Optimization on production of konjac oligo‐glucomannan and their effect on the gut microbiota

Abstract Konjac glucomannan (KGM) is a polysaccharide extracted from Amorphophallus konjac, and its degradation product is konjac oligo‐glucomannan (KOG). The aim of this study was to produce KOG from KGM and to evaluate its effect on the gut microbiota in fecal batch culture. KOG was produced by enzymatic hydrolysis using β‐mannanase. The optimum conditions were as follows: reaction temperature of 48°C, reaction time of 4 hr, pH of 5.5 and E/S of 0.05% followed by purification step using 3,000 NMWC ultrafiltration (UF) membrane pore size. The effect of KOG on changes in human fecal bacterial populations and short‐chain fatty acids (SCFAs) production was evaluated. The results showed that low‐molecular weight KOG (LKOG) from purification step with concentration of 9.54 mg/ml, and a prebiotic index (PI) of 0.76 was successfully produced. LKOG can enhance the production of butyric acid in the colon with the highest concentration (8.24 mM) found at 72 hr fermentation.

| 789 ARIESTANTI ET Al. Zhang, Zhao, & Ding, 2016). Enzymatic hydrolysis has been the most widely used in the degradation of polysaccharides due to it can be safely produced oligosaccharides in room temperature (Liu et al., 2015). β-mannanase randomly hydrolyze mannosidic linkages found in glucomannan (Jian, Zhu, Zhang, Sun, & Jiang, 2013). Response surface methodology (RSM) is a statistical protocol designed for experiments in which several factors can be varied concurrently.
Box-Behnken design (BBD) is one of the experimental designs in analytical chemistry. BBD is efficient when large number of variables are involved (Bezerra, Santelli, Oliveira, Villar, & Escaleira, 2008). As a dietary fiber, KGM and KOG are indigestible in small intestine and then partially or completely fermented by colonic microflora; therefore, it can act as prebiotic (Mudgil & Barak, 2013). The ability of KOG to promote the growth of lactobacilli and bifidobacteria has been demonstrated previously (Al-Ghazzewi et al., 2007). Al-Sheraji et al. (2013) and Slavin (2013) showed that KOG has prebiotic capacity. The authors ascribed this prebiotic potential to its ability to stimulate the growth of probiotic bacteria thereby enhancing short-chain fatty acids (SCFAs) production, reduce the number of pathogenic bacterial populations thereby improving the hosts immune system.
Commercial KOG with health benefit is not available in the market due to its level of purity (Connolly et al., 2010). Ultrafiltration (UF) is one of the methods that has found wide application in separation and purification of polysaccharides due to the advantages it confers on the final product (Xing & Li, 2009).
The aim of this study was to find the optimum conditions necessary for the production of KOG using β-mannanase and UF membrane, and then investigate its effect on the gut microbiota and in SCFAs.

| Production of konjac oligo-glucomannan
Konjac glucomannan (1% w/v) was added into 0.2 M CH 3 COONa (sodium acetate) buffer and then mixed with β-mannanase (E/S ranging from 0.025% to 0.1%, w/w) to start the reaction. The mixture was stirred continuously and incubated at pH 5.0-6.0 for 2-6 hr. The temperature of the water bath was maintained at 43-53°C. Enzyme was inactivated by boiling the samples for 10 min, and then, the samples were centrifuged at 8,000 g for 10 min. The supernatant was then analyzed using high-performance size exclusion chromatography (HPSEC; Agilent model 1200 series, CA, USA) for KOG concentration and molecular weight distribution.

| Experimental design
The hydrolysis parameters were optimized by RSM using four variables Box-Behnken factorial design (BBD) software version 16 (Minitab Pty Ltd, Sydney, Australia) to determine the optimum conditions for producing KOG. The parameters were temperature, time, pH, and E/S ratio.

| Purification of konjac oligo-glucomannan
Konjac oligo-glucomannan was filtered using polysulfone (PSU) ultrafiltration (UF) hollow fiber membrane (GE Healthcare Pte Ltd., Singapore) with pore size of 10,000 and 3,000 NMWC (membrane effective area was 110 and 140 cm 2 , respectively). Ten thousand NMWC membrane was operated for 6 hr, and 3,000 NMWC membrane was operated for 9 hr. Trans-membrane pressure (TMP) was set to 1.5 bar with feed tank temperature 50°C. Both retentate and permeate were collected and analyzed by HPSEC to determine the performance of the membrane in purification of KOG mixtures.
Purified KOG was dried by spray dryer. Volume concentration factor (VCF) and rejection value (R app ) were calculated according to the following equation (Millipore, 2003):

| Preparation of fecal slurry
Fecal samples were collected from six healthy male and female donors who had not consumed probiotics, prebiotics or received antibiotics treatment for at least 3 months and had no digestive system disease. Fresh fecal samples were put into an anaerobic chamber and mixed with 0.1 M phosphate-buffered saline (PBS; 8 g/L NaCl, 0.2 g/L KCl, 1.44 g/L Na 2 HPO 4 , and 0.24 g/L KH 2 PO 4 ) pH 7.4 for the final concentration 10% (w/v). The slurry was homogenized by stomacher for 2 min and filtered with dialysis bag before inoculated into vessels.

| In vitro fecal fermentation (batch culture) of konjac oligo-glucomannan
Water jacket glass vessels were used and filled with basal medium contained per liter: 0.9 g peptone water, 0.9 g yeast extract, 0.045 g NaCl, 0.018 g K 2 HPO 4 , 0.018 g KH 2 PO 4 , 0.0045 g MgSO 4 .7H 2 O, 0.0045 g CaCl 2 .6H 2 O, 0.9 g NaHCO 3 , 0.225 g L-cysteine. HCl, 0.225 g bile VCF = Total starting feed volume added to the operation∕current retentate volume R app = 1 − (concentration in filtrate∕concentration in feed) salts, 0.9 ml Tween 80, 4.5 μl vitamin K, 0.0225 g hemin, and 0.45 ml of 0.025% resazurin. The medium was dissolved in distilled water, adjusted to pH 7.0 and then sterilized by autoclave. Sterilized medium (90 ml) was placed and stirred magnetically into each of the vessels with temperature and pH controlled (37°C, pH 6.8) using circulated water bath and pH controller, respectively. The medium was maintained under an anaerobic environment by pumped in the nitrogen gas into vessel to imitate the conditions in colon (Connolly et al., 2010) overnight. The basal medium was inoculated with 10 ml of a 10% (w/v) fecal slurry. Substrates were added and dissolved in the basal medium with final concentration of 1% (w/v). The final volume of each culture was 100 ml. The sample (10 ml) from each vessel was taken at 0, 6, 12, 24, 48, and 72 hr for SCFAs analysis and enumeration of fecal bacteria by FISH technique. All samples were kept in −20°C.

| Short-chain fatty acids (SCFAs) analysis by high-performance liquid chromatography (HPLC)
Samples from batch culture (1,125 μl) were centrifuged at 13,000 g for 15 min at 4°C. The supernatant was obtained and filtered through a 0.22 μm membrane nylon filter. Samples were analyzed for SCFA concentration for acetic, propionic, and butyric acid through HPLC system (Agilent model 1200 series, CA, USA) using an Aminex HPX-87H ion exclusion column (Bio-Rad, Richmond CA, USA) with a 7.8 mm diameter and 300 mm length at 50°C. Diode array detector (DAD) was used in the HPLC system with the UV absorbance was set at 215.4 nm. Mobile phase was 0.005 M H 2 SO 4 with flow rate of 0.6 ml/min. The quantification of SCFAs of the samples were calculated from the calibration curves of acetic, propionic, and butyric acid.

| Enumeration of fecal bacteria by fluorescent in situ hybridization (FISH) technique
Fluorescent in situ hybridization is a technique using 16S rRNAtargeted oligonucleotide to evaluate the changes in different human fecal bacterial populations at six time points during fermentation of the substrates (Hugenholtz, Tyson, & Blackall, 2002).
Fecal batch culture samples (375 μl) were mixed and fixed for 4-16 hr in cold 4% paraformaldehyde (pH 7.2) at 4°C in a ratio 1:3 of sample to 4% paraformaldehyde (v/v). The samples were then centrifuged at 13,000 g for 20 min at 4°C. Supernatant was removed, and the pellet was washed twice by resuspending in 1 ml phosphate-buffered saline (PBS, pH 7.4) and centrifuging. The remaining pellet was resuspended in PBS/96% ethanol mix (1:1, v/v) and stored at −20°C for further use up to 3 months. The samples at the optimal dilution were mixed and dropped (20 μl) onto Teflon/ Poly-L-Lysine coated six well microscope slides (Tekdon Inc., Myakka City, FL, USA). The samples on the slide were dried using slide warmer (Digicon MD-700A, Thailand) at 46°C for 15 min and then dehydrated using alcohol series (50%, 80% and 96%) for 3 min in each concentration. Cells targeted with Lab158 probe was slightly modified by dropping 20 μl lysozyme into each well at room temperature for 15 min and the slides then washed in cool water (3 s) before dehydrated in alcohol series. After the evaporation of ethanol, the slides were returned to the slide warmer. Then 50 μl mixture of probe and hybridization buffer (1:9; v/v) were added to each well. Hybridization was performed in hybridization oven (Boekel Scientific InSlide Out Slide Hybridizer 241000, PA, USA) for 4 hr at appropriate temperature for the probes. The slides were washed with 50 ml of wash buffer for 15 min, then rinsed with cold water, and dried using compressed air. Antifade agent (10 μl; Invitrogen, USA) was applied to the surface of each well and covered with cover slip. The bacterial group was enumerated by fluorescence microscopy (Nikon E400 Eclipse, Nikon Instruments, Inc., New York, USA) and 15 randomized views from each well (Wichienchot, Prakobpran, Ngampanya, & Jaturapiree, 2017) were counted using NIS-Elements BR 3.00, SP6 software (Nikon Instruments, Inc., New York, USA) for each sample.
Prebiotic candidate can be evaluated with the value of its prebiotic index (PI). Calculation of PI (Palframan, Gibson, & Rastall, 2003) is described below: where Bif is number of bifidobacteria, Lac is number of lactobacilli, Bac is number of bacteroides, and Clos is number of clostridia.

| Statistical analysis
Statistical analysis was performed using the SPSS version 21.0 for windows (IBM Corp. New York, USA). The variation within the samples was analyzed using one-way analysis of variance (ANOVA).
Duncan's test was used to compare the means at 95% confidence level.

| Production of konjac oligo-glucomannan
In this study, KOG was produced by enzymatic hydrolysis of KGM using β-mannanase. The BBD experimental design was used to optimize the hydrolysis temperature, hydrolysis time, pH, and the initial enzyme to substrate ratio (E/S). The concentration of oligosaccharides after hydrolysis from each variation of the factors was analyzed by HPSEC, and the results are shown in Table 1. The highest oligosaccharides concentration was 9.21 mg/ml with the molecular weight of 1,552 Da ± 2.83 found at the condition of 48°C, pH of 5.5 and the E/S ratio of 0.05% (w/w) after 4 hr hydrolysis. The optimum conditions predicted by BBD for oligosaccharides and polysaccharides concentrations were 7.76 and 7.67 mg/ml, respectively.
The predicted result was lower compared to one of the treatments described above, probably because the pH (6.0) and temperature (50.37°C) were higher than the optimum working conditions of the enzyme (pH 5.5, temperature 48°C). Chen et al. (2013) found that the increase of hydrolysis time, the activity of β-mannanase decreased correspondingly. This is similar to our result. This study was successfully produced KOG by hydrolysis of KGM using β-mannanase. Despite this method can be applied in the production of KOG, the product still contained unhydrolyzed KGM shown by the concentration of polysaccharides (data not shown). The results showed that β-mannanase was able to hydrolyze KGM within 4 hr. The main reason for purifying KOG is to improve its market value and health benefits. Therefore, purification step is needed to purify KOG mixture to get the higher amount of oligosaccharide in the final product.

| Purification of konjac oligo-glucomannan
Konjac oligo-glucomannan was purified using membrane technique based on the size exclusion mechanism. The purpose for this step was to purify the mixture of KOG by rejecting the polysaccharides in the product using the membrane pore size of 10,000 NMWC and 3,000 NMWC. As a result of the molecular weight of KGM (2.39 × 10 5 Da), polysaccharides and oligosaccharides were obtained in retentate and permeate, respectively. Both the retentate and permeate were analyzed by HPSEC to ascertain the level of purification obtained for KOG mixtures.
At the optimum KOG production (48°C, pH of 5.5, and E/S 0.05% for 4 hr), the initial product mixture contained 12.81 and 9.21 mg/ ml of polysaccharides and oligosaccharides, respectively. The final concentration of polysaccharides and oligosaccharides during the purification step is shown in Table 2. The results indicate that the concentration of oligosaccharides in the permeate increased from 9.21 to 9.24 mg/ml with 10,000 NMWC membrane and from 9.21 to 9.54 mg/ml with 3,000 NMWC membrane. Additionally, the increased filtration time also improved the concentration of oligosaccharides in permeate. Volume concentration factor (VCF) of this study was 2.14 with the rejection value (R app ) for 10,000 NMWC membrane pore size was −0.07 for oligosaccharides and 0.42 for TA B L E 1 Concentration of oligosaccharides from enzymatic hydrolysis process polysaccharides. The R app of oligosaccharides and polysaccharides from 3,000 NMWC membrane pore size was −0.12 and 0.58, respectively. The VCF represents the amount that the feed stream has been reduced in volume from the initial volume. Meanwhile, R app represents the fraction of a particular substrate that is retained by the membrane. In order to obtain high yield of oligosaccharides, a low VCF and high R app values are desirable (Millipore, 2003).
Purification yield of oligosaccharides after UF step was 0.33% for 10,000 NMWC and 3.58% for 3,000 NMWC. In this study, the membrane with 10,000 NMWC pore size might not be effective to purify KOG because the concentration of oligosaccharides in the permeate was lower than the concentration obtained from the permeate of 3,000 NMWC. The results showed that the 3,000 NMWC   In 2018, the prebiotics market is approximated to reach $ 4.5 billion (Plongbunjong, Graidist, Knudsen, & Wichienchot, 2017).
Prebiotics are functional food ingredients that modify the gut microbiota thus improve the host health (Sarbini & Rastall, 2011).
Previous studies have investigated the health properties of KGM as a potential prebiotic with application in food and healthcare/pharmaceutical product (Behera & Ray, 2016;Jiang, Li, Shi, & Xu, 2018). Yang et al. (2017) showed KOG produced by enzymatic hydrolysis has the prebiotic effect by showing the ability to support the growth of lactobacilli and bifidobacteria. This result is comparable with ours.
Recently Mao, Song, Yao, and Wu (2018) showed that KGM, and its degradation products almost have no influence to the growth of five strain of bifidobacteria. In this study, LKOG promoted the growth of bifidobacteria compared to their study, but its effect for the specific strain of bifidobacteria need further investigation. Previous study about prebiotic activity of PGM in vivo showed that PGM has prebiotic activity by suppressing the growth of E. coli (Harmayani, Aprilia, & Marsono, 2014). Compared to their study, our result showed the ability of PGM to promote the growth of bifidobacteria and significantly reduced the number of clostridia. In the present study, LKOG fermentation in the fecal batch culture could increase the number of beneficial bacteria and decrease pathogenic bacterial populations.
Prebiotic activities may have a contribution for health by promoting the growth of bifidobacteria and/or lactobacilli (Connolly et al., 2010). PI value can be used to evaluate the prebiotic activities by giving the quantitative score. Association between changes in beneficial bacteria with unwanted bacteria related to their beginning level was described as PI (Thitiratsakul & Anprung, 2014). Table 3

| Short-chain fatty acids (SCFAs) production
Short-chain fatty acids are the major products of bacterial fermentation from carbohydrates in the large intestine. 2016). In our study, acetic acid was the highest SCFAs produced in all substrates during fermentation. LKOG produced the highest acetic acid compared to other substrates. Propionic acid production required more specific substrate and bacterial group compared to acetic acid (Morrison & Preston, 2016). Propionic acid production increased significantly (p < 0.05) at 6 hr for LKOG and at 12 hr for inulin. Butyric acid is the main source of energy for colonocytes (Al-Sheraji et al., 2013). Butyric acid was produced after 6 hr from LKOG and increased significantly during the fermentation. The highest concentration of butyric acid from LKOG was 8.24 ± 0.12 mM at 72 hr, while inulin fermentation did not produce butyric acid. An increased of butyric acid in the gut has been linked to colon cancer and colonic inflammation reduction (Al-Sheraji et al., 2013).
The butyric acid produced from LKOG fermentation is higher than the butyric acid produced from the report of Connolly et al. (2010) probably because of the fermentation time from this study was longer and the MW of LKOG was lower than their study. The concentration of butyric acid from LKOG was higher than propionic acid. SCFAs production was in the order of acetic, butyric, and propionic acid. It is suggested that LKOG has different benefits with KGM and may have specific health function related to butyric acid production. Therefore, LKOG demonstrated prebiotic potential comparable to inulin.

| CON CLUS IONS
Konjac oligo-glucomannan was successfully produced by hydrolysis of KGM using β-mannanase with the highest concentration of 9.54 mg/ml obtained under the conditions of 48°C, pH 5.5, hydrolysis time 4 hr, and E/S 0.05%. Low-molecular weight KOG produced from UF membrane was selectively fermented by beneficial bacteria and can enhance the production of butyric acid during fecal fermentation. High concentration of KOG obtained under the optimum conditions showed the potential of this method to produce KOG for industrial scale. Further study of KOG to investigate its beneficial effect in human should be performed.

This work was supported by the Thailand's Education Hub for
Southern Region of ASEAN Countries (TEH-AC) scholarship and research funding from Graduate School Prince of Songkla University.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L R E V I E W
This study does not involve any human or animal testing.