Sediment samples were collected from the Qiaoquan hot spring in the Tengchong Rehai Geothermal Area, Yunnan Province (24°56′38.4″N, 98°26′24″E). Samples underwent enrichment culture for 1 month using filter paper as the sole carbon source. Total DNA was extracted using the MOBIO DNeasy PowerSoil Kit (United States) according to the manufacturer’s instructions. Metagenomic sequencing was performed by GENEWIZ (Suzhou, China) on the Illumina HiSeq 2,500 platform. De novo assembly was conducted with Velvet v1.2.08 (Zerbino and Birney, 2008). Assembled contigs were preliminarily annotated on the JGI IMG/M ER platform¹. Functional annotation of genes and open reading frames (ORFs), and domain identification, were performed using the COG (Tatusov, 2001), KEGG (Nakaya et al., 2012), and Pfam (Finn et al., 2007) databases.

Candidate xylanase genes were identified based on KEGG and COG annotations integrated with Pfam domain information, yielding two novel GH10 family xylanases, Tc15-Xyn6 and Tc15-Xyn10. Their coding sequences were optimized for Escherichia coli codon usage for subsequent cloning and expression. Gene and protein sequence homology was assessed using BLASTx and BLASTp, respectively². Signal peptides were predicted with SignalP³. Physicochemical properties were evaluated using ExPASy ProtParam. Closely related sequences from the same or neighboring genera were selected as reference sequences via NCBI BLASTp, and sequences from more distant genera were included as outgroups. Multiple sequence alignments were performed in ClustalX (Thompson et al., 1997), and a phylogenetic tree was constructed in MEGA7 using the Maximum Likelihood method with the Poisson correction model (Kumar S. et al., 2016; Kumar V. et al., 2016). Alignments were visualized with ESPript (Larkin et al., 2007), and three-dimensional homology models were built using SWISS-MODEL. Structural visualization and analysis were conducted in PyMOL.

The DNA and protein sequences of tc15-xyn6 (Accession No.: PX522158) and tc15-xyn10 (Accession No.: PX522161) have been deposited in the NCBI GenBank database. For seamless cloning, 5′ primer extensions homologous to the pSHY211 vector ends (linearized with EcoRI and HindIII) were added to each primer. The primers were: Tc15-Xyn6-F (CATCATCATCATCATCATGAA CTGGCGCTGTGTCTGCTCTCCT), Tc15-Xyn6-R (GTGCTCGAGTGCGGCCGCAAG CTCGAGCCGCACACGAACTAC), and Tc15-Xyn10-F (CATCATCATCATCATCATGAA ATGGCCCTGGCCCAGTCTG), Tc15-Xyn10-R (GTGCTCGAGTGCGGCCGCAAG GGGTCGTTGCAAGGCTTGCT). The underlined part of the sequence is the homologous sequence of the restriction endonuclease digestion vector. The underlined 5′ segments (not shown here) are homologous to the EcoRI/HindIII-linearized pSHY211 ends. The PCR program was: 95 °C for 180 s; 10 cycles of 98 °C for 20 s and 68 °C for 150 s; 30 cycles of 98 °C for 20 s, 55 °C for 30 s, and 72 °C for 150 s; final extension at 72 °C for 10 min. PCR fragments were assembled into pSHY211 using the pEASY-Uni Seamless Cloning and Assembly Kit (TransGen Biotech, China) to generate pSHY211-Tc15-Xyn6 and pSHY211-Tc15-Xyn10. Recombinant plasmids were transformed into E. coli DH5α for cloning and expression. Transformants were selected on LB agar containing kanamycin (50 μg/mL). Plasmid DNA was purified using a commercial kit (Sangon Biotech, China) and sequence-verified.

A single E. coli DH5α colony harboring the confirmed expression plasmid was inoculated into 100 mL LB medium containing 50 μg/mL kanamycin and incubated at 37 °C, 180 rpm for approximately 7 h. The temperature was then reduced to 25 °C, and incubation continued at 180 rpm for an additional 12 h to promote soluble expression. Cells were harvested by centrifugation at 4,000 × g for 20 min at 4 °C and resuspended in lysis buffer (prepared according to the manufacturer’s instructions or established protocols). Cells were disrupted by sonication on ice, and the lysate was clarified by centrifugation at 12,000 × g for 15 min at 4 °C to obtain the crude enzyme supernatant.

His-tagged recombinant proteins were purified using a Ni²⁺-chelating affinity column (TransGen Biotech, China) following the manufacturer’s protocol and as described by Yin et al. (2017). Protein concentration was determined by the Bradford assay (Sangon Biotech, Cat. No. C503031) using bovine serum albumin as the standard. Purity and apparent molecular mass were assessed by 12% SDS-PAGE.

Xylanase activity of Tc15-Xyn6 and Tc15-Xyn10 was measured following Yin et al. (2023). Reducing sugars were quantified by the 3,5-dinitrosalicylic acid (DNS) method with xylose as the standard (Miller, 1959). Unless otherwise specified, reactions contained 1% (w/v) beechwood xylan in appropriate buffer and were incubated at the enzyme’s optimal temperature under initial-rate conditions. After adding DNS reagent, mixtures were boiled for 5 min, cooled to room temperature, and absorbance was measured at 540 nm. Reducing sugar concentrations were calculated from a xylose calibration curve. One unit (U) of xylanase activity was defined as the amount of enzyme releasing 1 μmol of xylose equivalents per minute under standard assay conditions. Each assay included a heat-inactivated enzyme control and was performed with three biological replicates.

Optimal temperature and thermostability: Relative activity to temperature was measured at pH 7.0 over 20–80 °C to determine the temperature optimum. Thermostability was evaluated by preincubating enzymes at the indicated temperatures for 0, 20, 40, 60, 80, 100, and 120 min, followed by measurement of residual activity under standard conditions.

Optimal pH and pH Stability: Relative activity in response to pH was determined using two buffer systems tailored to their effective pH ranges: sodium citrate–dibasic sodium phosphate buffer (pH 3.0–8.0) and glycine-NaOH buffer (pH 8.0–11.0). For pH stability analysis, purified enzymes were incubated at 4 °Cfor 12 h and 24 h in the aforementioned buffers spanning pH 3.0–11.0 (i.e., sodium citrate-dibasic sodium phosphate buffer for pH 3.0–8.0 and glycine–NaOH buffer for pH 8.0–11.0). Residual enzyme activity was then measured under standard assay conditions.

Effects of metal ions and chemical reagents: Metal ions (K⁺, Mg²⁺, Fe³⁺, Ca²⁺, Zn²⁺, Co²⁺, Cu²⁺, Ag⁺, Mn²⁺, Pb²⁺, Ni²⁺, Ba²⁺, Cd²⁺, Al³⁺) were added individually at final concentrations of 1 mM and 10 mM. Chemical reagents (EDTA, PMSF, Tween 80, methanol, ethanol, SDS, urea, β-mercaptoethanol, CTAB, DTT, Triton X-100, and isopropanol) were added at 0.1 and 1% (w/v or v/v, as appropriate). Relative activities were determined under standard assay conditions, with reactions lacking added ions or reagents serving as controls.

Substrate specificity and kinetic parameters: Relative activity was assessed using 1% (w/v) substrates: beechwood xylan, corn cob xylan, sugarcane bagasse xylan, carboxymethylcellulose sodium (CMC-Na), Avicel, and cellobiose, under each enzyme’s optimal temperature and pH. For kinetics, substrate concentrations ranged from 0.1 to 20 mg/mL under optimal conditions. Michaelis–Menten parameters (K_m and V_max) were calculated from Lineweaver–Burk double-reciprocal plots.

To profile hydrolysis products, 1% (w/v) beechwood xylan was incubated with 10 μg of purified enzyme (Tc15-Xyn6 or Tc15-Xyn10) for 12 h under the respective optimal conditions. Aliquots were spotted onto Silica gel 60 plates (Merck, Darmstadt, Germany). Chromatography was performed with n-butanol/acetic acid/water (2:1:1, v/v/v) as the mobile phase. Plates were sprayed with 5% (v/v) H₂SO₄ in ethanol and baked at 120 °C for 10 min to visualize sugars. Xylose (X1), xylobiose (X2), xylotriose (X3), and xylotetraose (X4) served as standards.

Raw material pretreatment and xylan extraction: Sugarcane bagasse, wheat bran, wheat straw, poplar wood residue, corn stover, corn cob, rice straw, and pine wood residue were washed thoroughly, air-dried overnight, ground, and sieved to 2–10 mm. For each material, 50 g was extracted in 1.5 L of 2% (w/v) NaOH at 80 °C for 90 min. After cooling, the slurry was filtered through cotton cloth. The filtrate was adjusted to pH 5.0 with glacial acetic acid, and 95% (v/v) ethanol was added to a final concentration of 50% (v/v) to precipitate xylan. After standing overnight, the precipitate was collected by centrifugation, washed with distilled water, and dried at 50 °C to constant weight to obtain alkali-extracted xylan (Faryar et al., 2015). For high-temperature water-treated substrates, NaOH was replaced with an equal volume of water; remaining steps were unchanged.

Enzymatic hydrolysis: For each substrate, 0.2 g xylan was incubated with either 80 μg purified Tc15-Xyn6 (enzyme solution 108.69 μg/mL) or 80 μg purified Tc15-Xyn10 (enzyme solution 357.14 μg/mL) in a total volume of 5 mL pH 6.6 buffer. Reactions were mixed thoroughly and incubated at 55 °C. Samples were collected every 2 h up to 12 h, and then every 12 h up to 48 h. Supernatants were obtained by centrifugation at 10,000 rpm for 3 min, and reducing sugars were quantified by the DNS method. For each substrate, a no-enzyme blank was included. All experiments were performed with at least three independent biological replicates.

Prebiotic product preparation and monoculture evaluation: Enzymatic hydrolysates generated in Section 2.8 were used as prebiotic products. An equal volume of hydrolysate was added to 10 mL LB medium, followed by inoculation with 200 μL of a lactic acid bacterium seed culture (e.g., Lactococcus lactis). LB medium without prebiotic served as the control. Cultures were incubated at 37 °C with shaking. OD600 was recorded every 2 h up to 12 h, and then every 12 h up to 72 h to construct growth curves. All assays were performed in triplicate.

Simulated gut co-culture: Equal inocula of EGFP-labeled E. coli and lactic acid bacteria were co-cultured in LB medium with or without the prebiotic product. The experimental group contained prebiotic; the negative control contained no prebiotic. Cultures were incubated at 37 °C with shaking. OD600 and fluorescence intensity (EGFP, excitation 485 nm) were measured at the same time points as above. Each condition was performed in triplicate to assess the impact of the prebiotic on competitive dynamics and proliferation of lactic acid bacteria versus other community members.

The Tc15-Xyn6 protein was used as the receptor, with small-molecule xylooligosaccharides (XOS) X2 and X4 as docking ligands. Simultaneously, Tc15-Xyn10 protein served as the receptor, with X3 as the docking ligand for molecular docking assays. The three-dimensional (3D) structures of Tc15-Xyn6 and Tc15-Xyn10 were constructed using SWISS-MODEL, and small-molecule XOS (X2, X3, and X4) were utilized as docking ligands (Yang and Han, 2018). Enzyme-ligand docking simulations were performed using the AutoDock Vina 1.5.6 platform. The docking grid parameters applied for all XOS are provided in the Supplementary material. This grid box encompassed the entire active site cleft, with the +1 and −1 subsites located at its center. Ligand atoms possessed high degrees of freedom; during ligand parameterization, active torsion selection and torsion root identification were conducted. In this process, the rotatable and non-rotatable bonds present in the substrate molecules were evaluated. Based on the sequence identity and query coverage of the protein complexes, the ligand conformation most compatible with the protein complex was selected. The binding energy scores of the obtained docking structures were ranked. The Protein-Ligand Interaction Profiler (PLIP) was utilized to optimize the active site residues. The criterion for determining hydrogen bonds was that the maximum distance between donor and acceptor atoms should be less than 3.4 Å. To confirm the interactions between the enzyme and substrate, visual analysis was performed using PyMOL.

Unless otherwise specified, all experiments were conducted with three biological replicates, and mean values were used for analysis. Statistical analyses were performed in SPSS 20.0. Data are reported as mean ± standard error of the mean (SEM). One-way ANOVA was used, followed by Tukey’s post hoc test for multiple comparisons. In all cases, p < 0.05 was considered statistically significant.

Metagenomic sequencing and mining of enriched cultures and total DNA from Qiaoquan hot spring sediments led to the identification of two putative xylanase genes, Tc15-Xyn6 and Tc15-Xyn10. Tc15-Xyn6 is 1,212 bp in length and encodes a 400-amino-acid protein. A signal peptide was predicted at residues 1–21, and the GH10 catalytic domain was mapped to residues 104–354. Tc15-Xyn10 is 1,305 bp and encodes a 315-amino-acid protein with a GH10 domain spanning residues 53–315; SignalP did not detect a typical N-terminal signal peptide. Theoretical molecular masses were 44.64 kDa (Tc15-Xyn6) and 35.84 kDa (Tc15-Xyn10), with predicted pI values of 5.96 and 7.77, respectively. BLAST and phylogenetic analyses showed that Tc15-Xyn6 clusters with a Bryobacteraceae bacterium xylanase (accession MCS7041942.1; 87.37% identity), while Tc15-Xyn10 clusters with a Deinococcota bacterium xylanase (accession RMH54196.1; 99.37% identity) (Figure 1). Homology modeling indicated that both proteins adopt the canonical (α/β)8 TIM-barrel fold characteristic of GH10 enzymes (Figure 2).

Tc15-Xyn6 and Tc15-Xyn10 were cloned into pSHY211 as His-tag fusions and verified by sequencing. Following Ni²⁺-NTA affinity purification, 12% SDS-PAGE revealed single bands at approximately 44.64 kDa (Tc15-Xyn6) and 35.84 kDa (Tc15-Xyn10), consistent with theoretical masses and indicating successful expression and purification (Figure 3).

At both 1 mM and 10 mM, Ag⁺ strongly inhibited both enzymes, whereas Co²⁺ showed an activating effect. For Tc15-Xyn6, activity increased in the presence of K⁺, but 10 mM Cu²⁺ reduced relative activity to ~10%; most other ions had weak effects. For Tc15-Xyn10, relative activity decreased to ~34.55% with 10 mM Fe²⁺ and ~28.26% with 10 mM Zn²⁺, while inhibition by most other ions was <50%. Among chemical reagents, SDS, β-mercaptoethanol, and PMSF significantly inhibited both enzymes. Tween 80 and DMSO showed activation, and Triton X-100 also activated Tc15-Xyn10. At 10% (v/v), isopropanol or DTT reduced Tc15-Xyn10 activity to <50%, and CTAB strongly inhibited activity. Notably, both enzymes retained >50% activity in 10% (v/v) methanol or ethanol, indicating a degree of alcohol tolerance (Figure 6).

Substrate specificity was evaluated using multiple polysaccharides (Table 1). Tc15-Xyn6 showed a specific activity of 90.80 μmol/min/mg on beechwood xylan, and 46.84, 18.06, and 30.51 μmol/min/mg on corncob xylan, bagasse xylan, and CMC-Na, respectively. No activity was detected on Avicel or cellobiose. Tc15-Xyn10 exhibited 34.91 μmol/min/mg on beechwood xylan, and 14.68, 6.75, and 8.26 μmol/min/mg on corncob xylan, bagasse xylan, and Avicel, respectively; no activity was detected on CMC-Na or cellobiose. Thus, both enzymes preferentially hydrolyze xylan substrates from various sources, with minimal or no activity on cellulosic substrates and cellobiose.

Kinetic analysis with beechwood xylan as substrate showed that Tc15-Xyn6 had K_m = 4.675 mg/mL, V_max = 125 μmol/min/mg, and K_cat = 93 s⁻¹, whereas Tc15-Xyn10 had K_m = 9.36 mg/mL, V_max = 59.52 μmol/min/mg, and K_cat = 35.55 s⁻¹ (Table 2). These results indicate higher substrate affinity and catalytic efficiency for Tc15-Xyn6 relative to Tc15-Xyn10.

Using beechwood xylan as substrate, TLC revealed that Tc15-Xyn6 primarily produced xylobiose (X2) and xylotetraose (X4), whereas Tc15-Xyn10 predominantly generated xylotriose (X3) (Figure 7). No corresponding oligosaccharide bands were observed in inactivated enzyme controls or substrate blanks, confirming the enzymatic origin of the products.

Both enzymes yielded low reducing sugar levels from poplar and pine residues. Among eight hot water–pretreated substrates, most produced <2 μmol/mL reducing sugars, with wheat bran exceeding 4 μmol/mL. By contrast, among eight hot alkali–pretreated substrates, wheat straw, rice straw, and corn stover each produced >2 μmol/mL (data not shown). Based on these observations, hot water–pretreated wheat bran and hot alkali–pretreated wheat straw, rice straw, and corn stover were selected for 48 h time-course analysis.

For Tc15-Xyn6, reducing sugar production showed similar trends across the four substrates: prior to 36 h, wheat straw and corn stover yielded comparable, higher sugar levels than rice straw and wheat bran. At 36 h, wheat straw and wheat bran reached peak yields and then slightly declined, while rice straw and corn stover continued to increase (Figure 8A). For Tc15-Xyn10, prior to 36 h, sugar yields from rice straw, corn stover, and hot water–pretreated wheat bran were all lower than from wheat straw. By 48 h, yields from wheat bran and rice straw tended to stabilize, while those from corn stover and wheat straw continued to rise (Figure 8B). These results suggest that hot alkali–pretreated wheat straw and hot water–pretreated wheat bran are relatively suitable direct hydrolysis substrates for both enzymes.

Further TLC analysis of 48 h hydrolysates from the four representative substrates (hot alkali–pretreated wheat straw, rice straw, and corn stover; hot water–pretreated wheat bran) revealed xylooligosaccharide bands of higher degrees of polymerization for both enzymes, with no cellooligosaccharide (G2–G4) bands detected (Figure 9). This indicates no significant cellulolytic activity under the tested conditions.

Xylan hydrolysates generated by Tc15-Xyn6 and Tc15-Xyn10 promoted the growth of Lactococcus lactis. In the control without prebiotics, L. lactis displayed a typical growth pattern: a lag phase during the first 12 h, entry into exponential growth between 12 and 36 h, and stabilization thereafter (Figures 10A,B). In cultures supplemented with Tc15-Xyn6-derived prebiotics, growth accelerated between 12 and 24 h, and biomass stabilized at higher OD600 values from 24 to 48 h relative to the control. With Tc15-Xyn10-derived prebiotics, L. lactis exhibited a sustained increase over 12–72 h, with a pronounced acceleration between 12 and 36 h, followed by a moderate rise from 36 to 72 h (Figures 10A,B).

Co-culture assays with EGFP-labeled E. coli and L. lactis NZ9000 were used to simulate gut-like competitive dynamics. In the control, EGFP fluorescence increased continuously, indicating progressive E. coli expansion and concomitant suppression of lactic acid bacteria. In contrast, Tc15-Xyn6 prebiotics produced a marked increase in total biomass (OD600) and a substantially lower fluorescence signal than the control, consistent with a higher L. lactis abundance and reduced E. coli proliferation. Tc15-Xyn10 prebiotics similarly elevated OD600 while lowering EGFP fluorescence relative to the control, indicating increased lactic acid bacteria counts and reduced E. coli abundance (Figures 10C,D). Collectively, these results demonstrate that Tc15-Xyn6- and Tc15-Xyn10-derived xylan hydrolysates act as prebiotics that selectively promote lactic acid bacteria growth and favorably modulate competitive community dynamics.

From a deployment perspective, temperature and pH optima critically determine enzyme operability and cost effectiveness on complex, partially pretreated substrates and under non-sterile conditions. While many reported xylanases exhibit optima between 30 and 60 °C and pH 4.0–9.0 (Dhiman and Mukherjee, 2018), representative neutral-preferring, moderately thermophilic enzymes such as AWS-2X, PW-xyl9, and PW-xyl37 operate at 55–60 °C (Wang et al., 2017, 2021). Tc15-Xyn6 and Tc15-Xyn10 show superior stability in the 55–65 °C range: Tc15-Xyn6 has a half-life of ~2 h at 65 °C, and Tc15-Xyn10 has a half-life of ~1 h at 60 °C, with substantial activity retention at slightly lower temperatures. These properties are advantageous for XOS production: operation at 55–60 °C improves xylan solubility and mass transfer, reduces viscosity, and suppresses microbial contamination, enabling stable continuous or semi-continuous processes without stringent sterilization. Tc15-Xyn6’s tolerance to short exposures at even higher temperatures provides flexibility for intensified hydrolysis with high substrate loading and short residence times.

Both enzymes exhibit neutral pH preference and broad pH tolerance. The optimal pH values were 6.6 (Tc15-Xyn6) and 6.0 (Tc15-Xyn10), with Tc15-Xyn6 retaining >60% activity across pH 4.0–10.0 and Tc15-Xyn10 retaining >50% activity across pH 5.0–10.0. This broad “stability window,” reported for several GH10 xylanases (e.g., Xyl10B, Xyn30Y5-SLH, Af-XYLA) (Bernardi et al., 2021; Lai et al., 2021; Mon et al., 2022), is valuable for food-grade XOS production. Operating near neutral pH (5–7) minimizes acid/base consumption and neutralization costs, mitigates sugar degradation (e.g., furfural formation), and facilitates integration with mild upstream pretreatments (hot water or dilute alkali).

Both enzymes preferentially hydrolyzed xylan from multiple sources and showed negligible activity on Avicel (Tc15-Xyn6) or CMC-Na (Tc15-Xyn10), with no activity detected on cellobiose. On beechwood xylan, Tc15-Xyn6 exhibited higher catalytic efficiency and affinity (K_m 4.675 mg/mL; V_max 125 μmol/min/mg; K_cat 93 s⁻¹) than Tc15-Xyn10 (K_m 9.36 mg/mL; V_max 59.52 μmol/min/mg; K_cat 35.55 s⁻¹). The study revealed that the K_m and V_max values of CrXyn toward beechwood xylan were 5.98 g/L and 179.9 μmol/min/mg, respectively, while those of XynB for the same substrate were 5.82 mg/mL and 380 μmol/min/mg (Zhu et al., 2022; Guo et al., 2013). The K_m value of Tc15-Xyn6 was essentially comparable to those of the aforementioned enzymes, although XynB exhibited a higher V_max; overall, Tc15-Xyn6 displayed substrate affinity similar to that of other reported xylanases. For BhS7Xyl, its K_m and V_max toward beechwood xylan were determined to be 32.1 mg/mL and 41.9 μmol/min/mg. Tc15-Xyn10 showed superior overall catalytic properties compared to BhS7Xyl but was inferior to Tc15-Xyn6. Xiao characterized the polysaccharide CVP-2 derived from Christia vespertilionis and demonstrated that its branched chain structures (e.g., α-l-arabinofuranosyl branches) can enhance its binding capacity to biological substrates, while the stable backbone structure (→5-α-l-arabinofuranosyl) ensures structural stability during the catalytic process (Xiao et al., 2025). It is speculated that the K_m values and V_max of Tc15-Xyn6 and Tc15-Xyn10 toward xylan may be associated with similar branched chain structural characteristics and structural stability.

TLC profiles revealed enzyme-specific XOS signatures: Tc15-Xyn6 predominantly yielded xylobiose (X2) and xylotetraose (X4), whereas Tc15-Xyn10 produced xylotriose (X3). Such product selectivity is important for tailoring XOS composition to target microbial beneficiaries and functional outcomes in feeds or functional foods.

Hydrolysis of pretreated substrates showed that hot alkali–pretreated wheat straw and hot water–pretreated wheat bran produced the highest sugar yields across time courses, supporting their use as practical feedstocks. The absence of detectable cellooligosaccharides by TLC suggests minimal cellulase activity under test conditions, which is desirable for XOS-centric processes that aim to limit glucose release and downstream interference.

Crucially, the resulting XOS-rich hydrolysates significantly promoted L. lactis growth and shifted co-culture dynamics toward lactic acid bacteria dominance by suppressing E. coli expansion. These results corroborate the prebiotic potential of Tc15-derived XOS and highlight their suitability for functional feed applications that aim to support gut health and reduce opportunistic pathogens.

The influence of metal ions and chemical additives on enzyme activity provides essential guidance for process design and upstream media purification. Both Tc15-Xyn6 and Tc15-Xyn10 were strongly inhibited by Ag⁺, consistent with the broad-spectrum antimicrobial and protein-denaturing properties of silver ions (Qadeer et al., 2024). It is plausible that Ag⁺ interacts with acid/base catalytic residues in the active site or with structural elements critical for proper folding, inducing conformational perturbations that hinder substrate binding. Although Ag⁺ contamination is uncommon in food-grade or conventional biomanufacturing systems, care should be taken to avoid contact with silver-containing antimicrobial materials or disinfectants; if necessary, trace heavy metals should be removed via water pretreatment, chelation, or adsorption.

SDS also strongly inhibited both enzymes, consistent with its function as an anionic denaturant that disrupts native protein conformations (Parker and Song, 1992; Srivastava and Alam, 2020; Krishnamani et al., 2012; Lefebvre-Cases et al., 2001; Minai-Tehrani et al., 2012). While SDS is unlikely to be present in food-grade processes, these results underscore the need to avoid contamination by strong surfactants and detergents. Notably, both xylanases retained more than 50% of their activity in 10% (v/v) methanol or ethanol, demonstrating tolerance to polar small-molecule alcohols (Benatti and Polizeli, 2023). This compatibility facilitates integration with solvent-based pretreatments (e.g., organosolv delignification or hemicellulose separation) and alcohol-containing steps (e.g., ethanol washing for lignin removal, or residual ethanol at the end of fermentation), and supports cross-application in alcohol-involving industries such as papermaking and biofuels.

Enzyme–ligand docking was performed using AutoDock Tools 1.5.6 (ADT) and visualized in PyMOL (Ellilä et al., 2019). The results provide structural rationales for the observed product selectivity and stability profiles. For Tc15-Xyn6, minimum binding energies were −6.7 kcal/mol for xylobiose (X2) and −8.1 kcal/mol for xylotetraose (X4). Hydrogen-bonding interactions for X2 involved Lys84, His113, Glu164, Gln237, Glu267, and Arg310; for X4 they involved Lys84, Asn202, Ala206, His239, and Arg310. For Tc15-Xyn10, the minimum binding energy for xylotriose (X3) was −8.7 kcal/mol, with hydrogen bonds primarily mediated by Glu50, Asn51, Lys54, His87, Glu135, Tyr176, Gln208, Glu240, and Trp281 (Figure 11). Consistent with Hassan et al. (2023), biomolecular structural stability is closely linked to the strength and organization of intramolecular interactions. Covalent disulfide bonds can rigidify protein conformations, while strengthened hydrophobic packing limits exposure of nonpolar cores and mitigates heat-induced unfolding. In the present study, in silico analyses predict that extensive hydrogen-bond networks—together with complementary hydrophobic contacts and potential salt bridges—attenuate thermal denaturation, providing a plausible molecular basis for the observed thermostability of the xylanases. Dense hydrogen-bond networks are associated with increased protein rigidity and enhanced conformational stability at elevated temperatures (Günther et al., 2023; Su et al., 2021). We hypothesize that the number of hydrogen bonds may affect the structural stability of the enzyme, resulting in differences in the thermotolerance of xylanases. Temperature tolerance experiments demonstrated that Tc15-Xyn10 maintained more than 80% activity after 6 h at 55 °C, whereas Tc15-Xyn6 decreased to approximately 36% after 6 h at 55 °C (Supplementary Figure S1). Molecular docking suggests the number of hydrogen bonds formed in Tc15-Xyn10 is 4/3 times that in Tc15-Xyn6. Tc15-Xyn10 forms a greater number of hydrogen bonds relative to Tc15-Xyn6, consistent with the observed long-term stability of Tc15-Xyn10 at 55 °C. Conversely, Tc15-Xyn6 exhibits superior short-term thermostability at higher temperatures, aligning with its kinetic performance and product profile. Both temperature tolerance experiments and molecular docking have indirectly verified that the number of hydrogen bonds influences the thermotolerance of xylanases; however, to fully clarify the role of hydrogen bonds, further validation through molecular dynamics simulations and mutagenesis experiments is required.

The dense architecture of lignocellulose and tight cross-linking between lignin and hemicellulose limit xylanase accessibility and hydrolysis efficiency. Appropriate pretreatment is therefore essential (Ramesh et al., 2018; Alawad and Ibrahim, 2022). In this study, wheat bran, wheat straw, rice straw, and corn stover were subjected to high-temperature (HT) and high-temperature alkaline (HTA) pretreatments. Results were consistent with established consensus: alkaline pretreatment promotes delignification, cleaves ester bonds and acetyl groups, and increases porosity through cell wall swelling, thereby improving hemicellulose accessibility and xylanase performance (Wu X. et al., 2022; Wu Z. et al., 2022). Compared with high-temperature water-only pretreatment, HTA substrates exhibited looser structures that enabled Tc15-Xyn6 and Tc15-Xyn10 to release more reducing sugars. Both enzymes also showed excellent hydrolysis on high-temperature water-pretreated wheat bran, suggesting that for raw materials rich in soluble or weakly cross-linked xylan (e.g., wheat bran), simple HT pretreatment can provide sufficient accessibility for neutral xylanases.

Relative to reported enzymatic XOS production from straw (e.g., XynB for garlic straw, Xyln-SH1 for wheat straw, MxynB-8 for poplar sawdust, as detailed in Table 3) Tc15-Xyn6 and Tc15-Xyn10 offer three practical advantages: (1) Neutral pH optima with moderate-to-high temperature tolerance, facilitating integration with mild pretreatments and reducing neutralization and cooling costs. (2) Alcohol tolerance, enhancing compatibility with alcohol-containing pretreatments and solvent recycling. (3) Product spectra dominated by X2–X4, aligning with desirable prebiotic XOS profiles.

It is important to emphasize that straw type, pretreatment conditions, and solids loading strongly influence hydrolysis. Future work should systematically evaluate yield, product selectivity, and energy consumption at higher solids (>10–15% w/w) and under continuous or semi-continuous operation to ensure comparability with published processes and identify optimal operating windows.

Xylanase treatment mitigates anti-nutritional effects of dietary xylan, disrupts cell wall structures, and frees encapsulated starch, proteins, and micronutrients, thereby enhancing digestibility in monogastric animals (Adebowale, 1985; Zhang et al., 2012; Bimrew, 2014). Wheat bran, a major feed byproduct, is rich in NSPs—especially xylan—that increase digesta viscosity, hinder enzyme–substrate contact, and impair nutrient absorption (Swiatkiewicz et al., 2015). Tc15-Xyn6 and Tc15-Xyn10 showed strong hydrolysis of HT-treated wheat bran, highlighting their potential for targeted enzymatic modification. Two application strategies emerge: (1) In-plant enzymatic pretreatment to partially hydrolyze bran, yielding “functionalized wheat bran” or blended liquid/wet feed ingredients with other byproducts. (2) Direct inclusion in diets, leveraging gastrointestinal pH (5–7) for optimal activity. Given transient high temperatures during pelleting, the thermostability of Tc15-Xyn6 and Tc15-Xyn10 at 60–65 °C is advantageous; further tolerance can be pursued via encapsulation or carrier immobilization. Improved digestibility and nutrient release, coupled with reduced excretion of undigested material, are expected to lower environmental burdens in livestock and poultry production (Lian et al., 2024), promoting high-value utilization and a circular economy for straw byproducts.

Tc15-Xyn6 primarily produced X2 and X4, while Tc15-Xyn10 predominantly yielded X3; products were concentrated in the X2–X4 range. XOS within this DP range exhibit typical prebiotic traits—low effective dosage, high selectivity—and are efficiently utilized by Lactobacillus and Bifidobacterium (Maria et al., 2014). The XOS generated here promoted L. lactis growth and suppressed E. coli, consistent with reports that dietary XOS produced in situ by xylanases favor beneficial bacteria (e.g., Streptococcus lactis, Lactobacillus bulgaricus) and inhibit opportunistic pathogens (e.g., Klebsiella pneumoniae) (Ding et al., 2018; Kim et al., 2011; Yadav and Jha, 2019; Ramatsui et al., 2024). In poultry, in ovo injection of X2 and X3 improves growth and intestinal morphology (Palaniappan et al., 2021). In mice, XOS alleviated diarrhea via gut microbiota modulation (Singh et al., 2015). In human nutrition, XOS are incorporated into functional foods (e.g., dairy, baked goods) with reported immunomodulatory, preservative, and antioxidant effects (Prete et al., 2021; Kumar et al., 2025; Das et al., 2024). Collectively, these findings support a conversion pathway from raw biomass to enzymatic hydrolysate to XOS to functional products, with Tc15-Xyn6 and Tc15-Xyn10 as enabling biocatalysts.