Galactan, galactomannan, arabinoxylan, arabinan, β-glucan, and D-Xylose and D-Glucose (GOPOD Format) Assay Kits were obtained from Megazyme International (Wicklow, Ireland). 2-Nitrophenyl-β-D-glucopyranoside, 4-methylumbelliferyl-β-D-glucopyranoside (MUG), 4-methylumbelliferyl-β-D-cellobioside (MUC), 4-methylumbelliferyl-β-D-manopyranoside (MUM), 4-methylumbelliferyl-β-D-galactopyranoside (MUGal) and 4-methylumbelliferyl-β-D-xylopyranoside (MUX) were obtained from Research Products International Corp (Mt. Prospect, IL). 20. Geneticin (G418), 4-nitrophenyl-β-D-galactopyranoside, 4-nitrophenyl-β-D-mannopyranoside, Avicel PH-101, Whatman 1 filter paper, glucose, and xylose were purchased from Sigma-Aldrich (St. Louis, MO). Xylooligosaccharides were obtained from Cascade Analytical Reagents and Biochemicals (Corvallis, Oregon). Xylanase 4 (Cthe_2972), α-arabinofuranoside 1 (G11MC16_1557), β-glucosidase 1 (Aaz_81839), β-glucosidase 2 (Cthe_0212, Cthe BglA),

β-glucosidase 4 (Bcel_0282), β-glucosidase 5 (Bcel_0705), cellulase 4, (Cthe_0797), curdlanase 1 (Bcell_0683),β-xylosidase 1 (Y412MC61_2711), A. niger α-xylosidase 21, xyloglucanase 2, Sfla_0771, 0797are products of C5-6 Technologies LLC (Fitchburg, WI). Barley flour, oatmeal and oat bran, wheat flour and wheat bran (Bob s Red Mill, Milwaukie, OR) were obtained from local grocery stores. Phosphoric acid swollen cellulose (PASC) was prepared from Avicel PH-101 by completely dissolving the cellulose in concentrated phosphoric acid 2022, followed by precipitation into 10 volumes of deionized water. All other chemicals were analytical grade and used without any further purification.

Glucose production was measured using the Megazyme GOPOD glucose kit and xylose production was measured using the Megazyme K-XYLOSE xylose kit, both using the manufacturer s instructions. For rate determinations, monosaccharide release was determined using 200 ml of either 1 mM solutions of oligosaccharides or 1% solutions of saccharides in 100 mM acetate buffer, pH 5.0 and 50°C. Substrates were incubated with 85 micrograms of pure enzyme for 10 minutes, the reaction was stopped by incubation at 95°C, and monosaccharides were measured using the appropriate Megazyme kit. For xylooligosaccharide hydrolysis experiments, monosaccharide release was determined using 1000 ml of 1% in 100 mM acetate buffer, pH 5.8 and 50°C. For polysaccharide hydrolysis experiments, monosaccharide release was determined using 1000 ml of 0.2% polysaccharide in 100 mM acetate buffer, pH 5.8 and 50°C. For grain and grain fractions, 10 g samples of material were incubated in 90 ml of 50mM acetate buffer, pH 5.8, at 50°C and 60 rpm in 250 ml bottles. In addition, all long-term reactions contained 5mg/ml vancomycin and 10mg/ml carbenicillin to prevent microbial growth. Enzyme dosing unless noted elsewhere was 20 mg/ml of pure enzyme. Aliquots of the reaction mixture were removed, the reaction was stopped by incubation at 95°C, and monosaccharides were measured using the appropriate Megazyme kit.

Protein purity was determined by SDS PAGE. The specific activity of the purified enzymes was determined using 2-NP-β-D-glucopyranoside at pH 5.0 and 50°C 23; one β-glucosidase unit produces 1 micromole of o-nitrophenol per minute at 50°C from a 1mM solution of 2-nitrophenyl-β-D-glucopyranoside. Protein concentration were determined by A₂₈₀ spectrophotometry. The extinction coefficient for the enzymes was calculated using ProtParam (https://web.expasy.org/protparam/ ).

DNA2.0 (now Atum) optimized the CBM1-Bgluc ORF for expression in Pichia pastoris and synthesized the gene based on the sequence determined in previous work 24. The optimized ORF for the CBM1-Bgluc was sub-cloned into BioGrammatics pJANaMF (AOX1 promoter, Nat resistance, alpha Mating Factor, aMF) as a single fragment with BsaIsites.

The 6His-Bgluc was sub-cloned into pJAGaMF (AOX1 promoter, G418 resistance, aMF) as two fragments. The 3’ fragment came from the DNA2.0 clone and extends from a native NcoI site 21 bases 3’ from the end of the CBM1 coding sequence, to the BsaI site at the stop codon. The 6His-Bgluc 5’ fragment was created by annealing the following primers (lower case nucleotides remain single stranded for subcloning):

Bgluc5frag-top:5-ggctCACCACCATCACCACCATGTCGTTCCTCCAGCTGGTACTC-3’

Bgluc5frag-bot:5’-catgGAGTACCAGCTGGAGGAACGACATGGTGGTGATGGTGGTG-3’

The entire N-term fusion and Bgluc coding region and cloning junctions of both expression vectors were sequenced and found to be as predicted for both expression vectors.

Sequencing primers:

1) Bgluc-1860F: 5’- ACACTATTGCTAAGTCGCC-3’

2) Bgluc-1240F: 5’- TGCTAGAGATGGTATTGTG-3’

3) Bgluc-660F: 5’- GCGTTGACCCTTATCTGA-3’

4) Bgluc-1540R: 5’-CAACATCTTTACCTCTAGCG-3’

5) Bgluc-920R: 5’- CCCAACTGATCCTTCAACAC-3’

6) Bgluc-360R: 5’- TCCATTCCATCCAACGCC-3’

BioGrammatics wild type Pichia pastoris strain Bg10 was transformed with both expression constructs (Bg10 E-comp cells MP0002-BR002, ~200 ng of each vector DNA independently) (Transformation by electroporation in 1 mm cuvettes, 1150 Volts, 600 Ohms, and 10 uF capacitance). pJANaMF-CBM1-Bgluc transformants were selected on YPD-agar plates with Noursethricin (Nat, 100 ug/ml). pJAGaMF-6His-Bgluc transformants were selected on plates with G418 (800 ug/ml). Thousands of transformants were obtained for each construct. Cells from distinct colonies of greater than 200 transformants were picked from the lowest dilution selection plate and patched to a similar plate with the same selective drug (Nat or G418).

Initially, 96 well deep plates were used to screen the transformants, using the original Tre Bgl1 clone as a control; 4-methylumbelliferyl-β-D-cellobioside (MUC) was used to assay samples of each well. This growth format proved ineffective, as methanol induction failed to induce production of enzyme in any of the wells, including the control. Subsequently, individual clones were grown in 15 ml falcon tube cultures with daily additions of methanol. This screening method proved effective, demonstrating that essentially all clones produced active protein at similar levels. Clones were grown in 3 liter fermentors using BSM defined media 25. Cell mass was developed using glycerol as the primary feed for 24 hours, then switched to methanol to induce enzyme production. Fermentations were pH controlled and feed was added based on dissolved oxygen levels. Fermentations were run for four days on the methanol feed.The protein was purified using a method previously developed for purification of Bgl1 produced in fermentors using the MSU clone 24. The amino acid sequences of the two constructs (Bgl1H and Bgl1X) are shown in Figure 1. CBM1 (Bgl1X) and 6His-tag (Bgl1H) are shown in bold.

The two new strains produced more than 2-fold higher levels of protein than the MSU clone 24 in 3-liter fermentations. Purification of the new proteins was performed using the method developed for the MSU clone of Bgl1. After fermentation was complete, yeast cells were removed by centrifugation, and the clarified fermentation broths were concentrated using a Sartorius cross-flow ultrafiltration unit equipped with 20K cut-off membranes. After ultrafiltration, green colored material (alcohol oxidase microcrystals) was removed from the concentrate by cross-flow filtration through a 0.2 m membrane. The resulting material was concentrated and further purified by size exclusion chromatography on Sephacryl S100. This purification method was used to purify both BglH and Bgl1X, resulting in a purity of >95% for the two proteins (Figure 2). Attempts to purify either Bgl1H and BglX from centrifuged fermentation broth using Ni affinity chromatography or ion exchange chromatography was unsuccessful.

The enzymatic activity of Bgl1H was investigated using 4-methylumbelliferyl and 2-nitrophenyl monosaccharides. As expected, the enzyme was active on both 2-nitrophenyl-β-D-glucopyranoside, 4-methylumbelliferyl-β-D-glucopyranoside (MUG). The specific activity of the enzyme was 11 u/mg at 50°C, pH 5.0, using 2-nitrophenyl-β-D-glucopyranoside as substrate. Bgl1H was also active on 4-methylumbelliferyl-β-D-xylopyranoside (MUX) and 2- nitrophenyl-β-D-xylopyranoside with a specific activity of 1.0 u/mg on 2- nitrophenyl-β-D-xylopyranoside. The enzyme showed no activity on the mannose or galactose-containing substrates. Hydrolysis of 2- nitrophenyl-β-D-glucopyranoside was inhibited by 10 mM glucose or cellobiose or 100 mM xylose. The inhibition by xylose confirms that a single active site is responsible for both hydrolysis of 2-nitrophenyl-β-D-glucopyranoside and other glucans and 2-nitrophenyl-β-D-xylopyranoside and other xylans. Assays of Bgl1X yielded specific activities statistically indistinguishable from the values obtained for Bgl1H, indicating the CBM1 module does not influence the rate on small substrates. Based on these results, Bgl1H was used for all kinetic studies.

Previous work 15 reported on the ability of the enzyme to produce glucose from a number of low molecular weight substrates, with a clear specificity for β-(1,3)-laminarbiose 15. In this work, glucose production by Bgl1H was determined using an expanded range of substrates including cellodextrins, reduced cellodextrins, and polysaccharides. The rate of glucose production increased with the chain length of cellodextrin, (Table 1) with an approximate three-fold increase from cellobiose to cellohexaose. The reduced cellodextrins showed a similar behavior, with the rate of glucose formation from each reduced cellodextrin corresponding to the non-reduced cellodextrin with one fewer glucose unit. The rate of glucose production from laminarbiose (G3Gp) was similar to that of cellobiose (G4Gp). The rates of glucose production from the polysaccharide substrates were unusually high. The rate of glucose production from curdlan was approximately three-fold higher than from laminarbiose, and the rate of glucose production from β-glucan and lichenan was ten-fold higher than from cellobiose. These results indicate that the preferred substrates for Bgl1H are β-(1,3)- and β-(1,4)-polysaccharides and not laminarbiose or another disaccharide substrate as indicated previously.

The xylosidase activity of Bgl1H was evaluated using xylobiose, xylotriose, and xylotetraose. Hydrolysis of xylobiose by BGL1H was extremely slow (4% of the rate of cellobiose hydrolysis), but the hydrolysis rates increased rapidly with xylotriose and xylotetraose as substrates to approximately 25% of the rates of the corresponding cellobiosides (Table 2). Higher oligomers are not available, a result of the substitution on the xylose backbone of the xylan used for preparation of the oligoxylans.

The conversion of β-glucan to glucose by Bgl1H and four bacterial β-glucosidases with similar temperature and pH optima was determined (Table 3). Bgl1H showed 68% conversion of the β-glucan to glucose at 66 hours, and extended incubation resulted in 92% conversion of the β-glucan to glucose. The bacterial enzymes tested did not generate more than 2% glucose under the same conditions. Under the same conditions, Bgl1H produced 10.3% glucose from PASC, while the bacterial enzymes produced no significant amount of glucose. The bacterial β-glucosidases appear to be unable to digest either glucan polysaccharides, while Bgl1H is able to extensively degrade these partially soluble polysaccharides containing β-(1,4)-linkages, and mixtures of β-(1,4)-linkages and β-(1,3)-linkages, an activity not previously seen with this enzyme or other β-glucosidases.

When evaluated using either of two insoluble celluloses, Whatman 1 filter paper or Avicel PH-10, Bgl1H and BglX produced no measurable glucose from these crystalline (insoluble) β-(1,4)-linked celluloses. Addition of Trichodermaendo- and exo-cellulases to Bgl1H results in almost complete conversion of these substrates to glucose 15.

When evaluated on insoluble β-(1,3)-linked curdlan, a maximum of approximately 20% hydrolysis to glucose was obtained using Bgl1H or Bgl1X, significantly better conversion than observed with insoluble crystalline or acid-swollen celluloses. Addition of a curdlanase, (Bcell_068326) to conversion reactions resulted in hydrolysis of approximately 70% of the curdlan to glucose. This increase in glucose production suggests crystalline regions within the curdlan may limit hydrolysis.

Xyloglucan differs from other glucans in possessing three out of four glucose residues substituted with α-(1,6)-linked xylose or substituted xylose residues. Even with extended incubations, Bgl1H produced less than 1% of the available glucose from tamarind xyloglucan. Addition of A. niger α-xylosidase (αXyl) 21, an enzyme capable of removing α-(1,6)-linked xylose from xyloglucan raised production of glucose to 4% of the available glucose (Table 4). Addition of an endo-acting xyloglucanase, Sfla_0771, or an endo-acting cellulase, Cthe_0797 increased glucose release to approximately 50% of the glucose present in the xyloglucan. Addition of the combination of a cellulase and xyloglucanase did not further increase glucose production.

The enzyme did not release glucose from dextran, pullulan or starch, indicating an absolute specificity for β-linked glucan. Bgl1H produced no measurable glucose from glucomannan when incubated in the absence or presence of endo-mannanase, indicating the enzyme was unable to cleave glucose β-(1,4)-linked to mannose.

Incubation of arabinoxylooligosaccharides with Bgl1H resulted in no release of xylose, the result of α-2 and α-3-arabinose substitutions on the xylan backbone. When supplemented with α-arabinofuranoside, (G11MC16_1557), Bgl1H converted 29% of a 1.0% solution of xylooligosaccharides to xylose in 42 hours, a result superior to three of four thermostable bacterial xylosidases (Table 5). Prolonged incubation showed Bcel_0705 peaked at 42% conversion to xylose, while Tre BglA continued to convert the xylooligosaccharides, reaching 57% conversion.

Bgl1 produced xylose from xylan when supplemented with xylanase (Cthe_2972) and α-arabinofuranoside (G11MC16_1557). Using 0.2% birchwood xylan, the combination of Bgl1H + xylanase + α-arabinofuranoside released 31% of the available xylose after 18 hr incubation, comparable to the 39% released using a combination of β-xylosidase (Y412MC61_2711) + xylanase + α-arabinofuranoside.

To determine if Tre Bgl1H could hydrolyze β-glucan and xylan present in whole grain samples, release of glucose and xylose were measured using commercial samples. Incubation for 115 hours resulted in significant formation of glucose in all samples, with barley and wheat flours being highest (Table 6). No glucose was formed in samples incubated identically without enzyme addition. Addition of thermostable β-glucanase (G11MC16_1587) and Bgl1 to an identical barley flour reaction resulted in production of 650 mg of glucose from 10g (6.5% by weight) of barley flour. The extractable β-glucan content of barley ranges from 1.8 to 5.4% 27, indicating the hydrolysis is able to convert all extractable and unextractable β-glucan to glucose.

This work reports the cloning, purification and first full characterization of Trichoderma reesei (Hypocreajecorina) Bgl1as both a his-tagged version and a version containing an N-terminal CBM1 module. The codon-optimized constructs yielded higher protein concentrations in the fermentation, and were significantly easier to purify than the original construct. Both forms were purified to greater than 95% purity, and both possessed essentially identical activities. The CBM1 module did not change either the rate of hydrolysis of tested substrates, and the CBM1 module did not improve conversion of crystalline or amorphous cellulose or other substrates. Bgl1 possesses a number of unique properties that are not found in other microbial β-glucosidases. The rate of Bgl1 increases with increasing saccharide chain length; most microbial β-glucosidases have maximum rates on disaccharides, with little or no activity on tetrasaccharides or higher polymers. The rates of glucose production from the polysaccharide substrates were unusually high. The rate of glucose production from curdlan was approximately three-fold higher than from laminarbiose, and the rate of glucose production from β-glucan and lichenan was ten-fold higher than from cellobiose. These results indicate that the preferred substrates for Bgl1H are β-(1,3)- and β-(1,4)-polysaccharides and not laminarbiose or another disaccharide substrate as indicated previously 15. Bgl1 also possesses β-xylosidase activity, again with a rate increasing with chain length.

The basis of the increasing rate with saccharide chain length is unclear. The x-ray structure (Figure 3) shows the active site in a cleft between two domains of the enzyme and clearly shows three substrate binding sites (-1. +1 and +2). No clearly define polysaccharide binding sites outside the active site are mapped, however, the crystal structure was not determined using a polysaccharide substrate. It is conceivable that this cleft binds to polysaccharides via a number of weak hydrophobic interactions or hydrogen bonds that keep the polysaccharide in the cleft and thereby increase the rate of hydrolysis via a processive mechanism 28. Additional structural and kinetic experiments are need to further clarify the action of this unique enzyme.