A Novel Saponin Hydrolase from Neocosmospora vasinfecta var. vasinfecta

Abstract

We isolated a soybean saponin hydrolase from Neocosmospora vasinfecta var. vasinfecta PF1225, a filamentous fungus that can degrade soybean saponin and generate soyasapogenol B. This enzyme was found to be a monomer with a molecular mass of about 77 kDa and a glycoprotein. Nucleotide sequence analysis of the corresponding gene (sdn1) indicated that this enzyme consisted of 612 amino acids and had a molecular mass of 65,724 Da, in close agreement with that of the apoenzyme after the removal of carbohydrates. The sdn1 gene was successfully expressed in Trichoderma viride under the control of the cellobiohydrolase I gene promoter. The molecular mass of the recombinant enzyme, about 69 kDa, was smaller than that of the native enzyme due to fewer carbohydrate modifications. Examination of the degradation products obtained by treatment of soyasaponin I with the recombinant enzyme showed that the enzyme hydrolyzed soyasaponin I to soyasapogenol B and triose [α-l-rhamnopyranosyl (1→2)-β-d-galactopyranosyl (1→2)-d-glucuronopyranoside]. Also, when soyasaponin II and soyasaponin V, which are different from soyasaponin I only in constituent saccharides, were treated with the enzyme, the ratio of the reaction velocities for soyasaponin I, soyasaponin II, and soyasaponin V was 2,680:886:1. These results indicate that this enzyme recognizes the fine structure of the carbohydrate moiety of soyasaponin in its catalytic reaction. The amino acid sequence of this enzyme predicted from the DNA sequence shows no clear homology with those of any of the enzymes involved in the hydrolysis of carbohydrates.

Saponins are glycosides widely distributed among plants; “saponin” is a general term for a group of compounds having steroids, steroidal alkaloids, and triterpenoids as aglycones. Many plants produce saponins with antifungal activities (17), and they are speculated to protect the plants from phytopathogenic fungi. The major mechanism of the antifungal activities of saponins apparently involves their ability to complex with sterols in fungal membranes and to cause a loss of membrane integrity (10, 21). Some phytopathogenic fungi, however, are known to have enzymes that convert saponins to derivatives with weaker antifungal activities, as their carbohydrate moieties are removed by the enzymes (18, 22, 25, 26, 29). A decrease in the antifungal activities of saponins facilitates the ability of phytopathogenic fungi to infect host plants. Therefore, these enzymes probably determine the range of plants susceptible to phytopathogenic fungi (4).

Saponins and their aglycones are known to have many useful physiological and pharmacological activities. For example, glycyrrhetic acid, obtained from glycyrrhizin, a saponin derived from the roots of licorice, is used as an antiulcer agent (31). Also, soyasapogenol B, obtained from soybean saponin, is known to have hepatoprotective, antimutagenic, antivirus, and anti-inflammatory activities (1, 2, 9, 11, 16, 30). These aglycones are produced by acid hydrolysis of saponins, but there have been reports of aglycone production by microorganisms or enzymes. Tanaka et al. isolated eight strains of microorganisms that produced glycyrrhetic acid from glycyrrhizin and reported that Pseudomonas saccharophila was the most effective and had the highest β-glucuronidase activity (32). Muro et al. reported that glycyrrhizin hydrolase, with a molecular weight of 150,000 and obtained from Aspergillus niger GRM3, hydrolyzed glycyrrhizin to glycyrrhetic acid and a carbohydrate moiety (19). Kudou et al. cultured 158 strains of the genus Aspergillus in a medium containing soybean saponin and reported that 26 of them had marked soybean saponin hydrolase activity (13). Purified soybean saponin hydrolase from A. oryzae KO-2 was a heterotetrameric protein having subunits of 35 and 45 kDa (14).

In the course of our study of the mechanism of degradation of soybean saponin by filamentous fungi, we discovered that Neocosmospora vasinfecta var. vasinfecta PF1225, which was isolated from a soil sample, specifically produced soyasapogenol B during cultivation in a soybean saponin-containing medium (35). N. vasinfecta is a widespread fungus belonging to the Ascomycetes, and some strains of N. vasinfecta are known to be able to infect soybean (33). Here, we purified and characterized the soyasapogenol B-producing enzyme and isolated the corresponding gene. This enzyme was found to be a soybean saponin hydrolase, to have a molecular weight and a substrate specificity different from those of the reported soybean saponin hydrolases, and to have no significant homology with any of the reported carbohydrate hydrolases.

MATERIALS AND METHODS

Chemicals.

Soybean extract I (content of soybean saponin, 32%; purchased from Tokiwa Phytochemical Co. Ltd., Chiba, Japan) was used as soybean saponin. Soyasapogenol B was prepared from soybean saponin as described previously (35). Glycyrrhizic acid monoammonium salt n-hydrate, glycyrrhetic acid 3-O-glucuronide, glycyrrhetinic acid, 4-methylumbelliferyl-β-d-glucuronide, and 4-methylumbelliferone were purchased from Wako Pure Chemical Co., Ltd., Osaka, Japan.

Fungal and bacterial strains and culture media.

N. vasinfecta var. vasinfecta PF1225, a stock strain in our collection, was used for purification of the soyasapogenol B-producing enzyme. This strain had been maintained on PDA medium (Difco) and was cultured in TS medium (20 g of soluble starch, 10 g of glucose, 5 g of Polypeptone, 6 g of wheat germ, 3 g of yeast extract, and 2 g of CaCO₃ in 1 liter [pH 7.0]) and MY medium [40 g of malt extract, 20 g of yeast extract, 2 g of KH₂PO₄, 2 g of (NH₄)₂SO₄, 0.3 g of MgSO₄ · 7H₂O, 0.3 g of CaCl₂ · 2H₂O, and 10 g of soybean saponin in 1 liter (pH 7.0)] for production of the soyasapogenol B-producing enzyme. Trichoderma viride strain 2, a uracil-requiring strain obtained by treating T. viride MC300-1 (a stock strain in our collection) with UV rays and selecting mutants for 5-fluoroortic acid resistance (3), was used as the host for heterologous expression of the sdn1 gene. T. viride was cultured in mycelium formation medium (25 g of glucose, 10 g of yeast extract, 10 g of malt extract, 20 g of Polypeptone, 1 g of K₂HPO₄, and 0.5 g of MgSO₄ · 7H₂O in 1 liter [pH 7.0]), minimum medium [5 g of glucose, 4 g of (NH₄)₂SO₄, 0.3 g of urea, 2 g of KH₂PO₄, 0.3 g of MgSO₄ · 7H₂O, 0.3 g of CaCl₂, 5 mg of FeSO₄ · 7H₂O, 1.56 mg of MnSO₄ · 5H₂O, 1.4 mg of ZnSO₄ · 7H₂O, 2 mg of CoCl₂, and 20 g of purified agar (Sigma-Aldrich Fine Chemicals) in 1 liter], and production medium [10 g of glucose, 40 g of lactose, 20 g of soybean lees, 10 g of yeast extract, 2 g of (NH₄)₂SO₄, 5 g of KH₂PO₄, 2 g of CaCO₃, and 0.3 g of MgSO₄ · 7H₂O in 1 liter]. Escherichia coli DH5α (supE44 ΔlacU169 [φ80 lacZΔM15] hsdR17 recA1 endA1 gyrA96 thi-1 relA1) was cultured in Luria-Bertani medium supplemented with ampicillin for the preparation and propagation of plasmids. E. coli XL1Blue MRA (ΔmcrA183 ΔmcrCB-hsdSMR-mrr173 endA1 supE44 thi-1 gyrA96 relA1 lac) was cultured in Luria-Bertani medium supplemented with 0.2% maltose and 10 mM MgSO₄ to prepare a λ phage library.

Enzyme purification.

N. vasinfecta var. vasinfecta PF1225 cultured on PDA medium was inoculated into 100 ml of TS medium prepared in a 500-ml conical flask and cultured with shaking at 25°C for 3 days. Then, 4 ml of the culture broth was transferred to each of 10 conical flasks containing 100 ml of MY medium and cultured with shaking at 25°C for 3 days. Of the culture broth obtained, about 1,000 ml was centrifuged, and a supernatant was collected. The pellet was resuspended in 500 ml of water, the suspension was centrifuged, and another supernatant was collected. The pellet was washed three times, and all supernatants (1,800 ml) were pooled and used as a crude enzyme fraction. Ammonium sulfate was dissolved in the crude enzyme fraction to a concentration of 20% saturation, the solution was centrifuged, and a supernatant was collected by removing the pellet. Ammonium sulfate was added to 75% saturation, the solution was centrifuged, and the precipitate was recovered. This precipitate was dissolved in 150 ml of 50 mM Tris-HCl (pH 7.5) containing 1 M ammonium sulfate; the solution was applied to a Toyopearl Butyl-650S (Tosoh, Tokyo, Japan) column (26 by 350 mm) which had been equilibrated with the same buffer as the sample and then was eluted with 600 ml of 50 mM Tris-HCl (pH 7.5) containing 1 M ammonium sulfate. This fraction was collected and concentrated to about 2 ml with Pellicon XL (Millipore) and Ultrafree 15 (Millipore) filters. After 1 ml of 0.3 M sodium phosphate (pH 5.8) containing 1 M ammonium sulfate was added to the concentrated fraction, the solution was applied to a Resource PHE (Amersham Bioscience) column (6 ml) which had been equilibrated with 0.1 M sodium phosphate (pH 5.8) containing 1 M ammonium sulfate and then was eluted with 30 ml of the same buffer. This fraction was concentrated to 3.5 ml with Ultrafree 15 filters, desalted by gel filtration on a PD-10 column (Amersham Bioscience) with 50 mM Tris-HCl (pH 7.5), applied to a Resource Q (Amersham Bioscience) column (6 ml) which had been equilibrated with 50 mM Tris-HCl (pH 7.5), and eluted with 18 ml of 50 mM Tris-HCl (pH 7.5). This fraction was concentrated with Ultrafree 15 filters, applied to a Superdex 200 preparation-grade (Amersham Bioscience) column (16 by 600 mm), and eluted with 50 mM sodium phosphate buffer containing 0.15 M sodium chloride (pH 7.0). Finally, the active fraction was eluted at an apparent molecular mass of about 77 kDa by using catalase (232,000 Da), aldolase (158,000 Da), bovine serum albumin (67,000 Da), ovalbumin (43,000 Da), and chymotrypsinogen A (25,000 Da) as standard proteins.

Detection of enzymatic activity.

Soybean saponin hydrolase activity was detected as follows. Thirty microliters of enzyme solution was mixed with 50 μl of 2% soybean saponin suspension and 20 μl of 0.5 M sodium phosphate (pH 5.8), and the reaction mixture was incubated at 37°C for 1 h. The reaction mixture was developed by thin-layer chromatography with a solvent system of chloroform-methanol (95:5), and the enzymatic activity was confirmed by detecting soyasapogenol B at an R_f of 0.35 by spraying 5% vanillin in 50% H₂SO₄ and heating at 120°C for 5 min.

Measurement of enzymatic activity.

Soybean saponin hydrolase activity was measured as follows. To 50 μl of 2% soybean saponin suspended in 0.2 M sodium phosphate (pH 5.8), 50 μl of enzyme solution was added, and the mixture was allowed to react at 37°C for 30 min. Reaction products were extracted with 100 μl of ethyl acetate, and 50 μl was diluted with 450 μl of the mobile phase. Ten microliters of this dilution was analyzed by high-pressure liquid chromatography (HPLC), and the quantity of soyasapogenol B in the sample was determined by comparison with authentic soyasapogenol B. HPLC was performed with a Shimadzu (Kyoto, Japan) model LC-VP apparatus under the following conditions: column, Inertsil ODS-2 (4.6 by 250 mm; GL Sciences, Tokyo, Japan); column temperature, 40°C; mobile phase, acetonitrile-methanol-water (50:35:15); flow rate, 0.8 ml/min; and UV wavelength, 210 nm. Glycyrrhizin and glycyrrhetic acid monoglucuronide hydrolase activities were measured by determining the quantity of glycyrrhetic acid under the conditions described for the measurement of soybean saponin hydrolase activity, except that the mobile phase for HPLC was acetonitrile-methanol-water-trifluoroacetic acid (50:35:15:0.05). β-Glucuronidase activity was measured as follows. To 50 μl of 1 mM 4-methylumbelliferyl-β-d-glucuronide in 0.2 M sodium phosphate (pH 5.8), 50 μl of enzyme solution was added, and the mixture was incubated at 37°C for 30 min. After the reaction was stopped by the addition of 50 μl of 2 M Na₂CO₃, 4-methylumbelliferone was quantified by measuring the intensity of fluorescence emitted from the reaction fluid at an excitation wavelength of 340 nm and an emission wavelength of 450 nm. One unit of activity of each enzyme is defined as the amount of enzyme that produces 1 nmol of aglycone per min from the substrate.

Protein determination and SDS-PAGE.

Protein contents of pooled fractions and purified proteins were measured by using a protein assay kit (Bio-Rad) with bovine gamma globulin as the standard. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was performed by using an SDS-PAGEmini apparatus (TEFCO, Tokyo, Japan) according to the manufacturer's instructions. Proteins were stained with SYPRO ruby gel stain (Molecular Probes) and detected with Molecular Imager FX (Bio-Rad).

Analysis of amino acid sequences.

Analysis of amino acid sequences was performed with an Applied Biosystems model 492 protein sequencer. The N-terminal amino acid sequence of the enzyme was analyzed by using the purified protein blotted on Immobilon-P^SQ (Millipore). Peptides generated by digesting the purified protein with trypsin in an SDS-polyacrylamide gel (27) were separated by HPLC, and the N-terminal amino acid sequences were analyzed.

Cloning of the sdn1 gene and analysis of its nucleotide sequence.

The genomic DNA of N. vasinfecta var. vasinfecta PF1225 was prepared from mycelia cultured in TS medium by the method of Horiuchi et al. (8). Primer N1 (5′-CCIGCITCNGTNCCNAA-3′) and primer 4A (5′-ACICCYTCNGGRTTRTG-3′) were prepared on the basis of the amino acid sequences of the N terminus of the purified protein and peptide Trp32.07, respectively. PCR was performed as follows with Takara Taq (Takara Bio, Shiga, Japan) and the genomic DNA as a template. After heat denaturation at 94°C for 1 min, a cycle of 94°C for 30 s, 45°C for 30 s, and 55°C for 3 min was repeated 10 times, and a cycle of 94°C for 30 s, 47°C for 30 s, and 60°C for 3 min was repeated 20 times. The amplified DNA fragment was purified by agarose gel electrophoresis and used as a probe for screening of the genomic DNA library. The genomic DNA library of N. vasinfecta var. vasinfecta PF1225 was constructed by ligating the genomic DNA partially digested with Sau3AI to vector EMBLIII (Stratagene) which had been digested with BamHI and EcoRI. The genomic DNA library was screened with a digoxigenin High Prime DNA labeling and detection starter kit (Roche Diagnostics). The nucleotide sequence was determined by the chain termination method (28).

5′ RACE and 3′ RACE analyses.

Analysis by rapid amplification of cDNA ends (RACE) was carried out as follows. mRNA was purified from total RNA prepared with Isogen (Wako Pure Chemical Co.) from the mycelia of N. vasinfecta var. vasinfecta PF1225 cultured in MY medium by using an <Oligotex-dT30> Super kit (Roche Diagnostics, Tokyo, Japan). 5′ RACE and 3′ RACE analyses were performed with a 5′/3′RACE kit (Roche Diagnostics) according to the manufacturer's instructions.

Plasmid construction.

The coding region of the sdn1gene was amplified by PCR with cDNA from N. vasinfecta var. vasinfecta PF1225 as a template. The primers used for PCR were as follows: 5′-GGGCCCGGGGCGCATCATGCACTTCTTTGACAAAGCGAC-3′ and 5′-GGGCTGCAGTTAAGTGCCGCTCTGAGGACT-3′. The PCR product was digested with SmaI and PstI and ligated to pCB1-M2 (34) which had been digested with StuI and PstI, yielding plasmid pCB-SB. After pFB6 (containing the pyr4 gene derived from Neurospora crassa) (5) was digested with Bg1II and then partially digested with HindIII, a 1.9-kbp Bg1II-HindIII fragment was recovered and ligated to LITMUS28 (New England BioLabs) which had been digested with Bg1II and HindIII, yielding plasmid pPYR4LIT281. pPYR4LIT281 was digested with Bg1II, blunted with T4 DNA polymerase, and ligated to an XbaI linker (5′-CTCTAGAG-3′), yielding plasmid pPYR4. After pPYR4 was digested with XbaI, a 1.9-kbp XbaI fragment containing the pyr4 gene was purified by agarose gel electrophoresis and inserted into the XbaI site of pCB-SB, yielding plasmid pCB-SBe.

Transformation of T. viride strain 2.

T. viride strain 2 was cultured with shaking in 50 ml of mycelium formation medium containing 100 μg of uridine/ml for 2 days, and the mycelia were collected by centrifugation and washed with 0.5 M sucrose solution. The washed mycelia were suspended in 3 mg of β-glucuronidase (Sigma-Aldrich Fine Chemicals)/ml-1 mg of Zymolyase 20T (Seikagaku Corporation, Tokyo, Japan)/ml-1 mg of chitinase (Sigma-Aldrich Fine Chemicals)/ml-0.5 M sucrose solution, and protoplasts were generated by incubation at 30°C for 90 min with gentle shaking. The protoplasts were collected by centrifugation, washed with SUTC buffer (10 mM Tris-HCl, 10 mM CaCl₂, 0.5 M sucrose [pH 7.5]), and suspended in 500 μl of SUTC buffer. To 100 μl of the protoplast suspension, 10 μl of plasmid DNA solution was added. The mixture was allowed to stand on ice for 5 min, mixed with 400 μl of PEG solution (60% polyethylene glycol 4000, 10 mM Tris-HCl, 10 mM CaCl₂ [pH 7.5]), and again placed on ice for 20 min. The PEG solution-treated protoplasts were washed with SUTC buffer and layered on plates with minimum medium soft agar. After incubation at 28°C for 5 days, colonies that appeared were transferred to fresh minimum medium, and colonies that could grow on fresh minimum medium were regarded as transformants.

Purification of the recombinant enzyme.

The culture broth of T. viride SDNrec9/pCB-SBe (see Results) was centrifuged, and a supernatant was obtained. The supernatant, supplemented to contain 50 mM Tris-HCl (pH 7.5) and 1 M ammonium sulfate, was applied to a Toyopearl Butyl-650C column (50 by 160 mm) which had been equilibrated with 50 mM Tris-HCl (pH 7.5) containing 1 M ammonium sulfate and then was eluted with 50 mM Tris-HCl (pH 7.5) containing ammonium sulfate at a linear concentration gradient of 1 to 0 M. The fractions that eluted at ammonium sulfate concentrations of 1 to 0.2 M were pooled. The pooled fraction was applied to a Toyopearl Butyl-650S column (26 by 200 mm) which had been equilibrated with 0.1 M sodium phosphate (pH 5.8) containing 1 M ammonium sulfate and then was eluted with 0.1 M sodium phosphate (pH 5.8) containing ammonium sulfate at a linear concentration gradient of 1 to 0 M. The nonadsorbed fraction and the fraction that eluted at 1 to 0.7 M were pooled. The pooled fraction was concentrated with Centricon Plus-20 (Millipore), applied to a Superdex 200 preparation-grade column (26 by 600 mm), and eluted with 0.05 M sodium phosphate (pH 7.0).

Deglycosylation of the purified enzyme.

Deglycosylation of the purified enzyme was done with an N-glycosidase F deglycosylation kit (Roche Diagnostics) according to the manufacturer's instructions.

Purification of soybean saponin.

Sixty grams of soybean saponin was mixed with silica gel 60 (Kanto Kagaku, Tokyo, Japan); the gel was placed in a column (inner diameter, 120 mm), washed with 12 liters of chloroform, and then eluted with chloroform-methanol-water (70:25:4). Three fractions of soyasaponins (soyasaponin I, 9.2 g; soyasaponin II, 4.9 g; and soyasaponin V, 1.5 g) were obtained. Each fraction was subjected to reverse-phase chromatography (mobile phase, acetonitrile-water-trifluoroacetic acid [40:60:0.05]) with Develosil ODS-15/30 (Nomura Chemical, Aichi, Japan), and 920, 490, and 150 mg of soyasaponins I, II, and V, respectively, was fractionated. The structures of soyasaponins I and II (12) and soyasaponin V (6) were confirmed by liquid chromatography-mass spectrometry (Fig. 1).

FIG.1.

Structures of soyasaponins (A) and glycyrrhizins (B). GlcA, glucuronic acid; Gal, galactose; Rha, rhamnose; Ara, arabinose; Glc, glucose.

NMR analysis of enzyme reaction products.

Nuclear magnetic resonance (NMR) analysis was carried out as follows. After 30 mg of soyasaponin I was dissolved in 6 ml of 0.1 M sodium phosphate buffer (pH 5.8), 150 μl of enzyme solution (1 mg/ml) was added, and the mixture was incubated at 37°C for 18 h. The reaction mixture was extracted with 6 ml of butyl acetate after the addition of 2 ml of 5 M NaCl. The water layer was lyophilized, dissolved in 2 ml of water, and subjected to gel filtration (Toyopearl HW-40; 16 by 600 mm); 10 mg of water-soluble enzyme reaction product was obtained. Structural elucidation of the sample was performed by ¹³C NMR spectroscopy (JNM-AL400; JEOL) with D₂O.

Nucleotide sequence accession number.

The nucleotide sequence reported in this study has been submitted to the DDBJ, EMBL, and GenBank databases under accession number AB110615.

RESULTS

Purification and characterization of a soyasapogenol B-producing enzyme and isolation of the gene encoding it.

Soyasapogenol B-producing enzymatic activity was detected in the supernatant obtained by culturing N. vasinfecta var. vasinfecta PF1225 in soybean saponin-containing medium but not in medium without soybean saponin, suggesting that the enzyme is induced by soybean saponin. After N. vasinfecta var. vasinfecta PF1225 was cultured in MY medium containing soybean saponin, the soyasapogenol B-producing enzyme was purified from the supernatant (Table 1; see also Materials and Methods) until a single band of 77 kDa was observed on SDS-PAGE (Fig. 2, lane 2). Judging from SDS-PAGE analysis, the protein was pure, and the enzymatic activity eluted as a protein having a mass of about 77 kDa during gel filtration, indicating that it was a monomer.

TABLE 1.

Purification of soybean saponin hydrolase from N. vasinfecta var. vasinfecta PF1225

Purification step	Protein (μg)	Total activity (U)	Sp act (U/μg)	Purifica- tion factor (fold)	Recovery (%)
Culture supernatant	145,100	16,380	0.11	1	100
Ammonium sulfate precipitation	75,700	7,920	0.11	1.0	48
Toyopearl Butyl-650S	13,452	6,574	0.49	4.5	40
Resource PHE	2,930	2,685	0.92	8.4	16
Resource Q	381	2,260	5.93	53.9	14
Superdex 200 prep- aration grade	178	1,620	9.10	82.7	10

FIG. 2.

SDS-PAGE analysis of purified saponin hydrolases. Purified saponin hydrolases treated or not treated with N-glycosidase F were subjected to SDS-PAGE, stained with SYPRO ruby gel stain, and detected with Molecular Imager FX. An arrow indicates the position of N-glycosidase F. Lane 1, molecular mass markers: phosphorylase b (97.0 kDa), albumin (66.0 kDa), ovalbumin (45.0 kDa), carbonic anhydrase (30.0 kDa), and trypsin inhibitor (20.1 kDa). Lane 2, saponin hydrolase purified from N. vasinfecta var. vasinfecta PF1225. Lane 3, N-glycosidase F-treated saponin hydrolase from N. vasinfecta var. vasinfecta PF1225. Lane 4, saponin hydrolase purified from T. viride. Lane 5, N-glycosidase F-treated saponin hydrolase from T. viride.

The N-terminal amino acid sequence of the purified enzyme was ASPPASVPNPPSPEPITLKQ. The amino acid sequences of five peptides obtained by digestion of the enzyme with trypsin were LVFNPSPK (Trp26.8), WNVAADGSGPSGEIR (Trp27.59), VTILHNPEGVAPITAK (Trp32.07), EHSDTIPWGVPYVPGSQ (Trp39.43), and LTDYSFDWYSDIR (Trp41.3). In order to isolate the gene encoding this enzyme, two degenerate primers were designed on the basis of the N-terminal amino acid sequence and the amino acid sequence of peptide Trp32.07. PCR was performed with genomic DNA from N. vasinfecta var. vasinfecta PF1225 as a template; a 0.7-kbp DNA fragment was specifically amplified. With this DNA fragment as a probe, a genomic DNA library constructed in EMBLIII was screened, and a positive clone was obtained.

Southern blot analyses of the phage DNA prepared from the positive clone and the genomic DNA from N. vasinfecta var. vasinfecta PF1225 indicated that a 7-kbp PstI fragment with which the probe hybridized was present in both DNA samples. The 7-kbp PstI fragment was subcloned in pUC18. By sequencing of the region that hybridized with the probe, a nucleotide sequence of 2,220 bp was determined (Fig. 3); it contained an open reading frame (ORF) of 1,914 bp. On the basis of 5′ RACE and 3′ RACE analyses (28), the start point of transcription was estimated to be 54 bp upstream of the putative initiation codon, and the poly(A) addition was estimated to be 129 bp downstream of the putative termination codon. These results suggest that the estimation of the ORF is correct. The 3′ RACE analysis with the primer (5′-GTCTTGGGAATTCTCTCGGCAAG-3′) corresponding to the 5′-terminal region of the estimated ORF confirmed that no in-frame intron was present. The amino acid sequence deduced from the ORF contained the N-terminal and five internal amino acid sequences determined for the purified enzyme. In addition, a sequence of 26 amino acids that was present upstream of the N-terminal amino acid sequence of the purified enzyme showed characteristics of a typical signal sequence, probably involved in secretion of this enzyme into culture medium. Based on its predicted sequence, the mature enzyme consisted of 612 amino acids and had a molecular mass of 65,724 Da. The fact that this value was smaller than the molecular mass of the purified enzyme determined by SDS-PAGE (about 77 kDa) suggested that the purified enzyme was modified by glycosylation. Therefore, the purified enzyme was treated with endoglycosidase F, which removes carbohydrate chains attached to asparagine residues, and analyzed by SDS-PAGE. The deduced mass of the deglycosylated enzyme, about 67 kDa (Fig. 2, lane 3), was in close agreement with the molecular mass estimated from the amino acid sequence. There are three potential asparagine-linked glycosylation sites (Asn-X-Ser or Thr; X is not Pro) in the deduced amino acid sequence, and all or most may be modified by asparagine-linked glycosylation. From these results, we tentatively concluded that this gene codes for a soyasapogenol B-producing enzyme of N. vasinfecta var. vasinfecta PF1225 and named it sdn1.

FIG. 3.

Nucleotide and deduced amino acid sequences of saponin hydrolase from N. vasinfecta var. vasinfecta PF1225. Nucleotide sequence numbering defines the adenine of the translational initiation codon as 1. Underlining indicates the amino acid sequences determined by N-terminal amino acid sequence analyses of the purified enzyme and peptides, and double underlining indicates the amino acid sequences used for the design of PCR primers. Lowercase letters indicate a putative signal sequence. Asparagine residues in bold type indicate potential asparagine-linked glycosylation sites.

BLAST analyses with the entire amino acid sequence or portions of it deduced from the sdn1 gene were carried out against protein databases, but no homologous sequence exceeding a score of 42 could be found. Moreover, no significant motifs were found when Motiffind, Profilefind, BLIMPS, BLASTP, or Hmmpfam was used. Against the translated databases, however, one homologous sequence (score, 173; expect value, 1e-41) was found; it represents an expressed sequence tag (EST) clone, VD0107B03, derived from Verticillium dahliae (GenBank accession no. BQ110383) and for which the function is unknown (20).

Heterologous expression of sdn1.

T. viride strain 2, which was used as the host, is a uracil-requiring mutant strain and is complemented by the pyr4 gene derived from N. crassa. We constructed plasmid pCB-SBe for the expression of the sdn1 gene in T. viride as described in Materials and Methods. This plasmid contained the pyr4 gene as a selectable marker of transformants; the sdn1 gene was under the control of the promoter and terminator of the cbh1 gene derived from the host (34). Following transformation, three transformants per 1 μg of plasmid DNA were obtained.

Randomly selected transformants (25 independently isolated) were cultured under conditions that induced the cbh1 promoter, and the culture supernatants were examined for soyasapogenol B-producing activity. Enzymatic activities were detected in the supernatants of 16 strains. The highest enzymatic activity (in strain SDNrec9; 25,722 U/ml) was 2,820 times higher than the activity in the culture supernatant of N. vasinfecta var. vasinfecta PF1225. The recombinant enzyme was purified from the culture supernatant of strain SDNrec9 as described in Materials and Methods, and the purified recombinant enzyme showed a single band on SDS-PAGE (Fig. 2, lane 4). The N-terminal amino acid sequence of the purified recombinant enzyme, ASPPASV, was the same as that of the endogenous enzyme from N. vasinfecta var. vasinfecta PF1225. The molecular mass of the recombinant enzyme determined by SDS-PAGE was about 69 kDa—smaller than that of the natural enzyme (77 kDa)—and decreased to about 67 kDa after treatment with endoglycosidase F, in agreement with the mass of the natural enzyme after treatment with endoglycosidase F (Fig. 2, lanes 3 and 5). The recombinant enzyme was estimated to be less modified by carbohydrate chains than the natural enzyme but to have the same polypeptide chain as the natural enzyme. Purified recombinant and natural enzymes displayed indistinguishable pH and temperature optima when soybean saponin was used as the substrate. Both enzymes were most active at pH 5.6 and 40 to 42°C. When the specific activities of both enzymes were measured at pH 5.8 and 37°C, they were nearly the same, 9.10 and 9.57 U/μg for the natural and recombinant enzymes, respectively. Since the properties of the natural and recombinant enzymes were nearly identical, the difference in the degree of modification of the carbohydrate chains apparently did not affect enzymatic activity.

Substrate specificity.

The substrate specificity of the purified recombinant enzyme was measured with soyasaponin I, soyasaponin II, and soyasaponin V (Table 2). The specific activities with soyasaponins I, II, and V were 16.3, 5.05, and 0.0057 U/μg, respectively, and the ratio of the activities with soyasaponins I, II, and V was 2,860:886:1. This enzyme also can act on glycyrrhizin and glycyrrhetic acid monoglucuronide when glycyrrhetic acid, which is an oleanane-type triterpenoid and which is similar in structure to soyasapogenol B, is the aglycone; its specific activities with these substrates were 0.371 and 0.088 U/μg, respectively, values which were 65 and 15 times higher than the value with soyasaponin V. On the other hand, when the substrate was 4-methylumbelliferyl-β-glucuronide, which is a common substrate for β-glucuronidase, its activity was lower than the detection limit (<0.000014 U/μg).

TABLE 2.

Substrate specificity of soyasaponin hydrolase

Substrate	Sp act (U/min/μg)	Relative activitya
Soyasaponin I	16.3	2,680
Soyasaponin II	5.05	886
Soyasaponin V	0.0057	1
Glycyrrhizin	0.371	65
Glycyrrhetic acid 3-O-glucuronide	0.088	15
4-Methylumbelliferyl-β-d-glucuronide	Not detected

^aRatio of specific activity of substrate to that of soyasaponin V.

Mode of action of the enzyme.

In order to clarify the mode of action of this enzyme, soyasaponin I was treated with the recombinant enzyme, and the reaction products were analyzed by thin-layer chromatography. With our solvent system (n-butanol-acetic acid-water [4:3:1]), an unknown spot with an R_f value of about 0.19 was detected in addition to the aglycone, soyasapogenol B. To examine the structure of the unknown compound, it was purified from the enzyme reaction mixture. After treating about 30 mg of soyasaponin I with 1,505 U of recombinant enzyme, the reaction mixture was extracted with butyl acetate to remove soyasapogenol B. The water layer was concentrated by lyophilization and subjected to gel filtration chromatography (Toyopearl HW-40), and 10 mg of the purified unknown compound was obtained. It was dissolved with deuterium oxide and analyzed by ¹³C NMR. Carbonyl carbons at 178.1 and 179.1 ppm, a hydroxymethyl group at 63.5 and 63.7 ppm, and a methyl group at 19.3 ppm were detected. These signals were in agreement with those derived from glucuronic acid, galactose, and rhamnose. Furthermore, three anomeric carbons were observed at 94.6 and 97.2 ppm (from glucuronic acid), 103.6 and 105.4 ppm (from galactose), and 103.2 and 103.3 ppm (from rhamnose). From these results, the unknown compound was identified to be α-l-rhamnopyranosyl (1→2)-β-d-galactopyranosyl (1→2)-d-glucuronopyranoside. Therefore, this enzyme was shown to be a hydrolase generating soyasapogenol B and triose from soyasaponin.

DISCUSSION

In this study, we purified Sdn1, a soybean saponin hydrolase from N. vasinfecta var. vasinfecta PF1225, and defined its primary structure by analyzing its corresponding gene. We also purified the recombinant enzyme produced by a T. viride transformant and analyzed it to clarify its enzymatic properties.

Similarly, Kudou et al. purified an enzyme that hydrolyzes soybean saponin from A. oryzae KO-2 and reported its properties (14). This enzyme was similar to Sdn1 in that it hydrolyzed soyasaponin I to soyasapogenol B and triose. However, while soyasaponin I is the best substrate for our enzyme, glycyrrhizin is a better substrate than soyasaponin I for the enzyme derived from A. oryzae KO-2. Moreover, while our enzyme proved to be a monomer with a molecular mass of about 77 kDa, the enzyme derived from A. oryzae KO-2 is a heterotetramer consisting of subunits with molecular masses of 45 and 35 kDa. Thus, the enzyme described in this report differed from the enzyme derived from A. oryzae KO-2 in substrate specificity and subunit structure. In addition, an enzyme that hydrolyzes glycyrrhizin has been purified from A. niger GRM3, and its properties have been reported (19). Although the soyasaponin-hydrolyzing activity of this enzyme was not mentioned, its molecular mass determined by gel filtration was about 150 kDa, different from the 77-kDa mass of Sdn1. Therefore, the enzyme described in this study is considered to be different from the enzyme derived from A. niger GRM3 as well and to be a novel saponin hydrolase that can hydrolyze soybean saponin.

The molecular mass of native Sdn1 from N. vasinfecta var. vasinfecta PF1225 was different from that of recombinant Sdn1 from T. viride. Both enzymes were shown to be glycosylated by the data obtained with endoglycosidase F treatment. Fungal asparagine-linked carbohydrate chains are composed of 2 N-acetylglucosamines and 6 to 11 mannose residues, in some cases more glucose or galactose residues, and the structures of asparagine-linked carbohydrate chains vary from strain to strain (15). The difference in molecular masses between native Sdn1 and recombinant Sdn1 is due to differences in the numbers and structures of the asparagine-linked carbohydrate chains. Although the carbohydrate chains of many glycoproteins confer important physical properties, such as conformational stability, protease resistance, charge, and water-binding capacity (23), the difference in the degree of modification of the carbohydrate chains apparently did not affect the enzymatic activity of Sdn1.

By purifying soyasaponins I, II, and V from soybean saponin and treating them with the enzyme, we found that the velocity of hydrolysis was substrate dependent: soyasaponin V was the least hydrolyzable, followed by soyasaponin II, and soyasaponin I was the most hydrolyzable. These substrates have the same aglycone skeleton modified by a glucuronic acid at the C-3 position but are different in two of the three other saccharide modifications. The fact that the velocity of hydrolysis varied with these substrates suggested that the enzyme recognizes the carbohydrate moiety of the substrates, that it has subsites which recognize at least three saccharides, and that it shows different hydrolysis kinetics because of differences in the affinities of the subsites for monosaccharides. On the other hand, this enzyme also could hydrolyze glycyrrhizin and glycyrrhetic acid monoglucuronide. While the aglycone parts of these substrates have structures similar to soyasapogenol B, the carbohydrate moieties are not trioses but rather are diglucuronide and monoglucuronide in glycyrrhizin and glycyrrhetic acid monoglucuronide, respectively. These observations suggest that, in the hydrolysis reaction of this enzyme, the carbohydrate moiety need not be a triose but that the C-1 position of β-glucuronic acid must be attached to the aglycone.

As described above, since this enzyme recognizes the β-anomer of glucuronic acid and hydrolyzes its bond, it may be regarded as a β-glucuronidase. However, as no enzymatic activity was detected when a typical substrate of β-glucuronidase, 4-methylumbelliferyl-β-d-glucuronide, was used as the substrate, the structure of the aglycone also is presumed to affect enzymatic activity. Furthermore, the primary structure of this enzyme showed no significant homology with any of the amino acid sequences registered in databases, including those of most known β-glucuronidases and carbohydrate hydrolases. Thus, this enzyme may have evolved independently.

Many plants produce saponins with antifungal activities, such as α-tomatine of tomato and avenacin A-1 of oats, that are effective for the prevention of infection by phytopathogenic fungi (17). On the other hand, phytopathogenic fungi produce enzymes that hydrolyze the carbohydrate moieties of saponins (17). Since deglycosylated saponins show reduced antifungal activities, saponin hydrolases of phytopathogenic fungi appear to detoxify saponins and facilitate their virulence in host plants. Since some strains of N. vasinfecta are able to infect soybean (33), the enzyme described in this study also may have evolved for the detoxification of saponins in host plants. The only sequence homologous to the enzyme in this study is derived from an EST clone of V. dahliae, which is known to be a phytopathogenic fungus for a variety of crops (7, 24). It is quite possible that the EST clone is part of a saponin detoxification enzyme. The fact that the saponin hydrolase analyzed here is homologous to the EST clone supports the hypothesis described above. This hypothesis may explain why the enzyme recognizes the aglycone derived from saponins in its hydrolysis reaction and why it shows no homology in the primary structure with carbohydrate hydrolases that have evolved for the acquisition of nutritional sources, such as β-glucuronidases, which are widely present in bacteria. On the other hand, tomatinase produced by the wilt pathogen Fusarium oxysporum f. sp. lycopersici hydrolyzes tomatine to aglycone and a carbohydrate moiety, but this enzyme shares sequence homology with xylanases, although it exhibits no xylanase activity (26).

In this study, we discovered a novel soybean saponin hydrolase and defined its primary structure. This enzyme recognizes the carbohydrate moiety of saponin, catalyzes a hydrolysis reaction that generates aglycone and a carbohydrate moiety but, interestingly, shows no significant homology with any of the known enzymes involved in carbohydrate hydrolysis. It will be interesting to clarify what kind of fungus has this type of enzyme, what the three-dimensional structure of the enzyme is, and how the enzyme recognizes substrates and catalyzes hydrolysis.