Heterologous expression and characterization of two novel glucanases derived from sheep rumen microbiota

β-Glucanases are a suite of glycoside hydrolases that depolymerize β-glucan into cellooligosaccharides and/or monosaccharides and have been widely used as feed additives in livestock. In this study, two novel glucanase genes, IDSGluc5-26 and IDSGluc5-37, derived from sheep rumen microbiota, were expressed and functionally characterized. The optimal temperatures/pH of recombinant IDSGLUC5-26 and IDSGLUC5-37 were 50 °C/5.0 and 40 °C/6.0, respectively. Notably, IDSGLUC5-26 showed considerable stability under acidic conditions. Both IDSGLUC5-26 and IDSGLUC5-37 showed the highest activities toward barley β-glucan, with Vmax values of 89.96 ± 9.19 µmol/min/mg and 459.50 ± 25.02 µmol/min/mg, respectively. Additionally, these two glucanases demonstrated hydrolysis of Icelandic moss lichenan and konjac gum, IDSGLUC5-26 releasing cellobiose (G2; occupying 17.37% of total reducing sugars), cellotriose (G3; 23.97%), and cellotetraose (G4; 30.93%) from barley β-glucan and Icelandic moss lichenan after 10 min and suggestive of a typical endo-β-1,4-glucanase (EC.3.2.1.4). In contrast, IDSGLUC5-37 was capable of liberating dominant G3 (64.11% or 67.55%) from barley β-glucan or Icelandic moss lichenan, suggesting that the enzyme was likely an endo-β-1,3 − 1,4-glucanases/lichenase (EC3.2.1.73). These findings describe the expression and characterization of two novel glucanase genes from sheep rumen microbiota. The two recombinant enzymes, particularly the acid-stable IDSGLUC5-26, will be of interest for potential application in food-/feed-additive development.

Non-starch polysaccharides (NSPs), including β-glucan, xylan, mannan, pectin, and cellulose, are normally regarded as antinutritional factors in plant-derived feedstuffs. Due to the lack of NSP-degrading enzymes in the digestive tract, single-stomach animals, such as pigs and chickens or even young ruminants with incomplete gastrointestinal function, cannot effectively digest NSP-like feedstuffs.
Consequently, increased chyme viscosity directly interferes with the contact between digestive enzymes and their substrates, leading to decreased feed utilization (Beckmann et al. 2006; Romero et al. 2016).
Glucanases, particularly endo-β-glucanase, play critical roles in breaking down plant-derived diets. Feed diets supplemented with glucanases reduce the viscosity of digestive-tract contents, improve the morphological structure of the intestine, and promote growth performance of animals. Previous studies reported that glucanase improved the conjugated bile acid fraction in intestinal contents and increased the villus length of the small intestine wall and nutrient digestion in broiler chickens fed a rye-based diet (Mathlouthi et al. 2002;Yu et al. 2002). In recent decades, this enzyme has been used in feed additives and contributed greatly to animal husbandry production. β-Glucanase is widely found in bacteria, fungi, algae, higher plants, and insects, with a majority of commercial glucanases obtained from microorganisms, especially bacteria and fungi. The main glucanase sources in bacteria include Bacillus spp., such as Bacillus subtilis (Qiao et al. 2009 (Chen et al. 2014) are common fungi that produce βglucanases. Furthermore, a large number of rumen microorganisms in the rumen of cattle, sheep, and other ruminants are also believed to produce abundant CAZymes, including β-glucanase ). For example, glucanases from Streptococcus bovis (Ekinci et al. 1997), Clostridium thermocellum (Costa et al. 2014), and Ruminococcus avofaciens (Mondal et al. 2021) have been partially characterized. However, due to the required long-term adaptation to the complicated and restricted growth environment of the rumen, most rumen microorganisms cannot be purely cultured, making it di cult to directly obtain β-glucanase from rumen using conventional pure-culture methods.
Phosphoric acid-swollen cellulose (PASC) was prepared from Avicel treated with 85% phosphoric acid according to a previously described method (Zhang et al. 2006).
Phylogeny analysis was performed using the Maximum Likelihood statistical method based on the WAG correction model. The test to assess the phylogeny used was performed by the bootstrap method with 1000 bootstrap replications.

Expression and puri cation of recombinant enzymes
Recombinant E. coli strains BL21/pET28a/IDSGluc5-26 and BL21/pET28a/IDSGluc5-37 were cultured, induced with IPTG, and sonicated as previously described (Sun et al. 2019). Crude enzyme was obtained by centrifugation at 12,000×g, 4°C for 10 min, and the supernatant were subjected to a nity puri cation using a HisTrap FF column (GE Healthcare Biosciences, Pittsburgh, PA, USA). The imidazole gradient in the elution buffer ranged from 20 mM to 300 mM. Puri ed IDSGLUC5-26 and IDSGLUC5-37 were analyzed by SDS-PAGE (12% running gel and 4% stacking gel) (Laemmli. 1970). After electrophoresis, the gel was stained with Coomassie Brilliant Blue G250 and destained with 15% methanol and 5% acetic acid. The substrate activities of enzymes were analyzed using agar plates containing 0.2% (w/v) barely β-glucan, lichenan, and konjac gum. Proteins (~20 U) were dotted on agar plates, incubated at 25°C for 16 h, followed by staining with 0.1% (w/v) Congo Red for 20 min, solution and destaining with 1 mol/L NaCl until transparent bands appeared.

Circular dichroism
Circular dichroism analysis was conducted using a JASCO CD spectrophotometer under constant N 2 ush (J-1500, Japan) and a 0.1 cm path-length cuvette at wavelengths ranging from 190 to 260 nm. The protein concentration was 0.2 mg/mL in sterile ddH 2 O.
After cooling to room temperature, absorbance was determined spectrophotometrically at 540 nm. One unit (U) of glucanase activity was de ned as the amount of enzyme that released 1 µmol of reducing sugar equivalent to glucose per min. Enzyme concentration was measured by the Bradford method (Bradford. 1976) and using bovine serum albumin as the standard. Approximately 12 U of puri ed enzyme was used for all experiments unless otherwise noted. All assays were performed in quadruplicate.
TLC was conducted using a silica gel-coated aluminum plate (Merck, Darmstadt, Germany) with a solvent mixture of n-butanol/acetic acid/water (2:1:1, v/v/v). Developed sugars were visualized after spraying with visualization reagent (sulfuric acid/ethanol = 5:95, v/v). Plates were then air-dried in a hood and heated for 10 min at 100°C to develop the chromatogram.

Expression and enzymatic properties of recombinant glucanases
Two E. coli transformants harboring IDSGluc5-26 and IDSGluc5-37 were induced at 16°C with 1 mmol/L IPTG for 16 h. After 6 × His-tagged a nity puri cation, we observed two distinguished bands of ~33 kDa and ~75 kDa ( Fig. 2a and b), which were consistent with their theoretical molecular weights, respectively. Circular dichroism analysis demonstrated that both IDSGLUC5-26 and IDSGLUC5-37 possess α-helix with typical double minima in ellipticity at 208 and 222 nm (Fig. 2c), which was consistent with homology modeling structure ( Fig. S1 and S2). Both recombinant IDSGLUC5-26 and IDSGLUC5-37 exhibited hydrolytic activities against barley β-glucan, Icelandic moss lichenan, and konjac gum (Fig. 2d-f) on plates. The optimal reaction temperatures of IDSGLUC5-26 and IDSGLUC5-37 were 50°C and 40°C, respectively ( Fig. 3a and b), and the optimal pH values were 5.0 and 6.0, respectively ( Fig. 3c and d). Both enzymes were relatively stable at < 40°C and maintained a majority of their activities after preincubation for 1 h at 40°C (Fig. 4a and b). However, both enzymes degenerated rapidly at temperatures > 50°C, with neither IDSGLUC5-26 nor IDSGLUC5-37 active after preincubation for 40 min at 50°C. IDSGLUC5-26 was stable at pH ranges of ~4.0 to ~8.0 (retained >70% residual activity) after pretreatment at the indicated pH for 1 h, whereas IDSGLUC5-37 showed stability only between pH 5.0 and pH 6.0 ( Fig. 4c and d). The pH-stability assay demonstrated that IDSGLUC5-26 showed a relatively wider pH-resilience range as compared with IDSGLUC5-37. Notably, after preincubation at pH 3.5 or pH 4.0 for 1 h, IDSGLUC5-26 retained 82.49% or 98.41% of its initial activity, respectively, exhibiting a half-life of 3.44 ± 0.14 h at pH 3.5.

Substrate speci cities of IDSGLUC5-26 and IDSGLUC5-37
To investigate the substrate spectrum of the two enzymes, we estimated their catalytic activities toward various polysaccharides (Fig. 5). The results showed that both IDSGLUC5-26 and IDSGLUC5-37 were robust in catalyzing barley β-glucan, Icelandic moss lichenan, and konjac gum. Notably, IDSGLUC5-26 also exhibited faint activity against pNPC, CMC-Na, and PASC, whereas IDSGLUC5-37 did not. Both enzymes were inactive on other substrates tested in this study.
We then determined the kinetic parameters of IDSGLUC5-26 and IDSGLUC5-37 toward the three preferable substrates. As shown in Table 1, both glucanases showed the highest activities toward barley β-glucan, with V max values of 89.96 ± 9.19 µmol/min/mg and 459.50 ± 25.02 µmol/min/mg, respectively.
Notably, IDSGLUC5-37 had a K m of 3.30 ± 0.49 mg/mL. For lichenan and konjac gum substrates, both IDSGLUC5-26 and IDSGLUC5-37 showed degenerated catalytic effects as compared with those determined in the presence of barley β-glucan. Data represent the mean ± SD (n = 4).
We then conducted HPLC analysis to investigate the detailed compositions of the hydrolyzed products.

Discussion
Mammalian herbivores, such as cattle and sheep, directly digest plant-derived feedstuffs mainly through rumen microorganisms in order to obtain nutrition and energy. The faunal composition of rumen microorganisms includes rumen bacteria, rumen protozoa, and fungi. Each gram of rumen content contains ~10 11 bacteria, ~10 4 to ~10 6 ciliates, and ~10 2 to Determination of the substrate spectrum of the enzymes revealed that both IDSGLUC5-26 and IDSGLUC5-37 show robust hydrolytic activity against barley β-glucan, Icelandic moss lichenan, and konjac gum, all of which include either β-1,3-1,4-or β-1,4-glycoside bonds; however, both are inactive toward β-1,3-glucan ( Fig. 5 and Table 1), indicating that the two enzymes are glucanases that target β-1,4-glycoside bonds between sugar rings. Subsequently, TLC and HPLC analyses clari ed the modes of action, revealing that IDSGLUC5-26 mainly produced G2, G3, and G4 (G3 > G4 > G2) as nal products from both β-glucan and lichenan (Figs. 6a and b; and 7), which is consistent with the hydrolytic patterns of reported endo-β-1,4glucanases, such as EGV from Talaromyces  2:1, respectively. Interestingly, after a 10-min hydrolysis, IDSGLUC5-37 generated 13.51 ± 0.19 mmol/L G3 (64.11%) and 3.44 ± 0.19 mmol/L G4 (16.32%) from barley β-glucan, exhibiting a G3:G4 ratio of 3.93, which was also observed in endo-β-1,3-1,4-glucanases from Rhizomucor miehei (Tang et al. 2012), A. niger (Elgharbi et al. 2013), and Malbranchea cinnamomea (Yang et al. 2014). In contrast, IDSGLUC5-37 almost exclusively produced G3 [4.82 ± 0.01 mmol/L (67.55%)] and negligible G1, G2, G4, or G5 from lichenan. Given the fundamental structure of MLG-like substrates and hydrolytic patterns, we hypothesize that IDSGLUC5-37 is an endo-β-1,3-1,4-glucanase/lichenase (EC 3.2.1.73) speci cally restricted to acting on the β-1,4-linkages adjacent to a β-1,3-glycoside bond within main chains of MLGs. Notably, trace amounts of G5 and G2 were also detected during substrate hydrolysis by either IDSGLUC5-26 or IDSGLUC5-37 ( Figs. 7 and 8). Moreover, substrate spectrum analysis indicated that in addition to MLGs, IDSGLUC5-37 exhibited considerable activity on konjac gum ( Fig. 5 and Table 1), which mainly comprises β-1,4-likages between glucose/glucose, glucose/mannose, or mannose/mannose, suggesting that IDSGLUC5-37 also demonstrates endo-β-1,4-glucanase activity. The appearance of G5 could be attributed to incomplete enzymatic digestion at the early stage, after which G5 was further converted into cellooligosaccharides with lower DPs. However, glucose reached its maximum yield at 10 min or 2 h of glucan decomposition by IDSGLUC5-26 or IDSGLUC5-37 and then decreased or even vanished after 12 h or 24 h. Although glucose commonly presents as an end product of GH hydrolysis, it is believed to be further consumed by transglycosylation in order to alleviate its inhibitory effect on enzyme activity.  2012) reportedly obtained the glucanase BT-01 derived from the buffalo rumen metagenomic library, with this enzyme capable of maintaining > 70% activity after pretreatment at pH 2.6 for 3 h. However, the enzyme showed poor thermostability and almost no activity, even at 40°C. Additionally, recent studies reported glucanases, such as CoCel5A from Colletotrichum orchidophilum ) and NMgh45 from saline-alkaline lake soil microbial metagenome (Zhao et al. 2018), that demonstrated stability from pH 3.0 to pH 4.0. However, enzymes derived from exogenous microbes, let alone those from extreme conditions, may cause unexpected risks to immune-response-related processes in hosts. In the present study, we found that IDSGLUC5-26 was stable and showed sustained catalytic activity within a pH range of 3.5 to 7.0 ( Figs. 3c and 4c). Thus, IDSGLUC5-26 derived from native microbiota in rumen uid will be of great interest for potential application in feed-additive development. Furthermore, heat-challenge assays demonstrated that both IDSGLUC5-26 and IDSGLUC5-37 are mesophilic enzymes that showed rapid decreases in activities at temperatures > 50 ℃ (Figs. 3 and 4). Because rumen temperature normally ranges from 38°C to 41°C, the reaction temperatures of IDSGLUC5-26 and IDSGLUC5-37 likely represent microbial accommodation to the gastrointestinal environment of ruminants (Ko et Cao et al. 2021). Surprisingly, a highly-thermostable endo-β-1,4-glucanase (PersiCel4) was obtained from a sheep rumen metagenome (Ariaeenejad et al. 2020) and showed both stability and activity at an optimal temperature of 85°C (retaining > 75% activity) and even after storage for 150 h at 85°C. Due to its poor thermostability and relatively low activity, the acid-adapted IDSGLUC5-26 requires molecular modi cation to meet the demands of industrial application. Protein-engineering strategies, such as site-directed mutagenesis ( In summary, these ndings demonstrated the successful mining of novel CAZymes from uncultured rumen microbes and provide a deeper understanding of the catalytic modes of endo-β-1,4-glucanases and lichenase. The two functionally characterized glucanases, and particularly the acid-adapted IDSGLUC5-26, will be of great interest for potential feed-additive development.