Introduction

β-1,3–1,6-Glucans consist of glucose residues linked by β-1,3 bonds with attached side chain glucose residues joined by β-1,6 linkages, and are found in a variety of natural resources such as fungi, yeasts, bacteria and seaweeds (Chen and Sevior 2007; Bobadilla et al. 2013; Zykova et al. 2014; Kuda et al. 2015). The importance of β-1,3–1,6-glucans as food supplements and clinical materials is increasing because of their biofunctional activities (Chen and Sevior 2007; Dalonso et al. 2015), such as antitumor, immunomodulatory, and antioxidant properties. In particular, the clinical potential of water-soluble β-1,3–1,6-glucans (laminaran from seaweed Kuda et al. 2015 and schizophyllan from Schizophyllum commune Zhong et al. 2015), and alkaline water-soluble lentinan from Lentinula edodes, (Sun et al. 2015) has been reported to date.

β-1,3–1,6-Glucan is also obtained by fermentation using the black yeast Aureobasidium pullulans, and its antitumor (Kimura et al. 2006), immunomodulatory (Tada et al. 2008; Le et al. 2010; Tanioka et al. 2013), food allergy inhibitory (Kimura et al. 2007a), and other medicinal activities (Kubala et al. 2003; Kimura et al. 2007b; Shin et al. 2007; Joo-Wan et al. 2012; Guzman-Villanueva et al. 2014; Kim et al. 2015) have been reported. A. pullulans β-1,3–1,6-glucan cannot be hydrolyzed completely by enzymes such as β-1,3-glucanase and/or β-1,6-glucanase because it is not water-soluble, which presents challenges for its utilization as a medical and/or bio-industrial material. Some research has been conducted on an partially isolated water-soluble, low-molecular weight fraction of A. pullulans β-1,3–1,6-glucan (Kimura et al. 2006, 2007a, b). However, there are no reports dealing with the enzymatic hydrolysis of intact water-soluble A. pullulans β-1,3–1,6-glucan.

In this paper, we first characterized and confirmed the chemical structure of hydrothermally treated β-1,3–1,6-glucan from A. pullulans using techniques such as GC/MS and NMR. Next, we investigated the water-solubility of the samples. Water solubility is important because it greatly impacts the utility of this functional carbohydrate in many industrial fields.

Materials and methods

Preparation of hydrothermally treated A. pullulans β-1,3–1,6-glucan

Aureobasidium pullulans ATCC 20524 was cultured in liquid medium consisting of 0.6% (w/v) sucrose, 0.2% (w/v) rice bran and 0.2% (w/v) ascorbic acid at 23 °C for 72 h. β-1,3–1,6-Glucan was then harvested by the addition of burnt alum (Bulson et al. 1984; Savangikar et al. 1985; Asif et al. 2016). The obtained β-1,3–1,6-glucan was subjected to hydrothermal treatment (180 °C for 15 min at pH 5.5), concentrated by ultrafiltration, autoclaved for 15 min at 121 °C, and lyophilized for 48 h.

Monosaccharide compositional analysis

Sample (10 mg) was hydrolyzed with 3 mL of 2 N H2SO4 at 105 °C for 6 h and then 15 mL distilled water was added. After neutralization with BaCO3 and centrifugation (13,000×g, 15 min), the solution was evaporated to dryness. The residue was dissolved in 2 mL of distilled water and the product was identified using high performance liquid chromatography (HPLC, Shimadzu LC-10A; Shimadzu, Kyoto, Japan, fitted with a Sugar-D column 4.6 × 25 mm; Nacalai Tesque, Kyoto, Japan) under the following conditions: temperature, 30 °C; mobile phase, acetonitrile:water (80:20, v/v); flow rate, 1 mL/min; and RI detector. The sample was also analyzed using thin layer chromatography (TLC) on silica gel plates (Silica Gel 60, Merck Co., Darmstadt, Germany) using a mixed solvent (n-butanol:ethanol:water = 5:2:1, v/v) for development.

Fourier transform infrared spectrometry (FT-IR) analysis

A KBr tablet containing sample powder (approx. 1%, w/w) was analyzed using FT-IR (FT/IR-300, JASCO Corporation, Tokyo, Japan).

Methylation and gas chromatograph/mass spectrometry (GC/MS) analysis

Sample (5 mg) was methylated three times with 1 mL sodium hydroxide/dimethyl sulphoxide (40 mg/mL) and 0.5 mL methyl iodide at 30 °C for 30 min. The methylated product was isolated by partitioning between CHCl3 and H2O (3:1, v/v). The organic layer containing product was washed with 3 mL of water three times and dried. The resulting partially methylated product was hydrolyzed, reduced, acetylated and analyzed using a Shimadzu GCMS-QP2010SE with an InerCap Rtx-5MS capillary column (60 m × 0.25 mm × 0.25 μm). The oven temperature was initially 180 °C during injection for 5 min, then was increased at 2 °C/min to 230 °C and held at this temperature for 5 min. The partially methylated alditol acetates were identified by their relative retention times on GC and by their fragment ions in EI-MS using NIST08 Mass Spectral Library.

Nuclear magnetic resonance (NMR) analysis

The sample was dissolved in D2O (30 mg/mL) to exchange the active hydrogens and was then lyophilized. This process was repeated three times. The 1H NMR and 13C NMR spectra were obtained on a Bruker DPX 400 MHz NMR for δH and a 100 MHz NMR for δC at room temperature (25 °C) in D2O. HSQC was carried out using standard Bruker software. Chemical shifts are given as δ values with reference to 1,4-dioxsane used as an internal standard, and coupling constants are given in Hz.

Measurement of solubility

Sample solution (1.2 g/10 mL) was kept at each temperature for 30 min with reference to The Japanese Pharmacopeia Sixteen Edition and then centrifuged at 2220×g for 30 min at the same temperature. The obtained supernatants were dried overnight at 105 °C and the solubilities of the residues were determined as % (w/w).

Molecular weight estimation

The molecular weight of each sample (10 mg/mL) was estimated by high performance gel permeation chromatography (HPGPC, Shimadzu LC-10A; Shimadzu) fitted with a TSKgel G5000PWxL-CP column (7.8 × 30 mm; Tosoh Corporation, Tokyo, Japan) under the following conditions: temperature, 40 °C; mobile phase, 20 mmol/L phosphate buffer, pH 6.8; flow rate, 0.5 mL/min; and RI detector. The average molecular weight was measured using standard polysaccharide pullulans (Shodex STADARD P Series, SHOWA DENKO K. K., Tokyo, Japan).

Enzymatic hydrolysis

Lysing enzymes (primarily β-glucanase, cellulase, protease and chitinase; Sigma-Aldrich Japan, Tokyo, Japan) and Uskizyme (primarily β-1,3-glucanase and chitinase; Wako Pure Chemical Industries Ltd, Osaka, Japan) were used. Lysing enzymes were dialyzed against MacIlvain buffer (pH 7.5) to remove glucose from the preparation prior to sample hydrolysis. The reaction mixture containing 500 μL of sample as substrate (6% w/v) in 150 mmol/L MacIlvain buffer (pH 5.5) and 500 μL of Lysing enzymes (80 U) or Uskizyme (25 U) was incubated for 24 h at 40 °C. The obtained glucose and gentiobiose were determined using HPLC using the same conditions as described above. Total yields of glucose and gentiobiose are shown as the percentage of the initial substrate concentration in the reaction mixture.

Results

Monosaccharide component

The results of monosaccharide component analysis using HPLC are shown in Fig. 1. A single peak was detected from the hydrolyzed sample and the retention time was 17.35 min, essentially the same as that of standard glucose (17.24 min). The yield of glucose from the hydrolyzed sample was 98% (w/w). The hydrolyzed sample also provided a single spot by TLC whose retention factor value (R f value) was 0.65, the same as that of standard glucose (0.65) (Fig. 2). From these results, the monosaccharide component of the sample was concluded to be glucose.

Fig. 1
figure 1

Monosaccharide component analysis using HPLC

Fig. 2
figure 2

Monosaccharide component analysis using TLC

FT-IR analysis

The IR spectrum of the sample is shown in Fig. 3. Absorbances due to β-glucosidic linkage (886–900 and 950–954 cm−1 Mohacek-Grosev et al. 2001, 889–890 cm−1 Tanioka et al. 2013 and 892 cm−1 Novak et al. 2012), pyranose ring (1022–1079 cm−1 Novak et al. 2012, 1200–1206 cm−1 Mohacek-Grosev et al. 2001, and 2853–2922 cm−1 Novak et al. 2012) and hydroxyl group (3383 cm−1 Novak et al. 2012) were observed and identified by reference to the values provided in previous reports. From the above results, the sample was concluded to be β-glucan.

Fig. 3
figure 3

IR spectra of the sample

GC/MS analysis

The total ion monitor and analytical data of the methylated sample obtained by GC/MS analysis are shown in Fig. 4 and Table 1, respectively. As shown in Table 1, the linkages between the glucosyl residues were determined to be Glc-(1→, →3)-Glc-(1→, →3,6)-Glc-(1→ according to the values of mass fragments (m/z) described in previous reports (Mizuno et al. 1999; Rout et al. 2008). From the above results, the sample was concluded to be 1,3-1,6-glucan.

Fig. 4
figure 4

Total ion monitor of GC/MS analysis of the methylated sample

Table 1 Linkage analysis of the sample

NMR analysis

The HSQC (1H–13C) spectra of the sample are shown in Fig. 5 and the values of the chemical shifts in the anomeric region of the sample are summarized in Table 2. From the chemical shifts of the anomeric carbons, the mode of the glycosidic linkages of Glc-(1→, →3)-Glc-(1→, →3,6)-Glc-(1→ were determined to be β due to the detection of cross-peaks of anomeric carbons at around 103 ppm and anomeric protons at around 4.5 or 4.7 ppm. From these results, the sample was concluded to be β-1,3–1,6-glucan according to the values of chemical shifts described in previous reports (Kimura et al. 2007a; Pramanik et al. 2007; Santos-Neves et al. 2008; Sumidele et al. 2008; Tada et al. 2008; Carbonero et al. 2012; Fang et al. 2012; Ruthes et al. 2013; Dalonso et al. 2015).

Fig. 5
figure 5

HSQC NMR analysis of the sample

Table 2 Chemical shifts of anomeric carbons and protons for each type of glucose residue in the sample

Solubility

The solubility of hydrothermally treated A. pullulans β-1,3–1,6-glucan is shown in Fig. 6. Although an increase in solubility was not seen as the temperature increased, the maximum solubility was approximately 10% (w/w) between 50 and 70 °C.

Fig. 6
figure 6

Solubility of the sample

Molecular weight

The HPGPC chromatogram of the sample is shown in Fig. 7. A wide range of molecular weights was observed, with the average value estimated to be approximately 128,000 Da.

Fig. 7
figure 7

Molecular weight estimation using HPGPC

Enzymatic hydrolysis

The time course of enzymatic hydrolysis of a 3% (w/v) (30 mg/mL) sample as substrate by Lysing enzymes is shown in Fig. 8. The % glucose yield to the initial substrate concentration, i.e., the hydrolysis efficiency, reached 102.2% after 1 h reaction at 40 °C, indicating complete hydrolysis of the sample. The amount of gentiobiose produced during the reaction decreased, and it was completely hydrolyzed after 24 h reaction with Lysing enzymes. After 1 h reaction, the amounts of glucose and gentiobiose reached 9.4 mg/mL (52.2 mmol/L) and 21.3 mg/mL (62.2 mmol/L), respectively. The yield of gentiobiose was 69.5% and the molar ratio of gentiobiose to glucose was 1.19.

Fig. 8
figure 8

Time course of the enzymatic hydrolysis of the sample by Lysing enzymes

The time course of the enzymatic hydrolysis of 3% (w/v) (30 mg/mL) samples by Uskizyme is shown in Fig. 9. The hydrolysis efficiency was 94.6% after 24 h reaction at 40 °C, indicating that the principal product was glucose and thus the sample was considered to be hydrolyzed by Uskizyme. After 3 h reaction, the levels of glucose and gentiobiose reached 14.6 mg/mL (81.0 mmol/L) and 11.5 mg/mL (33.6 mmol/L), respectively, with a total yield of 86.6%. The molar ratio of gentiobiose to glucose (0.41) was less than that shown in Fig. 8.

Fig. 9
figure 9

Time course of the enzymatic hydrolysis of the sample by Uskizyme

Discussion

Previously, Zhang et al. (2002) reported that anomeric carbons underwent a chemical shift from 104.1 ppm to 103.1 ppm in the 13C-NMR spectrum when the triple-helical chain of β-1,3–1,6-glucan from Lentinus edodes was changed to a single-flexible chain in DMSO. It is possible that the current samples contain single flexible chains, given the observed resonances below 130 ppm. However, further investigations to clarify the structure of the sample are necessary. It may be possible to manufacture pharmaceutical products and functional food materials using hydrothermally treated A. pullulans β-1,3–1,6-glucan, similar to the use of polysaccharides such as Laminallan and Schizophyllan, because hydrothermally treated A. pullulans β-1,3–1,6-glucan changed from being water-insoluble to water-soluble, which is a very important and valuable characteristic especially in the bio-industry.

The molecular weight of sample was slightly larger than that of Schizophyllan (76,800) and Lentinun (94,700), and smaller than that of Curdlan (>136,000) (Tanioka 2010). The molecular weight distribution of polysaccharides such as β-glucan is variable (Tsubaki et al. 2008) and wide (300,000–700,000, Lentinun CAS 37339-90-5). A wide distribution of molecular weights after hydrothermal treatment might be a feature of the heat treatment of polysaccharides.

The hydrolysis (15.9, 24 and 33%) of acidic polysaccharide from Aureobasidium by Kitalase (exo-β-1,3-glucanase) was reported previously, but no detailed reaction conditions such as the concentrations of substrate and products were provided (Hamada and Tsujisaka 1983). The detection of gentiobiose and glucose as hydrolysis products generated by Kitalase were also described previously (Kimura et al. 2006), but again, no information regarding the reaction conditions or the results were given. From the molar ratio of gentiobiose to glucose (1.19), the sample is likely highly branched, with branches originating from the C6 of glucose every one or two units of the main β-1,3-glucan chain. The above results suggest that the sample has a similar branching pattern to that previously reported for β-1,3–1,6-glucan from A. pullulans (Tada et al. 2008), and that the chemical structure of the sample is similar to that of the low-molecular weight water-soluble β-1,3–1,6-glucan from A. pullulans 1A1 strain (Kimura et al. 2006, 2007a). From the above results, it is concluded that the water-solubility of the sample rendered it sensitive to enzymatic catalysis. This property is very useful for the production of functional oligosaccharides such as gentiobiose because the yield of gentiobiose after 1 h reaction was approximately 70%.

In the present experiment, gentiobiose produced from the sample was also hydrolyzed by Uskizyme (primarily β-1,3-glucanase and chitinase). It is considered that Uskizyme also contained β-1,6-glucanase/β-1,6-glucosidase activities because we confirmed the complete hydrolysis of commercial gentiobiose (2% w/v, Wako Pure Chemical Industries Ltd) to glucose by the enzyme (data not shown). It would be far preferable to use purified enzymes such as β-1,3-glucanase/β-1,3-glucosidase and/or β-1,6-glucanase/β-1,6-glucosidase for the hydrolysis experiments, but these enzymes are currently not available commercially. In the near future we intend to purify the hydrolyzing enzymes from Lysing enzymes and/or Uskizyme to conduct our hydrolysis experiments more precisely.

From the above results, it is proposed that hydrothermally treated β-1,3–1,6-glucan from A. pullulans can be efficiently hydrolyzed by enzymatic methods. Although there are several reports on the hydrolysis of β-1,3–1,6-glucan, as described in the Introduction, this is the first evidence of the complete enzymatic hydrolysis of β-1,3–1,6-glucan from A. pullulans. Investigations to clarify why hydrothermal treatment rendered A. pullulans β-1,3–1,6-glucan water-soluble and amenable to enzymatic hydrolysis will be carried out in the near future. It is also considered that the sensitivity of the present sample to hydrolyzing enzymes is useful for the production of functional oligosaccharides such as gentiooligosaccharides (Takahashi et al. 2014; Unno et al. 2005) from A. pullulans β-1,3–1,6-glucan.

In the present report, we noted and described the newly obtained superior characteristics of hydrothermally treated β-1,3–1,6-glucan from A. pullulans and suggest that this material will help contribute to bio-industries utilizing functional carbohydrates.