Heterologous Expression and Characterization of a Thermostable Exo-β-D-Glucosaminidase from Aspergillus oryzae

Abstract

An exo-β-D-glucosaminidase (AorCsxA) from Aspergillus oryzae FL402 was heterologously expressed and purified. The deduced amino acid sequence indicated that AorCsxA belonged to glycoside hydrolase family 2. AorCsxA digested colloid chitosan into glucosamine but not into chitosan oligosaccharides, demonstrating exo-β-D-glucosaminidase (CsxA) activity. AorCsxA exhibited optimal activity at pH 5.5 and 50℃; however, the enzyme expressed in Pichia pastoris (PpAorCsxA) showed much stronger thermostability at 50℃ than that expressed in Escherichia coli (EcAorCsxA), which may be related to glycosylation. AorCsxA activity was inhibited by EDTA and most of the tested metal ions. A single amino acid mutation (F769W) in AorCsxA significantly enhanced the specific activity and hydrolysis velocity as revealed by comparison of V_max and k_cat values with those of the wild-type enzyme. The three-dimensional structure suggested the tightened pocket at the active site of F769W enabled efficient substrate binding. The AorCsxA gene was heterologously expressed in P. pastoris, and one transformant was found to produce 222 U/ml activity during the high-cell-density fermentation. This AorCsxA-overexpressing P. pastoris strain is feasible for large-scale production of AorCsxA.

Introduction

Glucosamine is a preferred substrate for the biosynthesis of glycosaminoglycan chains and subsequently for the production of aggrecan and other proteoglycans that constituted cartilage [25]. Chitosan, a polysaccharide consisting of randomly distributed β-(1-4)-linked N-glucosamine and β-(1-4)-linked N-acetylglucosamine, is the most important material in the production of glucosamine formed through a hydrolysis reaction.

The two most widely used methods for chitosan hydrolysis include chemical treatment and enzymatic digestion [8]. The hydrolysates comprise chitosan oligosaccharides (COS) and glucosamine. Because glucosamine is important in the treatment of osteoarthritis [5,24], there is a great demand for its production using a method safer than the concentrated acid hydrolysis of chitosan. For specific enzymatic digestion, two types of catalysts are required: endo-chitosanase and exo-chitosanase (exo-β-D-glucosaminidase, CsxA). It was previously reported that crude enzymes with endo- and exo-chitosanase or chitinase and chitobiase activities were prepared for the production of glucosamine (or N-acetyl-D-glucosamine) [1,14,23,27]. However, the enzymatic processes are limited by the lack of exo-chitosanase and low supply of both specific enzymes. Thus, establishing methods for the industrial production of endo- and exo-chitosanases is important for the enzymatic production of glucosamine [1].

Since the first CsxA was discovered in 1990 [18], several types of CsxAs have been characterized from different organisms, including bacteria [2,7,9], fungi [3,10,12,15,19,20,30], and archaea [16,28,29]. However, only seven csxA genes have been cloned and sequenced. With the exception of one study reporting the catalytic site of a CsxA from Amycolatopsis orientalis [7], no other studies have evaluated the effect of structure on the specific activity of this enzyme.

In this study, we cloned a csxA gene from Aspergillus oryzae FL402, and then heterologously expressed it in Escherichia coli and Pichia pastoris. We found that a single amino acid mutation (F769W) of AorCsxA significantly influenced the specific activity and hydrolysis velocity. Furthermore, we successfully obtained large-scale production of a genetically engineered CsxA through high-cell-density fermentation.

 

Materials and Methods

Strains and Culture Conditions

A. oryzae FL402 (CCTCC AF2014001), used to clone the csxA gene, was cultured on potato dextrose agar plates at 28℃ for 3 days. Conidia were collected by washing with sterilized water followed by inoculation in Czapek-Dox medium at 28℃ for 12 h with vigorous shaking (200 rpm) using the sugars colloid chitosan, acetylglucosamine, or glucosamine (GlcN) as inducers. After incubation, the culture supernatant was obtained by centrifugation (8,000 ×g) and dialysis w ith 5 0mM sodium a cetate b uf fer ( pH 5.5); it was then used for the chitosanase assay by thin-layer chromatography (TLC). Mycelia from the cultures containing chitosanase activity were collected by filtration, which was immediately followed by RNA extraction. E. coli DH5α cells used f or D NA c loning were c ultured at 3 7℃ in L uria-Bertani ( LB) medium, and a final concentration of 30 μg/ml kanamycin was added to the medium as required. E. coli Rosetta (DE3) pLysS cells (Novagen, Darmstadt, Germany) were used as hosts for protein production and were cultured in LB medium containing 30 μg/ml kanamycin. P. pastoris GS115 cells (Invitrogen, Carlsbad, CA, USA) were used as the hosts for protein production and grown in BMGY (2% (w/v) peptone, 1% (w/v) yeast extract, 1.34% (w/v) yeast nitrogen base, 1% (v/v) glycerol, 0.4 mg biotin/l, and 100 mM potassium phosphate buffer (pH 6.0)) medium and induced in BMMY (0.5% (v/v) methanol added instead of glycerol in BMGY) medium. The pPIC9k-based expression vector was electroporated into P. pastoris GS115 according to the pPIC9K vector user manual (Invitrogen).

Total RNA Preparation and RT-PCR

RNA was extracted from A. oryzae FL402 using TRIzol reagent (Invitrogen). DNase I (New England Biolabs, MA, USA) was used to remove trace DNA in the RNA sample at 37℃ for 30 min. After treatment, 0.5 μl of 25 mM EDTA was added to 20 μl of the reaction mixture and subsequently heated at 65℃ for 10 min to inactivate the DNase I. First-strand cDNA was synthesized by RT-PCR using Oligo(dT)15 (Fermentas, Vilnius, Lithuania) using the purified total RNA as a template.

Gene Cloning and Construction of Expression Plasmids

Using the csxAF/csxAR and csxPAF/csxPAR primer pairs (Table 1) designed according to the csx gene sequences from the closely related Aspergillus sp. CJ22-326 (GenBank Accession No. FJ449570), the csxA gene was cloned into the E. coli expression plasmid pET30a and the P. pastoris expression plasmid pPIC9K, both at the EcoRI/NotI sites, respectively. A csxA point mutation was created by inverse-PCR using the primers csxA769WF and csxA769WR (Table 1). The point mutation was included in the csxA769WF primer using pET30a-csxA plasmid DNA as the template and high-fidelity Phusion DNA polymerase (New England Biolabs). The whole length of linear double-stranded DNAs (7,992 base pair) was purified from an agarose gel, phosphorylated using T4 polynucleotide kinase (Fermentas), andthen self-ligated using T4 DNA ligase (Fermentas). The products were diluted and transformed into E. coli DH5α. The csxA mutant gene was amplified using primers csxPAF/csxPAR from the mutant E. coli expression plasmid and subcloned into the yeast expression vector pPIC9k at the EcoRI and NotI sites. After verification of the sequences by DNA sequencing, the E. coli expression plasmid and the P. pastoris expression plasmid were transformed into Rosetta (DE3) pLysS cells and P. pastoris GS115 cells.

Table 1.Restriction enzyme recognition sites (underlined) used for cloning were EcoRI and NotI.

Heterologous Expression and Purification of AorCsxA

AorCsxA gene heterologous expression in P. pastoris was carried out as described previously [22]. The E. coli transformants carrying the csxA and csxA mutant expression plasmids were grown to an OD600 value of 0.8 at 37℃ under shaking in LB medium and then induced with 0.5 mM (final concentration) isopropyl β-D-1-thiogalactopyranoside at 16℃ overnight with shaking (200 rpm). After induction, the cells were harvested and the resuspended cells were sonicated after adding phenylmethanesulfonyl fluoride (Sigma, St. Louis, MO, USA) to 0.5 mM. After centrifugation, the recombinant protein in each supernatant was purified using nickel-chelating resin followed by a UNO-Q1 ion-exchange column (Bio-Rad, Hercules, CA, USA). The protein concentration was measured with a MicroBCA kit (Pierce, Rockford, IL, USA) after dialysis against 50 mM sodium acetate buffer (pH 5.5). Deglycosylation of PpAorCsxA was carried out using glycopeptidase F (Takara, Dalian, China). Briefly, 25 μg of recombinant CsxA was incubated with 1 mU of glycopeptidase F under native conditions (100 mM Tris–HCl, pH 8.6) at 37℃ for 24 h, followed by SDS-PAGE analysis.

Three-Dimensional Protein Structure Modeling

The SWISS-MODEL server (http://swissmodel.expasy.org) was used to rebuild the three-dimensional (3D) protein structure model. CsxA of A. orientalis (PDB code: 2VZS) was used as a template to model the 3D structure of its mutant W781F.

TLC and High-Performance Liquid Chromatography

TLC was used to analyzed the hydrolysis products, as described previously [22]. The products were also analyzed by high-performance liquid chromatography (HPLC). The HPLC assay was carried out using a Shimadzu LC-20AT (Shimadzu, Kyoto, Japan) system equipped with an autoinjector, a column oven, and an RI detector. Chromatographic separations were performed on a Luna NH2 100A column (250 × 4.60 mm, 5 μm) from Phenomenex. The mobile phase contained acetonitrile and water at 3:1 ratio, and the flow rate was 1 ml/min.

Enzyme Activity Assay

The exo-chitosanase activity assay was done as described previously with slight modification [11]. Generally, the reaction mixtures consisting of 0.5 ml of 1% colloidal chitosan and 150 μl of enzyme solution in 50 mM acetate buffer (pH 5.5) were incubated at 50℃ for 30 min. The reaction was stopped by the addition of potassium ferricyanide and boiling for 15 min. The reaction mixtures were centrifuged (8,000 ×g) and the supernatants were left. The quantity of reducing sugar was determined at 420 nm using the modified Schales method [11]. To determine the metal ion dependence of AorCsxA, different metal ions were added into the reaction mixture. To determine the pH dependence of AorCsxA activity, acetate buffer (pH 3.6–5.8) and phosphate buffer (pH 6.0–8.0) were used to test different pH values in the reaction mixtures. To determine the thermostability of the enzymes expressed in E. coli (EcAorCsxA) and P. pastoris (PpAorCsxA), the enzymes were incubated at 50℃ for 25 h and then enzyme activities were determined at intervals during incubation. One unit of activity was defined as the amount of enzyme that liberated 1 μM of reducing sugar from the substrate per minute; glucosamine (Sigma) was used as the standard. To determine kinetic constants, 150 μl of enzyme (0.3 U/ml) was incubated with colloidal chitosan, ranging from 0.5 to 5.0 mg/ml, at 50℃ and pH 5.5 for 10 min. The maximum velocity (Vmax) and Michaelis-Menten constant (Km) were measured by Lineweaver-Burk transformation. All experiments were conducted in triplicates.

High-Cell-Density Fermentation for PpAorCsxA

A 50 L fermentation tank (Gaoji Bio-Engineering Co., Ltd., Shanghai, China) containing 20 L of basal salt medium was used to culture P. pastoris transformants for PpAorCsxA as described previously [22]. The fermentation culture was sampled at intervals after methanol induction, and the specific CsxA activity and wet cell weight were measured.

Statistical Analyses

The significance of enzyme activity in the metal ion experiment was analyzed using one-way ANOVA. The significance of Vmax, kcat, and Km of CsxA between wild-type E. coli and mutant (F769W) was analyzed using the independent-sample t-test. Analyses were conducted using SPSS ver. 13.0.

 

Results and Discussion

Cloning, Expression, and Purification of Exo-β-D-Glucosaminidase of A. oryzae FL402

The csxA transcript (2,643 base pairs) was amplified by RT-PCR from total RNA of the strain induced by acetylglucosamine or GlcN, and was deposited into the GenBank database under the accession number KJ133556. The deduced amino acid sequence showed high identities to the exo-β-D-glucosaminidase from Aspergillus sp. CJ22-326 (61% identity) [15]. Purified EcAorCsxA showed a single sharp band at approximately 100 kDa on SDS-PAGE (Fig. 1A), which conformed with the deduced molecular mass of this enzyme (99.2 kDa); however, AorCsxA expressed in P. pastoris GS115 cells (PpAorCsxA) was 110 kDa in mass, as indicated by SDS-PAGE (Fig. 1A). The molecular mass of PpAorCsxA was larger than that of EcAorCsxA because of glycosylation, as PpAorCsxA treated with glycopeptidase F showed the same molecular mass as EcAorCsxA, according to the SDS-PAGE results (Fig. 1A).

Fig. 1.Purification of recombinant exo-β-D-glucosaminidase and determination of its reaction pattern. (A) M, protein marker, with the molecular mass indicated on the left; lane 1, PpAorCsxA, the AorCsxA purified from P. pastoris; lane 2, EcAorCsxA, the AorCsxA purified from E. coli; lane 3, deglycosylated PpAorCsxA. (B) TLC analysis of degradation products from chitosan by recombinant AorCsxA. Lane M, standard glucosamine oligomers mixture (GlcN)1 to (GlcN)5; lane S, chitosan hydrolysate by recombinant AorCsxA. (C) HPLC profiles of chitosan hydrolysate. Solid and dashed lines represent the standard mixture (GlcN)1–5 and chitosan hydrolysate produced by recombinant AorCsxA, respectively. P1 through P5 refer to (GlcN)1, (GlcN)2, (GlcN)3, (GlcN)4, and (GlcN)5, respectively.

Identification of the General Properties of AorCsxA

Reaction pattern. The reaction product of EcAorCsxA using 80–95% deacetylated colloidal chitosan was detected by TLC and HPLC. On the TLC plate, only one spot (corresponding to the glucosamine standard) was detected, but no spots corresponding to the chitosan oligosaccharide standards were detected (Fig. 1B). Furthermore, the HPLC profile of the hydrolysate produced by EcAorCsxA showed a single peak, which corresponded to glucosamine (Fig. 1C). Extension of the reaction time (30 min to 6 h) yielded glucosamine as the only hydrolysis product. PpAorCsxA showed the same reaction pattern (data not shown). These results indicate that AorCsxA acted as an exo-β-D-glucosaminidase.

Optimal reaction pH, temperature, and metal ion dependence. As shown in Fig. 2A, the relative activities of the AorCsxA enzymes from E. coli and P. pastoris were nearly abolished at pH values lower than 3.5 and higher than 8. AorCsxA functioned at pH values between 4 and 6, with the optimum pH at 5.5. Both enzymes functioned at temperatures between 30℃ and 70℃, and the optimal temperature at pH 5.5 was 50℃ (Fig. 2B). PpAorCsxA showed higher activity than EcAorCsxA at non-optimal pH and temperature values. For the effect of metal ions on enzyme activity, as shown in Table 2, EDTA and all metal ions tested, except Ca2+ and Ni2+ and Ba2+, decreased the activity of AorCsxA at the evaluated concentrations. Notably, Mn2+ (5 and 10 mg/ml), EDTA (5 and 10 mg/ml), Ag+ (10 mg/ml), Cu2+ (10 mg/ml), Co2+ (10 mg/ml), and Fe2+ (10 mg/ml) were found to strongly inhibit EcAorCsxA activity. This result differed from those of previous studies, which showed that Mn2+ typically activated enzyme activity (Table 4) [4,13]. However, Eom and Lee [6] also found that the activity of an exo-β-D-glucosaminidase from A. fumigatus KB-1 was inhibited by Mn2+.

Fig. 2.pH, temperature, and thermostability profiles of exo-β-D-glucosaminidase from A. oryzae FL402 (AorCsxA) expressed in P. pastoris (PpAorCsxA) and E. coli (EcAorCsxA). (A) pH profile of PpAorCsxA and EcAorCsxA. (B) Temperature profile of both enzymes. (C) Thermostability (at 50℃) profile of both enzymes.

Table 2.*Different superscripts in one column indicate significant difference (p < 0.05, N = 3, one-way ANOVA).

Thermostability. The temperature stability test showed that PpAorCsxA had much higher thermostability than EcAorCsxA at 50℃ (Fig. 2C). During the first 15 h, the enzyme activity of PpAorCsxA remained stable, and thereafter was slowly reduced to 80% at 25 h. In contrast, the activity of EcAorCsxA was sharply reduced to 0.48% within 3 h. These results indicate that AorCsxA from P. pastoris was much more thermostable than the enzyme from E. coli. Furthermore, after treatment with glycopeptidase F, deglycosylated PpAorCsxA was not stable at 50℃ (data not shown), and the enzyme showed the same molecular mass as EcAorCsxA (Fig. 1A). This result demonstrated that glycosylation was important for the thermostability of PpAorCsxA. Here, we first reported that glycosylation of PpAorCsxA expressed in P. pastoris enhanced the thermostability of exo-β-D-glucosaminidase. We found that PpAorCsxA, which had been modified by glycosylation, retained 80% activity at 50℃, pH 5.5 for 25 h; thus, the thermostable exo-β-D-glucosaminidase is advantageous for use in the industrial production of glucosamine.

Genetic Engineering of AorCsxA for Enhanced Specific Activity

Seven amino acid sequences of CsxA enzymes from different strains were aligned using AlignX (Fig. 3A). Nearly all sequences shared the conserved amino acid residue Trp/W, corresponding to W781 of the enzyme from A. orientalis. However, the AorCsxA sequence from A. oryzae FL402 encoded F769 (Phe) at this site. Thus, a mutated enzyme (F769W) was generated to clarify the role of the amino acid at this position and this enzyme was expressed in E. coli (Fig. 3B). The kinetic constants for wild-type EcAorCsxA and F769W were determined using colloid chitosan as the substrate (Fig. 3C, Table 3). F769W showed higher Vmax, Km, kcat, and kcat/Km values (17.58 ± 1.04 × 10-3 mM s-1, 32.9 ± 1.7 mM, 102.6 ± 3.3 s-1, and 3.11 ± 0.06 mM-1 s-1, respectively) than the wild type (8.28 ± 0.09 × 10-3 mM s-1, 18.4 ± 0.5 mM, 23.8 ± 1.4 s-1, and 1.29 ± 0.04 mM-1 s-1, respectively) (Table 3). Moreover, the specific activities of F769W and wild-type EcAorCsxA were 14.0 U/mg and 5.8 U/mg, respectively. Circular dichroism spectrum profiles of F769W and the wild-type enzyme were nearly identical, indicating that the mutation of F769 to tryptophan did not affect the secondary structure of the enzyme (Fig. S1). 3D structure remodeling suggested that the engineered amino acid residue tryptophan at site 769 formed a tighter cleft for substrate binding, resulting in the higher specific activity (Fig. 4). Fungal exo-β-D-glucosaminidases purified from their native producing strains have been characterized in previous studies [4,12,30]; however, only two studies characterized recombinant exo-β-D-glucosaminidase expressed in E. coli [15] or in P. pastoris [10]. In this study, the activity of A. oryzae FL402 exo-β-D-glucosaminidase expressed in E. coli and P. pastoris was characterized and compared. Moreover, a single mutation in the enzyme was found to enhance the enzyme activity by 2.4-fold.

Fig. 3.Identification of a crucial amino acid residue for the specific activity and substrate binding of exo-β-D-glucosaminidase from A. oryzae FL402 (AorCsxA). (A) Amino acid sequence alignment of AorCsxA with five reported exo-β-D-glucosaminidases. The alignment includes exo-β-D-glucosaminidase from A. orientalis (SwissProt sequence code: Q56F26), Hypocrea jecorina PC-3-7 (Q4R1C4), Streptomyces avermitilis (Q82NR8), Aspergillus sp. CJ22-326 (C3U4R9), and Hypocrea virens (C0LRA7). The conserved tryptophan located at the edge of the cleft is marked with an asterisk (*) above the sequences. Red letters with yellow background represent conserved residues among the CsxA enzymes compared. (B) SDS-PAGE analysis of recombinant AorCsxA and the mutation F769W. Lane M, protein marker and the bands corresponding to 120, 100, and 85 kDa are indicated; lane 1, AorCsxA; lane 2, phenylalanine at the 769 site of AorCsxA changed to tryptophan. (C) Kinetic parameters of AorCsxA and the single-mutation F769W. The Lineweaver-Burk curves of AorCsxA and the mutation F769W are represented by lines with squares and diamonds, respectively. The Michaelis-Menten constant (Km) and maximum velocity (Vmax) were determined by Lineweaver-Burk transformation. kcat and kcat/Km were calculated and the results are shown in Table 3.

Table 3.*Different superscripts in one column indicate significant difference (p < 0.01,N = 3, independent samples t-test).

Fig. 4.3D models of CsxA W781F (A) and W781 (B) from A. orientalis, which represent CsxA F769 and F769W of A. oryzae FL402, respectively. The red carbon ring represents the glucosamine unit of chitosan; the blue surface represents the amino acid residues F769 and F769W, respectively. 3D remodeling was conducted in SWISS-MODEL (http://swissmodel.expasy.org/).

Enhanced Production of AorCsxA F769W in High-Cell-Density Fermentation

Although enzymatic production of COS and glucosamine (or N-acetyl-D-glucosamine) from chitosan (or chitin) has been reported previously [1,21,26,27], large-scale fermentation of the enzyme is in great demand for the industrial production of COS and glucosamine using enzymatic methods. In this study, production of the engineered enzyme (AorCsxA F769W) with higher (2.4-fold) specific activity than the wild-type enzyme was enhanced in P. pastoris and AorCsxA F769W. Production reached 222 U/ml culture in high-cell-density fermentation (Fig. 5), which was 10-fold higher than that produced in flask fermentation using both P. pastoris and E. coli hosts. Exo-β-D-glucosaminidase has not been applied in industry for GlcN production because of the low expression of exo-β-D-glucosaminidase [17]. Previous studies reported that the enzyme specific activity in the culture filtrate was always below 2 U/mg (Table 4), which did not meet industry requirements. Recently, Nidheesh et al. [19] reported that Penicillium decumbens CFRNT15 can produce highly active exo-β-D-glucosaminidase (2,335 ± 12 U/g initial dry substrate) by solid-state fermentation. In the present study, recombinant P. pastoris was fermented in a 50 L fermenter in the liquid state, and 222 U/ml enzyme activity in the culture filtrate was reached (Fig. 5). This is the highest reported activity in liquid-state fermentation.

Table 4.aAorCsx expression in P. pastoris by glycosylation. bSolid substrate fermentation. cLiquid fermentation in a 50 L fermenter.

Fig. 5.Fermentation kinetics of recombinant AorCsxA F769W production from P. pastoris. Methanol was fed at 0 h for 54 h. The wet cell weights and CsxA activities were determined during the experiment. Rectangles represent wet cell weight (g/l) and triangles represent CsxA activity (U/ml fermentation culture).

In conclusion, an exo-β-D-glucosaminidase (AorCsxA) gene (cxsA) was cloned from A. oryzae FL402 and then heterologously expressed and purified in E. coli and P. pastoris. The optimal pH and temperature for activity of AorCsxA was 5.5 and 50℃; these values were the same for the enzyme expressed in E. coli and P. pastoris. However, PpAorCsxA was more thermostable than EcAorCsxA. Interestingly, a single amino acid mutation (F769W) in AorCsxA produced through genetic engineering significantly enhanced the specific activity from 5.8 to 14 U/mg compared with that of the wild-type enzyme. Importantly, the engineered enzyme (AorCsxA F769W) showed enhanced production in P. pastoris, reaching 222 U/ml culture in high-cell-density fermentation. Hence, the recombinant P. pastoris strain is fit for industrial-scale production of AorCsxA.