prolonged reaction, X 4 , X 3 , and X 2 were accumulated as the major products ( Figure 5c). Interestingly, oat-spelt xylan was a poor substrate for XynRA2∆CBM (Figure 5d), as reported for other xylanases [39,40,57,58]. A lower activity against oat-spelt xylan might be due to the inefficient binding onto the substrate, as a result of CBM truncation in XynRA2∆CBM. It has also been recurrently reported that the truncation of CBMs affects catalytic efficiency of GHs towards other insoluble substrates but not the soluble counterparts [18,19].
We compared the product formation pattern of XynRA2 and XynRA2ΔCBM by reacting the purified enzymes with beechwood xylan and oat-spelt xylan ( Figure 5). Upon reacting XynRA2 with beechwood xylan, the products constituted a mixture of XOs ranging X6, X5, X4, X3, and X2 at the beginning of the reaction (15 min). After a prolonged hydrolysis (24 h), xylobiose (X2) was accumulated as the primary product together with detectable X3 and X1 (Figure 5a), and the product formation pattern for XynRA2ΔCBM against beechwood xylan was shown in Figure 5b. For reactions of 15 min and 24 h, the product profile for XynRA2ΔCBM was similar to that of XynRA2. Yet, the ratio of X3 and X2 was slightly different in the 1 h and 3 h reactions. Previous reports on xylanase rXTMA from Thermotoga maritima and xylanase A from Caldibacillus cellulovorans also showed a variation in the profiles of XOs produced by native and CBM-depletion xylanases [55,56].

Cloning of Xylanases
Genomic DNA of R. sacchariphilus RA was extracted using DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). The gene sequence of xynRA2 was amplified from the genomic DNA using a forward primer GH10F (5 -AGCCATATGCGTGCGCAGAGCAACACCA-3 ) and a reverse primer GH10R (5 -CGATGGGTACTGGTCCGCCTCGAGCACC-3 ). The underlined sequences represent NdeI and XhoI restriction sites respectively. N-terminal signal peptide was not included in the recombinant enzymes. The truncated gene xynRA2∆CBM was amplified using primer GH10F-LC (5 -AGCCATATGCCCCTGGCGGGAGC-3 ) and GH10R.
Both the gene fragments were amplified using Q5 ® High-Fidelity PCR kit (NEB, Ipswich, MA, USA). The PCR products were digested with NdeI and XhoI followed by ligation into pET28a(+) (Novagen, Madison, WI, USA) at the corresponding sites. The recombinant plasmids (pET28a_xynRA2 and pET28a_xynRA2∆CBM) were separately transformed into E. coli BL21 (DE3) competent cells using the heat shock method. The transformants were grown in Luria-Bertani (LB) medium containing 50 µg/mL kanamycin at 37 • C for 18 h. Transformants harboring the recombinant plasmid were identified by restriction digestion and DNA sequencing.

Expression and Purification of Xylanases
The transformed cells were grown in LB medium containing 50 µg/mL kanamycin at 37 • C to an A 600 nm of 0.6. Protein expression was induced by addition of isopropyl-β-d-thiogalactopyranoside (IPTG) at a final concentration of 0.4 mM at 25 • C for 18 h. The cells were harvested by centrifugation (6000× g, 4 • C, 10 min) and lysed using B-PER™ Direct Bacterial Protein Extraction Kit (Thermo Scientific, Waltham, MA, USA) The crude enzyme was collected (12,000× g, 4 • C, 10 min) and dialyzed against 20 mM sodium phosphate buffer (pH 7.4) at 4 • C overnight in a SnakeSkin™ Dialysis Tubing 10k MWCO (Thermo Scientific, Waltham, MA, USA). To purify the His-tagged proteins, the crude enzyme was loaded onto a Ni-NTA Superflow column (Qiagen, Hilden, Germany) equilibrated with 20 mM sodium phosphate buffer (pH 7.4) and 50 mM imidazole. The enzymes were eluted with a linear gradient of 50-500 mM imidazole in 20 mM phosphate buffer (pH 7.4) containing 500 mM NaCl. Upon elution, fractions containing the XynRA2 and XynRA2∆CBM were respectively pooled and dialyzed against 20 mM sodium phosphate buffer (pH 7.4) at 4 • C overnight to remove the remaining salts. The purity and apparent molecular mass of XynRA2 and XynRA2∆CBM were validated by SDS-PAGE. The activity of the purified enzymes was assayed as described below.

Xylanase Assay
The xylanase activity of XynRA2 and XynRA2∆CBM was calculated by measuring the reducing sugars released from substrates using 3,5-dinitrosalicylic acid (DNS) method. The reaction mixtures contained 50 µL of appropriately diluted enzymes and 500 µL of 1% (w/v) beechwood xylan (Megazyme, Bray, County Wicklow, Ireland) in 0.1 M Tris-HCl buffer (pH 8.5). The enzymatic reaction was carried out at 70 • C for 15 min, stopped with 500 µL DNS reagent and boiled for 5 min. The absorbance at 540 nm was measured when the reaction mixture is cooled to room temperature. The amount of sugar released was estimated using a standard curve of d-xylose (Sigma-Aldrich, St. Louis, MO, USA). One unit (U) of xylanase activity was defined as 1 µmol of reducing sugars released from substrate per minute per mL of enzyme under the assay condition. The enzyme activity was calculated by this standard procedure unless otherwise noted. All reactions were performed in at least triplicate.

Biochemical Characterization of XynRA2 and XynRA2∆CBM
The optimum pH of XynRA2 was determined in a range of 2-11 at 50 • C. The buffers used were 0.1 M of glycine HCl (pH 2-3), sodium acetate (pH 4-6), Tris-HCl (pH 7-9), and glycine-NaOH (pH 10-11) containing 1% (w/v) purified beechwood xylan (Megazyme, Bray, County Wicklow, Ireland). The optimum temperature of the enzyme was determined over a range of temperature from 20 to 90 • C in Tris-HCl buffer (pH 8.5). The optimum pH of XynRA2∆CBM was determined at 70 • C and the optimum temperature was determined in acetate buffer (pH 6.0). Thermostability of XynRA2 and XynRA2∆CBM were determined by measuring the residual activity of the enzyme after pre-incubation in 0.1 M Tris-HCl buffer (pH 8.5) and 0.1 M acetate buffer (pH 6.0), respectively, at 70 • C without substrate for 2 h. The initial activity of enzymes without pre-incubation was set as 100%.
The effect of NaCl on the activity of XynRA2 and XynRA2∆CBM was determined at 70 • C in 0.1 M Tris-HCl buffer (pH 8.5) and 0.1 M sodium acetate buffer (pH 6.0), respectively, in the presence of up to 5.0 M NaCl.
To determine the specific activities of purified XynRA2 and XynRA2∆CBM, the enzyme activities were determined using a xylanase assay as described above, and the protein concentration was determined by Pierce™ BCA Protein Assay kit (Thermo Scientific, Waltham, MA, USA) using BSA as a standard. To determine the turnover rate (k cat ) of XynRA2 and XynRA2∆CBM, the respective enzymatic reaction was carried out at 70 • C in 0.1 M Tris-HCl buffer (pH 8.5) and 0.1 M sodium acetate buffer (pH 6.0) containing 0.1-1.5% (w/v) of purified beechwood xylan. The k cat of the enzymes were determined based on non-linear regression using PRISM 7 software (GraphPad Software Inc., San Diego, CA, USA).
To analyze the hydrolysis products of XynRA2 and XynRA2∆CBM, the reaction mixtures with 3 U of purified enzymes and 1% (w/v) beechwood xylan and oat-spelt xylan were incubated at 70 • C in Tris-HCl buffer (pH 8.5) for 24 h. The hydrolysis products were eluted using Rezex™ RSO-oligosaccharides Ag + (4%) column (Phenomenex, Torrance, CA, USA) at a flow rate of 1.0 mL/min at 80 • C for 80 min and detected using 1260 Infinity ELSD (Agilent Technologies, Santa Clara, CA, USA). Xylo-oligosaccharides (X 2 -X 6 ) and xylose were used as its product standards.

Conclusions
Roseithermus is a newly proposed genus in family Rhodothermaceae affiliated to order Rhodothermales. Currently, the whole taxonomic order is comprised of only 14 type strains. So far, xylanase from Rhodothermus marinus is the only enzyme that was well characterized. This study described for the first time the biochemical properties of a xylanase from Roseithermus. The native XynRA2 was active at alkaline pH and elevated temperature (pH 8.5 and 70 • C) while retaining excellent activity even at 5.0 M NaCl. Such properties make XynRA2 a potential candidate for applications involving an alkaline environment, elevated temperature, and high salinity. In a separate part of the study, the CBM4_9 domain was removed. The data elucidated that CBM truncation affected enzyme specific activity, turnover rate, pH optimum, and NaCl tolerance, with an additional marginal effect on thermostability.