In Vitro Infant Faecal Fermentation of Low Viscosity Barley β-Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics

Ka-Lung Lam; Kin-Chun Ko; Xiaojie Li; Xinxin Ke; Wai-Yin Cheng; Tianfeng Chen; Lijun You; Hoi-Shan Kwan; Peter  Chi-Keung Cheung

doi:10.3390/molecules24050828

In Vitro Infant Faecal Fermentation of Low Viscosity Barley β-Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics

Lam, Ka-Lung;Ko, Kin-Chun;Li, Xiaojie;Ke, Xinxin;Cheng, Wai-Yin;Chen, Tianfeng;You, Lijun;Kwan, Hoi-Shan;Cheung, Peter Chi-Keung 2019-02-26 00:00:00 molecules Article In Vitro Infant Faecal Fermentation of Low Viscosity Barley -Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics 1 1 1 1 1 2 Ka-Lung Lam , Kin-Chun Ko , Xiaojie Li , Xinxin Ke , Wai-Yin Cheng , Tianfeng Chen , 3 1 1 , Lijun You , Hoi-Shan Kwan and Peter Chi-Keung Cheung * Food and Nutritional Sciences, School of Life Sciences, The Chinese University of Hong Kong, Shatin, Hong Kong; [email protected] (K.-L.L.); [email protected] (K.-C.K.); [email protected] (X.L.); [email protected] (X.K.); [email protected] (W.-Y.C.); [email protected] (H.-S.K.) Department of Chemistry, Jinan University, Guangzhou 510632, China; [email protected] School of Food Science and Engineering, South China University of Technology, Guangzhou 510640, China; [email protected] * Correspondence: [email protected]; Tel.: +852-39436144 Academic Editor: Chi-Fai Chau Received: 1 February 2019; Accepted: 20 February 2019; Published: 26 February 2019 Abstract: Barley contains high level of -1,3-1,4-glucans (BBGs) which can be fermented by microbes and are a potential prebiotic. In the present study, native BBG with low viscosity and a MW of 319 kDa was depolymerized by acid hydrolysis to produce a series of four structurally characterized fragments with MWs ranging from 6–104 kDa. In vitro fermentation of these BBG samples by infant faecal microbiome was evaluated using a validated deep-well plate protocol as parallel miniature bioreactors. Microbial taxa were identiﬁed using 16S amplicon sequencing after 40 h of anaerobic fermentation. Bioinformatics analysis including diversity indexes, predicted metagenomic KEGG functions and predicted phenotypes were performed on the sequenced data. Short chain fatty acids and dissolved ammonia were quantiﬁed and the SCFAs/NH ratio was used to evaluate the eubiosis/dysbiosis potential. Correlation analysis showed that most of the parameters investigated showed a parabolic function instead of a monotonous function with the BBG samples having different MWs. Among the ﬁve BBGs, it was concluded that BBG with an intermediate MW of 28 kDa is the most promising candidate to be developed as a novel prebiotic. Keywords: acid hydrolysis; depolymerization; infant faecal microbiome; short chain fatty acids; ammonia; SCFAs/NH 1. Introduction Oat and barley are widely consumed worldwide in different forms, including direct consumption, ground powders, fermented beverages and so on, with the world oat consumption being 22,777 thousand metric tons and that of barley, 142,581 thousand metric tons in December 2018 [1]. -1,3-1,4-Glucans (BBGs) are naturally present in oat and barley as hemi-cellulose and they are now being examined as a functional food with prebiotic potential [2]. However, -glucans differ from well-known prebiotics such as inulin, FOS, GOS due to their high intrinsic physico-chemical property heterogeneity, including differences in molecular weight (MW), glycosidic linkages, degree and length of branching, water solubility, etc. [3]. The MW of -glucans is one of the important attributes towards their biological function as it affects solubility, viscosity, and enzymatic breakdown efﬁciency [4]. Currently, high purity (>95%) commercial native -1,3-1,4-glucans with different MW from barley (BBGs) and oat are available from Megazyme (County Wicklow, Ireland). For example, a Molecules 2019, 24, 828; doi:10.3390/molecules24050828 www.mdpi.com/journal/molecules Molecules 2019, 24, 828 2 of 20 total of ﬁve different MW -1,3-1,4-glucans have been used for investigation of their effects on in vitro fermentation using human faecal microbiota [5] and three BBGs of different MW used in in vitro fermentation with porcine faeces [6]. There is also another publication that employed -1,3-1,4-glucans extracted from barley with different molecular weights but with a lower purity [7]. In this study, we employed low viscosity BBG from Megazyme and its acid hydrolyzed derivatives, with reported purity >95%. In the present study, we aimed at evaluating the enhancement of the in vitro infant faecal fermentation of BBG with lower MW than the native ones produced by controlled acid hydrolysis. 16S amplicon proﬁling was determined to assess the microbial composition/diversity. Human infant faecal samples were chosen due to the high viscosity and texture similarity when in a slurry form with in vitro culture medium. Microbial fermentation is catalytically multi-fold, including saccharolytic, proteolytic, lipolytic types and so on [8]. In this study, BBGs of different MW were added as sole carbon sources to the culture medium, commonly containing partially digested proteins. Therefore, saccharolytic and proteolytic fermentation are mainly involved in this investigation. The effect of polymeric -glucans and its monomer glucose on microbial proﬁle was also compared and the correlations and insights from the phenotypic data, microbial proﬁle and MWs of BGGs were discussed. 2. Results 2.1. Chemical Composition and Structual Characterization of BBG Samples All BBG samples were predominantly carbohydrate in nature (97–98% by weight), with glucose as the only monosaccharide and contained only trace amounts of protein (less than 1% by weight) without signiﬁcant difference (Table 1). The monosaccharide proﬁle of native BBG and depolymerized BBG_0.2(2) shown in Figure S1A indicated that the only sugars found were glucose and allose (internal standard). The FT-IR spectra (Figure S1B) and the proton NMR spectra (Figure S1C) conﬁrmed the chemical structure of BBG and BBG_0.2(2) indicated that acid hydrolysis did not introduce any structural changes to the functional groups of the BBG derivatives. Increasing acid concentration and time of hydrolysis reduced the MW of BBG derivatives as shown in Table 1. Table 1. Preparation and structural characteristics of barley -glucans (BBG) and its acid hydrolyzed derivatives. Samples * Carbohydrate (% by wt) Protein (% by wt) Molecular Weight (kDa) BBG 98.0 0.7 0.2 0.5 319 BBG_0.05 98.2 0.3 0.2 0.7 104 BBG_0.1 97.6 0.3 0.2 0.1 28 BBG_0.2 98.1 0.5 0.2 0.3 11 BBG_0.2(2) 97.9 0.4 0.2 0.1 6 * The preparation conditions were: BBG: original, no hydrolysis; BBG_0.05: 0.05M H SO , 10 min at 80 C; BBG_0.1: 2 4 0.10M H SO , 60 min at 80 C; BBG_0.2: 0.20M H SO , 60 min at 80 C; BBG_0.2(2): 0.20M H SO , 120 min at 80 C. 2 4 2 4 2 4 All depolymerized BBG derivatives had a MW higher than the glucose monomer with BBG_0.2(2) having the lowest MW of 6 kDa as shown in Figure 1. Linkage analysis of methylated BBG samples by GC-MS indicated the presence of three major partially methylated alditol acetates (PMAAs) including 2,3,4,6-Me -Glc (terminal glucose), 2,4,6-Me -Glc (1,3-linked glucose) and 2,3,6-Me -Glc (1,4-linked 4 3 3 glucose) (Table 2). These linkage results were consistent with the typical mixed-1,3-1,4 linkages found in BBG. Based on the molar ratios of these three PMAAs, the change in the ratio of 1,3- and 1,4-glycosidic linkages before and after various acid hydrolysis was approximately 1:6, suggesting that the acid hydrolysis has no preference for -1,3 or -1,4 linkages (Table 2). In conclusion, controlled partial acid hydrolysis is an effective means of MW reduction in BBG. Molecules 2019, 24, 828 3 of 20 Molecules 2019, 24, x FOR PEER REVIEW 3 of 20 Figure 1. SEC chromatograms showing the MW distribution of BBGs and its depolymerized derivatives. Figure 1. SEC chromatograms showing the MW distribution of BBGs and its depolymerized Glucose was used as a monomer control. derivatives. Glucose was used as a monomer control. Table 2. GC-MS linkage analysis of barley -glucan (BBG) and its acid hydrolyzed derivatives. Table 2. GC-MS linkage analysis of barley β-glucan (BBG) and its acid hydrolyzed derivatives. Peak Area (%) c c Relative Molar Ratio Relative Molar Ratio Samples a c c Peak Area (%) Rela T-Glc tivep Mo :1,3-Glc lar Ra p:1,4-Glc tio p T Re -Glc lative p:1,3-Glc Mola p:1,4-Glc r Ratio p T-Glcp 1,3-Glcp 1,4-Glcp Samples T-Glcp 1,3-Glcp 1,4-Glcp T-Glcp:1,3-Glcp:1,4-Glcp T-Glcp:1,3-Glcp:1,4-Glcp BBG 3.99 1.00 14.25 0.15 81.76 0.85 1:3.57:20.49 0.28:1:5.74 BBG 3.99 ± 1.00 14.25 ± 0.15 81.76 ± 0.85 1:3.57:20.49 0.28:1:5.74 BBG_0.05 3.76 0.94 14.15 0.14 82.09 0.80 1:3.77:21.84 0.27:1:5.80 BBG_0.05 3.76 ± 0.94 14.15 ± 0.14 82.09 ± 0.80 1:3.77:21.84 0.27:1:5.80 BBG_0.1 13.22 2.99 13.11 0.45 73.68 2.54 1:0.99:5.58 1.01:1:5.62 BBG_0.1 13.22 ± 2.99 13.11 ± 0.45 73.68 ± 2.54 1:0.99:5.58 1.01:1:5.62 BBG_0.2 22.46 4.55 9.94 0.58 67.60 3.97 1:0.44:3.01 2.26:1:6.80 BBG_0.2 22.46 ± 4.55 9.94 ± 0.58 67.60 ± 3.97 1:0.44:3.01 2.26:1:6.80 BBG_0.2(2) 32.28 5.73 9.24 0.78 58.48 4.94 1:0.29:1.81 3.49:1:6.33 BBG_0.2(2) 32.28 ± 5.73 9.24 ± 0.78 58.48 ± 4.94 1:0.29:1.81 3.49:1:6.33 a b c a: calculated from the peak area in the GC-MS total ion chromatograms. : Glcp, glucopyranose. b : molar ratio was c : calculated from the peak area in the GC-MS total ion chromatograms. : Glcp, glucopyranose. : calculated from the ratio of the (peak area)/(MW) of individual PMAAs. molar ratio was calculated from the ratio of the (peak area)/(MW) of individual PMAAs. 2.2. BBG In Vitro Fermentation Using Infant Faecal Inocula 2.2. BBG In Vitro Fermentation Using Infant Faecal Inocula 2.2.1. Fermentation of BBG Samples Evaluated by Total Bacterial Count and Total DNA Extracted 2.2.1. Fermentation of BBG Samples Evaluated by Total Bacterial Count and Total DNA Extracted Results of the bacterial total plate count (TPC) in terms of colony formation units (CFUs) and Results of the bacterial total plate count (TPC) in terms of colony formation units (CFUs) and the the fold of population increase relative to glucose are shown in Table 3 and a representative plate fold of population increase relative to glucose are shown in Table 3 and a representative plate of each of each of the ﬁve BBG samples and glucose monomer is shown in Figure S2. In the present results, of the five BBG samples and glucose monomer is shown in Figure S2. In the present results, the the concentration of the extracted microbial DNA was accessed by Nano-drop (Thermo Scientiﬁc, concentration of the extracted microbial DNA was accessed by Nano-drop (Thermo Scientific, Waltham, MA, USA) and the fold changes of the ﬁve BBG samples relative to glucose are shown in Waltham, MA, USA) and the fold changes of the five BBG samples relative to glucose are shown in Table 3. All BBG samples supported the growth of more microbes than glucose, which was probably Table 3. All BBG samples supported the growth of more microbes than glucose, which was probably due to the larger amount of monomeric glucose produced after enzymatic hydrolysis of the BGG due to the larger amount of monomeric glucose produced after enzymatic hydrolysis of the BGG samples on molar basis. TPC results indicated that BBG had the highest CFUs (3.7-fold increase), samples on molar basis. TPC results indicated that BBG had the highest CFUs (3.7-fold increase), followed by BBG_0.2 (2.9-fold), BBG_0.1 (2.3-fold), BBG_0.2(2) (2.2-fold), and BBG_0.05 (1.0-fold) when followed by BBG_0.2 (2.9-fold), BBG_0.1 (2.3-fold), BBG_0.2(2) (2.2-fold), and BBG_0.05 (1.0-fold) compared to glucose. Total DNA results indicated that the increase of biomass was in a descending when compared to glucose. Total DNA results indicated that the increase of biomass was in a order of: BBG_0.1 (4.2-fold), BBG_0.05 (3.9-fold), BBG_0.2 (3.6-fold), BBG (3.4-fold) and BBG_0.2(2) descending order of: BBG_0.1 (4.2-fold), BBG_0.05 (3.9-fold), BBG_0.2 (3.6-fold), BBG (3.4-fold) and (2.8-fold). There seemed to have a nonlinear relationship of increase in microbial populations with BBG_0.2(2) (2.8-fold). There seemed to have a nonlinear relationship of increase in microbial a decrease in MW of the BGG samples. The discrepancy between the results of TPC and total DNA populations with a decrease in MW of the BGG samples. The discrepancy between the results of TPC might be explained by the fact the TPC is the ﬁnal outcome of complex interactions of all microbes and total DNA might be explained by the fact the TPC is the final outcome of complex interactions including self/mutual promotion and/or inhibition which might not only depend on the MW of the of all microbes including self/mutual promotion and/or inhibition which might not only depend on carbon source. the MW of the carbon source. Molecules 2019, 24, 828 4 of 20 Table 3. Bacterial total plate count (CFUs) and concentration of total DNA extracted from culture broth Molecules 2019, 24, x FOR PEER REVIEW 4 of 20 of the infant faecal fermentation of ﬁve BBG samples, glucose monomer and time-0. Total Plate Ratio of CFUS Ratio of [DNA] Table 3. Bacterial total plate count (CFUs) and concentration of total DNA extracted from culture Samples [DNA] (g/mL) * Count (CFU) (Relative to Glucose) (Relative to Glucose) broth of the infant faecal fermentation of five BBG samples, glucose monomer and time-0. a,b BBG 8.7 104 3.7 3.4 67.3 7.4 Total Plate Ratio of CFUS Ratio of [DNA] Samples [DNA] (µg/mL) * c,d BBG_0.05 2.3 104 1.0 76.4 4.0 3.9 Count (CFU) (Relative to Glucose) (Relative to Glucose) a,b a,e,f BBG 8.7 × 104 3.7 67.3 ± 7.4 3.4 BBG_0.1 5.4 104 2.3 83.5 9.1 4.2 c,d BBG_0.05 2.3 × 104 1.0 76.4 ± 4.0 3.9 g,h BBG_0.2 6.8 104 2.9 3.6 71.9 4.1 a,e,f BBG_0.1 5.4 × 104 2.3 83.5 ± 9.1 4.2 c,g,i BBG_0.2(2) 5.2 104 2.2 54.8 0.6 2.8 g,h BBG_0.2 6.8 × 104 2.9 71.9 ± 4.1 3.6 b,d,e,f,h,i c,g,i BBG_0.2(2) GLC 5.2 × 2.4 104 104 2.2 1 19.7 54.8 ± 2.4 0.6 2.8 1 b,d,e,f,h,i GLC 2.4 × 104 1 19.7 ± 2.4 1 T0 6.4 103 0.3 4.3 0.2 0.2 T0 6.4 × 103 0.3 4.3 ± 0.2 0.2 a–i * Mean values with different superscripts ( ) represent signiﬁcant difference by one-way ANOVA (Tukey HSD a–i * Mean values with different superscripts ( ) represent significant difference by one-way ANOVA post-hoc test), p < 0.05. : Time-0, T0 group is statistically signiﬁcant different compared to all other groups. (Tukey HSD post-hoc test), p < 0.05. : Time-0, T0 group is statistically significant different compared to all other groups. Compared to TPC, the amount of total DNA extracted in a microbial fermentation could better reﬂect the total bacterial biomass supported by the carbon source since it is not selective and Compared to TPC, the amount of total DNA extracted in a microbial fermentation could better medium-independent reflect the total bacterial biomass suppo [9]. Therefore, DNA rted by the c concentration arbon source was since it is not se selected to repr lective esent and m total edium bacteria - independent [9]. Therefore, DNA concentration was selected to represent total bacteria count of count of samples in the following analysis. samples in the following analysis. 2.2.2. Short Chain Fatty Acids and Dissolved Ammonia Content after In Vitro Fermentation 2.2.2. Short Chain Fatty Acids and Dissolved Ammonia Content after In Vitro Fermentation The highest level of short chain fatty acids (SCFAs) produced was found in BBG_0.2 which The highest level of short chain fatty acids (SCFAs) produced was found in BBG_0.2 which was was ﬁve times that of glucose as shown in Figure 2A. Glucose only generated acetic acid and its five times that of glucose as shown in Figure 2A. Glucose only generated acetic acid and its concentration was lower than all the ﬁve BBGs. This might be explained by the fact that infant faecal concentration was lower than all the five BBGs. This might be explained by the fact that infant faecal inoculum contained a myriad of bacteria that not only converted glucose into SCFAs, but also other inoculum contained a myriad of bacteria that not only converted glucose into SCFAs, but also other metabolites. Acetic acid was the major SCFAs produced by all BBG samples, followed by propionic metabolites. Acetic acid was the major SCFAs produced by all BBG samples, followed by propionic acid and then butyric acid. The distribution of SCFAs produced by different BBG samples also varied. acid and then butyric acid. The distribution of SCFAs produced by different BBG samples also varied. The highest acetic acid production was found in BBG_0.1 and BBG_0.2, while BBG_0.2(2) having the The highest acetic acid production was found in BBG_0.1 and BBG_0.2, while BBG_0.2(2) having the lowest MW had the highest concentration of propionic acid (Figure 2A). lowest MW had the highest concentration of propionic acid (Figure 2A). Figure 2. (A) Short chain fatty acid proﬁles and (B) ammonia concentration of the ﬁve BBG samples Figure 2. (A) Short chain fatty acid profiles and (B) ammonia concentration of the five BBG samples a–h and glucose monomer. Different superscripts ( ) represent signiﬁcant difference by one-way ANOVA a–h and glucose monomer. Different superscripts ( ) represent significant difference by one-way (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically signiﬁcant different from all other groups. ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically significant different from all other groups. The highest ammonia production was obtained by BBG_0.05 samples with original BBG produced the least among the ﬁve BBGs (Figure 2B). All BBGs produced more ammonia compared to glucose. As The highest ammonia production was obtained by BBG_0.05 samples with original BBG shown in Figure 2A, glucose produced the highest level of propionic acid, which might explain its least produced the least among the five BBGs (Figure 2B). All BBGs produced more ammonia compared ammonia production as propionate ion was shown to be a growth inhibitor to pathogenic/spoilage to glucose. As shown in Figure 2A, glucose produced the highest level of propionic acid, which might microbes [10]. explain its least ammonia production as propionate ion was shown to be a growth inhibitor to pathogenic/spoilage microbes [10]. Molecules 2019, 24, 828 5 of 20 Molecules 2019, 24, x FOR PEER REVIEW 5 of 20 2.3. Microbiome Proﬁle Changes after BBG Fermentation 2.3. Microbiome Profile Changes after BBG Fermentation 2.3.1. Change of Infant Faecal Microbiome Proﬁle by Fermentation of BBG Samples 2.3.1. Change of Infant Faecal Microbiome Profile by Fermentation of BBG Samples The microbial taxonomy proﬁles of the ﬁve BBG samples, glucose and time-0 (T0) are shown The microbial taxonomy profiles of the five BBG samples, glucose and time-0 (T0) are shown in in Figure 3A with abundance data scaled according to sample total DNA concentration, reﬂecting a Figure 3A with abundance data scaled according to sample total DNA concentration, reflecting a quasi-absolute taxa amount. quasi-absolute taxa amount. Figure 3. 16S amplicon sequencing results of fermentation of BBG samples by infant faecal inoculum. Figure 3. 16S amplicon sequencing results of fermentation of BBG samples by infant faecal inoculum. (A) Bar plot of the major identiﬁed bacterial taxonomy scaled to the total extracted DNA concentration, (A) Bar plot of the major identified bacterial taxonomy scaled to the total extracted DNA noted as quasi-absolute bacterial amount; (B) Enlarged view of the relative abundance of the taxon concentration, noted as quasi-absolute bacterial amount; (B) Enlarged view of the relative abundance >1% of T0 group at the beginning of fermentation; (C) Enlarged view of the relative abundance of the of the taxon >1% of T0 group at the beginning of fermentation; (C) Enlarged view of the relative taxon >1% of the ﬁve BBG groups and glucose monomer, Glc; (D) Venn diagram showing the similar abundance of the taxon >1% of the five BBG groups and glucose monomer, Glc; (D) Venn diagram and different number of bacteria taxonomy identiﬁed; (E) -diversity analysis showing the principle showing the similar and different number of bacteria taxonomy identified; (E) β-diversity analysis component showing the principle analysis of the com ﬁve ponBBGs ent analy and sis glucose of the fi monomer ve BBGs and g . lucose monomer. It was clearly shown that T0 group had the least taxa abundance in the beginning. Moreover, the It was clearly shown that T0 group had the least taxa abundance in the beginning. Moreover, the biological triplicates among each group gave similar microbiome proﬁles, demonstrating a sample biological triplicates among each group gave similar microbiome profiles, demonstrating a sample Molecules 2019, 24, 828 6 of 20 size of three could be sufﬁcient for gaining some insights. Figure 3B showed the relative abundance of T0 group with distinct taxa (average relative taxa abundance >1%) compared to the BBGs samples. Taxa unique to T0 group were noted with "*" as shown in the ﬁgure legend, namely, Prevotella sp., Lachnospira sp., Bacillus sp., other genus in Lachnospiraceae family, Ruminococcus sp., Parabacteroides sp. and Streptococcus sp. These taxa abundance contributed less than 1% after BBGs/glucose fermentation. Figure 3C shows that all the BBGs shared similar microbial taxa distribution while glucose monomer though had similar taxa but with a distinctive distribution. The Venn diagram of the microbiome proﬁles of the ﬁve BBG samples indicates that they all shared 57 core microbial taxa and some group-speciﬁc unique taxa (Figure 3D). The number of group-speciﬁc taxa ranged from 1 to 6 and these speciﬁc taxa (family/genus) were Actinomycetales, Oscillospira, and Peptoniphilus for BBG; Bacilli for BBG_0.05; Cellulomonas, Aerococcus and Leuconostoc for BBG_0.1; Rothia, Pediococcus, Roseburia, Eubacterium, Burkholderiaceae, and Oceanospirillaceae for BBG_0.2; Anoxybacillus and Weissella for BBG_0.2(2). These speciﬁc taxa are highlighted in Table S1. All these unique taxa were composed of less than 0.001% of total taxa, implying their minor role in BBG metabolism and SCFAs/NH production. -Diversity, depicted in Figure 3E, shows that all the BBG samples were clustered as a group with glucose monomer as a distinctive group and T0 as another distinct group. Figure 4 shows the results from BugBase with the upper panel being the predicted relative abundance of aerobes, anaerobes and facultative anaerobes. The relative abundance of aerobes of 5 BBGs were similar to T0 but not the glucose group. As for the predicted anaerobe proportion, T0 group contained the highest abundance. Since BBGs had similar proportion of aerobes with T0 but different proportion of anaerobes with T0, the difference should be contributed by the proportion of facultative anaerobes. As shown, the ﬁve BBGs had various proportions of facultative anaerobes with BBG_0.2 having the highest and BBG the least. T0 group had the least proportion of facultative anaerobes among all. The lower panel of Figure 4 refers to the predicted Gram-positive, Gram-negative and proportion of taxa capable of forming bioﬁlms. For the ﬁve BBGs, a reduction of proportion of Gram-positive taxa and an increase of proportion of Gram-negative taxa compared with glucose was observed. The predicted phenotype of T0 group was of less interest as shown in Figure 3A, while the quasi-absolute number of all bacteria was very low, thus making the comparison of proportion less biologically meaningful. But interesting phenomena could still be observed in the ﬁve BBGs and glucose groups. Figure 5 shows the signiﬁcantly different phylum of the ﬁve BBGs and glucose with descending abundance: Firmicutes (Figure 5A), Proteobacteria (Figure 5D), Actinobacteria (Figure 5B) and Bacteroidetes (Figure 5C). Glucose could signiﬁcantly provide a higher growth support to Firmicutes and Actinobacteria while BBGs allowed enrichment of Bacteroidetes (well-known for complex polysaccharide breakdown such -glucans [11]) and Proteobacteria. Molecules Molecules 2019 2019,, 24 24,, x FO 828 R PEER REVIEW 7 of 7 of 20 20 Molecules 2019, 24, x FOR PEER REVIEW 7 of 20 Figure 4. Predicted output of selected phenotype by BugBase. Relative abundance of aerobes, Figure 4. Predicted output of selected phenotype by BugBase. Relative abundance of aerobes, anaerobes Figure 4. Predicted output of selected phenotype by BugBase. Relative abundance of aerobes, anaerobes and facultative anaerobes are in the upper panel. Gram-positive bacteria, Gram-negative and facultative anaerobes are in the upper panel. Gram-positive bacteria, Gram-negative bacteria and anaerobes and facultative anaerobes are in the upper panel. Gram-positive bacteria, Gram-negative bacteria and taxa that are predicted to form biofilm are given in the lower panel. Significant difference taxa that are predicted to form bioﬁlm are given in the lower panel. Signiﬁcant difference analyzed by bacteria and taxa that are predicted to form biofilm are given in the lower panel. Significant difference analyzed by one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically significant one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically signiﬁcant different from analyzed by one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically significant different from all other groups. all other groups. different from all other groups. Figure 5. Signiﬁcantly different taxa in phyla level identiﬁed using ANOVA (Tukey HSD post-hoc Figure 5. Significantly different taxa in phyla level identified using ANOVA (Tukey HSD post-hoc Figure 5. Significantly different taxa in phyla level identified using ANOVA (Tukey HSD post-hoc test) with Benjamini-Hochberg FDR multiple test correction. (A) Firmicutes; (B) Actinobacteria; test) with Benjamini-Hochberg FDR multiple test correction. (A) Firmicutes; (B) Actinobacteria; (C) test) with Benjamini-Hochberg FDR multiple test correction. (A) Firmicutes; (B) Actinobacteria; (C) (C) Bacteroidetes; and (D) Proteobacteria. These 4 major taxa accounted for 99.90 0.07% of total Bacteroidetes; and (D) Proteobacteria. These 4 major taxa accounted for 99.90 ± 0.07% of total taxa Bacteroidetes; and (D) Proteobacteria. These 4 major taxa accounted for 99.90 ± 0.07% of total taxa taxa identiﬁed. identified. identified. Figure 6 depicts some host beneﬁcial taxa that were identiﬁed in the ﬁve BBGs and glucose. Figure 6 depicts some host beneficial taxa that were identified in the five BBGs and glucose. Glucose Figu supported re 6 depicmor ts some host e probiotic benefic Lactobacillus ial taxa t sp.hwhile at were low identified MW BBGin out-performed the five BBGs and g high MW luco BBG se. Glucose supported more probiotic Lactobacillus sp. while low MW BBG out-performed high MW BBG Glucose supported more probiotic Lactobacillus sp. while low MW BBG out-performed high MW BBG (Figure 6A), a similar phenomenon was also observed in Lactococcus sp. (Figure 6C). Another probiotic (Figure 6A), a similar phenomenon was also observed in Lactococcus sp. (Figure 6C). Another Biﬁdobacterium (Figure 6A), a sp. sim was ilar ph morenomenon w e selectivelya enriched s also obser by glucose ved in than Lactococcus BBGs sp. but (F among igure the 6C ﬁve5 ). An BBGs, other probiotic Bifidobacterium sp. was more selectively enriched by glucose than BBGs but among the five5 probiotic Bifidobacterium sp. was more selectively enriched by glucose than BBGs but among the five5 intermediately MW such as BBG_0.1 was more favorable (Figure 6B). On the other hand, BBGs showed BBGs, intermediately MW such as BBG_0.1 was more favorable (Figure 6B). On the other hand, BBGs better BBGs, interm support ediately towards MW Veillonella such as BBG_0.1 w sp. (higher as more efﬁciency favor inapr ble opionate (Figure 6B pr). oduction On the ot [12 her ], hand Figur , e BBGs 6D), showed better support towards Veillonella sp (higher efficiency in propionate production [12], Figure showed better support towards Veillonella sp (higher efficiency in propionate production [12], Figure Enterobacteriaceae family (common enteric microbe in animal gut, Figure 6E), and Bacteroides (boom 6D), Enterobacteriaceae family (common enteric microbe in animal gut, Figure 6E), and Bacteroides 6D), Enterobacteriaceae family (common enteric microbe in animal gut, Figure 6E), and Bacteroides with diet rich in complex polysaccharides, Figure 6F). (boom with diet rich in complex polysaccharides, Figure 6F). (boom with diet rich in complex polysaccharides, Figure 6F). Molecules 2019, 24, 828 8 of 20 Molecules 2019, 24, x FOR PEER REVIEW 8 of 20 Molecules 2019, 24, x FOR PEER REVIEW 8 of 20 Figure Figure 6. 6. Signiﬁcantly Significantly d dif iffere ferent nt taxa in genus lev taxa in genus level el identiﬁed identified using ANO using ANOV VA A (Tukey HSD post-hoc (Tukey HSD post-hoc Figure 6. Significantly different taxa in genus level identified using ANOVA (Tukey HSD post-hoc test) test) with with Benj Benjamini-Hochber amini-Hochberg g FFDR DR m multiple ultiple tetest st correct correction. ion. (A) (A La ) ct Lactobacillus obacillus; (B;) (Bifido B) Biﬁdobacterium bacterium; (C) ; test) with Benjamini-Hochberg FDR multiple test correction. (A) Lactobacillus; (B) Bifidobacterium; (C) (Lactococcus C) Lactococcus ; (D ;) (D Veill ) Veillonella onella; (E ; ) (E E )n Enter teroba obacteriaceae cteriaceae; an ; and d (F () F)Ba Bacter cteroides oides . These taxa were considered . These taxa were considered Lactococcus; (D) Veillonella; (E) Enterobacteriaceae; and (F) Bacteroides. These taxa were considered beneﬁcial to the hosts. beneficial to the hosts. beneficial to the hosts. On the contrary, Figure 7 depicts some neutral/opportunistically pathogenic taxa. Among the On the contrary, Figure 7 depicts some neutral/ opportunistically pathogenic taxa. Among the selected, On the contrary, BBGs reduced Fig Bacillus ure 7 depict (Figur se some 7A), Faecalibacterium neutral/ opportunist (Figur ice ally 7B), pat Lachnospiraceae hogenic taxa. Among (Figure t 7h D) e selected, BBGs reduced Bacillus (Figure 7A), Faecalibacterium (Figure 7B), Lachnospiraceae (Figure 7D) and selecLeuconostocaceae ted, BBGs reduc(Figur ed Bacillus e 7E). (Fi However gure 7A , BBGs ), Faecalibacterium augmented (F the igu grr owth e 7B),of Lac Klebsiella hnospiraceae (Figur (Figu e 7C) re 7 with D) and Leuconostocaceae (Figure 7E). However, BBGs augmented the growth of Klebsiella (Figure 7C) with some and Leuconos of its species tocaceae r egar (Figur ded e 7E). Howev as emerging er, BBGs pathogens augm [13 ented ]. Figur the gr e 7owth F shows of Klebsiella Enterococcus (Figu as re 7 the C) w most ith some of its species regarded as emerging pathogens [13]. Figure 7F shows Enterococcus as the most abundant some of its species r taxa, commonly egarded as emerg found in animal ing pathogens [13]. F intestines. igure 7F shows Enterococcus as the most abundant taxa, commonly found in animal intestines. abundant taxa, commonly found in animal intestines. Figure 7. Significantly different taxa in genus level identified using ANOVA (Tukey HSD post-hoc Figure Figure 7. 7. Signiﬁcantly Significantly d dififffere erent nt taxa in genus lev taxa in genus level e identiﬁed l identified using ANO using ANOVAV (T A (Tukey HSD post-hoc ukey HSD post-hoc test) test) with Benjamini-Hochberg FDR multiple test correction. (A) Bacillus; (B) Faecalibacterium; (C) with test) w Benjamini-Hochber ith Benjamini-Hochberg FDR mu g FDR multiple test ltiple tes correction. t correctio (A) Bacillus n. (A) ; ( Ba B)ciFaecalibacterium llus; (B) Faecaliba ; (C cterium ) Klensiella ; (C) ; Klensiella; (D) Lachnospiraceae; (E) Leuconostocaceae; and (F) Enterococcus. These taxa were considered (D) Lachnospiraceae; (E) Leuconostocaceae; and (F) Enterococcus. These taxa were considered either neutral Klensiella; (D) Lachnospiraceae; (E) Leuconostocaceae; and (F) Enterococcus. These taxa were considered either neutral or opportunistic pathogenic to the hosts. or opportunistic pathogenic to the hosts. either neutral or opportunistic pathogenic to the hosts. Table 4 shows one phylogeny-based α-diversity index (PD whole tree) and 5 non-phylogeny- Table 4 shows one phylogeny-based -diversity index (PD whole tree) and 5 non-phylogeny-based Table 4 shows one phylogeny-based α-diversity index (PD whole tree) and 5 non-phylogeny- based indices including Fisher alpha, Berger-Parker d, Shannon, Simpson evenness, and Simpson indices including Fisher alpha, Berger-Parker d, Shannon, Simpson evenness, and Simpson reciprocal. based indices including Fisher alpha, Berger-Parker d, Shannon, Simpson evenness, and Simpson reciprocal. Briefly, BBG_0.1 showed significantly higher α-diversity than BBG for PD whole tree, Brieﬂy, BBG_0.1 showed signiﬁcantly higher -diversity than BBG for PD whole tree, Shannon, reciprocal. Briefly, BBG_0.1 showed significantly higher α-diversity than BBG for PD whole tree, Shannon, Simpson evenness, and Simpson reciprocal. While BBG_0.2(2) showed significantly higher Simpson evenness, and Simpson reciprocal. While BBG_0.2(2) showed signiﬁcantly higher value than Shannon, Simpson evenness, and Simpson reciprocal. While BBG_0.2(2) showed significantly higher value than BBG in terms of Fisher alpha, BBG had a significant higher value of Berger-Parker-d index. value than BBG in terms of Fisher alpha, BBG had a significant higher value of Berger-Parker-d index. Overall, BBG_0.1 was shown to have the highest increase in α-diversity. Overall, BBG_0.1 was shown to have the highest increase in α-diversity. Molecules 2019, 24, 828 9 of 20 BBG in terms of Fisher alpha, BBG had a signiﬁcant higher value of Berger-Parker-d index. Overall, BBG_0.1 was shown to have the highest increase in -diversity. Table 4. Rareﬁed -diversity indices of the infant faecal fermentation of 5 BBG samples and glucose monomer. Phylogeny-Based Non-Phylogeny-Based Samples Simpson PD Whole Tree Fisher Alpha Berger Parker D Shannon Simpson_E Reciprocal a,b a,b,c a,b,c,d,e a,b,c,d a a,b,c,d BBG 17.68 0.77 90.81 4.66 0.18 0.01 4.87 0.08 0.026 0.001 13.17 0.84 a,f a,e b a,e BBG_0.05 18.69 0.50 100.40 1.89 5.22 0.07 16.96 0.92 0.14 0.01 0.031 0.002 a a b,g b,f a,c b,f BBG_0.1 19.95 0.59 103.76 5.26 0.15 0.01 5.31 0.11 0.033 0.001 18.27 1.15 c,g c,g b c,h d BBG_0.2 19.41 0.95 105.66 2.98 0.13 0.01 5.21 0.04 0.031 0.000 17.69 0.51 b c d,i d d,h BBG_0.2(2) 19.85 0.53 107.50 5.94 0.15 0.01 5.12 0.08 0.028 0.004 16.14 1.50 e,f,g,h,i e,f,g b,c,d e,f,g,h GLC 19.34 0.81 101.11 5.89 0.22 0.01 4.93 0.02 0.023 0.002 12.63 0.25 a–i Different superscripts represent signiﬁcant difference by one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. 2.3.2. Metagenome Prediction of KEGG Functional Annotation Kyoto Encylopedia of Genes and Genomes (KEGG) functional annotation was done using PICRUSt with standard analysis including 16S copy number normalization and whole metagenome prediction with raw data output are shown in Table S2. The NSTI scores of PICRUSt metagenome prediction for the BBG samples were: BBG (0.032); BBG_0.05 (0.032); BBG_0.1 (0.032); BBG_0.2 (0.033); BBG_0.2(2) (0.031); and that of Glucose (0.034) T0 (0.0589). An NSTI score of 0.03 indicates that the average microbe in your sample can be predicted using a relative from the same (97%) species [14]. All BBG samples had higher number of predicted genes than that of glucose with BBG_0.05 having the highest number (Table S2). “Metabolism" was the highest class of annotated genes in Level 1 as shown in Figure S4A, followed by "Environmental information processing" and “Genetic information processing”. Most genes annotated in “Metabolism” at Level 2 were related to “Carbohydrate metabolism”, followed by “Amino acid metabolism” and “Energy metabolism” with “Lipid metabolism” ranked the sixth as shown in Figure S4B. All BBGs had a higher proportion of predicted KEGG class of "Carbohydrate Metabolism" than glucose monomer and T0, reﬂecting extra machinery was required to breakdown and absorb -glucans. Since BBG metabolism, short chain fatty acid production and ammonia production were our targets of investigation, related KEGG classes were single out for more in-depth study using the Statistical Analysis of Metagenomic Proﬁles (STAMP) approach. Figure S5A shows the “Carbohydrate metabolism” KEGG class. All BBG samples had a signiﬁcant increase in the number of predicted genes in carbohydrate metabolism, compared to glucose. The ﬁve BBG groups also have higher proportion of predicted KEGG class ABC Transporter (Figure S5B), Bacterial Secretion System (Figure S5C) and Glycan biosynthesis and metabolism (Figure S5D). All the four predicted KEGG classes actually coordinated among themselves, such as secretion system to secrete extracellular enzymes, ABC transporters to uptake carbon/glycan with the expense of cellular energy and glycan metabolism to breakdown complex glycan or use the carbon source for bacterial glycan formation. Although “Fatty acid metabolism” does not only involve SCFAs, this class of keys could also be used to infer the potential in SCFAs production. Among the ﬁve BBG samples, BBG_0.1 had the highest number of predicted genes annotated to “Fatty acid metabolism”, followed by BBG_0.05, BBG_0.2, BBG_0.2(2) and the native BBG having the lowest number (Figure S6A). Glucose had a similar proportion of predicted genes in fatty acid metabolism with BBG_0.2. Not only with saccharolytic potential and lipolytic potential, we also extracted some ammonia production related KEGG classes. For example, BBGs stimulated less “Protein digestion and absorption” with medium MW BBG such as BBG_0.005 and BBG_0.1 being more favorable (Figure Molecules 2019, 24, 828 10 of 20 S6B). However, the BBG groups had a signiﬁcant higher relative abundance of KEGG classes “Amino acid metabolism” (Figure S6C) and "Nitrogen metabolism" (Figure S6D). 2.3.3. Microbial Biomarkers Identiﬁcation Molecules 2019, 24, x FOR PEER REVIEW 10 of 20 Microbial biomarkers were identiﬁed using the LEfSe. Table S3 shows the raw data output Microbial biomarkers were identified using the LEfSe. Table S3 shows the raw data output of of signiﬁcant biomarkers identiﬁed with the ﬁve BBG samples and glucose. Brieﬂy, glucose had significant biomarkers identified with the five BBG samples and glucose. Briefly, glucose had the the highest number of biomarkers compared to the ﬁve BBG samples, including Lactobacillus highest number of biomarkers compared to the five BBG samples, including Lactobacillus and and Biﬁdobacterium, while BBG_0.1 gave the highest number of biomarkers among the 5 BBGs, Bifidobacterium, while BBG_0.1 gave the highest number of biomarkers among the 5 BBGs, including including Proteobacteria and Bacteroides. Figure 8 shows the cladogram of biomarkers with Bacteroides, Proteobacteria and Bacteroides. Figure 8 shows the cladogram of biomarkers with Bacteroides, Parabacteroides, Bacillus, Enterococcus, Lactobacillus, Veillonella, Sutterella, and Klebsiella being the Parabacteroides, Bacillus, Enterococcus, Lactobacillus, Veillonella, Sutterella, and Klebsiella being the dominant ones. LEfSe analysis presented some statistically signiﬁcant biomarkers, from which we dominant ones. LEfSe analysis presented some statistically significant biomarkers, from which we choose some biologically signiﬁcant taxa for further comparative analysis. Among these biomarkers, choose some biologically significant taxa for further comparative analysis. Among these biomarkers, excluding the group-speciﬁc unique taxa in the Venn diagram analysis (Figure 3D), some interested excluding the group-specific unique taxa in the Venn diagram analysis (Figure 3D), some interested taxa were chosen for correlation analysis with BBG samples having different MWs. taxa were chosen for correlation analysis with BBG samples having different MWs. Figure 8. Cladogram of biomarkers identiﬁed in the infant faecal fermentation of 5 BBG samples Figure 8. Cladogram of biomarkers identified in the infant faecal fermentation of 5 BBG samples against glucose monomer using LEfSe. against glucose monomer using LEfSe. 3. Discussion 3. Discussion 3.1. Efﬁcacy of Controlled Acid Hydrolysis in MW Reduction 3.1. Efficacy of Controlled Acid Hydrolysis in MW Reduction One of the major focus of this investigation was to investigate the effects of MW reduction in One of the major focus of this investigation was to investigate the effects of MW reduction in BBG on microbial fermentation. We have showed that acid hydrolysis in a controlled manner could BBG on microbial fermentation. We have showed that acid hydrolysis in a controlled manner could produce BBG with reduced molecular weight (Figure 1, Tables 1 and 2) without the modiﬁcation of produce BBG with reduced molecular weight (Figure 1, Table 1,2) without the modification of functional groups or glycosidic linkages. Controlled acid hydrolysis was also employed for the MW functional groups or glycosidic linkages. Controlled acid hydrolysis was also employed for the MW reduction of other -glucans such as -13,16-glucans [15]. Acid hydrolysis of -1,3-1,4-glucan into reduction of other β-glucans such as β-13,16-glucans [15]. Acid hydrolysis of β-1,3-1,4-glucan into smaller fragments had also been conducted with different acids, with results indicated that H SO 2 4 smaller fragments had also been conducted with different acids, with results indicated that H2SO4 produced the least glucose monomer and the highest proportion of oligosaccharides than TFA while produced the least glucose monomer and the highest proportion of oligosaccharides than TFA while HCl produce only monomers at low concentration (0.1 or 0.05 M) [16]. Thus, H SO generates the 2 4 HCl produce only monomers at low concentration (0.1 or 0.05 M) [16]. Thus, H2SO4 generates the production of a wider spectrum of smaller fragments. As shown in Figure 1, a lateral transverse production of a wider spectrum of smaller fragments. As shown in Figure 1, a lateral transverse of peaks from high MW to low MW BBGS and then to glucose monomer was observed in the size exclusion chromatography (SEC). Although the peaks of the five BBGs only differed in MW, they all had similar peak height, skewness and kurtosis. Therefore, we speculate that acid hydrolysis is a ”gain-by-loss” method, i.e., lower MW samples were obtained by random hydrolysis with steric advantage from both ends. Furthermore, tiny fragments such as mono-, di-, tri-, tetramers, might be gradually released and removed by dialysis. Therefore, controlled acid hydrolysis might not be very Molecules 2019, 24, 828 11 of 20 of peaks from high MW to low MW BBGS and then to glucose monomer was observed in the size exclusion chromatography (SEC). Although the peaks of the ﬁve BBGs only differed in MW, they all had similar peak height, skewness and kurtosis. Therefore, we speculate that acid hydrolysis is a ”gain-by-loss” method, i.e., lower MW samples were obtained by random hydrolysis with steric advantage from both ends. Furthermore, tiny fragments such as mono-, di-, tri-, tetramers, might Molecules 2019, 24, x FOR PEER REVIEW 11 of 20 be gradually released and removed by dialysis. Therefore, controlled acid hydrolysis might not be very efﬁcient in producing very short -glucans fragments such as oligo- -glucans which could pass efficient in producing very short β-glucans fragments such as oligo-β-glucans which could pass through dialysis tubing or membrane easily. On the contrary, enzymatic degradation using glycosidic through dialysis tubing or membrane easily. On the contrary, enzymatic degradation using linkage/chain-length speciﬁc enzymes could selectively cut the polymeric chains into fragments glycosidic linkage/chain-length specific enzymes could selectively cut the polymeric chains into to produce oligomers instead of monomers [17,18]. Therefore, controlled acid hydrolysis serves an fragments to produce oligomers instead of monomers [17,18]. Therefore, controlled acid hydrolysis excellent purpose of reducing the MW of -glucans and other polysaccharides extracted from raw serves an excellent purpose of reducing the MW of β-glucans and other polysaccharides extracted materials down to the MW cut-off of tubing/membranes before further reduction into oligomers. from raw materials down to the MW cut-off of tubing/membranes before further reduction into oligome We have rs. tried to correlate different parameters with the MWs of depolymerized BBG samples and glucose We ha(180 ve triDa) ed to correl as the monomer ate different pa of BBGs. ramete In rs wi Figur th the e 9 M A,W we s of observe depolyme an ri incr zed B ease BGof saterminal mples and glucose (180 Da) as the monomer of BBGs. In Figure 9A, we observe an increase of terminal glucose with MW reduction in log scale. The correlation could be well ﬁtted by a power equation glucose with MW reduction in log scale. The correlation could be well fitted by a power equation with an R value = 0.947. Since the 1-3 and 1-4 linkage has an approximate ratio of 1:6, therefore, with with an R value = 0.947. Since the 1-3 and 1-4 linkage has an approximate ratio of 1:6, therefore, with this power equation, we could extrapolate the linkage of BBGs with further MW reduction. On the this power equation, we could extrapolate the linkage of BBGs with further MW reduction. On the other hand, Figure 9B shows total biomass (bacteria) after 40-hour fermentation with different MW other hand, Figure 9B shows total biomass (bacteria) after 40-hour fermentation with different MW of BBGs. We could observe an increase of biomass as MW was increased due to more compact and of BBGs. We could observe an increase of biomass as MW was increased due to more compact and dense energy provided by the longer chain -glucan and then a decrease as higher requirement of dense energy provided by the longer chain β-glucan and then a decrease as higher requirement of energy/time/machinery in breakdown the long compact chain into smaller unit for absorption. Thus, energy/time/machinery in breakdown the long compact chain into smaller unit for absorption. Thus, we observe a turning point with MW close to BBG_0.1 (28 kDa). BBG_0.1 allow the most carrying we observe a turning point with MW close to BBG_0.1 (28 kDa). BBG_0.1 allow the most carrying capacity in equal mass of carbon source added. capacity in equal mass of carbon source added. Figure 9. Correlation analysis of selected parameters against BBGs of different MWs with glucose as Figure 9. Correlation analysis of selected parameters against BBGs of different MWs with glucose as monomer. (A) Terminal glucose molar ratio identiﬁed via linkage analysis using GC-MS; (B) total monomer. (A) Terminal glucose molar ratio identified via linkage analysis using GC-MS; (B) total biomass support by BBGs and glucose. DNA concentration was used to reﬂect the quasi-absolute biomass support by BBGs and glucose. DNA concentration was used to reflect the quasi-absolute bacterial biomass. bacterial biomass. 3.2. Polymeric BBGs and Monomeric Glucose on Microbial Taxa 3.2. Polymeric BBGs and Monomeric Glucose on Microbial Taxa Glucose (Glc) was added as a sole carbon source for the evaluation of monomeric and polymeric Glucose (Glc) was added as a sole carbon source for the evaluation of monomeric and polymeric glucose chain. Some selected taxa among the ﬁve BBGs and glucose have been depicted in Figures 5–7. glucose chain. Some selected taxa among the five BBGs and glucose have been depicted in Figures 5– Here, we evaluate the correlation of Firmicutes/Bacteroides (F/B) ratio as one of the indicators in 7. Here, we evaluate the correlation of Firmicutes/Bacteroides (F/B) ratio as one of the indicators in microbiome of gut eubiosis and dysbiosis. It was proposed that a lower of F/B ratio boost eubiosis microbiome of gut eubiosis and dysbiosis. It was proposed that a lower of F/B ratio boost eubiosis and promote resistance to certain syndromes. In Figure 10A, we observed a U-shape pattern of F/B and promote resistance to certain syndromes. In Figure 10A, we observed a U-shape pattern of F/B ratio against the increase of MW. BBG_0.1 had the lowest F/B ratio among the ﬁve BBGs and glucose ratio against the increase of MW. BBG_0.1 had the lowest F/B ratio among the five BBGs and glucose monomer. Actually, glucose had the highest F/B ratio, in accordance to previous reports, showing monomer. Actually, glucose had the highest F/B ratio, in accordance to previous reports, showing high glucose/fructose consumption would lead to increase of F/B ratio in mice without body weight high glucose/fructose consumption would lead to increase of F/B ratio in mice without body weight changes [19]. Figure 10B,C show two common -diversity indexes including the Shannon index and changes [19]. Figure 10B and Figure 10C show two common α-diversity indexes including the Simpson index. Shannon index emphasizes on both abundance and evenness, while the Simpson Shannon index and Simpson index. Shannon index emphasizes on both abundance and evenness, while the Simpson index is a dominance index because it gives more weight to common or dominant species [20]. A similar pattern between the two index and the most diversifying carbon source was found in BBG_0.1 that gave peak value in both indexes with an inverted-U shape being observed. Molecules 2019, 24, 828 12 of 20 index is a dominance index because it gives more weight to common or dominant species [20]. A similar pattern between the two index and the most diversifying carbon source was found in BBG_0.1 Molecules 2019, 24, x FOR PEER REVIEW 12 of 20 that gave peak value in both indexes with an inverted-U shape being observed. Molecules 2019, 24, x FOR PEER REVIEW 12 of 20 Figure 10. Correlation analysis of selected parameters against BBGs of different MWs with glucose as Figure 10. Correlation analysis of selected parameters against BBGs of different MWs with glucose as monomer. (A) Firmicutes/Bacteroidetes ratio; (B) Shannon -diversity; (C) Simpson -diversity. Figure 10. Correlation analysis of selected parameters against BBGs of different MWs with glucose as monomer. (A) Firmicutes/Bacteroidetes ratio; (B) Shannon α-diversity; (C) Simpson α-diversity. monomer. (A) Firmicutes/Bacteroidetes ratio; (B) Shannon α-diversity; (C) Simpson α-diversity. 3.3. Effect of Polymeric BBGs and Monomeric Glucose on Saccharolytic and Proteolytic Proﬁles 3.3. Effect of Polymeric BBGs and Monomeric Glucose on Saccharolytic and Proteolytic Profiles 3.3. Effect of Polymeric BBGs and Monomeric Glucose on Saccharolytic and Proteolytic Profiles Correlation analysis of the MW of BBG samples together with glucose shows a parabolic function Correlation analysis of the MW of BBG samples together with glucose shows a parabolic with total short chain fatty acids with the highest concentration found at a MW closed to BBG_0.2 Correlation analysis of the MW of BBG samples together with glucose shows a parabolic function with total short chain fatty acids with the highest concentration found at a MW closed to (11 kDa) (Figure 11A), showing BBG_0.2 had the highest saccharolytic potential among the ﬁve BBGs function with total short chain fatty acids with the highest concentration found at a MW closed to BBG_0.2 (11 kDa) (Figure 11A), showing BBG_0.2 had the highest saccharolytic potential among the though this relationship was not linear. The culture medium (mMCB) used in fermentation contained BBG_0.2 (11 kDa) (Figure 11A), showing BBG_0.2 had the highest saccharolytic potential among the five BBGs though this relationship was not linear. The culture medium (mMCB) used in fermentation partially digested protein such as peptone, tryptone, yeast extract and soy peptone. There have been five BBGs though this relationship was not linear. The culture medium (mMCB) used in fermentation contained partially digested protein such as peptone, tryptone, yeast extract and soy peptone. There publication showing that more saccharolytic microbial activity and less proteolytic activity confer contained partially digested protein such as peptone, tryptone, yeast extract and soy peptone. There have been publication showing that more saccharolytic microbial activity and less proteolytic activity eubiosis [21]. Moreover, this inverted-U shape pattern was also found in a similar article using oat have been publication showing that more saccharolytic microbial activity and less proteolytic activity confer eubiosis [21]. Moreover, this inverted-U shape pattern was also found in a similar article using -1,3-1,4-glucans confer eubiosis [of 21] six . Moreo differ vent er, tMW his inver with ted-U sh the original ape pat -glucan tern was and also fo theund lowest in a simi MWla fraction r article us gave ingthe oat β-1,3-1,4-glucans of six different MW with the original β-glucan and the lowest MW fraction gave least oat total β-1,3SCF -1,4-gl Asupr caoduction ns of six dif [22 ferent MW wi ]. We have also th the ori investigated ginal β-glu the can a concentration nd the lowest MW of dissolved fraction ga ammonia ve the least total SCFAs production [22]. We have also investigated the concentration of dissolved the least total SCFAs production [22]. We have also investigated the concentration of dissolved after 40-h fermentation (Figure 11B). Generally, hydrolyzed BBGs produced more dissolved ammonia ammonia after 40-hour fermentation (Figure 11B). Generally, hydrolyzed BBGs produced more ammonia after 40-hour fermentation (Figure 11B). Generally, hydrolyzed BBGs produced more with the native BBG and glucose monomer having similar concentration. Here, we denoted the dissolved ammonia with the native BBG and glucose monomer having similar concentration. Here, dissolved ammonia with the native BBG and glucose monomer having similar concentration. Here, SCF we denoted the SCFAs/NH As/NH ratio as an indicator 3 ratio to as evaluate an indic the ator to evaluate t saccharolytic and he sacch proteolytic arolytic and pro potential. tSince eolytic the we denoted the SCFAs/NH3 ratio as an indicator to evaluate the saccharolytic and proteolytic total potenti SCFAs al. Si and nce the tota NH data l SCFAs are in and dif NH fer 3ent data scales, are in different scales, both data were bo ﬁrst th da scaled ta were fi using rst scaled usi the min-max ng potential. Since the total SCFAs and NH3 data are in different scales, both data were first scaled using the min-max approach before their ratio was taken. As shown in Figure 11C, glucose monomer gave approach before their ratio was taken. As shown in Figure 11C, glucose monomer gave the least the min-max approach before their ratio was taken. As shown in Figure 11C, glucose monomer gave the least SCFAs/NH3 ratio while the native BBG gave the highest SCFAs/NH3 ratio, followed by SCFAs/NH ratio while the native BBG gave the highest SCFAs/NH ratio, followed by BBG_0.2. 3 3 the least SCFAs/NH3 ratio while the native BBG gave the highest SCFAs/NH3 ratio, followed by BBG_0.2. The use of SCFAs/NH3 ratio allowed signal to be observed that could otherwise not be The use of SCFAs/NH ratio allowed signal to be observed that could otherwise not be found in BBG_0.2. The use of SCFAs/NH3 ratio allowed signal to be observed that could otherwise not be found in Figure 11A and 11B. Eubiosis could be achieved by promoting more saccharolysis or by Figure 11A,B. Eubiosis could be achieved by promoting more saccharolysis or by diminishing the found in Figure 11A and 11B. Eubiosis could be achieved by promoting more saccharolysis or by diminishing the proteolysis. Therefore, we proposed that SCFAs/NH3 ratio could serve as a better proteolysis. Therefore, we proposed that SCFAs/NH ratio could serve as a better indicator than just diminishing the proteolysis. Therefore, we proposed that SCFAs/NH3 ratio could serve as a better indicator than just total SCFAs alone. total SCFAs alone. indicator than just total SCFAs alone. Figure 11. Correlation analysis of selected parameters against BBGs of different MWs with glucose as Figure 11. Correlation analysis of selected parameters against BBGs of different MWs with glucose Figure 11. Correlation analysis of selected parameters against BBGs of different MWs with glucose as 3 monomer. (A) Total short chain fatty acids (SCFAs); (B) total dissolved ammonia (NH ); (C) as monomer. (A) Total short chain fatty acids (SCFAs); (B) total dissolved ammonia (NH ); (C) monomer. (A) Total short chain fatty acids (SCFAs); (B) total dissolved ammonia (NH ); (C) SCFAs/NH3 ratio after data scaling to both using the min-max approach. SCFAs/NH ratio after data scaling to both using the min-max approach. SCFAs/NH3 ratio after data scaling to both using the min-max approach. 3.4. Insights from the Phenotypic Data, Microbial Profiles against MW Reduction of BBGs 3.4. Insights from the Phenotypic Data, Microbial Proﬁles against MW Reduction of BBGs 3.4. Insights from the Phenotypic Data, Microbial Profiles against MW Reduction of BBGs With sufficient carbon sources, microbes can grow better in substrates with lower MW because With sufﬁcient carbon sources, microbes can grow better in substrates with lower MW because With sufficient carbon sources, microbes can grow better in substrates with lower MW because of the extra energy/time spent in high MW polymeric breakdown and absorption. Monosaccharides of the extra energy/time spent in high MW polymeric breakdown and absorption. Monosaccharides of the extra energy/time spent in high MW polymeric breakdown and absorption. Monosaccharides such as glucose, and simple sugars such as lactose and sucrose are therefore widely used as sole such as glucose, and simple sugars such as lactose and sucrose are therefore widely used as sole carbon such as glucose, and simple sugars such as lactose and sucrose are therefore widely used as sole carbon sources in microbiological culture media preparation. However, the advantage of MW carbon sources in microbiological culture media preparation. However, the advantage of MW reduction does not follow a linear trend as with MW as demonstrated by Figures 9–11. It is because reduction does not follow a linear trend as with MW as demonstrated by Figures 9–11. It is because fermentation in pure bacterial culture does not consider the interactions between microbes. For fermentation in pure bacterial culture does not consider the interactions between microbes. For Molecules 2019, 24, 828 13 of 20 sources in microbiological culture media preparation. However, the advantage of MW reduction does not follow a linear trend as with MW as demonstrated by Figures 9–11. It is because fermentation in pure bacterial culture does not consider the interactions between microbes. For example, although strain speciﬁc, the in vitro fermentation of oat -1,3-1,4-glucan and its hydrolyzed fragments with pure probiotic culture Lactobacillus helveticus, Lactobacillus rhamnosus and Biﬁdobacterium longum showed that growth and total SCFAs production was highest with glucose, followed by hydrolyzed derivative and then the original oat -glucan [23]. Important elements of research in microbial community ecology include the analysis of functional pathways for nutrient resource and energy ﬂows, mechanistic understanding of interactions between microbial populations and their environment, and the emergent properties of the complex community [24]. Microbial community interaction is multi-fold, including competition, mutualism, inhibition, promotion and so on. Therefore, the use bacterial community in fermentation leads to a diverse results. In the present study, four very different correlation patterns resulting from fermentation with BBGs with different MWs were extracted and discussed as examples. Reduction of the MW in BBGs seemed to increase the population of Lactobacillus gradually with glucose having the maximum increase (Figure 12A). Though it was not linear, a monotonous trend was still observed. Such trend suggested that further reduction in the MW of BBGs would probably increase the population growth of Lactobacillus among the whole microbial community. On the other hand, Figure 12B shows the pattern of Biﬁdobacterium, with the highest population increase in Biﬁdobacterium being with BBG_01. An inverted-U shape pattern was also recognized among the ﬁve BBGs. With reference to glucose, we expected a further reduction in the MW of BBG samples would ﬁrstly reduce the population of Biﬁdobacterium followed by an increase again. Actually, glucose monomer induced particularly Biﬁdobacterium growth (a little bit more relative abundance than BBG_0.01). Veillonella might have an exceptional ability in fermenting BBG with MW lower than BBG_0.2(2) as shown in Figure 12C. Veillonella is a group of microbes well-known for its ability to convert lactate into acetate and propionate during fermentation [12]. This might also contribute to the trend in propionic acid production (Figure 2A) as the growth pattern of Veillonella matched well with the propionic acid production pattern. On the contrary, the growth pattern of Veillonella was a mirror reﬂection of Biﬁdobacterium with Veillonella having a slight phase difference. This could possibly explain the reduction of Biﬁdobacterium with BBG having a MW lower than BBG_0.2(2) (6 kDa) as Veillonella might be able to utilize BBGs with a MW lower than 6 kDa. Figure 12D shows the correlation pattern of the Enterococcus which contributed about 60% of total bacterial abundance in all the ﬁve BBG samples. It was observed that the population of Enterococcus had a moderate increase when the MW was above that of BBG_0.2(2) but decreased drastically when the MW was below that of BBG_0.2(2) (Figure 12D). Growth of Enterococcus was supported by polymeric BBGs while glucose monomer, and perhaps dimers and oligomers are extrapolated to be less supportive under in vitro condition. Molecules 2019, 24, 828 14 of 20 Molecules 2019, 24, x FOR PEER REVIEW 14 of 20 Figure 12. Correlation analysis of selected taxa at genus level against BBGs of different MWs with Figure 12. Correlation analysis of selected taxa at genus level against BBGs of different MWs with glucose as monomer. (A) Lactobacillus; (B) Biﬁdobacterium; (C) Veillonella; and (D) Entrococcus. glucose as monomer. (A) Lactobacillus; (B) Bifidobacterium; (C) Veillonella; and (D) Entrococcus. 4. Materials and Methods 4. Materials and Methods 4.1. Materials 4.1. Materials All chemicals were obtained from Sigma-Aldrich (St. Louis, Missouri, USA) unless All chemicals were obtained from Sigma-Aldrich (St. Louis, Missouri, USA) unless otherwise otherwise speciﬁed. specified. 4.2. Infant Faecal Sample Collection 4.2. Infant Faecal Sample Collection Sample collection protocols were approved by the Clinical Research Ethics Committee of the Sample collection protocols were approved by the Clinical Research Ethics Committee of the Joint Chinese University of Hong Kong (CREC Ref. No: 2015.206). Brieﬂy, infant faecal samples were Joint Chinese University of Hong Kong (CREC Ref. No: 2015.206). Briefly, infant faecal samples were obtained from four 9–15 months old infants (3–15 month is known as the microbiome developmental obtained from four 9–15 months old infants (3–15 month is known as the microbiome developmental phase [25]) without any antibiotic or medicine intake at least three months prior to the collection. All phase [25]) without any antibiotic or medicine intake at least three months prior to the collection. All the infants were delivered by natural birth and undertook solid food and breast milk simultaneously the infants were delivered by natural birth and undertook solid food and breast milk simultaneously at the at the ti time me of the experi of the experiments, ments, gi giving ving p possibly ossibly th the e h highest ighest microb microbial ial diver diversity sity (s(solid olid food on food only ly and and breast-feeding breast-feedin only g only gave gadif ve fer different d ent diversity iversity [25]).[25] The ). The faecalfa samples ecal samples were collected were collect in sample ed in sample collection collection tubes diluted at 1:1 v/v with stabilizing buffer containing 30% glycerol (Thermo Scientific tubes diluted at 1:1 v/v with stabilizing buffer containing 30% glycerol (Thermo Scientiﬁc Pierce, Pierce, Waltham, MA, USA, to maintain bacteria viability during cold storage) supplemented with Waltham, MA, USA, to maintain bacteria viability during cold storage) supplemented with 0.1% 0.1% cysteine hydrochloride (a reducing agent to maintain anaerobicity). Faecal samples were stored cysteine hydrochloride (a reducing agent to maintain anaerobicity). Faecal samples were stored in in −20 °C during the delivery to the laboratory and were kept at −80 °C upon arrival. The viability of 20 C during the delivery to the laboratory and were kept at80 C upon arrival. The viability of the the faecal bacteria was checked by using total plate count (TPC) with reinforced clostridial broth faecal bacteria was checked by using total plate count (TPC) with reinforced clostridial broth (Oxoid (Oxoid Thermo, Waltham, MA, USA) with 2% agar (RCA) to ensure a sufficient number of viable Thermo, Waltham, MA, USA) with 2% agar (RCA) to ensure a sufﬁcient number of viable bacteria at a bacteria at a dilution of 10 . Any faecal sample with a TPC of less than 300 Colony Forming Units dilution of 10 . Any faecal sample with a TPC of less than 300 Colony Forming Units (CFUs) at the (CFUs) at the dilution of 10 was rejected. All faecal samples from the four infants were then pooled dilution of 10 was rejected. All faecal samples from the four infants were then pooled together with together with equal TPC CFU ratio to consolidate the generalizability of findings. equal TPC CFU ratio to consolidate the generalizability of ﬁndings. 4.3. Depolymerization of Barley β-Glucans 4.3. Depolymerization of Barley b-Glucans Low viscosity barley β-1,3-1,4-glucans (BBGs) was obtained from Megazyme (product code: P- Low viscosity barley -1,3-1,4-glucans (BBGs) was obtained from Megazyme (product code: BGBL, >95% purity as stated). Acid hydrolysis was performed with different concentration of sulfuric P-BGBL, >95% purity as stated). Acid hydrolysis was performed with different concentration of acid (0.05, 0.10 and 0.20 M) for various time periods (10, 60, and 120 min) to obtain a range of BBG sulfuric acid (0.05, 0.10 and 0.20 M) for various time periods (10, 60, and 120 min) to obtain a range of Molecules 2019, 24, 828 15 of 20 BBG samples with different MWs. Brieﬂy, an aqueous solution of BBG at a concentration of 10 mg/mL was prepared and maintained at 80 C, followed by the addition of an appropriate amount of sulfuric acid (8M) to give a ﬁnal concentration of 0.05, 0.1 and 0.2 M. The samples were stirred at 80 C for the various time periods. The preparation scheme for BBG is shown in Table 1. A total of ﬁve BBG samples, including the native BBG and four depolymerized derivatives were obtained after the different acid hydrolysis conditions and they were denoted as BBG, BBG_0.05, BBG_0.1, BBG_0.2, and BBG_0.2(2). These sample solutions were neutralized with NaOH, dialyzed in Spectra/Por tubing with molecular weight cut off (MWCO) 3 kDa (Repligen, Waltham, MA, USA) for desalting in Milli-Q water for 24 h with three changes of water until the conductivity of the Milli-Q water reached single digit (Ciba-Corning Cond/TDS, Corning, NY, USA). All dialyzed BBG samples were then freeze-dried. 4.4. Chemical Composition and Linkage Analysis of BBGs The total carbohydrate and total protein contents of the ﬁve BBG samples were determined by the phenol-sulfuric acid method and the bicinchoninic acid (BCA) protein assay (Thermo Scientiﬁc Pierce, Waltham, MA, USA), respectively by protocols described previously [26]. The monosaccharide composition of the native BBG and its depolymerized derivatives were determined by gas chromatographic analysis of the alditol acetate derivatives of the sugars formed by sequential acid hydrolysis, reduction and acetylation described previously [27]. The glycosidic linkages of the ﬁve BBG samples were determined by methylation analysis. In brief, partially methylated alditol acetate (PMAA) derivatives of the samples were prepared using CH I and solid NaOH in dry DMSO described in a modiﬁed protocol reported previously. A gas chromatograph (6890N, Agilent Technologies, Santa Clara, CA, USA) coupled to a mass spectrometer (5973N, Agilent Technologies) (GC-MS) was used for analysis of both monosaccharides as their alditol acetate derivatives and linkage positions between sugar residues as their PMAA derivatives by protocols described previously [26]. 4.5. Molecular Weight Distribution of BBGs The MW distribution of the ﬁve BBG samples was determined using size exclusion chromatography (SEC) (Waters e2695 HPLC system) coupled with a refractive index (RI) detector (Waters 2414, Waters Inc., Milford, MA, USA) [28]. In brief, a size exclusion column TSK gel G5000 PW (30 cm 7.5 mm i.d., Supelco, Bellefonte, PA, USA) was used to determine the MW proﬁle. A range of dextran standards (5, 25, 80, 150, 410, 670 kDa) were used for calibrating the MW in the SEC analysis. The samples were dissolved in Milli-Q water to a concentration of 10 mg/mL and ﬁltered through 0.22 mm nylon syringe ﬁlter before injection. Water was applied as mobile phase at a ﬂow-rate of 0.7 mL/min at 30 C. 4.6. Spectroscopic Methods of BBGs The infrared spectra of the ﬁve BBG samples were recorded with a Fourier transform infrared spectrometer (FT-IR, Nicolet 670, Thermo Scientiﬁc, Waltham, MA, USA) in the range 4000–400/cm using the KBr-disk method [29]. One-dimensional proton NMR measurements of the samples were analyzed on a 700 MHz NMR spectrometer (Bruker Advanced III HD, Bruker, Billerica, MA, USA) at 60 C using DMSO-d as the solvent [30]. 4.7. In Vitro Deep Well-Plate Fermentation Modiﬁed medium for colonic broth (mMCB) was used as the basal medium for the fermentation experiments as previously reported [29] and contained (per litre) 6.5 g of bacteriological peptone, 5.0 g of soy peptone, 2.5 g of tryptone, 3.0 g of yeast extract, 2.0 g of KCl, 0.2 g of NaHCO , 4.5 g of NaCl, 0.5 g of MgSO 7H O, 0.45 g of CaCl 2H O, 0.2 g of MnSO H O, 0.005 g of FeSO 7H O, 4 2 2 2 4 2 4 2 0.005 g of ZnSO 7H O, 0.4 g of cysteine–HCl, 0.005 g of hemin, 0.005 g of menadione, 0.5 mL of 4 2 H PO , and 2 mL of Tween 80. No added carbon sources. All ﬁve BBG samples as well as glucose 3 4 (as -glucan monomer) were added as the sole carbon source at a concentration of 2% (w/v). Pooled Molecules 2019, 24, 828 16 of 20 infant faecal inoculum was mixed with mMCB having individual carbon sources at a volume ratio of 1:9 [31]. All fermentation mixtures (carbon sources in mMCB with infant faecal inoculum) were done in biological triplicates for all ﬁve BBGs and glucose monomer. Fermentation was conducted in deep well plates (Thermo Scientiﬁc) with each well composed of 2 mL of fermentation mixtures in an anaerobic jar containing AnaeroGen (Oxoid Thermo) and incubated at 37 C for 40 h [31]. At the end of the fermentation period, aliquots of the fermentation samples (from the triplicates of each of the 6 carbon sources) were collected for microbial plating (100 L), DNA extraction (200 L), short chain fatty acid proﬁling (600 L) and ammonia determination (100 L). 4.8. Standard Bacterial Plate Count Standard bacterial plate count was done using Reinforced Clostridia Broth (RCB) (Oxoid Thermo) plus 2% Bacteriological Agar (Oxoid Thermo), maintained under anaerobic conditions in duplicate using standard protocols of serial dilution with sterile PBS [31]. The growth of the total bacteria in the fermentation mixture after 40-h fermentation was expressed as CFUs after 48 h anaerobic incubation at 37 C. 4.9. Short Chain Fatty Acids and Dissolved Ammonia Determination SCFAs were extracted and analyzed using GC-FID with slight modiﬁcations [27]. In brief, 30 L of methylvaleric acid (100 mg/mL) was added as internal standard and a hydrogen ﬂow rate of 0.5 mL/min were used in the present method. A mixture of individual SCFA standards including acetic, propionic, butyric, valeric and caproic acid (SCFA standards kit, Alltech Inc., Nicolasville, KY, USA), and 4-methylvaleric acid (as internal standard) was prepared in 25% metaphosphoric acid at a ﬁnal concentration of 10.0 mg/mL for identiﬁcation and quantitation. The amount of the SCFAs after 24 h of fermentation was expressed in mM. Ammonia was determined using Ammonia Rapid Assay (K-AMIAR, Megazyme). It is a highly speciﬁc assay involving the enzymatic conversion of 2-oxoglutarate, NADPH and dissolved ammonia into L-glutamic acid, NADP and water by glutamate dehydrogenase. Brieﬂy according to manual, 100 L fermentation samples were centrifuged at 14000 rpm (Spectrafuge 24D, Labnet, Woodbridge, NJ, USA) for 5 min to obtain supernatant, then 2 L of samples, ammonia standard (0.04 mg/mL) and water as blank were added with 30 L Buffer and 20 L NADPH with 208 L distilled water replaced by PBS pH 9.8 to maintain solubility of ammonia in reaction. Absorbance (A1) was taken at 340 nm with microplate reader (Molecular Devices SpectraMax Plus 384, San Jose, CA, USA). Then 2 L glutamate dehydrogenase was added to each mixture and allowed to react for 5 min, followed by second absorbance (A2) at 340 nm. Calibration was also performed using a serial 2-fold dilution of ammonia standards. The ﬁnal ammonia concentration (mg/mL) was calculated as: (Sample [A1 A2]/Standard[A1 A2]) mg/mL Standard. 4.10. DNA Extraction, 16S Amplicon Sequencing and Bioinformatics There was a total of 18 sequencing samples consisted of BBG, BBG_0.05, BBG_0.1, BBG_0.2, BBG_0.2(2), and glucose (each in biological triplicate). DNA extraction was conducted using EZNA Stool DNA Kits (Omega Bio-tek Inc., Norcross, GA, USA) in accordance with the manufacturer ’s manual with some modiﬁcations [32]. These included an enzymatic pre-treatment step (mutanolysin and lysozyme) and a bead-beading step using a mix of 0.1 mm and 0.5 mm beads (Bioprep-24 Homogenizer & bead beater) to enhance DNA release from Gram positive bacteria. 16S amplicons were generated with Phusion polymerase (New England BioLabs, Ipswich, MA, USA) with barcoded forward primers targeting 16S V3 region [32] based on manufacturer ’s protocols. After PCR, sharp DNA bands appeared at approximately 200 bp were cut and gel puriﬁed using QiaQuick Gel Extraction kit (Qiagen, Venlo, Netherlands) followed by AMPure treatment using the Agencourt AMPure XP system (Beckman Coulter, Brea, CA, USA). The quantity and quality of DNA after AMPure treatment were accessed using a Qubit Fluorometer (Thermo Scientiﬁc, Waltham, MA, USA) and Bio-Analyzer Molecules 2019, 24, 828 17 of 20 (Agilent, Santa Clara, CA, USA). Sequencing library was prepared by pooling the 18 samples in equal molar ratio. Then, emulsion PCR and enrichment were done using the Ion PGM Template OT2 400 Kit in Ion OneTouch 2 System (Thermo Scientiﬁc, Waltham, MA, USA). The DNA sequencing was done using the Ion PGM Sequencing 400 Kit on Ion torrent PGM using Chip 318v2 in the Core Facilities of the Chinese University of Hong Kong. After sequencing, the individual sequence reads were ﬁltered by the PGM software to remove low quality and polyclonal sequences. Sequences matching the PGM 3 adaptor were also automatically trimmed. All PGM quality-approved, trimmed and ﬁltered data were exported as .sff ﬁles. Bioinformatics analysis was done using Qiime-v1.9.1 pipeline [33]. Greengenes-v13.8 was used as reference for bacterial 16S sequence database [34]. Sequences were trimmed for primers and barcodes. The cleaned sequences were then clustered at 97% similarity as Operational Taxonomic Unites (OTUs) followed by deletion of chimeras and singleton reads. and diversity were then analyzed. KEGG prediction was carried out using PICRUSt-v1.0.0 [14] with NSTI score output included and statistically analyzed using STAMP-v2.1.3 [35]. Bacterial biomarkers unique to groups were identiﬁed using LEfSe [36]. Signiﬁcant microbial taxa were also identiﬁed using LEfSe. High level microbial phenotypes that could not predicted by PICRUSt were analyzed using BugBase and default BugBase traits were predicted [37]. All computational analysis was carried out on a local ® ® Dell Workstation with Quad Core Intel Xeon Processor and 32G RAM. 4.11. Integration and Statistical Analysis Integration analysis was done by ﬁtting the trends of different parameters (DNA amount/terminal glucose/SCFAs/NH3/microbial taxa) investigated in this study against increase in MW modelled with regression equation and coefﬁcient of determination provided. Unless otherwise speciﬁed, R software-v3.5.2 [38] was used for statistical calculation and independent pairwise student’s t-test or ANOVA with Tukey HSD post-hoc adjustment (CI = 0.95) was used. Statistical test parameters for bioinformatics tools include: STAMP (Statistical test: ANOVA with Tukey-Kramer post-hoc test CI = 0.95, multiple test correction: Benjamini-Hochberg FDR); LEfSe (alpha = 0.05 for factorial Kruskal-Wallis test; alpha = 0.05 for pairwise Wilcoxon test; threshold = 2.0 for logarithmic LDA score) and BugBase (threshold values were automatically deﬁned with highest variance across all sample, output to be <0.03, ANOVA with Tukey HSD post-hoc adjustment (CI = 0.95) was used). 5. Conclusions Our present results have demonstrated a partial acid hydrolysis under controlled conditions seemed to be an effective means to depolymerize BBG into derivatives with lower MW. It is a versatile method suitable for screening and puriﬁcation after hydrolysis. We have also demonstrated that the platform used for infant faecal sample collection and storage as well as using the deep well plate to serving as multiple miniature bioreactors, is applicable for in vitro fermentation. We have also coined the SCFAs/NH ratio as a new indicator of eubiosis/dysbiosis in terms of saccharolytic and proteolytic fermentation. Correlation analysis had shown that most of the fermentation parameters investigated showed a parabolic function instead of a linear function with the BBG samples having different MW. Microbial community investigation is challenging as there is a myriad of interactions among all the microbes. We have exempliﬁed this idea with the monotonous pattern, U-shape pattern, inverted-U shape pattern, wave-like pattern of some selected taxa. Among the ﬁve BBGs, we concluded that the BBG_0.1 with intermediate MW of 28 kDa showed the most promising lead to be developed as a potential novel prebiotic. The data of this project has been submitted to NCBI SRA and available under the BioProject ID: PRJNA487303, SRA accession: SRP158602. Supplementary Materials: The following are available online, Figure S1, (A) GC monosaccharide proﬁle; (B) FTIR spectrum; and (C) NMR 1H spectrum of BBG and BBG_0.2(2). Figure S2, Representative bacterial total plate count of the 5 BBG samples and glucose monomer. Figure S3, GC-FID chromatograms (A) Short chain fatty acid external standards, C2, C3, C4, C5, C6 and internal standard, methyl-valeric acid; (B) short chain fatty acid proﬁles of the BBG after 40 h of fermentation. Figure S4, Overall distribution of metagenome KEGG prediction output of the infant faecal fermentation of 5 BBG samples, glucose monomer and T0 group using PICRUSt. (A) Molecules 2019, 24, 828 18 of 20 KEGG level 1 class distribution; (B) expanded KEGG level 2 class distribution from level 1 “Metabolism” group. Figure S5, Selected metagenome KEGG prediction output of the infant faecal fermentation of 5 BBG samples and glucose monomer using PICRUSt and statistically analysis using STAMP. (A) Carbohydrate metabolism; (B) ABC transporters; (C) Bacterial secretion system; and (D) Glycan biosynthesis and metabolism. Figure S6, Selected metagenome KEGG prediction output of the infant faecal fermentation of 5 BBG samples and glucose monomer using PICRUSt and statistically analysis using STAMP. (A) Lipid metabolism; (B) Protein digestion and absorption; (C) Amino acid metabolism; and (D) Nitrogen metabolism. Table S1, All the identiﬁed microbial taxa of the 5 BBGs (XLSX format). Table S2, The output of PICRUSt metagenome prediction of infant faecal fermentation of 5 BBG samples and glucose (XLSX format). Table S3, The output of LEfSe of infant faecal fermentation of 5 BBGs and glucose (XLSX format). Author Contributions: K.-L.L., K.-C.K., T.C., H.-S.K., L.Y. and P.C.-K.C. designed the experiments. K.-L.L. played the major role, with assistance of K.-C.K., X.L., X.K., W.-Y.C. for performing all experiments, analysis and bioinformatics; K.-L.L. wrote the manuscript with advice from by P.C.-K.C. All authors contributed to interpreting the results, critically revising the manuscript for important intellectual content, and approving the ﬁnal manuscript. Funding: This research was funded by RGC Research Grant General Research Fund, grant number 14114115. Acknowledgments: The authors express their gratitude to the breast-feeding mother in Hong Kong for their participation in this project, collection and donation of infant faecal samples. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Grain: World Markets and Trade (December report by United States Department of Agriculture Foreign Agricultural Service). Available online: https://downloads.usda.library.cornell.edu/usda-esmis/ﬁles/ zs25x844t/8g84mr14m/0k225g11x/grain.pdf (accessed on 1 February 2019). 2. Lam, K.L.; Cheung, P.C.K. Non-digestible long chain -glucans as novel prebiotics. Bioact. Carbohydr. Dietary Fibre. 2013, 2, 45–64. [CrossRef] 3. Jayachandran, M.; Chen, J.; Chung, S.S.M.; Xu, B. A critical review on the impacts of -glucans on gut microbiota and human health. J. Nutr. Biochem. 2018, 61, 101–110. [CrossRef] [PubMed] 4. Zhu, F.; Du, B.; Xu, B. A critical review on production and industrial applications of -glucans. Food Hydrocoll. 2016, 52, 275–288. [CrossRef] 5. Hughes, S.A.; Shewry, P.R.; Gibson, G.R.; McCleary, B.V.; Rastall, R.A. In vitro fermentation of oat and barley derived -glucans by human faecal microbiota. FEMS Microbiol. Ecol. 2008, 64, 482–493. [CrossRef] [PubMed] 6. Williams, B.A.; Mikkelsen, D.; le Paih, L.; Gidley, M.J. In vitro fermentation kinetics and end-products of cereal arabinoxylans and (1,3;1,4)- -glucans by porcine faeces. J. Cereal Sci. 2011, 53, 53–58. [CrossRef] 7. Mikkelsen, M.S.; Jensen, M.G.; Nielsen, T.S. Barley -glucans varying in molecular mass and oligomer structure affect cecal fermentation and microbial composition but not blood lipid proﬁles in hypercholesterolemic rats. Food Funct. 2017, 8, 4723–4732. [CrossRef] [PubMed] 8. Korpela, K. Diet, microbiota, and metabolic health: Trade-off between saccharolytic and proteolytic fermentation. Annu. Rev. Food Sci. Technol. 2018, 9, 65–84. [CrossRef] [PubMed] 9. Tsai, Y.L.; Olson, B.H. Rapid method for direct extraction of DNA from soil and sediments. Appl. Environ. Microbiol. 1991, 57, 1070–1074. [PubMed] 10. Vogt, S.L.; Pena-Díaz, J.; Finlay, B.B. Chemical communication in the gut: Effects of microbiota generated metabolites on gastrointestinal bacterial pathogens. Anaerobe 2015, 34, 106–115. [CrossRef] [PubMed] 11. Tamura, K.; Hemsworth, G.R.; Dejean, G.; Rogers, T.E.; Pudlo, N.A.; Urs, K.; Jain, N.; Davies, G.J.; Martens, E.C.; Brumer, H. Molecular mechanism by which prominent human gut bacteroidetes utilize mixed-linkage -glucans, major health-promoting cereal polysaccharides. Cell Rep. 2017, 21, 417–430. [CrossRef] [PubMed] 12. Riíos-Coviaán, D.; Ruas-Madiedo, P.; Margolles, A.; Gueimonde, M.; de los Reyes-Gavilaán, C.G.; Salazar, N. Intestinal Short Chain Fatty Acids and their Link with Diet and Human Health. Front. Microbiol. 2016, 7, 185. [CrossRef] [PubMed] 13. Seidel, J.; Haller, S.; Eckmanns, T.; Harder, T. Routine screening for colonization by Gram-negative bacteria in neonates at intensive care units for the prediction of sepsis: Systematic review and meta-analysis. J. Hosp. Infect. 2018, 99, 367–380. [CrossRef] [PubMed] Molecules 2019, 24, 828 19 of 20 14. Langille, M.G.I.; Zaneveld, J.; Caporaso, J.G.; McDonald, D.; Knights, D.; a Reyes, J.; Clemente, J.C.; Burkepile, D.E.; Vega Thurber, R.L.; Knight, R.; et al. Predictive functional proﬁling of microbial communities using 16S rRNA marker gene sequences. Nat. Biotech. 2013, 31, 814–821. [CrossRef] [PubMed] 15. Ishimoto, Y.; Ishibashi, K.; Yamanaka, D.; Adachi, Y.; Kanzaki, K.; Iwakura, Y.; Ohno, N. Production of low-molecular weight soluble yeast -glucan by an acid degradation method. Int. J. Biol. Macromol. 2018, 107, 2269–2278. [CrossRef] [PubMed] 16. Johansson, L.; Virkki, L.; Anttila, H.; Esselstrom, H.; Tuomainen, P.; Sontag-Strohm, T. Hydrolysis of -glucan. Food Chem. 2006, 97, 71–79. [CrossRef] 17. Cho, H.J.; Jang, W.J.; Moon, S.Y.; Lee, J.M.; Kim, J.H.; Han, H.S.; Kim, K.W.; Lee, B.J.; Kong, I.S. Immobilization of -1,3-1,4-glucanase from Bacillus sp. on porous silica for production of -glucooligosaccharides. Enzyme Microb. Technol. 2018, 110, 30–37. [CrossRef] [PubMed] 18. Shi, Y.; Liu, J.; Yan, Q.; You, X.; Yang, S.; Jiang, Z. In vitro digestibility and prebiotic potential of curdlan (1-3)- -D-glucan oligosaccharides in Lactobacillus species. Carbohydr. Polym. 2018, 188, 17–26. [CrossRef] [PubMed] 19. Do, M.H.; Lee, E.; Oh, M.J.; Kim, Y.; Park, H.Y. High-glucose or-fructose diet cause changes of the gut microbiota and metabolic disorders in mice without body weight change. Nutrients 2018, 10, 761. [CrossRef] [PubMed] 20. Morris, E.K.; Caruso, T.; Buscot, F.; Fischer, M.; Hancock, C.; Maier, T.S.; Meiners, T.; Muller, C.; Obermaier, E.; Prati, D.; et al. Choosing and using diversity indices: Insights for ecological applications from the german biodiversity exploratories. Ecol. Evol. 2014, 4, 3514–3524. [CrossRef] [PubMed] 21. Makki, K.; Deehan, E.C.; Walter, J.; Backhed, F. The impact of dietary ﬁber on gut microbiota in host health and disease. Cell Host Microbe 2018, 23, 705–715. [CrossRef] [PubMed] 22. Kim, H.J.; White, P.J. Optimizing the molecular weight of oat -glucan for In vitro bile acid binding and fermentation. J. Agric. Food Chem. 2011, 59, 10322–10328. [CrossRef] [PubMed] 23. Dong, J.; Yu, X.; Dong, L.; Shen, R. In vitro fermentation of oat -glucan and hydrolysates by fecal microbiota and selected probiotic strains. J. Sci. Food Agric. 2017, 97, 4198–4203. [CrossRef] [PubMed] 24. Konopka, A. What is microbial community ecology? ISME J. 2009, 3, 1223–1230. [CrossRef] [PubMed] 25. Stewart, C.J.; Ajami, N.J.; O’Brien, J.L.; Hutchinson, D.S.; Smith, D.P.; Wong, M.C.; Ross, M.C.; Lloyd, R.E.; Doddapaneni, H.; Metcalf, G.A.; et al. Temporal development of the gut microbiome in early childhood from the teddy study. Nature 2018, 562, 583–588. [CrossRef] [PubMed] 26. Li, X.; Zhou, J.; Liu, C.; Xiong, Q.; Duan, H.; Cheung, P.C.K. Stable and biocompatible mushroom -glucan modiﬁed gold nanorods for cancer photothermal therapy. J. Agric. Food Chem. 2017, 65, 9529–9536. [CrossRef] [PubMed] 27. Zhao, J.; Cheung, P.C. Fermentation of -glucans derived from different sources by biﬁdobacteria: Evaluation of their biﬁdogenic effect. J. Agric. Food Chem. 2011, 59, 5986–5992. [CrossRef] [PubMed] 28. Liu, C.; Choi, M.W.; Li, X.; Cheung, P.C.K. Immunomodulatory effect of structurally-characterized mushroom sclerotial polysaccharides isolated from Polyporus rhinocerus on human monoctyes THP-1. J. Funct. Foods 2018, 41, 90–99. [CrossRef] 29. Liu, C.R.; Chen, J.L.; Chen, L.; Huang, X.S.; Cheung, P.C.K. Immunomodulatory activity of polysaccharide-protein complex from the mushroom sclerotia of Polyporus rhinocerus in murine macrophages. J. Agric. Food Chem. 2016, 64, 3206–3214. [CrossRef] [PubMed] 30. Chen, L.; Zhang, B.; Chen, J.; Cheung, P.C.K. Cell wall structure of mushroom sclerotium (Pleurotus tuber-regium): Part 2. Fine structure of a novel alkali-soluble hyper-branched cell wall polysaccharide. Food Hydrocoll. 2014, 38, 48–55. [CrossRef] 31. Lam, K.L.; Keung, H.Y.; Ko, K.C.; Kwan, H.S.; Cheung, P.C.K. In vitro fermentation of -glucans and other selected carbohydrates by infant fecal inoculum: An evaluation of their potential as prebiotics in infant formula. Bioact. Carbohydr. Dietary Fibre 2018, 14, 20–24. [CrossRef] 32. Milani, C.; Hevia, A.; Foroni, E.; Duranti, S.; Turroni, F.; Lugli, G.A.; Sanchez, B.; Martin, R.; Gueimonde, M.; van Sinderen, D.; et al. Assessing the fecal microbiota: An optimized ion torrent 16S rRNA gene-based analysis protocol. PLoS ONE 2013, 8, 68739. [CrossRef] [PubMed] 33. Caporaso, J.G.; Kuczynski, J.; Stombaugh, J.; Bittinger, K.; Bushman, F.D.; Costello, E.K.; Fierer, N.; Pena, A.G.; Goodrich, J.K.; Gordon, J.I.; et al. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 2010, 7, 335–336. [CrossRef] [PubMed] Molecules 2019, 24, 828 20 of 20 34. DeSantis, T.Z.; Hugenholtz, P.; Larsen, N.; Rojas, M.; Brodie, E.L.; Keller, K.; Huber, T.; Dalevi, D.; Hu, P.; Andersen, G.L. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Appl. Environ. Microbiol. 2006, 72, 5069–5072. [CrossRef] [PubMed] 35. Parks, D.H.; Tyson, G.W.; Hugenholtz, P.; Beiko, R.G. STAMP: Statistical analysis of taxonomic and functional proﬁles. Bioinformatics 2014, 30, 3123–3124. [CrossRef] [PubMed] 36. Segata, N.; Izard, J.; Walron, L.; Gevers, D.; Miropolsky, L.; Garrett, W.; Huttenhower, C. Metagenomic biomarker discovery and explanation. Genome Biol. 2011, 12, R60. [CrossRef] [PubMed] 37. Ward, T.; Larson, J.; Meulemans, J.; Hillmann, B.; Lynch, J.; Sidiropoulos, D.; Spear, J.; Caporaso, G.; Blekhman, R.; Knight, R.; et al. BugBase predicts organism level microbiome phenotypes. bioRxiv 2017, 133462. [CrossRef] 38. R Core Team. A language and environment for statistical computing. R Foundation for Statistical Computing. Available online: http://www.R-project.org/ (accessed on 1 February 2019). Sample Availability: Samples of the compounds BBGs are available from the authors. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Molecules Multidisciplinary Digital Publishing Institute http://www.deepdyve.com/lp/multidisciplinary-digital-publishing-institute/in-vitro-infant-faecal-fermentation-of-low-viscosity-barley-glucan-and-B0WehtfhP4

Loading next page...

References (41)

A. Konopka (2009)
What is microbial community ecology?
The ISME Journal, 3
Lei Chen, Bo-Bo Zhang, Jialun Chen, P. Cheung (2014)
Cell wall structure of mushroom sclerotium (Pleurotus tuber-regium): Part 2. Fine structure of a novel alkali-soluble hyper-branched cell wall polysaccharide
Food Hydrocolloids, 38
Fengmei Zhu, B. Du, Baojun Xu (2016)
A critical review on production and industrial applications of beta-glucans
Food Hydrocolloids, 52
J. Caporaso, Justin Kuczynski, Jesse Stombaugh, K. Bittinger, F. Bushman, E. Costello, N. Fierer, Antonio Peña, Julia Goodrich, J. Gordon, G. Huttley, S. Kelley, D. Knights, J. Koenig, R. Ley, C. Lozupone, Daniel McDonald, B. Muegge, M. Pirrung, Jens Reeder, Joel Sevinsky, P. Turnbaugh, William Walters, Jeremy Widmann, Tanya Yatsunenko, Jesse Zaneveld, R. Knight (2010)
QIIME allows analysis of high-throughput community sequencing data
Nature Methods, 7
Xiaojie Li, Jiajing Zhou, Chaoran Liu, Qirong Xiong, Hongwei Duan, P. Cheung (2017)
Stable and Biocompatible Mushroom β-Glucan Modified Gold Nanorods for Cancer Photothermal Therapy.
Journal of agricultural and food chemistry, 65 43
Kassem Makki, E. Deehan, J. Walter, F. Bäckhed (2018)
The Impact of Dietary Fiber on Gut Microbiota in Host Health and Disease.
Cell host & microbe, 23 6
Grain: World Markets and Trade (December report by United States Department of Agriculture Foreign Agricultural Service)
B. Williams, D. Mikkelsen, L. Paih, M. Gidley (2011)
In vitro fermentation kinetics and end-products of cereal arabinoxylans and (1,3;1,4)-β-glucans by porcine faeces
Journal of Cereal Science, 53
K. Korpela (2018)
Diet, Microbiota, and Metabolic Health: Trade-Off Between Saccharolytic and Proteolytic Fermentation.
Annual review of food science and technology, 9
Tonya Ward, Jake Larson, Jeremy Meulemans, Benjamin Hillmann, Joshua Lynch, Dimitri Sidiropoulos, J. Spear, G. Caporaso, R. Blekhman, R. Knight, R. Fink, D. Knights (2017)
BugBase predicts organism-level microbiome phenotypes
bioRxiv
C. Milani, A. Hevia, Elena Foroni, S. Duranti, F. Turroni, G. Lugli, B. Sánchez, R. Martín, M. Gueimonde, D. Sinderen, A. Margolles, M. Ventura (2013)
Assessing the Fecal Microbiota: An Optimized Ion Torrent 16S rRNA Gene-Based Analysis Protocol
PLoS ONE, 8
Hwa Cho, W. Jang, S. Moon, Jong Lee, Jang-Ho Kim, Hyonsob Han, K. Kim, Bong-Joo Lee, I. Kong (2018)
Immobilization of β-1,3-1,4-glucanase from Bacillus sp. on porous silica for production of β-glucooligosaccharides.
Enzyme and microbial technology, 110
Chaoran Liu, Jialun Chen, Lei Chen, Xuesong Huang, P. Cheung (2016)
Immunomodulatory Activity of Polysaccharide-Protein Complex from the Mushroom Sclerotia of Polyporus rhinocerus in Murine Macrophages.
Journal of agricultural and food chemistry, 64 16
M. Do, Eunjung Lee, Mi-Jin Oh, Yoonsook Kim, Ho‐Young Park (2018)
High-Glucose or -Fructose Diet Cause Changes of the Gut Microbiota and Metabolic Disorders in Mice without Body Weight Change
Nutrients, 10
Chaoran Liu, Man Choi, Xiaojie Li, P. Cheung (2018)
Immunomodulatory effect of structurally-characterized mushroom sclerotial polysaccharides isolated from Polyporus rhinocerus on human monoctyes THP-1
Journal of Functional Foods, 41
Jinyang Zhao, P. Cheung (2011)
Fermentation of β-glucans derived from different sources by bifidobacteria: evaluation of their bifidogenic effect.
Journal of agricultural and food chemistry, 59 11
(2006)
Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB
Appl. Environ. Microbiol., 72
J. Seidel, S. Haller, T. Eckmanns, T. Harder (2018)
Routine screening for colonization by Gram-negative bacteria in neonates at intensive care units for the prediction of sepsis: systematic review and meta-analysis.
The Journal of hospital infection, 99 4
L. Johansson, L. Virkki, H. Anttila, Heli Esselström, P. Tuomainen, T. Sontag-Strohm (2006)
Hydrolysis of -glucan
Food Chemistry
K. Tamura, G. Hemsworth, Guillaume Déjean, Theresa Rogers, Nicholas Pudlo, Karthik Urs, N. Jain, G. Davies, E. Martens, H. Brumer (2017)
Molecular Mechanism by which Prominent Human Gut Bacteroidetes Utilize Mixed-Linkage Beta-Glucans, Major Health-Promoting Cereal Polysaccharides.
Cell reports, 21 2
Y. Ishimoto, K. Ishibashi, D. Yamanaka, Y. Adachi, K. Kanzaki, Y. Iwakura, N. Ohno (2018)
Production of low-molecular weight soluble yeast β-glucan by an acid degradation method.
International journal of biological macromolecules, 107 Pt B
Ji-lin Dong, Xiao Yu, Liang-Er Dong, Ruiling Shen (2017)
In vitro fermentation of oat β-glucan and hydrolysates by fecal microbiota and selected probiotic strains.
Journal of the science of food and agriculture, 97 12
A language and environment for statistical computing. R Foundation for Statistical Computing
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution
K. Lam, Hoi-Yee Keung, Kin-Chun Ko, H. Kwan, P. Cheung (2017)
In vitro fermentation of beta-glucans and other selected carbohydrates by infant fecal inoculum: An evaluation of their potential as prebiotics in infant formula
Bioactive Carbohydrates and Dietary Fibre, 14
S. Vogt, J. Peña-Díaz, B. Finlay (2015)
Chemical communication in the gut: Effects of microbiota-generated metabolites on gastrointestinal bacterial pathogens.
Anaerobe, 34
H. Kim, P. White (2011)
Optimizing the molecular weight of oat β-glucan for in vitro bile acid binding and fermentation.
Journal of agricultural and food chemistry, 59 18
D. Ríos-Covian, P. Ruas-Madiedo, A. Margolles, M. Gueimonde, C. Reyes-Gavilán, N. Salazar (2016)
Intestinal Short Chain Fatty Acids and their Link with Diet and Human Health
Frontiers in Microbiology, 7
C. Stewart, N. Ajami, Jacqueline O’Brien, D. Hutchinson, Daniel Smith, M. Wong, M. Ross, R. Lloyd, H. Doddapaneni, G. Metcalf, D. Muzny, R. Gibbs, T. Vatanen, C. Huttenhower, R. Xavier, M. Rewers, W. Hagopian, J. Toppari, A. Ziegler, J. She, B. Akolkar, Å. Lernmark, H. Hyoty, K. Vehik, J. Krischer, J. Petrosino (2018)
Temporal development of the gut microbiome in early childhood from the TEDDY study
Nature, 562
K. Lam, P. Cheung (2013)
Non-digestible long chain beta-glucans as novel prebiotics
Bioactive Carbohydrates and Dietary Fibre, 2
M. Mikkelsen, M. Jensen, T. Nielsen (2017)
Barley beta-glucans varying in molecular mass and oligomer structure affect cecal fermentation and microbial composition but not blood lipid profiles in hypercholesterolemic rats.
Food & function, 8 12
E. Morris, T. Caruso, F. Buscot, M. Fischer, C. Hancock, Tanja Maier, T. Meiners, C. Müller, E. Obermaier, D. Prati, S. Socher, I. Sonnemann, Nicole Wäschke, T. Wubet, Susanne Wurst, M. Rillig (2014)
Choosing and using diversity indices: insights for ecological applications from the German Biodiversity Exploratories
Ecology and Evolution, 4
Simon Hughes, P. Shewry, G. Gibson, B. McCleary, R. Rastall (2008)
In vitro fermentation of oat and barley derived beta-glucans by human faecal microbiota.
FEMS microbiology ecology, 64 3
Donovan Parks, G. Tyson, P. Hugenholtz, R. Beiko (2014)
STAMP: statistical analysis of taxonomic and functional profiles
Bioinformatics, 30 21
T. DeSantis, P. Hugenholtz, N. Larsen, M. Rojas, Eoin Brodie, K. Keller, T. Huber, D. Dalevi, P. Hu, G. Andersen (2006)
Greengenes: Chimera-checked 16S rRNA gene database and workbench compatible in ARB
Lawrence Berkeley National Laboratory
report by United States Department of Agriculture Foreign Agricultural Service)
N. Segata, J. Izard, L. Waldron, D. Gevers, L. Miropolsky, W. Garrett, C. Huttenhower (2011)
Metagenomic biomarker discovery and explanation
Genome Biology, 12
Y. Tsai, B. Olson (1991)
Rapid method for direct extraction of DNA from soil and sediments
Applied and Environmental Microbiology, 57
M. Jayachandran, Jiali Chen, S. Chung, Baojun Xu (2018)
A critical review on the impacts of β-glucans on gut microbiota and human health.
The Journal of nutritional biochemistry, 61
Yuqin Shi, Jun Liu, Q. Yan, X. You, Shaoqing Yang, Zhengqiang Jiang (2018)
In vitro digestibility and prebiotic potential of curdlan (1 → 3)-β-d-glucan oligosaccharides in Lactobacillus species.
Carbohydrate polymers, 188
M. Langille, Jesse Zaneveld, J. Caporaso, Daniel McDonald, D. Knights, Joshua Reyes, J. Clemente, D. Burkepile, Rebecca Thurber, R. Knight, R. Beiko, C. Huttenhower (2013)
Predictive functional profiling of microbial communities using 16S rRNA marker gene sequences
Nature biotechnology, 31

Publisher: Multidisciplinary Digital Publishing Institute
Copyright: © 1996-2019 MDPI (Basel, Switzerland) unless otherwise stated
ISSN: 1420-3049
DOI: 10.3390/molecules24050828
Publisher site: See Article on Publisher Site

Abstract

molecules Article In Vitro Infant Faecal Fermentation of Low Viscosity Barley -Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics 1 1 1 1 1 2 Ka-Lung Lam , Kin-Chun Ko , Xiaojie Li , Xinxin Ke , Wai-Yin Cheng , Tianfeng Chen , 3 1 1 , Lijun You , Hoi-Shan Kwan and Peter Chi-Keung Cheung * Food and Nutritional Sciences, School of Life Sciences, The Chinese University of Hong Kong, Shatin, Hong Kong; [email protected] (K.-L.L.); [email protected] (K.-C.K.); [email protected] (X.L.); [email protected] (X.K.); [email protected] (W.-Y.C.); [email protected] (H.-S.K.) Department of Chemistry, Jinan University, Guangzhou 510632, China; [email protected] School of Food Science and Engineering, South China University of Technology, Guangzhou 510640, China; [email protected] * Correspondence: [email protected]; Tel.: +852-39436144 Academic Editor: Chi-Fai Chau Received: 1 February 2019; Accepted: 20 February 2019; Published: 26 February 2019 Abstract: Barley contains high level of -1,3-1,4-glucans (BBGs) which can be fermented by microbes and are a potential prebiotic. In the present study, native BBG with low viscosity and a MW of 319 kDa was depolymerized by acid hydrolysis to produce a series of four structurally characterized fragments with MWs ranging from 6–104 kDa. In vitro fermentation of these BBG samples by infant faecal microbiome was evaluated using a validated deep-well plate protocol as parallel miniature bioreactors. Microbial taxa were identiﬁed using 16S amplicon sequencing after 40 h of anaerobic fermentation. Bioinformatics analysis including diversity indexes, predicted metagenomic KEGG functions and predicted phenotypes were performed on the sequenced data. Short chain fatty acids and dissolved ammonia were quantiﬁed and the SCFAs/NH ratio was used to evaluate the eubiosis/dysbiosis potential. Correlation analysis showed that most of the parameters investigated showed a parabolic function instead of a monotonous function with the BBG samples having different MWs. Among the ﬁve BBGs, it was concluded that BBG with an intermediate MW of 28 kDa is the most promising candidate to be developed as a novel prebiotic. Keywords: acid hydrolysis; depolymerization; infant faecal microbiome; short chain fatty acids; ammonia; SCFAs/NH 1. Introduction Oat and barley are widely consumed worldwide in different forms, including direct consumption, ground powders, fermented beverages and so on, with the world oat consumption being 22,777 thousand metric tons and that of barley, 142,581 thousand metric tons in December 2018 [1]. -1,3-1,4-Glucans (BBGs) are naturally present in oat and barley as hemi-cellulose and they are now being examined as a functional food with prebiotic potential [2]. However, -glucans differ from well-known prebiotics such as inulin, FOS, GOS due to their high intrinsic physico-chemical property heterogeneity, including differences in molecular weight (MW), glycosidic linkages, degree and length of branching, water solubility, etc. [3]. The MW of -glucans is one of the important attributes towards their biological function as it affects solubility, viscosity, and enzymatic breakdown efﬁciency [4]. Currently, high purity (>95%) commercial native -1,3-1,4-glucans with different MW from barley (BBGs) and oat are available from Megazyme (County Wicklow, Ireland). For example, a Molecules 2019, 24, 828; doi:10.3390/molecules24050828 www.mdpi.com/journal/molecules Molecules 2019, 24, 828 2 of 20 total of ﬁve different MW -1,3-1,4-glucans have been used for investigation of their effects on in vitro fermentation using human faecal microbiota [5] and three BBGs of different MW used in in vitro fermentation with porcine faeces [6]. There is also another publication that employed -1,3-1,4-glucans extracted from barley with different molecular weights but with a lower purity [7]. In this study, we employed low viscosity BBG from Megazyme and its acid hydrolyzed derivatives, with reported purity >95%. In the present study, we aimed at evaluating the enhancement of the in vitro infant faecal fermentation of BBG with lower MW than the native ones produced by controlled acid hydrolysis. 16S amplicon proﬁling was determined to assess the microbial composition/diversity. Human infant faecal samples were chosen due to the high viscosity and texture similarity when in a slurry form with in vitro culture medium. Microbial fermentation is catalytically multi-fold, including saccharolytic, proteolytic, lipolytic types and so on [8]. In this study, BBGs of different MW were added as sole carbon sources to the culture medium, commonly containing partially digested proteins. Therefore, saccharolytic and proteolytic fermentation are mainly involved in this investigation. The effect of polymeric -glucans and its monomer glucose on microbial proﬁle was also compared and the correlations and insights from the phenotypic data, microbial proﬁle and MWs of BGGs were discussed. 2. Results 2.1. Chemical Composition and Structual Characterization of BBG Samples All BBG samples were predominantly carbohydrate in nature (97–98% by weight), with glucose as the only monosaccharide and contained only trace amounts of protein (less than 1% by weight) without signiﬁcant difference (Table 1). The monosaccharide proﬁle of native BBG and depolymerized BBG_0.2(2) shown in Figure S1A indicated that the only sugars found were glucose and allose (internal standard). The FT-IR spectra (Figure S1B) and the proton NMR spectra (Figure S1C) conﬁrmed the chemical structure of BBG and BBG_0.2(2) indicated that acid hydrolysis did not introduce any structural changes to the functional groups of the BBG derivatives. Increasing acid concentration and time of hydrolysis reduced the MW of BBG derivatives as shown in Table 1. Table 1. Preparation and structural characteristics of barley -glucans (BBG) and its acid hydrolyzed derivatives. Samples * Carbohydrate (% by wt) Protein (% by wt) Molecular Weight (kDa) BBG 98.0 0.7 0.2 0.5 319 BBG_0.05 98.2 0.3 0.2 0.7 104 BBG_0.1 97.6 0.3 0.2 0.1 28 BBG_0.2 98.1 0.5 0.2 0.3 11 BBG_0.2(2) 97.9 0.4 0.2 0.1 6 * The preparation conditions were: BBG: original, no hydrolysis; BBG_0.05: 0.05M H SO , 10 min at 80 C; BBG_0.1: 2 4 0.10M H SO , 60 min at 80 C; BBG_0.2: 0.20M H SO , 60 min at 80 C; BBG_0.2(2): 0.20M H SO , 120 min at 80 C. 2 4 2 4 2 4 All depolymerized BBG derivatives had a MW higher than the glucose monomer with BBG_0.2(2) having the lowest MW of 6 kDa as shown in Figure 1. Linkage analysis of methylated BBG samples by GC-MS indicated the presence of three major partially methylated alditol acetates (PMAAs) including 2,3,4,6-Me -Glc (terminal glucose), 2,4,6-Me -Glc (1,3-linked glucose) and 2,3,6-Me -Glc (1,4-linked 4 3 3 glucose) (Table 2). These linkage results were consistent with the typical mixed-1,3-1,4 linkages found in BBG. Based on the molar ratios of these three PMAAs, the change in the ratio of 1,3- and 1,4-glycosidic linkages before and after various acid hydrolysis was approximately 1:6, suggesting that the acid hydrolysis has no preference for -1,3 or -1,4 linkages (Table 2). In conclusion, controlled partial acid hydrolysis is an effective means of MW reduction in BBG. Molecules 2019, 24, 828 3 of 20 Molecules 2019, 24, x FOR PEER REVIEW 3 of 20 Figure 1. SEC chromatograms showing the MW distribution of BBGs and its depolymerized derivatives. Figure 1. SEC chromatograms showing the MW distribution of BBGs and its depolymerized Glucose was used as a monomer control. derivatives. Glucose was used as a monomer control. Table 2. GC-MS linkage analysis of barley -glucan (BBG) and its acid hydrolyzed derivatives. Table 2. GC-MS linkage analysis of barley β-glucan (BBG) and its acid hydrolyzed derivatives. Peak Area (%) c c Relative Molar Ratio Relative Molar Ratio Samples a c c Peak Area (%) Rela T-Glc tivep Mo :1,3-Glc lar Ra p:1,4-Glc tio p T Re -Glc lative p:1,3-Glc Mola p:1,4-Glc r Ratio p T-Glcp 1,3-Glcp 1,4-Glcp Samples T-Glcp 1,3-Glcp 1,4-Glcp T-Glcp:1,3-Glcp:1,4-Glcp T-Glcp:1,3-Glcp:1,4-Glcp BBG 3.99 1.00 14.25 0.15 81.76 0.85 1:3.57:20.49 0.28:1:5.74 BBG 3.99 ± 1.00 14.25 ± 0.15 81.76 ± 0.85 1:3.57:20.49 0.28:1:5.74 BBG_0.05 3.76 0.94 14.15 0.14 82.09 0.80 1:3.77:21.84 0.27:1:5.80 BBG_0.05 3.76 ± 0.94 14.15 ± 0.14 82.09 ± 0.80 1:3.77:21.84 0.27:1:5.80 BBG_0.1 13.22 2.99 13.11 0.45 73.68 2.54 1:0.99:5.58 1.01:1:5.62 BBG_0.1 13.22 ± 2.99 13.11 ± 0.45 73.68 ± 2.54 1:0.99:5.58 1.01:1:5.62 BBG_0.2 22.46 4.55 9.94 0.58 67.60 3.97 1:0.44:3.01 2.26:1:6.80 BBG_0.2 22.46 ± 4.55 9.94 ± 0.58 67.60 ± 3.97 1:0.44:3.01 2.26:1:6.80 BBG_0.2(2) 32.28 5.73 9.24 0.78 58.48 4.94 1:0.29:1.81 3.49:1:6.33 BBG_0.2(2) 32.28 ± 5.73 9.24 ± 0.78 58.48 ± 4.94 1:0.29:1.81 3.49:1:6.33 a b c a: calculated from the peak area in the GC-MS total ion chromatograms. : Glcp, glucopyranose. b : molar ratio was c : calculated from the peak area in the GC-MS total ion chromatograms. : Glcp, glucopyranose. : calculated from the ratio of the (peak area)/(MW) of individual PMAAs. molar ratio was calculated from the ratio of the (peak area)/(MW) of individual PMAAs. 2.2. BBG In Vitro Fermentation Using Infant Faecal Inocula 2.2. BBG In Vitro Fermentation Using Infant Faecal Inocula 2.2.1. Fermentation of BBG Samples Evaluated by Total Bacterial Count and Total DNA Extracted 2.2.1. Fermentation of BBG Samples Evaluated by Total Bacterial Count and Total DNA Extracted Results of the bacterial total plate count (TPC) in terms of colony formation units (CFUs) and Results of the bacterial total plate count (TPC) in terms of colony formation units (CFUs) and the the fold of population increase relative to glucose are shown in Table 3 and a representative plate fold of population increase relative to glucose are shown in Table 3 and a representative plate of each of each of the ﬁve BBG samples and glucose monomer is shown in Figure S2. In the present results, of the five BBG samples and glucose monomer is shown in Figure S2. In the present results, the the concentration of the extracted microbial DNA was accessed by Nano-drop (Thermo Scientiﬁc, concentration of the extracted microbial DNA was accessed by Nano-drop (Thermo Scientific, Waltham, MA, USA) and the fold changes of the ﬁve BBG samples relative to glucose are shown in Waltham, MA, USA) and the fold changes of the five BBG samples relative to glucose are shown in Table 3. All BBG samples supported the growth of more microbes than glucose, which was probably Table 3. All BBG samples supported the growth of more microbes than glucose, which was probably due to the larger amount of monomeric glucose produced after enzymatic hydrolysis of the BGG due to the larger amount of monomeric glucose produced after enzymatic hydrolysis of the BGG samples on molar basis. TPC results indicated that BBG had the highest CFUs (3.7-fold increase), samples on molar basis. TPC results indicated that BBG had the highest CFUs (3.7-fold increase), followed by BBG_0.2 (2.9-fold), BBG_0.1 (2.3-fold), BBG_0.2(2) (2.2-fold), and BBG_0.05 (1.0-fold) when followed by BBG_0.2 (2.9-fold), BBG_0.1 (2.3-fold), BBG_0.2(2) (2.2-fold), and BBG_0.05 (1.0-fold) compared to glucose. Total DNA results indicated that the increase of biomass was in a descending when compared to glucose. Total DNA results indicated that the increase of biomass was in a order of: BBG_0.1 (4.2-fold), BBG_0.05 (3.9-fold), BBG_0.2 (3.6-fold), BBG (3.4-fold) and BBG_0.2(2) descending order of: BBG_0.1 (4.2-fold), BBG_0.05 (3.9-fold), BBG_0.2 (3.6-fold), BBG (3.4-fold) and (2.8-fold). There seemed to have a nonlinear relationship of increase in microbial populations with BBG_0.2(2) (2.8-fold). There seemed to have a nonlinear relationship of increase in microbial a decrease in MW of the BGG samples. The discrepancy between the results of TPC and total DNA populations with a decrease in MW of the BGG samples. The discrepancy between the results of TPC might be explained by the fact the TPC is the ﬁnal outcome of complex interactions of all microbes and total DNA might be explained by the fact the TPC is the final outcome of complex interactions including self/mutual promotion and/or inhibition which might not only depend on the MW of the of all microbes including self/mutual promotion and/or inhibition which might not only depend on carbon source. the MW of the carbon source. Molecules 2019, 24, 828 4 of 20 Table 3. Bacterial total plate count (CFUs) and concentration of total DNA extracted from culture broth Molecules 2019, 24, x FOR PEER REVIEW 4 of 20 of the infant faecal fermentation of ﬁve BBG samples, glucose monomer and time-0. Total Plate Ratio of CFUS Ratio of [DNA] Table 3. Bacterial total plate count (CFUs) and concentration of total DNA extracted from culture Samples [DNA] (g/mL) * Count (CFU) (Relative to Glucose) (Relative to Glucose) broth of the infant faecal fermentation of five BBG samples, glucose monomer and time-0. a,b BBG 8.7 104 3.7 3.4 67.3 7.4 Total Plate Ratio of CFUS Ratio of [DNA] Samples [DNA] (µg/mL) * c,d BBG_0.05 2.3 104 1.0 76.4 4.0 3.9 Count (CFU) (Relative to Glucose) (Relative to Glucose) a,b a,e,f BBG 8.7 × 104 3.7 67.3 ± 7.4 3.4 BBG_0.1 5.4 104 2.3 83.5 9.1 4.2 c,d BBG_0.05 2.3 × 104 1.0 76.4 ± 4.0 3.9 g,h BBG_0.2 6.8 104 2.9 3.6 71.9 4.1 a,e,f BBG_0.1 5.4 × 104 2.3 83.5 ± 9.1 4.2 c,g,i BBG_0.2(2) 5.2 104 2.2 54.8 0.6 2.8 g,h BBG_0.2 6.8 × 104 2.9 71.9 ± 4.1 3.6 b,d,e,f,h,i c,g,i BBG_0.2(2) GLC 5.2 × 2.4 104 104 2.2 1 19.7 54.8 ± 2.4 0.6 2.8 1 b,d,e,f,h,i GLC 2.4 × 104 1 19.7 ± 2.4 1 T0 6.4 103 0.3 4.3 0.2 0.2 T0 6.4 × 103 0.3 4.3 ± 0.2 0.2 a–i * Mean values with different superscripts ( ) represent signiﬁcant difference by one-way ANOVA (Tukey HSD a–i * Mean values with different superscripts ( ) represent significant difference by one-way ANOVA post-hoc test), p < 0.05. : Time-0, T0 group is statistically signiﬁcant different compared to all other groups. (Tukey HSD post-hoc test), p < 0.05. : Time-0, T0 group is statistically significant different compared to all other groups. Compared to TPC, the amount of total DNA extracted in a microbial fermentation could better reﬂect the total bacterial biomass supported by the carbon source since it is not selective and Compared to TPC, the amount of total DNA extracted in a microbial fermentation could better medium-independent reflect the total bacterial biomass suppo [9]. Therefore, DNA rted by the c concentration arbon source was since it is not se selected to repr lective esent and m total edium bacteria - independent [9]. Therefore, DNA concentration was selected to represent total bacteria count of count of samples in the following analysis. samples in the following analysis. 2.2.2. Short Chain Fatty Acids and Dissolved Ammonia Content after In Vitro Fermentation 2.2.2. Short Chain Fatty Acids and Dissolved Ammonia Content after In Vitro Fermentation The highest level of short chain fatty acids (SCFAs) produced was found in BBG_0.2 which The highest level of short chain fatty acids (SCFAs) produced was found in BBG_0.2 which was was ﬁve times that of glucose as shown in Figure 2A. Glucose only generated acetic acid and its five times that of glucose as shown in Figure 2A. Glucose only generated acetic acid and its concentration was lower than all the ﬁve BBGs. This might be explained by the fact that infant faecal concentration was lower than all the five BBGs. This might be explained by the fact that infant faecal inoculum contained a myriad of bacteria that not only converted glucose into SCFAs, but also other inoculum contained a myriad of bacteria that not only converted glucose into SCFAs, but also other metabolites. Acetic acid was the major SCFAs produced by all BBG samples, followed by propionic metabolites. Acetic acid was the major SCFAs produced by all BBG samples, followed by propionic acid and then butyric acid. The distribution of SCFAs produced by different BBG samples also varied. acid and then butyric acid. The distribution of SCFAs produced by different BBG samples also varied. The highest acetic acid production was found in BBG_0.1 and BBG_0.2, while BBG_0.2(2) having the The highest acetic acid production was found in BBG_0.1 and BBG_0.2, while BBG_0.2(2) having the lowest MW had the highest concentration of propionic acid (Figure 2A). lowest MW had the highest concentration of propionic acid (Figure 2A). Figure 2. (A) Short chain fatty acid proﬁles and (B) ammonia concentration of the ﬁve BBG samples Figure 2. (A) Short chain fatty acid profiles and (B) ammonia concentration of the five BBG samples a–h and glucose monomer. Different superscripts ( ) represent signiﬁcant difference by one-way ANOVA a–h and glucose monomer. Different superscripts ( ) represent significant difference by one-way (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically signiﬁcant different from all other groups. ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically significant different from all other groups. The highest ammonia production was obtained by BBG_0.05 samples with original BBG produced the least among the ﬁve BBGs (Figure 2B). All BBGs produced more ammonia compared to glucose. As The highest ammonia production was obtained by BBG_0.05 samples with original BBG shown in Figure 2A, glucose produced the highest level of propionic acid, which might explain its least produced the least among the five BBGs (Figure 2B). All BBGs produced more ammonia compared ammonia production as propionate ion was shown to be a growth inhibitor to pathogenic/spoilage to glucose. As shown in Figure 2A, glucose produced the highest level of propionic acid, which might microbes [10]. explain its least ammonia production as propionate ion was shown to be a growth inhibitor to pathogenic/spoilage microbes [10]. Molecules 2019, 24, 828 5 of 20 Molecules 2019, 24, x FOR PEER REVIEW 5 of 20 2.3. Microbiome Proﬁle Changes after BBG Fermentation 2.3. Microbiome Profile Changes after BBG Fermentation 2.3.1. Change of Infant Faecal Microbiome Proﬁle by Fermentation of BBG Samples 2.3.1. Change of Infant Faecal Microbiome Profile by Fermentation of BBG Samples The microbial taxonomy proﬁles of the ﬁve BBG samples, glucose and time-0 (T0) are shown The microbial taxonomy profiles of the five BBG samples, glucose and time-0 (T0) are shown in in Figure 3A with abundance data scaled according to sample total DNA concentration, reﬂecting a Figure 3A with abundance data scaled according to sample total DNA concentration, reflecting a quasi-absolute taxa amount. quasi-absolute taxa amount. Figure 3. 16S amplicon sequencing results of fermentation of BBG samples by infant faecal inoculum. Figure 3. 16S amplicon sequencing results of fermentation of BBG samples by infant faecal inoculum. (A) Bar plot of the major identiﬁed bacterial taxonomy scaled to the total extracted DNA concentration, (A) Bar plot of the major identified bacterial taxonomy scaled to the total extracted DNA noted as quasi-absolute bacterial amount; (B) Enlarged view of the relative abundance of the taxon concentration, noted as quasi-absolute bacterial amount; (B) Enlarged view of the relative abundance >1% of T0 group at the beginning of fermentation; (C) Enlarged view of the relative abundance of the of the taxon >1% of T0 group at the beginning of fermentation; (C) Enlarged view of the relative taxon >1% of the ﬁve BBG groups and glucose monomer, Glc; (D) Venn diagram showing the similar abundance of the taxon >1% of the five BBG groups and glucose monomer, Glc; (D) Venn diagram and different number of bacteria taxonomy identiﬁed; (E) -diversity analysis showing the principle showing the similar and different number of bacteria taxonomy identified; (E) β-diversity analysis component showing the principle analysis of the com ﬁve ponBBGs ent analy and sis glucose of the fi monomer ve BBGs and g . lucose monomer. It was clearly shown that T0 group had the least taxa abundance in the beginning. Moreover, the It was clearly shown that T0 group had the least taxa abundance in the beginning. Moreover, the biological triplicates among each group gave similar microbiome proﬁles, demonstrating a sample biological triplicates among each group gave similar microbiome profiles, demonstrating a sample Molecules 2019, 24, 828 6 of 20 size of three could be sufﬁcient for gaining some insights. Figure 3B showed the relative abundance of T0 group with distinct taxa (average relative taxa abundance >1%) compared to the BBGs samples. Taxa unique to T0 group were noted with "*" as shown in the ﬁgure legend, namely, Prevotella sp., Lachnospira sp., Bacillus sp., other genus in Lachnospiraceae family, Ruminococcus sp., Parabacteroides sp. and Streptococcus sp. These taxa abundance contributed less than 1% after BBGs/glucose fermentation. Figure 3C shows that all the BBGs shared similar microbial taxa distribution while glucose monomer though had similar taxa but with a distinctive distribution. The Venn diagram of the microbiome proﬁles of the ﬁve BBG samples indicates that they all shared 57 core microbial taxa and some group-speciﬁc unique taxa (Figure 3D). The number of group-speciﬁc taxa ranged from 1 to 6 and these speciﬁc taxa (family/genus) were Actinomycetales, Oscillospira, and Peptoniphilus for BBG; Bacilli for BBG_0.05; Cellulomonas, Aerococcus and Leuconostoc for BBG_0.1; Rothia, Pediococcus, Roseburia, Eubacterium, Burkholderiaceae, and Oceanospirillaceae for BBG_0.2; Anoxybacillus and Weissella for BBG_0.2(2). These speciﬁc taxa are highlighted in Table S1. All these unique taxa were composed of less than 0.001% of total taxa, implying their minor role in BBG metabolism and SCFAs/NH production. -Diversity, depicted in Figure 3E, shows that all the BBG samples were clustered as a group with glucose monomer as a distinctive group and T0 as another distinct group. Figure 4 shows the results from BugBase with the upper panel being the predicted relative abundance of aerobes, anaerobes and facultative anaerobes. The relative abundance of aerobes of 5 BBGs were similar to T0 but not the glucose group. As for the predicted anaerobe proportion, T0 group contained the highest abundance. Since BBGs had similar proportion of aerobes with T0 but different proportion of anaerobes with T0, the difference should be contributed by the proportion of facultative anaerobes. As shown, the ﬁve BBGs had various proportions of facultative anaerobes with BBG_0.2 having the highest and BBG the least. T0 group had the least proportion of facultative anaerobes among all. The lower panel of Figure 4 refers to the predicted Gram-positive, Gram-negative and proportion of taxa capable of forming bioﬁlms. For the ﬁve BBGs, a reduction of proportion of Gram-positive taxa and an increase of proportion of Gram-negative taxa compared with glucose was observed. The predicted phenotype of T0 group was of less interest as shown in Figure 3A, while the quasi-absolute number of all bacteria was very low, thus making the comparison of proportion less biologically meaningful. But interesting phenomena could still be observed in the ﬁve BBGs and glucose groups. Figure 5 shows the signiﬁcantly different phylum of the ﬁve BBGs and glucose with descending abundance: Firmicutes (Figure 5A), Proteobacteria (Figure 5D), Actinobacteria (Figure 5B) and Bacteroidetes (Figure 5C). Glucose could signiﬁcantly provide a higher growth support to Firmicutes and Actinobacteria while BBGs allowed enrichment of Bacteroidetes (well-known for complex polysaccharide breakdown such -glucans [11]) and Proteobacteria. Molecules Molecules 2019 2019,, 24 24,, x FO 828 R PEER REVIEW 7 of 7 of 20 20 Molecules 2019, 24, x FOR PEER REVIEW 7 of 20 Figure 4. Predicted output of selected phenotype by BugBase. Relative abundance of aerobes, Figure 4. Predicted output of selected phenotype by BugBase. Relative abundance of aerobes, anaerobes Figure 4. Predicted output of selected phenotype by BugBase. Relative abundance of aerobes, anaerobes and facultative anaerobes are in the upper panel. Gram-positive bacteria, Gram-negative and facultative anaerobes are in the upper panel. Gram-positive bacteria, Gram-negative bacteria and anaerobes and facultative anaerobes are in the upper panel. Gram-positive bacteria, Gram-negative bacteria and taxa that are predicted to form biofilm are given in the lower panel. Significant difference taxa that are predicted to form bioﬁlm are given in the lower panel. Signiﬁcant difference analyzed by bacteria and taxa that are predicted to form biofilm are given in the lower panel. Significant difference analyzed by one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically significant one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically signiﬁcant different from analyzed by one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. @: denotes statistically significant different from all other groups. all other groups. different from all other groups. Figure 5. Signiﬁcantly different taxa in phyla level identiﬁed using ANOVA (Tukey HSD post-hoc Figure 5. Significantly different taxa in phyla level identified using ANOVA (Tukey HSD post-hoc Figure 5. Significantly different taxa in phyla level identified using ANOVA (Tukey HSD post-hoc test) with Benjamini-Hochberg FDR multiple test correction. (A) Firmicutes; (B) Actinobacteria; test) with Benjamini-Hochberg FDR multiple test correction. (A) Firmicutes; (B) Actinobacteria; (C) test) with Benjamini-Hochberg FDR multiple test correction. (A) Firmicutes; (B) Actinobacteria; (C) (C) Bacteroidetes; and (D) Proteobacteria. These 4 major taxa accounted for 99.90 0.07% of total Bacteroidetes; and (D) Proteobacteria. These 4 major taxa accounted for 99.90 ± 0.07% of total taxa Bacteroidetes; and (D) Proteobacteria. These 4 major taxa accounted for 99.90 ± 0.07% of total taxa taxa identiﬁed. identified. identified. Figure 6 depicts some host beneﬁcial taxa that were identiﬁed in the ﬁve BBGs and glucose. Figure 6 depicts some host beneficial taxa that were identified in the five BBGs and glucose. Glucose Figu supported re 6 depicmor ts some host e probiotic benefic Lactobacillus ial taxa t sp.hwhile at were low identified MW BBGin out-performed the five BBGs and g high MW luco BBG se. Glucose supported more probiotic Lactobacillus sp. while low MW BBG out-performed high MW BBG Glucose supported more probiotic Lactobacillus sp. while low MW BBG out-performed high MW BBG (Figure 6A), a similar phenomenon was also observed in Lactococcus sp. (Figure 6C). Another probiotic (Figure 6A), a similar phenomenon was also observed in Lactococcus sp. (Figure 6C). Another Biﬁdobacterium (Figure 6A), a sp. sim was ilar ph morenomenon w e selectivelya enriched s also obser by glucose ved in than Lactococcus BBGs sp. but (F among igure the 6C ﬁve5 ). An BBGs, other probiotic Bifidobacterium sp. was more selectively enriched by glucose than BBGs but among the five5 probiotic Bifidobacterium sp. was more selectively enriched by glucose than BBGs but among the five5 intermediately MW such as BBG_0.1 was more favorable (Figure 6B). On the other hand, BBGs showed BBGs, intermediately MW such as BBG_0.1 was more favorable (Figure 6B). On the other hand, BBGs better BBGs, interm support ediately towards MW Veillonella such as BBG_0.1 w sp. (higher as more efﬁciency favor inapr ble opionate (Figure 6B pr). oduction On the ot [12 her ], hand Figur , e BBGs 6D), showed better support towards Veillonella sp (higher efficiency in propionate production [12], Figure showed better support towards Veillonella sp (higher efficiency in propionate production [12], Figure Enterobacteriaceae family (common enteric microbe in animal gut, Figure 6E), and Bacteroides (boom 6D), Enterobacteriaceae family (common enteric microbe in animal gut, Figure 6E), and Bacteroides 6D), Enterobacteriaceae family (common enteric microbe in animal gut, Figure 6E), and Bacteroides with diet rich in complex polysaccharides, Figure 6F). (boom with diet rich in complex polysaccharides, Figure 6F). (boom with diet rich in complex polysaccharides, Figure 6F). Molecules 2019, 24, 828 8 of 20 Molecules 2019, 24, x FOR PEER REVIEW 8 of 20 Molecules 2019, 24, x FOR PEER REVIEW 8 of 20 Figure Figure 6. 6. Signiﬁcantly Significantly d dif iffere ferent nt taxa in genus lev taxa in genus level el identiﬁed identified using ANO using ANOV VA A (Tukey HSD post-hoc (Tukey HSD post-hoc Figure 6. Significantly different taxa in genus level identified using ANOVA (Tukey HSD post-hoc test) test) with with Benj Benjamini-Hochber amini-Hochberg g FFDR DR m multiple ultiple tetest st correct correction. ion. (A) (A La ) ct Lactobacillus obacillus; (B;) (Bifido B) Biﬁdobacterium bacterium; (C) ; test) with Benjamini-Hochberg FDR multiple test correction. (A) Lactobacillus; (B) Bifidobacterium; (C) (Lactococcus C) Lactococcus ; (D ;) (D Veill ) Veillonella onella; (E ; ) (E E )n Enter teroba obacteriaceae cteriaceae; an ; and d (F () F)Ba Bacter cteroides oides . These taxa were considered . These taxa were considered Lactococcus; (D) Veillonella; (E) Enterobacteriaceae; and (F) Bacteroides. These taxa were considered beneﬁcial to the hosts. beneficial to the hosts. beneficial to the hosts. On the contrary, Figure 7 depicts some neutral/opportunistically pathogenic taxa. Among the On the contrary, Figure 7 depicts some neutral/ opportunistically pathogenic taxa. Among the selected, On the contrary, BBGs reduced Fig Bacillus ure 7 depict (Figur se some 7A), Faecalibacterium neutral/ opportunist (Figur ice ally 7B), pat Lachnospiraceae hogenic taxa. Among (Figure t 7h D) e selected, BBGs reduced Bacillus (Figure 7A), Faecalibacterium (Figure 7B), Lachnospiraceae (Figure 7D) and selecLeuconostocaceae ted, BBGs reduc(Figur ed Bacillus e 7E). (Fi However gure 7A , BBGs ), Faecalibacterium augmented (F the igu grr owth e 7B),of Lac Klebsiella hnospiraceae (Figur (Figu e 7C) re 7 with D) and Leuconostocaceae (Figure 7E). However, BBGs augmented the growth of Klebsiella (Figure 7C) with some and Leuconos of its species tocaceae r egar (Figur ded e 7E). Howev as emerging er, BBGs pathogens augm [13 ented ]. Figur the gr e 7owth F shows of Klebsiella Enterococcus (Figu as re 7 the C) w most ith some of its species regarded as emerging pathogens [13]. Figure 7F shows Enterococcus as the most abundant some of its species r taxa, commonly egarded as emerg found in animal ing pathogens [13]. F intestines. igure 7F shows Enterococcus as the most abundant taxa, commonly found in animal intestines. abundant taxa, commonly found in animal intestines. Figure 7. Significantly different taxa in genus level identified using ANOVA (Tukey HSD post-hoc Figure Figure 7. 7. Signiﬁcantly Significantly d dififffere erent nt taxa in genus lev taxa in genus level e identiﬁed l identified using ANO using ANOVAV (T A (Tukey HSD post-hoc ukey HSD post-hoc test) test) with Benjamini-Hochberg FDR multiple test correction. (A) Bacillus; (B) Faecalibacterium; (C) with test) w Benjamini-Hochber ith Benjamini-Hochberg FDR mu g FDR multiple test ltiple tes correction. t correctio (A) Bacillus n. (A) ; ( Ba B)ciFaecalibacterium llus; (B) Faecaliba ; (C cterium ) Klensiella ; (C) ; Klensiella; (D) Lachnospiraceae; (E) Leuconostocaceae; and (F) Enterococcus. These taxa were considered (D) Lachnospiraceae; (E) Leuconostocaceae; and (F) Enterococcus. These taxa were considered either neutral Klensiella; (D) Lachnospiraceae; (E) Leuconostocaceae; and (F) Enterococcus. These taxa were considered either neutral or opportunistic pathogenic to the hosts. or opportunistic pathogenic to the hosts. either neutral or opportunistic pathogenic to the hosts. Table 4 shows one phylogeny-based α-diversity index (PD whole tree) and 5 non-phylogeny- Table 4 shows one phylogeny-based -diversity index (PD whole tree) and 5 non-phylogeny-based Table 4 shows one phylogeny-based α-diversity index (PD whole tree) and 5 non-phylogeny- based indices including Fisher alpha, Berger-Parker d, Shannon, Simpson evenness, and Simpson indices including Fisher alpha, Berger-Parker d, Shannon, Simpson evenness, and Simpson reciprocal. based indices including Fisher alpha, Berger-Parker d, Shannon, Simpson evenness, and Simpson reciprocal. Briefly, BBG_0.1 showed significantly higher α-diversity than BBG for PD whole tree, Brieﬂy, BBG_0.1 showed signiﬁcantly higher -diversity than BBG for PD whole tree, Shannon, reciprocal. Briefly, BBG_0.1 showed significantly higher α-diversity than BBG for PD whole tree, Shannon, Simpson evenness, and Simpson reciprocal. While BBG_0.2(2) showed significantly higher Simpson evenness, and Simpson reciprocal. While BBG_0.2(2) showed signiﬁcantly higher value than Shannon, Simpson evenness, and Simpson reciprocal. While BBG_0.2(2) showed significantly higher value than BBG in terms of Fisher alpha, BBG had a significant higher value of Berger-Parker-d index. value than BBG in terms of Fisher alpha, BBG had a significant higher value of Berger-Parker-d index. Overall, BBG_0.1 was shown to have the highest increase in α-diversity. Overall, BBG_0.1 was shown to have the highest increase in α-diversity. Molecules 2019, 24, 828 9 of 20 BBG in terms of Fisher alpha, BBG had a signiﬁcant higher value of Berger-Parker-d index. Overall, BBG_0.1 was shown to have the highest increase in -diversity. Table 4. Rareﬁed -diversity indices of the infant faecal fermentation of 5 BBG samples and glucose monomer. Phylogeny-Based Non-Phylogeny-Based Samples Simpson PD Whole Tree Fisher Alpha Berger Parker D Shannon Simpson_E Reciprocal a,b a,b,c a,b,c,d,e a,b,c,d a a,b,c,d BBG 17.68 0.77 90.81 4.66 0.18 0.01 4.87 0.08 0.026 0.001 13.17 0.84 a,f a,e b a,e BBG_0.05 18.69 0.50 100.40 1.89 5.22 0.07 16.96 0.92 0.14 0.01 0.031 0.002 a a b,g b,f a,c b,f BBG_0.1 19.95 0.59 103.76 5.26 0.15 0.01 5.31 0.11 0.033 0.001 18.27 1.15 c,g c,g b c,h d BBG_0.2 19.41 0.95 105.66 2.98 0.13 0.01 5.21 0.04 0.031 0.000 17.69 0.51 b c d,i d d,h BBG_0.2(2) 19.85 0.53 107.50 5.94 0.15 0.01 5.12 0.08 0.028 0.004 16.14 1.50 e,f,g,h,i e,f,g b,c,d e,f,g,h GLC 19.34 0.81 101.11 5.89 0.22 0.01 4.93 0.02 0.023 0.002 12.63 0.25 a–i Different superscripts represent signiﬁcant difference by one-way ANOVA (Tukey HSD post-hoc test), p < 0.05. 2.3.2. Metagenome Prediction of KEGG Functional Annotation Kyoto Encylopedia of Genes and Genomes (KEGG) functional annotation was done using PICRUSt with standard analysis including 16S copy number normalization and whole metagenome prediction with raw data output are shown in Table S2. The NSTI scores of PICRUSt metagenome prediction for the BBG samples were: BBG (0.032); BBG_0.05 (0.032); BBG_0.1 (0.032); BBG_0.2 (0.033); BBG_0.2(2) (0.031); and that of Glucose (0.034) T0 (0.0589). An NSTI score of 0.03 indicates that the average microbe in your sample can be predicted using a relative from the same (97%) species [14]. All BBG samples had higher number of predicted genes than that of glucose with BBG_0.05 having the highest number (Table S2). “Metabolism" was the highest class of annotated genes in Level 1 as shown in Figure S4A, followed by "Environmental information processing" and “Genetic information processing”. Most genes annotated in “Metabolism” at Level 2 were related to “Carbohydrate metabolism”, followed by “Amino acid metabolism” and “Energy metabolism” with “Lipid metabolism” ranked the sixth as shown in Figure S4B. All BBGs had a higher proportion of predicted KEGG class of "Carbohydrate Metabolism" than glucose monomer and T0, reﬂecting extra machinery was required to breakdown and absorb -glucans. Since BBG metabolism, short chain fatty acid production and ammonia production were our targets of investigation, related KEGG classes were single out for more in-depth study using the Statistical Analysis of Metagenomic Proﬁles (STAMP) approach. Figure S5A shows the “Carbohydrate metabolism” KEGG class. All BBG samples had a signiﬁcant increase in the number of predicted genes in carbohydrate metabolism, compared to glucose. The ﬁve BBG groups also have higher proportion of predicted KEGG class ABC Transporter (Figure S5B), Bacterial Secretion System (Figure S5C) and Glycan biosynthesis and metabolism (Figure S5D). All the four predicted KEGG classes actually coordinated among themselves, such as secretion system to secrete extracellular enzymes, ABC transporters to uptake carbon/glycan with the expense of cellular energy and glycan metabolism to breakdown complex glycan or use the carbon source for bacterial glycan formation. Although “Fatty acid metabolism” does not only involve SCFAs, this class of keys could also be used to infer the potential in SCFAs production. Among the ﬁve BBG samples, BBG_0.1 had the highest number of predicted genes annotated to “Fatty acid metabolism”, followed by BBG_0.05, BBG_0.2, BBG_0.2(2) and the native BBG having the lowest number (Figure S6A). Glucose had a similar proportion of predicted genes in fatty acid metabolism with BBG_0.2. Not only with saccharolytic potential and lipolytic potential, we also extracted some ammonia production related KEGG classes. For example, BBGs stimulated less “Protein digestion and absorption” with medium MW BBG such as BBG_0.005 and BBG_0.1 being more favorable (Figure Molecules 2019, 24, 828 10 of 20 S6B). However, the BBG groups had a signiﬁcant higher relative abundance of KEGG classes “Amino acid metabolism” (Figure S6C) and "Nitrogen metabolism" (Figure S6D). 2.3.3. Microbial Biomarkers Identiﬁcation Molecules 2019, 24, x FOR PEER REVIEW 10 of 20 Microbial biomarkers were identiﬁed using the LEfSe. Table S3 shows the raw data output Microbial biomarkers were identified using the LEfSe. Table S3 shows the raw data output of of signiﬁcant biomarkers identiﬁed with the ﬁve BBG samples and glucose. Brieﬂy, glucose had significant biomarkers identified with the five BBG samples and glucose. Briefly, glucose had the the highest number of biomarkers compared to the ﬁve BBG samples, including Lactobacillus highest number of biomarkers compared to the five BBG samples, including Lactobacillus and and Biﬁdobacterium, while BBG_0.1 gave the highest number of biomarkers among the 5 BBGs, Bifidobacterium, while BBG_0.1 gave the highest number of biomarkers among the 5 BBGs, including including Proteobacteria and Bacteroides. Figure 8 shows the cladogram of biomarkers with Bacteroides, Proteobacteria and Bacteroides. Figure 8 shows the cladogram of biomarkers with Bacteroides, Parabacteroides, Bacillus, Enterococcus, Lactobacillus, Veillonella, Sutterella, and Klebsiella being the Parabacteroides, Bacillus, Enterococcus, Lactobacillus, Veillonella, Sutterella, and Klebsiella being the dominant ones. LEfSe analysis presented some statistically signiﬁcant biomarkers, from which we dominant ones. LEfSe analysis presented some statistically significant biomarkers, from which we choose some biologically signiﬁcant taxa for further comparative analysis. Among these biomarkers, choose some biologically significant taxa for further comparative analysis. Among these biomarkers, excluding the group-speciﬁc unique taxa in the Venn diagram analysis (Figure 3D), some interested excluding the group-specific unique taxa in the Venn diagram analysis (Figure 3D), some interested taxa were chosen for correlation analysis with BBG samples having different MWs. taxa were chosen for correlation analysis with BBG samples having different MWs. Figure 8. Cladogram of biomarkers identiﬁed in the infant faecal fermentation of 5 BBG samples Figure 8. Cladogram of biomarkers identified in the infant faecal fermentation of 5 BBG samples against glucose monomer using LEfSe. against glucose monomer using LEfSe. 3. Discussion 3. Discussion 3.1. Efﬁcacy of Controlled Acid Hydrolysis in MW Reduction 3.1. Efficacy of Controlled Acid Hydrolysis in MW Reduction One of the major focus of this investigation was to investigate the effects of MW reduction in One of the major focus of this investigation was to investigate the effects of MW reduction in BBG on microbial fermentation. We have showed that acid hydrolysis in a controlled manner could BBG on microbial fermentation. We have showed that acid hydrolysis in a controlled manner could produce BBG with reduced molecular weight (Figure 1, Tables 1 and 2) without the modiﬁcation of produce BBG with reduced molecular weight (Figure 1, Table 1,2) without the modification of functional groups or glycosidic linkages. Controlled acid hydrolysis was also employed for the MW functional groups or glycosidic linkages. Controlled acid hydrolysis was also employed for the MW reduction of other -glucans such as -13,16-glucans [15]. Acid hydrolysis of -1,3-1,4-glucan into reduction of other β-glucans such as β-13,16-glucans [15]. Acid hydrolysis of β-1,3-1,4-glucan into smaller fragments had also been conducted with different acids, with results indicated that H SO 2 4 smaller fragments had also been conducted with different acids, with results indicated that H2SO4 produced the least glucose monomer and the highest proportion of oligosaccharides than TFA while produced the least glucose monomer and the highest proportion of oligosaccharides than TFA while HCl produce only monomers at low concentration (0.1 or 0.05 M) [16]. Thus, H SO generates the 2 4 HCl produce only monomers at low concentration (0.1 or 0.05 M) [16]. Thus, H2SO4 generates the production of a wider spectrum of smaller fragments. As shown in Figure 1, a lateral transverse production of a wider spectrum of smaller fragments. As shown in Figure 1, a lateral transverse of peaks from high MW to low MW BBGS and then to glucose monomer was observed in the size exclusion chromatography (SEC). Although the peaks of the five BBGs only differed in MW, they all had similar peak height, skewness and kurtosis. Therefore, we speculate that acid hydrolysis is a ”gain-by-loss” method, i.e., lower MW samples were obtained by random hydrolysis with steric advantage from both ends. Furthermore, tiny fragments such as mono-, di-, tri-, tetramers, might be gradually released and removed by dialysis. Therefore, controlled acid hydrolysis might not be very Molecules 2019, 24, 828 11 of 20 of peaks from high MW to low MW BBGS and then to glucose monomer was observed in the size exclusion chromatography (SEC). Although the peaks of the ﬁve BBGs only differed in MW, they all had similar peak height, skewness and kurtosis. Therefore, we speculate that acid hydrolysis is a ”gain-by-loss” method, i.e., lower MW samples were obtained by random hydrolysis with steric advantage from both ends. Furthermore, tiny fragments such as mono-, di-, tri-, tetramers, might Molecules 2019, 24, x FOR PEER REVIEW 11 of 20 be gradually released and removed by dialysis. Therefore, controlled acid hydrolysis might not be very efﬁcient in producing very short -glucans fragments such as oligo- -glucans which could pass efficient in producing very short β-glucans fragments such as oligo-β-glucans which could pass through dialysis tubing or membrane easily. On the contrary, enzymatic degradation using glycosidic through dialysis tubing or membrane easily. On the contrary, enzymatic degradation using linkage/chain-length speciﬁc enzymes could selectively cut the polymeric chains into fragments glycosidic linkage/chain-length specific enzymes could selectively cut the polymeric chains into to produce oligomers instead of monomers [17,18]. Therefore, controlled acid hydrolysis serves an fragments to produce oligomers instead of monomers [17,18]. Therefore, controlled acid hydrolysis excellent purpose of reducing the MW of -glucans and other polysaccharides extracted from raw serves an excellent purpose of reducing the MW of β-glucans and other polysaccharides extracted materials down to the MW cut-off of tubing/membranes before further reduction into oligomers. from raw materials down to the MW cut-off of tubing/membranes before further reduction into oligome We have rs. tried to correlate different parameters with the MWs of depolymerized BBG samples and glucose We ha(180 ve triDa) ed to correl as the monomer ate different pa of BBGs. ramete In rs wi Figur th the e 9 M A,W we s of observe depolyme an ri incr zed B ease BGof saterminal mples and glucose (180 Da) as the monomer of BBGs. In Figure 9A, we observe an increase of terminal glucose with MW reduction in log scale. The correlation could be well ﬁtted by a power equation glucose with MW reduction in log scale. The correlation could be well fitted by a power equation with an R value = 0.947. Since the 1-3 and 1-4 linkage has an approximate ratio of 1:6, therefore, with with an R value = 0.947. Since the 1-3 and 1-4 linkage has an approximate ratio of 1:6, therefore, with this power equation, we could extrapolate the linkage of BBGs with further MW reduction. On the this power equation, we could extrapolate the linkage of BBGs with further MW reduction. On the other hand, Figure 9B shows total biomass (bacteria) after 40-hour fermentation with different MW other hand, Figure 9B shows total biomass (bacteria) after 40-hour fermentation with different MW of BBGs. We could observe an increase of biomass as MW was increased due to more compact and of BBGs. We could observe an increase of biomass as MW was increased due to more compact and dense energy provided by the longer chain -glucan and then a decrease as higher requirement of dense energy provided by the longer chain β-glucan and then a decrease as higher requirement of energy/time/machinery in breakdown the long compact chain into smaller unit for absorption. Thus, energy/time/machinery in breakdown the long compact chain into smaller unit for absorption. Thus, we observe a turning point with MW close to BBG_0.1 (28 kDa). BBG_0.1 allow the most carrying we observe a turning point with MW close to BBG_0.1 (28 kDa). BBG_0.1 allow the most carrying capacity in equal mass of carbon source added. capacity in equal mass of carbon source added. Figure 9. Correlation analysis of selected parameters against BBGs of different MWs with glucose as Figure 9. Correlation analysis of selected parameters against BBGs of different MWs with glucose as monomer. (A) Terminal glucose molar ratio identiﬁed via linkage analysis using GC-MS; (B) total monomer. (A) Terminal glucose molar ratio identified via linkage analysis using GC-MS; (B) total biomass support by BBGs and glucose. DNA concentration was used to reﬂect the quasi-absolute biomass support by BBGs and glucose. DNA concentration was used to reflect the quasi-absolute bacterial biomass. bacterial biomass. 3.2. Polymeric BBGs and Monomeric Glucose on Microbial Taxa 3.2. Polymeric BBGs and Monomeric Glucose on Microbial Taxa Glucose (Glc) was added as a sole carbon source for the evaluation of monomeric and polymeric Glucose (Glc) was added as a sole carbon source for the evaluation of monomeric and polymeric glucose chain. Some selected taxa among the ﬁve BBGs and glucose have been depicted in Figures 5–7. glucose chain. Some selected taxa among the five BBGs and glucose have been depicted in Figures 5– Here, we evaluate the correlation of Firmicutes/Bacteroides (F/B) ratio as one of the indicators in 7. Here, we evaluate the correlation of Firmicutes/Bacteroides (F/B) ratio as one of the indicators in microbiome of gut eubiosis and dysbiosis. It was proposed that a lower of F/B ratio boost eubiosis microbiome of gut eubiosis and dysbiosis. It was proposed that a lower of F/B ratio boost eubiosis and promote resistance to certain syndromes. In Figure 10A, we observed a U-shape pattern of F/B and promote resistance to certain syndromes. In Figure 10A, we observed a U-shape pattern of F/B ratio against the increase of MW. BBG_0.1 had the lowest F/B ratio among the ﬁve BBGs and glucose ratio against the increase of MW. BBG_0.1 had the lowest F/B ratio among the five BBGs and glucose monomer. Actually, glucose had the highest F/B ratio, in accordance to previous reports, showing monomer. Actually, glucose had the highest F/B ratio, in accordance to previous reports, showing high glucose/fructose consumption would lead to increase of F/B ratio in mice without body weight high glucose/fructose consumption would lead to increase of F/B ratio in mice without body weight changes [19]. Figure 10B,C show two common -diversity indexes including the Shannon index and changes [19]. Figure 10B and Figure 10C show two common α-diversity indexes including the Simpson index. Shannon index emphasizes on both abundance and evenness, while the Simpson Shannon index and Simpson index. Shannon index emphasizes on both abundance and evenness, while the Simpson index is a dominance index because it gives more weight to common or dominant species [20]. A similar pattern between the two index and the most diversifying carbon source was found in BBG_0.1 that gave peak value in both indexes with an inverted-U shape being observed. Molecules 2019, 24, 828 12 of 20 index is a dominance index because it gives more weight to common or dominant species [20]. A similar pattern between the two index and the most diversifying carbon source was found in BBG_0.1 Molecules 2019, 24, x FOR PEER REVIEW 12 of 20 that gave peak value in both indexes with an inverted-U shape being observed. Molecules 2019, 24, x FOR PEER REVIEW 12 of 20 Figure 10. Correlation analysis of selected parameters against BBGs of different MWs with glucose as Figure 10. Correlation analysis of selected parameters against BBGs of different MWs with glucose as monomer. (A) Firmicutes/Bacteroidetes ratio; (B) Shannon -diversity; (C) Simpson -diversity. Figure 10. Correlation analysis of selected parameters against BBGs of different MWs with glucose as monomer. (A) Firmicutes/Bacteroidetes ratio; (B) Shannon α-diversity; (C) Simpson α-diversity. monomer. (A) Firmicutes/Bacteroidetes ratio; (B) Shannon α-diversity; (C) Simpson α-diversity. 3.3. Effect of Polymeric BBGs and Monomeric Glucose on Saccharolytic and Proteolytic Proﬁles 3.3. Effect of Polymeric BBGs and Monomeric Glucose on Saccharolytic and Proteolytic Profiles 3.3. Effect of Polymeric BBGs and Monomeric Glucose on Saccharolytic and Proteolytic Profiles Correlation analysis of the MW of BBG samples together with glucose shows a parabolic function Correlation analysis of the MW of BBG samples together with glucose shows a parabolic with total short chain fatty acids with the highest concentration found at a MW closed to BBG_0.2 Correlation analysis of the MW of BBG samples together with glucose shows a parabolic function with total short chain fatty acids with the highest concentration found at a MW closed to (11 kDa) (Figure 11A), showing BBG_0.2 had the highest saccharolytic potential among the ﬁve BBGs function with total short chain fatty acids with the highest concentration found at a MW closed to BBG_0.2 (11 kDa) (Figure 11A), showing BBG_0.2 had the highest saccharolytic potential among the though this relationship was not linear. The culture medium (mMCB) used in fermentation contained BBG_0.2 (11 kDa) (Figure 11A), showing BBG_0.2 had the highest saccharolytic potential among the five BBGs though this relationship was not linear. The culture medium (mMCB) used in fermentation partially digested protein such as peptone, tryptone, yeast extract and soy peptone. There have been five BBGs though this relationship was not linear. The culture medium (mMCB) used in fermentation contained partially digested protein such as peptone, tryptone, yeast extract and soy peptone. There publication showing that more saccharolytic microbial activity and less proteolytic activity confer contained partially digested protein such as peptone, tryptone, yeast extract and soy peptone. There have been publication showing that more saccharolytic microbial activity and less proteolytic activity eubiosis [21]. Moreover, this inverted-U shape pattern was also found in a similar article using oat have been publication showing that more saccharolytic microbial activity and less proteolytic activity confer eubiosis [21]. Moreover, this inverted-U shape pattern was also found in a similar article using -1,3-1,4-glucans confer eubiosis [of 21] six . Moreo differ vent er, tMW his inver with ted-U sh the original ape pat -glucan tern was and also fo theund lowest in a simi MWla fraction r article us gave ingthe oat β-1,3-1,4-glucans of six different MW with the original β-glucan and the lowest MW fraction gave least oat total β-1,3SCF -1,4-gl Asupr caoduction ns of six dif [22 ferent MW wi ]. We have also th the ori investigated ginal β-glu the can a concentration nd the lowest MW of dissolved fraction ga ammonia ve the least total SCFAs production [22]. We have also investigated the concentration of dissolved the least total SCFAs production [22]. We have also investigated the concentration of dissolved after 40-h fermentation (Figure 11B). Generally, hydrolyzed BBGs produced more dissolved ammonia ammonia after 40-hour fermentation (Figure 11B). Generally, hydrolyzed BBGs produced more ammonia after 40-hour fermentation (Figure 11B). Generally, hydrolyzed BBGs produced more with the native BBG and glucose monomer having similar concentration. Here, we denoted the dissolved ammonia with the native BBG and glucose monomer having similar concentration. Here, dissolved ammonia with the native BBG and glucose monomer having similar concentration. Here, SCF we denoted the SCFAs/NH As/NH ratio as an indicator 3 ratio to as evaluate an indic the ator to evaluate t saccharolytic and he sacch proteolytic arolytic and pro potential. tSince eolytic the we denoted the SCFAs/NH3 ratio as an indicator to evaluate the saccharolytic and proteolytic total potenti SCFAs al. Si and nce the tota NH data l SCFAs are in and dif NH fer 3ent data scales, are in different scales, both data were bo ﬁrst th da scaled ta were fi using rst scaled usi the min-max ng potential. Since the total SCFAs and NH3 data are in different scales, both data were first scaled using the min-max approach before their ratio was taken. As shown in Figure 11C, glucose monomer gave approach before their ratio was taken. As shown in Figure 11C, glucose monomer gave the least the min-max approach before their ratio was taken. As shown in Figure 11C, glucose monomer gave the least SCFAs/NH3 ratio while the native BBG gave the highest SCFAs/NH3 ratio, followed by SCFAs/NH ratio while the native BBG gave the highest SCFAs/NH ratio, followed by BBG_0.2. 3 3 the least SCFAs/NH3 ratio while the native BBG gave the highest SCFAs/NH3 ratio, followed by BBG_0.2. The use of SCFAs/NH3 ratio allowed signal to be observed that could otherwise not be The use of SCFAs/NH ratio allowed signal to be observed that could otherwise not be found in BBG_0.2. The use of SCFAs/NH3 ratio allowed signal to be observed that could otherwise not be found in Figure 11A and 11B. Eubiosis could be achieved by promoting more saccharolysis or by Figure 11A,B. Eubiosis could be achieved by promoting more saccharolysis or by diminishing the found in Figure 11A and 11B. Eubiosis could be achieved by promoting more saccharolysis or by diminishing the proteolysis. Therefore, we proposed that SCFAs/NH3 ratio could serve as a better proteolysis. Therefore, we proposed that SCFAs/NH ratio could serve as a better indicator than just diminishing the proteolysis. Therefore, we proposed that SCFAs/NH3 ratio could serve as a better indicator than just total SCFAs alone. total SCFAs alone. indicator than just total SCFAs alone. Figure 11. Correlation analysis of selected parameters against BBGs of different MWs with glucose as Figure 11. Correlation analysis of selected parameters against BBGs of different MWs with glucose Figure 11. Correlation analysis of selected parameters against BBGs of different MWs with glucose as 3 monomer. (A) Total short chain fatty acids (SCFAs); (B) total dissolved ammonia (NH ); (C) as monomer. (A) Total short chain fatty acids (SCFAs); (B) total dissolved ammonia (NH ); (C) monomer. (A) Total short chain fatty acids (SCFAs); (B) total dissolved ammonia (NH ); (C) SCFAs/NH3 ratio after data scaling to both using the min-max approach. SCFAs/NH ratio after data scaling to both using the min-max approach. SCFAs/NH3 ratio after data scaling to both using the min-max approach. 3.4. Insights from the Phenotypic Data, Microbial Profiles against MW Reduction of BBGs 3.4. Insights from the Phenotypic Data, Microbial Proﬁles against MW Reduction of BBGs 3.4. Insights from the Phenotypic Data, Microbial Profiles against MW Reduction of BBGs With sufficient carbon sources, microbes can grow better in substrates with lower MW because With sufﬁcient carbon sources, microbes can grow better in substrates with lower MW because With sufficient carbon sources, microbes can grow better in substrates with lower MW because of the extra energy/time spent in high MW polymeric breakdown and absorption. Monosaccharides of the extra energy/time spent in high MW polymeric breakdown and absorption. Monosaccharides of the extra energy/time spent in high MW polymeric breakdown and absorption. Monosaccharides such as glucose, and simple sugars such as lactose and sucrose are therefore widely used as sole such as glucose, and simple sugars such as lactose and sucrose are therefore widely used as sole carbon such as glucose, and simple sugars such as lactose and sucrose are therefore widely used as sole carbon sources in microbiological culture media preparation. However, the advantage of MW carbon sources in microbiological culture media preparation. However, the advantage of MW reduction does not follow a linear trend as with MW as demonstrated by Figures 9–11. It is because reduction does not follow a linear trend as with MW as demonstrated by Figures 9–11. It is because fermentation in pure bacterial culture does not consider the interactions between microbes. For fermentation in pure bacterial culture does not consider the interactions between microbes. For Molecules 2019, 24, 828 13 of 20 sources in microbiological culture media preparation. However, the advantage of MW reduction does not follow a linear trend as with MW as demonstrated by Figures 9–11. It is because fermentation in pure bacterial culture does not consider the interactions between microbes. For example, although strain speciﬁc, the in vitro fermentation of oat -1,3-1,4-glucan and its hydrolyzed fragments with pure probiotic culture Lactobacillus helveticus, Lactobacillus rhamnosus and Biﬁdobacterium longum showed that growth and total SCFAs production was highest with glucose, followed by hydrolyzed derivative and then the original oat -glucan [23]. Important elements of research in microbial community ecology include the analysis of functional pathways for nutrient resource and energy ﬂows, mechanistic understanding of interactions between microbial populations and their environment, and the emergent properties of the complex community [24]. Microbial community interaction is multi-fold, including competition, mutualism, inhibition, promotion and so on. Therefore, the use bacterial community in fermentation leads to a diverse results. In the present study, four very different correlation patterns resulting from fermentation with BBGs with different MWs were extracted and discussed as examples. Reduction of the MW in BBGs seemed to increase the population of Lactobacillus gradually with glucose having the maximum increase (Figure 12A). Though it was not linear, a monotonous trend was still observed. Such trend suggested that further reduction in the MW of BBGs would probably increase the population growth of Lactobacillus among the whole microbial community. On the other hand, Figure 12B shows the pattern of Biﬁdobacterium, with the highest population increase in Biﬁdobacterium being with BBG_01. An inverted-U shape pattern was also recognized among the ﬁve BBGs. With reference to glucose, we expected a further reduction in the MW of BBG samples would ﬁrstly reduce the population of Biﬁdobacterium followed by an increase again. Actually, glucose monomer induced particularly Biﬁdobacterium growth (a little bit more relative abundance than BBG_0.01). Veillonella might have an exceptional ability in fermenting BBG with MW lower than BBG_0.2(2) as shown in Figure 12C. Veillonella is a group of microbes well-known for its ability to convert lactate into acetate and propionate during fermentation [12]. This might also contribute to the trend in propionic acid production (Figure 2A) as the growth pattern of Veillonella matched well with the propionic acid production pattern. On the contrary, the growth pattern of Veillonella was a mirror reﬂection of Biﬁdobacterium with Veillonella having a slight phase difference. This could possibly explain the reduction of Biﬁdobacterium with BBG having a MW lower than BBG_0.2(2) (6 kDa) as Veillonella might be able to utilize BBGs with a MW lower than 6 kDa. Figure 12D shows the correlation pattern of the Enterococcus which contributed about 60% of total bacterial abundance in all the ﬁve BBG samples. It was observed that the population of Enterococcus had a moderate increase when the MW was above that of BBG_0.2(2) but decreased drastically when the MW was below that of BBG_0.2(2) (Figure 12D). Growth of Enterococcus was supported by polymeric BBGs while glucose monomer, and perhaps dimers and oligomers are extrapolated to be less supportive under in vitro condition. Molecules 2019, 24, 828 14 of 20 Molecules 2019, 24, x FOR PEER REVIEW 14 of 20 Figure 12. Correlation analysis of selected taxa at genus level against BBGs of different MWs with Figure 12. Correlation analysis of selected taxa at genus level against BBGs of different MWs with glucose as monomer. (A) Lactobacillus; (B) Biﬁdobacterium; (C) Veillonella; and (D) Entrococcus. glucose as monomer. (A) Lactobacillus; (B) Bifidobacterium; (C) Veillonella; and (D) Entrococcus. 4. Materials and Methods 4. Materials and Methods 4.1. Materials 4.1. Materials All chemicals were obtained from Sigma-Aldrich (St. Louis, Missouri, USA) unless All chemicals were obtained from Sigma-Aldrich (St. Louis, Missouri, USA) unless otherwise otherwise speciﬁed. specified. 4.2. Infant Faecal Sample Collection 4.2. Infant Faecal Sample Collection Sample collection protocols were approved by the Clinical Research Ethics Committee of the Sample collection protocols were approved by the Clinical Research Ethics Committee of the Joint Chinese University of Hong Kong (CREC Ref. No: 2015.206). Brieﬂy, infant faecal samples were Joint Chinese University of Hong Kong (CREC Ref. No: 2015.206). Briefly, infant faecal samples were obtained from four 9–15 months old infants (3–15 month is known as the microbiome developmental obtained from four 9–15 months old infants (3–15 month is known as the microbiome developmental phase [25]) without any antibiotic or medicine intake at least three months prior to the collection. All phase [25]) without any antibiotic or medicine intake at least three months prior to the collection. All the infants were delivered by natural birth and undertook solid food and breast milk simultaneously the infants were delivered by natural birth and undertook solid food and breast milk simultaneously at the at the ti time me of the experi of the experiments, ments, gi giving ving p possibly ossibly th the e h highest ighest microb microbial ial diver diversity sity (s(solid olid food on food only ly and and breast-feeding breast-feedin only g only gave gadif ve fer different d ent diversity iversity [25]).[25] The ). The faecalfa samples ecal samples were collected were collect in sample ed in sample collection collection tubes diluted at 1:1 v/v with stabilizing buffer containing 30% glycerol (Thermo Scientific tubes diluted at 1:1 v/v with stabilizing buffer containing 30% glycerol (Thermo Scientiﬁc Pierce, Pierce, Waltham, MA, USA, to maintain bacteria viability during cold storage) supplemented with Waltham, MA, USA, to maintain bacteria viability during cold storage) supplemented with 0.1% 0.1% cysteine hydrochloride (a reducing agent to maintain anaerobicity). Faecal samples were stored cysteine hydrochloride (a reducing agent to maintain anaerobicity). Faecal samples were stored in in −20 °C during the delivery to the laboratory and were kept at −80 °C upon arrival. The viability of 20 C during the delivery to the laboratory and were kept at80 C upon arrival. The viability of the the faecal bacteria was checked by using total plate count (TPC) with reinforced clostridial broth faecal bacteria was checked by using total plate count (TPC) with reinforced clostridial broth (Oxoid (Oxoid Thermo, Waltham, MA, USA) with 2% agar (RCA) to ensure a sufficient number of viable Thermo, Waltham, MA, USA) with 2% agar (RCA) to ensure a sufﬁcient number of viable bacteria at a bacteria at a dilution of 10 . Any faecal sample with a TPC of less than 300 Colony Forming Units dilution of 10 . Any faecal sample with a TPC of less than 300 Colony Forming Units (CFUs) at the (CFUs) at the dilution of 10 was rejected. All faecal samples from the four infants were then pooled dilution of 10 was rejected. All faecal samples from the four infants were then pooled together with together with equal TPC CFU ratio to consolidate the generalizability of findings. equal TPC CFU ratio to consolidate the generalizability of ﬁndings. 4.3. Depolymerization of Barley β-Glucans 4.3. Depolymerization of Barley b-Glucans Low viscosity barley β-1,3-1,4-glucans (BBGs) was obtained from Megazyme (product code: P- Low viscosity barley -1,3-1,4-glucans (BBGs) was obtained from Megazyme (product code: BGBL, >95% purity as stated). Acid hydrolysis was performed with different concentration of sulfuric P-BGBL, >95% purity as stated). Acid hydrolysis was performed with different concentration of acid (0.05, 0.10 and 0.20 M) for various time periods (10, 60, and 120 min) to obtain a range of BBG sulfuric acid (0.05, 0.10 and 0.20 M) for various time periods (10, 60, and 120 min) to obtain a range of Molecules 2019, 24, 828 15 of 20 BBG samples with different MWs. Brieﬂy, an aqueous solution of BBG at a concentration of 10 mg/mL was prepared and maintained at 80 C, followed by the addition of an appropriate amount of sulfuric acid (8M) to give a ﬁnal concentration of 0.05, 0.1 and 0.2 M. The samples were stirred at 80 C for the various time periods. The preparation scheme for BBG is shown in Table 1. A total of ﬁve BBG samples, including the native BBG and four depolymerized derivatives were obtained after the different acid hydrolysis conditions and they were denoted as BBG, BBG_0.05, BBG_0.1, BBG_0.2, and BBG_0.2(2). These sample solutions were neutralized with NaOH, dialyzed in Spectra/Por tubing with molecular weight cut off (MWCO) 3 kDa (Repligen, Waltham, MA, USA) for desalting in Milli-Q water for 24 h with three changes of water until the conductivity of the Milli-Q water reached single digit (Ciba-Corning Cond/TDS, Corning, NY, USA). All dialyzed BBG samples were then freeze-dried. 4.4. Chemical Composition and Linkage Analysis of BBGs The total carbohydrate and total protein contents of the ﬁve BBG samples were determined by the phenol-sulfuric acid method and the bicinchoninic acid (BCA) protein assay (Thermo Scientiﬁc Pierce, Waltham, MA, USA), respectively by protocols described previously [26]. The monosaccharide composition of the native BBG and its depolymerized derivatives were determined by gas chromatographic analysis of the alditol acetate derivatives of the sugars formed by sequential acid hydrolysis, reduction and acetylation described previously [27]. The glycosidic linkages of the ﬁve BBG samples were determined by methylation analysis. In brief, partially methylated alditol acetate (PMAA) derivatives of the samples were prepared using CH I and solid NaOH in dry DMSO described in a modiﬁed protocol reported previously. A gas chromatograph (6890N, Agilent Technologies, Santa Clara, CA, USA) coupled to a mass spectrometer (5973N, Agilent Technologies) (GC-MS) was used for analysis of both monosaccharides as their alditol acetate derivatives and linkage positions between sugar residues as their PMAA derivatives by protocols described previously [26]. 4.5. Molecular Weight Distribution of BBGs The MW distribution of the ﬁve BBG samples was determined using size exclusion chromatography (SEC) (Waters e2695 HPLC system) coupled with a refractive index (RI) detector (Waters 2414, Waters Inc., Milford, MA, USA) [28]. In brief, a size exclusion column TSK gel G5000 PW (30 cm 7.5 mm i.d., Supelco, Bellefonte, PA, USA) was used to determine the MW proﬁle. A range of dextran standards (5, 25, 80, 150, 410, 670 kDa) were used for calibrating the MW in the SEC analysis. The samples were dissolved in Milli-Q water to a concentration of 10 mg/mL and ﬁltered through 0.22 mm nylon syringe ﬁlter before injection. Water was applied as mobile phase at a ﬂow-rate of 0.7 mL/min at 30 C. 4.6. Spectroscopic Methods of BBGs The infrared spectra of the ﬁve BBG samples were recorded with a Fourier transform infrared spectrometer (FT-IR, Nicolet 670, Thermo Scientiﬁc, Waltham, MA, USA) in the range 4000–400/cm using the KBr-disk method [29]. One-dimensional proton NMR measurements of the samples were analyzed on a 700 MHz NMR spectrometer (Bruker Advanced III HD, Bruker, Billerica, MA, USA) at 60 C using DMSO-d as the solvent [30]. 4.7. In Vitro Deep Well-Plate Fermentation Modiﬁed medium for colonic broth (mMCB) was used as the basal medium for the fermentation experiments as previously reported [29] and contained (per litre) 6.5 g of bacteriological peptone, 5.0 g of soy peptone, 2.5 g of tryptone, 3.0 g of yeast extract, 2.0 g of KCl, 0.2 g of NaHCO , 4.5 g of NaCl, 0.5 g of MgSO 7H O, 0.45 g of CaCl 2H O, 0.2 g of MnSO H O, 0.005 g of FeSO 7H O, 4 2 2 2 4 2 4 2 0.005 g of ZnSO 7H O, 0.4 g of cysteine–HCl, 0.005 g of hemin, 0.005 g of menadione, 0.5 mL of 4 2 H PO , and 2 mL of Tween 80. No added carbon sources. All ﬁve BBG samples as well as glucose 3 4 (as -glucan monomer) were added as the sole carbon source at a concentration of 2% (w/v). Pooled Molecules 2019, 24, 828 16 of 20 infant faecal inoculum was mixed with mMCB having individual carbon sources at a volume ratio of 1:9 [31]. All fermentation mixtures (carbon sources in mMCB with infant faecal inoculum) were done in biological triplicates for all ﬁve BBGs and glucose monomer. Fermentation was conducted in deep well plates (Thermo Scientiﬁc) with each well composed of 2 mL of fermentation mixtures in an anaerobic jar containing AnaeroGen (Oxoid Thermo) and incubated at 37 C for 40 h [31]. At the end of the fermentation period, aliquots of the fermentation samples (from the triplicates of each of the 6 carbon sources) were collected for microbial plating (100 L), DNA extraction (200 L), short chain fatty acid proﬁling (600 L) and ammonia determination (100 L). 4.8. Standard Bacterial Plate Count Standard bacterial plate count was done using Reinforced Clostridia Broth (RCB) (Oxoid Thermo) plus 2% Bacteriological Agar (Oxoid Thermo), maintained under anaerobic conditions in duplicate using standard protocols of serial dilution with sterile PBS [31]. The growth of the total bacteria in the fermentation mixture after 40-h fermentation was expressed as CFUs after 48 h anaerobic incubation at 37 C. 4.9. Short Chain Fatty Acids and Dissolved Ammonia Determination SCFAs were extracted and analyzed using GC-FID with slight modiﬁcations [27]. In brief, 30 L of methylvaleric acid (100 mg/mL) was added as internal standard and a hydrogen ﬂow rate of 0.5 mL/min were used in the present method. A mixture of individual SCFA standards including acetic, propionic, butyric, valeric and caproic acid (SCFA standards kit, Alltech Inc., Nicolasville, KY, USA), and 4-methylvaleric acid (as internal standard) was prepared in 25% metaphosphoric acid at a ﬁnal concentration of 10.0 mg/mL for identiﬁcation and quantitation. The amount of the SCFAs after 24 h of fermentation was expressed in mM. Ammonia was determined using Ammonia Rapid Assay (K-AMIAR, Megazyme). It is a highly speciﬁc assay involving the enzymatic conversion of 2-oxoglutarate, NADPH and dissolved ammonia into L-glutamic acid, NADP and water by glutamate dehydrogenase. Brieﬂy according to manual, 100 L fermentation samples were centrifuged at 14000 rpm (Spectrafuge 24D, Labnet, Woodbridge, NJ, USA) for 5 min to obtain supernatant, then 2 L of samples, ammonia standard (0.04 mg/mL) and water as blank were added with 30 L Buffer and 20 L NADPH with 208 L distilled water replaced by PBS pH 9.8 to maintain solubility of ammonia in reaction. Absorbance (A1) was taken at 340 nm with microplate reader (Molecular Devices SpectraMax Plus 384, San Jose, CA, USA). Then 2 L glutamate dehydrogenase was added to each mixture and allowed to react for 5 min, followed by second absorbance (A2) at 340 nm. Calibration was also performed using a serial 2-fold dilution of ammonia standards. The ﬁnal ammonia concentration (mg/mL) was calculated as: (Sample [A1 A2]/Standard[A1 A2]) mg/mL Standard. 4.10. DNA Extraction, 16S Amplicon Sequencing and Bioinformatics There was a total of 18 sequencing samples consisted of BBG, BBG_0.05, BBG_0.1, BBG_0.2, BBG_0.2(2), and glucose (each in biological triplicate). DNA extraction was conducted using EZNA Stool DNA Kits (Omega Bio-tek Inc., Norcross, GA, USA) in accordance with the manufacturer ’s manual with some modiﬁcations [32]. These included an enzymatic pre-treatment step (mutanolysin and lysozyme) and a bead-beading step using a mix of 0.1 mm and 0.5 mm beads (Bioprep-24 Homogenizer & bead beater) to enhance DNA release from Gram positive bacteria. 16S amplicons were generated with Phusion polymerase (New England BioLabs, Ipswich, MA, USA) with barcoded forward primers targeting 16S V3 region [32] based on manufacturer ’s protocols. After PCR, sharp DNA bands appeared at approximately 200 bp were cut and gel puriﬁed using QiaQuick Gel Extraction kit (Qiagen, Venlo, Netherlands) followed by AMPure treatment using the Agencourt AMPure XP system (Beckman Coulter, Brea, CA, USA). The quantity and quality of DNA after AMPure treatment were accessed using a Qubit Fluorometer (Thermo Scientiﬁc, Waltham, MA, USA) and Bio-Analyzer Molecules 2019, 24, 828 17 of 20 (Agilent, Santa Clara, CA, USA). Sequencing library was prepared by pooling the 18 samples in equal molar ratio. Then, emulsion PCR and enrichment were done using the Ion PGM Template OT2 400 Kit in Ion OneTouch 2 System (Thermo Scientiﬁc, Waltham, MA, USA). The DNA sequencing was done using the Ion PGM Sequencing 400 Kit on Ion torrent PGM using Chip 318v2 in the Core Facilities of the Chinese University of Hong Kong. After sequencing, the individual sequence reads were ﬁltered by the PGM software to remove low quality and polyclonal sequences. Sequences matching the PGM 3 adaptor were also automatically trimmed. All PGM quality-approved, trimmed and ﬁltered data were exported as .sff ﬁles. Bioinformatics analysis was done using Qiime-v1.9.1 pipeline [33]. Greengenes-v13.8 was used as reference for bacterial 16S sequence database [34]. Sequences were trimmed for primers and barcodes. The cleaned sequences were then clustered at 97% similarity as Operational Taxonomic Unites (OTUs) followed by deletion of chimeras and singleton reads. and diversity were then analyzed. KEGG prediction was carried out using PICRUSt-v1.0.0 [14] with NSTI score output included and statistically analyzed using STAMP-v2.1.3 [35]. Bacterial biomarkers unique to groups were identiﬁed using LEfSe [36]. Signiﬁcant microbial taxa were also identiﬁed using LEfSe. High level microbial phenotypes that could not predicted by PICRUSt were analyzed using BugBase and default BugBase traits were predicted [37]. All computational analysis was carried out on a local ® ® Dell Workstation with Quad Core Intel Xeon Processor and 32G RAM. 4.11. Integration and Statistical Analysis Integration analysis was done by ﬁtting the trends of different parameters (DNA amount/terminal glucose/SCFAs/NH3/microbial taxa) investigated in this study against increase in MW modelled with regression equation and coefﬁcient of determination provided. Unless otherwise speciﬁed, R software-v3.5.2 [38] was used for statistical calculation and independent pairwise student’s t-test or ANOVA with Tukey HSD post-hoc adjustment (CI = 0.95) was used. Statistical test parameters for bioinformatics tools include: STAMP (Statistical test: ANOVA with Tukey-Kramer post-hoc test CI = 0.95, multiple test correction: Benjamini-Hochberg FDR); LEfSe (alpha = 0.05 for factorial Kruskal-Wallis test; alpha = 0.05 for pairwise Wilcoxon test; threshold = 2.0 for logarithmic LDA score) and BugBase (threshold values were automatically deﬁned with highest variance across all sample, output to be <0.03, ANOVA with Tukey HSD post-hoc adjustment (CI = 0.95) was used). 5. Conclusions Our present results have demonstrated a partial acid hydrolysis under controlled conditions seemed to be an effective means to depolymerize BBG into derivatives with lower MW. It is a versatile method suitable for screening and puriﬁcation after hydrolysis. We have also demonstrated that the platform used for infant faecal sample collection and storage as well as using the deep well plate to serving as multiple miniature bioreactors, is applicable for in vitro fermentation. We have also coined the SCFAs/NH ratio as a new indicator of eubiosis/dysbiosis in terms of saccharolytic and proteolytic fermentation. Correlation analysis had shown that most of the fermentation parameters investigated showed a parabolic function instead of a linear function with the BBG samples having different MW. Microbial community investigation is challenging as there is a myriad of interactions among all the microbes. We have exempliﬁed this idea with the monotonous pattern, U-shape pattern, inverted-U shape pattern, wave-like pattern of some selected taxa. Among the ﬁve BBGs, we concluded that the BBG_0.1 with intermediate MW of 28 kDa showed the most promising lead to be developed as a potential novel prebiotic. The data of this project has been submitted to NCBI SRA and available under the BioProject ID: PRJNA487303, SRA accession: SRP158602. Supplementary Materials: The following are available online, Figure S1, (A) GC monosaccharide proﬁle; (B) FTIR spectrum; and (C) NMR 1H spectrum of BBG and BBG_0.2(2). Figure S2, Representative bacterial total plate count of the 5 BBG samples and glucose monomer. Figure S3, GC-FID chromatograms (A) Short chain fatty acid external standards, C2, C3, C4, C5, C6 and internal standard, methyl-valeric acid; (B) short chain fatty acid proﬁles of the BBG after 40 h of fermentation. Figure S4, Overall distribution of metagenome KEGG prediction output of the infant faecal fermentation of 5 BBG samples, glucose monomer and T0 group using PICRUSt. (A) Molecules 2019, 24, 828 18 of 20 KEGG level 1 class distribution; (B) expanded KEGG level 2 class distribution from level 1 “Metabolism” group. Figure S5, Selected metagenome KEGG prediction output of the infant faecal fermentation of 5 BBG samples and glucose monomer using PICRUSt and statistically analysis using STAMP. (A) Carbohydrate metabolism; (B) ABC transporters; (C) Bacterial secretion system; and (D) Glycan biosynthesis and metabolism. Figure S6, Selected metagenome KEGG prediction output of the infant faecal fermentation of 5 BBG samples and glucose monomer using PICRUSt and statistically analysis using STAMP. (A) Lipid metabolism; (B) Protein digestion and absorption; (C) Amino acid metabolism; and (D) Nitrogen metabolism. Table S1, All the identiﬁed microbial taxa of the 5 BBGs (XLSX format). Table S2, The output of PICRUSt metagenome prediction of infant faecal fermentation of 5 BBG samples and glucose (XLSX format). Table S3, The output of LEfSe of infant faecal fermentation of 5 BBGs and glucose (XLSX format). Author Contributions: K.-L.L., K.-C.K., T.C., H.-S.K., L.Y. and P.C.-K.C. designed the experiments. K.-L.L. played the major role, with assistance of K.-C.K., X.L., X.K., W.-Y.C. for performing all experiments, analysis and bioinformatics; K.-L.L. wrote the manuscript with advice from by P.C.-K.C. All authors contributed to interpreting the results, critically revising the manuscript for important intellectual content, and approving the ﬁnal manuscript. Funding: This research was funded by RGC Research Grant General Research Fund, grant number 14114115. Acknowledgments: The authors express their gratitude to the breast-feeding mother in Hong Kong for their participation in this project, collection and donation of infant faecal samples. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Grain: World Markets and Trade (December report by United States Department of Agriculture Foreign Agricultural Service). Available online: https://downloads.usda.library.cornell.edu/usda-esmis/ﬁles/ zs25x844t/8g84mr14m/0k225g11x/grain.pdf (accessed on 1 February 2019). 2. Lam, K.L.; Cheung, P.C.K. Non-digestible long chain -glucans as novel prebiotics. Bioact. Carbohydr. Dietary Fibre. 2013, 2, 45–64. [CrossRef] 3. Jayachandran, M.; Chen, J.; Chung, S.S.M.; Xu, B. A critical review on the impacts of -glucans on gut microbiota and human health. J. Nutr. Biochem. 2018, 61, 101–110. [CrossRef] [PubMed] 4. Zhu, F.; Du, B.; Xu, B. A critical review on production and industrial applications of -glucans. Food Hydrocoll. 2016, 52, 275–288. [CrossRef] 5. Hughes, S.A.; Shewry, P.R.; Gibson, G.R.; McCleary, B.V.; Rastall, R.A. In vitro fermentation of oat and barley derived -glucans by human faecal microbiota. FEMS Microbiol. Ecol. 2008, 64, 482–493. [CrossRef] [PubMed] 6. Williams, B.A.; Mikkelsen, D.; le Paih, L.; Gidley, M.J. In vitro fermentation kinetics and end-products of cereal arabinoxylans and (1,3;1,4)- -glucans by porcine faeces. J. Cereal Sci. 2011, 53, 53–58. [CrossRef] 7. Mikkelsen, M.S.; Jensen, M.G.; Nielsen, T.S. Barley -glucans varying in molecular mass and oligomer structure affect cecal fermentation and microbial composition but not blood lipid proﬁles in hypercholesterolemic rats. Food Funct. 2017, 8, 4723–4732. [CrossRef] [PubMed] 8. Korpela, K. Diet, microbiota, and metabolic health: Trade-off between saccharolytic and proteolytic fermentation. Annu. Rev. Food Sci. Technol. 2018, 9, 65–84. [CrossRef] [PubMed] 9. Tsai, Y.L.; Olson, B.H. Rapid method for direct extraction of DNA from soil and sediments. Appl. Environ. Microbiol. 1991, 57, 1070–1074. [PubMed] 10. Vogt, S.L.; Pena-Díaz, J.; Finlay, B.B. Chemical communication in the gut: Effects of microbiota generated metabolites on gastrointestinal bacterial pathogens. Anaerobe 2015, 34, 106–115. [CrossRef] [PubMed] 11. Tamura, K.; Hemsworth, G.R.; Dejean, G.; Rogers, T.E.; Pudlo, N.A.; Urs, K.; Jain, N.; Davies, G.J.; Martens, E.C.; Brumer, H. Molecular mechanism by which prominent human gut bacteroidetes utilize mixed-linkage -glucans, major health-promoting cereal polysaccharides. Cell Rep. 2017, 21, 417–430. [CrossRef] [PubMed] 12. Riíos-Coviaán, D.; Ruas-Madiedo, P.; Margolles, A.; Gueimonde, M.; de los Reyes-Gavilaán, C.G.; Salazar, N. Intestinal Short Chain Fatty Acids and their Link with Diet and Human Health. Front. Microbiol. 2016, 7, 185. [CrossRef] [PubMed] 13. Seidel, J.; Haller, S.; Eckmanns, T.; Harder, T. Routine screening for colonization by Gram-negative bacteria in neonates at intensive care units for the prediction of sepsis: Systematic review and meta-analysis. J. Hosp. Infect. 2018, 99, 367–380. [CrossRef] [PubMed] Molecules 2019, 24, 828 19 of 20 14. Langille, M.G.I.; Zaneveld, J.; Caporaso, J.G.; McDonald, D.; Knights, D.; a Reyes, J.; Clemente, J.C.; Burkepile, D.E.; Vega Thurber, R.L.; Knight, R.; et al. Predictive functional proﬁling of microbial communities using 16S rRNA marker gene sequences. Nat. Biotech. 2013, 31, 814–821. [CrossRef] [PubMed] 15. Ishimoto, Y.; Ishibashi, K.; Yamanaka, D.; Adachi, Y.; Kanzaki, K.; Iwakura, Y.; Ohno, N. Production of low-molecular weight soluble yeast -glucan by an acid degradation method. Int. J. Biol. Macromol. 2018, 107, 2269–2278. [CrossRef] [PubMed] 16. Johansson, L.; Virkki, L.; Anttila, H.; Esselstrom, H.; Tuomainen, P.; Sontag-Strohm, T. Hydrolysis of -glucan. Food Chem. 2006, 97, 71–79. [CrossRef] 17. Cho, H.J.; Jang, W.J.; Moon, S.Y.; Lee, J.M.; Kim, J.H.; Han, H.S.; Kim, K.W.; Lee, B.J.; Kong, I.S. Immobilization of -1,3-1,4-glucanase from Bacillus sp. on porous silica for production of -glucooligosaccharides. Enzyme Microb. Technol. 2018, 110, 30–37. [CrossRef] [PubMed] 18. Shi, Y.; Liu, J.; Yan, Q.; You, X.; Yang, S.; Jiang, Z. In vitro digestibility and prebiotic potential of curdlan (1-3)- -D-glucan oligosaccharides in Lactobacillus species. Carbohydr. Polym. 2018, 188, 17–26. [CrossRef] [PubMed] 19. Do, M.H.; Lee, E.; Oh, M.J.; Kim, Y.; Park, H.Y. High-glucose or-fructose diet cause changes of the gut microbiota and metabolic disorders in mice without body weight change. Nutrients 2018, 10, 761. [CrossRef] [PubMed] 20. Morris, E.K.; Caruso, T.; Buscot, F.; Fischer, M.; Hancock, C.; Maier, T.S.; Meiners, T.; Muller, C.; Obermaier, E.; Prati, D.; et al. Choosing and using diversity indices: Insights for ecological applications from the german biodiversity exploratories. Ecol. Evol. 2014, 4, 3514–3524. [CrossRef] [PubMed] 21. Makki, K.; Deehan, E.C.; Walter, J.; Backhed, F. The impact of dietary ﬁber on gut microbiota in host health and disease. Cell Host Microbe 2018, 23, 705–715. [CrossRef] [PubMed] 22. Kim, H.J.; White, P.J. Optimizing the molecular weight of oat -glucan for In vitro bile acid binding and fermentation. J. Agric. Food Chem. 2011, 59, 10322–10328. [CrossRef] [PubMed] 23. Dong, J.; Yu, X.; Dong, L.; Shen, R. In vitro fermentation of oat -glucan and hydrolysates by fecal microbiota and selected probiotic strains. J. Sci. Food Agric. 2017, 97, 4198–4203. [CrossRef] [PubMed] 24. Konopka, A. What is microbial community ecology? ISME J. 2009, 3, 1223–1230. [CrossRef] [PubMed] 25. Stewart, C.J.; Ajami, N.J.; O’Brien, J.L.; Hutchinson, D.S.; Smith, D.P.; Wong, M.C.; Ross, M.C.; Lloyd, R.E.; Doddapaneni, H.; Metcalf, G.A.; et al. Temporal development of the gut microbiome in early childhood from the teddy study. Nature 2018, 562, 583–588. [CrossRef] [PubMed] 26. Li, X.; Zhou, J.; Liu, C.; Xiong, Q.; Duan, H.; Cheung, P.C.K. Stable and biocompatible mushroom -glucan modiﬁed gold nanorods for cancer photothermal therapy. J. Agric. Food Chem. 2017, 65, 9529–9536. [CrossRef] [PubMed] 27. Zhao, J.; Cheung, P.C. Fermentation of -glucans derived from different sources by biﬁdobacteria: Evaluation of their biﬁdogenic effect. J. Agric. Food Chem. 2011, 59, 5986–5992. [CrossRef] [PubMed] 28. Liu, C.; Choi, M.W.; Li, X.; Cheung, P.C.K. Immunomodulatory effect of structurally-characterized mushroom sclerotial polysaccharides isolated from Polyporus rhinocerus on human monoctyes THP-1. J. Funct. Foods 2018, 41, 90–99. [CrossRef] 29. Liu, C.R.; Chen, J.L.; Chen, L.; Huang, X.S.; Cheung, P.C.K. Immunomodulatory activity of polysaccharide-protein complex from the mushroom sclerotia of Polyporus rhinocerus in murine macrophages. J. Agric. Food Chem. 2016, 64, 3206–3214. [CrossRef] [PubMed] 30. Chen, L.; Zhang, B.; Chen, J.; Cheung, P.C.K. Cell wall structure of mushroom sclerotium (Pleurotus tuber-regium): Part 2. Fine structure of a novel alkali-soluble hyper-branched cell wall polysaccharide. Food Hydrocoll. 2014, 38, 48–55. [CrossRef] 31. Lam, K.L.; Keung, H.Y.; Ko, K.C.; Kwan, H.S.; Cheung, P.C.K. In vitro fermentation of -glucans and other selected carbohydrates by infant fecal inoculum: An evaluation of their potential as prebiotics in infant formula. Bioact. Carbohydr. Dietary Fibre 2018, 14, 20–24. [CrossRef] 32. Milani, C.; Hevia, A.; Foroni, E.; Duranti, S.; Turroni, F.; Lugli, G.A.; Sanchez, B.; Martin, R.; Gueimonde, M.; van Sinderen, D.; et al. Assessing the fecal microbiota: An optimized ion torrent 16S rRNA gene-based analysis protocol. PLoS ONE 2013, 8, 68739. [CrossRef] [PubMed] 33. Caporaso, J.G.; Kuczynski, J.; Stombaugh, J.; Bittinger, K.; Bushman, F.D.; Costello, E.K.; Fierer, N.; Pena, A.G.; Goodrich, J.K.; Gordon, J.I.; et al. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 2010, 7, 335–336. [CrossRef] [PubMed] Molecules 2019, 24, 828 20 of 20 34. DeSantis, T.Z.; Hugenholtz, P.; Larsen, N.; Rojas, M.; Brodie, E.L.; Keller, K.; Huber, T.; Dalevi, D.; Hu, P.; Andersen, G.L. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Appl. Environ. Microbiol. 2006, 72, 5069–5072. [CrossRef] [PubMed] 35. Parks, D.H.; Tyson, G.W.; Hugenholtz, P.; Beiko, R.G. STAMP: Statistical analysis of taxonomic and functional proﬁles. Bioinformatics 2014, 30, 3123–3124. [CrossRef] [PubMed] 36. Segata, N.; Izard, J.; Walron, L.; Gevers, D.; Miropolsky, L.; Garrett, W.; Huttenhower, C. Metagenomic biomarker discovery and explanation. Genome Biol. 2011, 12, R60. [CrossRef] [PubMed] 37. Ward, T.; Larson, J.; Meulemans, J.; Hillmann, B.; Lynch, J.; Sidiropoulos, D.; Spear, J.; Caporaso, G.; Blekhman, R.; Knight, R.; et al. BugBase predicts organism level microbiome phenotypes. bioRxiv 2017, 133462. [CrossRef] 38. R Core Team. A language and environment for statistical computing. R Foundation for Statistical Computing. Available online: http://www.R-project.org/ (accessed on 1 February 2019). Sample Availability: Samples of the compounds BBGs are available from the authors. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Journal

Molecules – Multidisciplinary Digital Publishing Institute

Published: Feb 26, 2019

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

In Vitro Infant Faecal Fermentation of Low Viscosity Barley β-Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics

In Vitro Infant Faecal Fermentation of Low Viscosity Barley β-Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

In Vitro Infant Faecal Fermentation of Low Viscosity Barley β-Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics

In Vitro Infant Faecal Fermentation of Low Viscosity Barley β-Glucan and Its Acid Hydrolyzed Derivatives: Evaluation of Their Potential as Novel Prebiotics

References (41)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies