Wood-Derived Dietary Fibers Promote Beneficial Human Gut Microbiota

The architecture of the gut bacterial ecosystem has a profound effect on the physiology and well-being of the host. Modulation of the gut microbiota and the intestinal microenvironment via administration of prebiotics represents a valuable strategy to promote host health. This work provides insights into the ability of two novel wood-derived preparations, AcGGM and AcAGX, to influence human gut microbiota composition and activity. These compounds were selectively fermented by commensal bacteria such as Bifidobacterium, Bacteroides-Prevotella, F. prausnitzii, and clostridial cluster IX spp. This promoted the microbial synthesis of acetate, propionate, and butyrate, which are beneﬁcial to the microbial ecosystem and host colonic epithelial cells. Thus, our results demonstrate potential prebiotic properties for both AcGGM and AcAGX from lignocellulosic feedstocks. These findings represent pivotal requirements for rationally designing intervention strategies based on the dietary supplementation of AcGGM and AcAGX to improve or restore gut health.

though there has been increasing interest in the structure and utilization of XOS/xylans and ␤-MOS/␤-mannans manufactured from woody biomass, limited information is available on their use as a potential resource with health-promoting properties. Previous studies investigating the prebiotic potential of XOS employed fractions obtained from cereals and fruit waste streams (corn cobs [13], corn straw [14], olive [15], and rice husk [16]), brewers' spent grains, and eucalyptus wood (17). The XOS mentioned above selectively stimulated bifidobacteria, which led to an increase in levels of SCFAs in fecal batch culture (13)(14)(15)(16) and in in vivo studies (18)(19)(20), thus showing prebiotic effects. The level of SCFA production and the fermentation rate were found to be dependent on XOS length and the presence of various substituents (15,17). On the other hand, only one study has evaluated the fermentability of glucomannan and galactoglucomannan mixtures from pine wood by human fecal microbiota, reporting increased bifidobacterial numbers and higher acetate levels (21). In addition, Polari et al. showed that a galactoglucomannan mixture prepared from spruce sapwood promoted the growth of bifidobacteria in pure cultures (22).
Understanding how novel wood biomass-derived oligosaccharides impact the growth of commensal gut bacteria could lead to selective manipulation of the microbiota through their supplementation in diets, eventually promoting host health. Thus, the aim of the present study was to provide an insight into the potential prebiotic properties of AcGGM and AcAGX from Norwegian lignocellulosic biomass. In vitro single and batch cultures were used to determine their ability to stimulate the growth of beneficial human gut bacterial groups and to determine the profile of fermentative SCFAs corresponding to fecal human microbiota inocula.

RESULTS AND DISCUSSION
Preparation and characterization of the AcGGM and AcAGX samples. The samples were prepared using methods similar to those described previously (11), except that the spruce wood was pretreated for 10 min at 200°C and then subjected to steam explosion treatment (23). Briefly, hemicelluloses from the pretreated spruce and birch wood samples were extracted by the use of warm water, and the extracts were filtered by bag filtering prior to ultrafiltration and nanofiltration. The retentate from the nanofiltration step was lyophilized, yielding AcGGM and AcAGX preparations from spruce and birch wood, respectively. Composition analysis showed that the molar ratio of mannose/glucose/galactose/acetyl units in the AcGGM preparation was 1:0.32:0.16: 0.28, whereas the AcAGX preparation consisted of xylosyl units with glucuronic (uronic) acid, arabinose, and acetyl substituents in a molar ratio of 1:0.08:0.01:0.54 ( Table 1). As expected, mannose (52.8%) and xylose (81.7%) were the most abundant monosaccharides in the AcGGM and AcAGX preparations, respectively. Arabinose (3.3%), uronic acids (2.4%), and xylose (16.2%) were also found in the AcGGM sample. These sugars likely resulted from the presence of a minor amount of xylan in the spruce wood, consistent with the results obtained in previous studies (24). Similarly, in the AcAGX sample, the levels of mannose (5.0%) and glucose (4.2%) indicated the presence of a minor amount of glucomannan in birch wood, according to data shown previously (25).
To look for additional information on mass distribution and structural features, the two preparations were analyzed by high-performance anion-exchange chromatography (HPAEC) coupled with a pulsed amperometric detector (HPAEC-PAD) and matrixassisted laser desorption-ionization time of flight mass spectrometry (MALDI-TOF) (Fig. 2). HPAEC-PAD analysis of the AcGGM preparation revealed the presence of ␤-MOS with DP of 2 to 10 and manno-polysaccharides (DP of Ն11) (Fig. 2a). The AcAGX preparation consisted predominantly of oligosaccharides with DP of 2 to 10 but also polymeric xylan (DP of Ն11) (Fig. 2d). Both preparations contained less than 1% monosaccharides. A deeper insight into the structure of the different components with low to medium molecular weight was obtained from the MALDI-TOF data. As shown in Fig. 2b and c, the AcGGM preparation contained a series of oligomers with DP ranging from 2 to 14 with single, double, triple, and quadruple acetylations. The degree of acetylation increased for higher-DP ␤-MOS; the most complex detectable components were DP 13 and DP 14 saccharides bearing four acetyl groups. The AcAGX preparation contained native and acetylated oligomers with DP ranging from 2 to 18 ( Fig. 2e and f). The degree of acetylation increased in the higher-DP XOS, where DP 17 to 18 had up to 10 acetylations.
Growth experiments with single bacterial cultures. The ability of the gut microbes to degrade and metabolize ␤-MOS/␤-mannans and XOS/xylans with various DP and substitutions has been reported for several numerically dominant and physiological relevant members of the gut microbiome, including members of the Bifidobacterium, Lactobacillus, and Bacteroides genera (22,26,27). In order to determine the fermentability of wood-derived AcGGM and AcAGX by these bacteria, independent in vitro culture experiments were conducted for a collection of 5 Bifidobacterium strains, 6 Lactobacillus strains, and 32 Bacteroides strains. Bacteria were cultured in either minimal medium (MM) (for Bacteroides) or semidefined medium (SDM) (for bifidobacteria and lactobacilli) supplemented with 0.5% (wt/vol) of either AcGGM (MM-AcGGM or SDM-AcGGM) or AcAGX (MM-AcAGX or SDM-AcAGX) separately, as the sole carbon source, and growth was evaluated as maximum optimal density at 600 nm (OD 600 ). Commercial polymeric ␤-mannans (carob glucomannan [CGM] and konjac galactoman- were calculated based on the total sugar molarity measured on each sample. c The uronic acid (U) content is based on the adapted colorimetric assay developed by Scott (63) described in Materials and Methods. d Man:Glc:Gal:Ac acetylation ratio (mol/mol) per mannose: 1:0.32:0.16:0.28. Please note that the degree of acetylation cannot be precisely evaluated since the fraction contains Xyl, which can be esterified by acetyl groups. e Xyl:U:Ara:Ac acetylation ratio (mol/mol) per xylose: 1:0.08:0.01:0.54. Please note that the degree of acetylation cannot be precisely evaluated since the fraction contains Man, which can be esterified by acetyl groups.
nan [KGM]) and xylans (wheat arabinoxylan [WAX]) were used as controls. All the bacteria displayed growth on MM or SDM supplemented with 0.5% (wt/vol) glucose (see Table S1 in the supplemental material). Bacteroides cellulosilyticus, B. ovatus, B. plebeius, B. uniformis, B. dorei, and B. xylanisolvens all grew in the presence of AcGGM, whereas they displayed various levels of the ability to metabolize either CGM or KGM (Fig. 3a). B. fragilis, B. clarus, B. intestinalis, B. oleiciplenus, and B. eggerthii achieved considerable growth on AcGGM but failed to grow in the presence of both CGM and xylo-oligosaccharides (DP ranging from 2 to 10) and xylo-polysaccharides (XPS) (DP, Ͼ10) present in the AcAGX preparation. Samples were analyzed with the following external standards: X 1 , xylose; X 2 , xylobiose; X 3 , xylotriose; X 4 , xylotetraose; X 5 , xylopentaose; X 6 , xylohexaose. (e and f) MALDI-TOF spectra of low-molecular-weight (e) to medium-molecular-weight (f) components present in the AcAGX preparation. Spectra show sodium adducts (Mass ϩ Na ϩ ). In panels b, c, d, and e, the following abbreviations are used: Ac, acetyl group; Hex, hexose; Me, methyl group; GlcA, glucuronic acid; Xyl, xylose. KGM (Fig. 3a). These data suggest that these strains might be able to utilize the low-molecular-weight ␤-MOS present in the AcGGM mixture but cannot metabolize the high-molecular-weight ␤-mannans CGM and KGM. AcAGX supported the growth of B. cellulosilyticus, B. ovatus, B. xylanisolvens, B. eggerthii, B. oleiciplenus, B. plebeius, B. intestinalis, B. dorei, and B. vulgatus (Fig. 3b). Under these conditions, most of the bacteria reached higher maximum OD 600 levels than those observed on WAX. Overall, several isolates did not show growth in MM-AcGGM or MM-AcAGX, consistent with the fact that the efficient utilization of ␤-mannans and xylans is not universal in the gut resident Bacteroides (28).
Bifidobacterium adolescentis, B. animalis subsp. lactis Bl-04, and B. dentium were capable of fermenting AcGGM and AcAGX, while all strains failed to grow on the pure polymeric substrates CGM, KGM, and WAX ( Fig. 3c and d). These results are in agreement with a previous study showing that only a few of the Bifidobacterium strains isolated from the human gut were capable of growth on high-molecular-weight XOS and MOS (22,29). Lactobacilli, in contrast, do not metabolize XOS and MOS significantly (30), with the exception of Lactobacillus brevis (31). However, they can utilize linear short-chain oligosaccharides (DP of 2 to 3), as exemplified by the uptake of mannobiose and mannotriose by L. plantarum WCFS1 (see Fig. S4 in the supplemental material). This is in agreement with our results, where L. brevis grew well on SDM-AcGGM and SDM-AcAGX whereas the other strains displayed limited growth on these substrates ( Fig. 3c and d). None of the lactobacilli displayed growth on CGM, KGM, and WAX ( Fig. 3c and d).
In contrast to Bacteroides, bifidobacteria lack extracellular endo-xylanases (GH10), thus exhibiting no hydrolytic capabilities with respect to polymeric xylans (39). However, previous studies have shown that several members of the Bifidobacterium genus are able to grow on XOS and that XOS uptake and metabolism rely on the presence of an ATP-binding cassette (ABC) transporter and cytoplasmic enzymes that process these to their monosaccharide components (40) (Fig. S3a). B. animalis subsp. lactis Bl-04 possesses an arabinoxylo-oligosaccharide (AXOS) utilization locus (Fig. S3a) consisting of a LacI transcriptional regulator, xylosidases (GH8, GH43, and GH120), arabinofuranosidases (GH43), and acetyl esterases for the removal of arabinosyl and acetyl substituents, as well as an ABC transporter (40,41). Consistent with the fact that this locus is conserved across different XOS-utilizing Bifidobacterium species, it was found in the genome of the B. adolescentis and B. dentium (40) (Fig. S3a). Some Bifidobacterium strains, including B. adolescentis, can ferment GGM, KGM, and glucomannan oligosaccharides (22,42); nevertheless, the ␤-mannan degradation system in bifidobacteria remains largely underexplored. A study reported by Kulcinskaja et al. (43) characterized a multimodular B. adolescentis ␤-mannanase (BaMan26A), although the ability of this bacterium to hydrolyze all of the potential linkages found in ␤-mannan has not been confirmed. BaMan26A homologs were found in the genome of ␤-mannan-utilizing B. animalis subsp. lactis Bl-04 and B. dentium (Table S1).
Carbohydrates are mainly imported by phosphoenolpyruvate (PEP):carbohydrate phosphotransferase system (PTS) transporters in lactobacilli (44). These systems consist of cytoplasmic components (enzyme E1 and histidine-phosphorylatable protein HPr) that lack sugar specificity and of membrane-associated enzymes (enzymes E2A, E2B, E2C, and sometimes E2D) that transfer a phosphate group from PEP to the incoming sugar, preventing it from diffusing back out of the cell. Glycans are then hydrolyzed by specific phospho-␤-glucosidases (GH1) into the glycolytic precursors glucose-6phosphate and glucose. These glycan utilization loci are conserved in the genome of L. plantarum WCFS1, L. helveticus, L. acidophilus, and L. gasseri (45) (Fig. S3b). This system would allow the bacteria to transport and metabolize short-chain undecorated XOS and ␤-MOS, whereas larger polysaccharides could not be processed. Indeed, polymeric CGM did not support growth of L. plantarum WCFS1, whereas the bacterium was able to grow and take up ␤-MOS derived from GH36/GH26 hydrolysis of CGM (Fig. S4). As an exception, L. brevis is capable of fermenting arabinoxylan-oligosaccharide (AXOS) through the action of arabinofuranosidases, including a GH43 and two GH51s (31,46). Overall, consistent with the absence of metabolism involving ␤-MOS/␤-mannans and XOS/xylans, no related gene clusters were found in the genome of the Bacteroides, Bifidobacterium, and Lactobacillus strains lacking the ability to grow on these hemicelluloses in vitro (Table S1).
In vitro fermentation of AcGGM and AcAGX. Batch pH-controlled fermentations were set up to investigate the prebiotic potential of the AcGGM and AcAGX mixtures. Samples were taken at designated time intervals for enumeration of physiologically relevant bacterial groups by fluorescent in situ hybridization (FISH) coupled with flow cytometry (FISH-FCM). For comparisons, commercial FOS was used as a positive control due to its established prebiotic properties (6). Results presented in Fig. 4 show an increase of the total bacterial population (enumerated using the EUB338 probe) after 10 h of fermentation in response to both substrates and FOS. In the absence of any carbon source (negative control), the total bacterial numbers decreased during fermentation, thus confirming the suitability of AcGGM and AcAGX as substrates for the metabolism of the human fecal microbiota. Comparisons of both substrates with FOS revealed that the total bacterial populations at the end of the fermentation were not significantly different (P Ͼ 0.05). Counts of bifidobacteria (probe BIF164) increased significantly after 5 h of fermentation in response to the positive control and to both AcGGM and AcAGX. At 24 h of fermentation, the highest increase (of 1.7 Ϯ 0.1 log 10 cells/ml) was observed in the samples supplemented with FOS, followed by increases of 1.1 Ϯ 0.2 and 1.1 Ϯ 0.3 log 10 cells/ml in AcGGM and AcAGX samples, respectively. Previous studies indicated that an increase of 0.5 to 1.0 log 10 cells/ml in the population of bifidobacteria elicits a major shift in the gut microbiota toward a healthier composition in infants and adults (47,48). Thus, both AcGGM and AcAGX can be considered bifidogenic under the conditions tested. Fermentation of AcGGM and AcAGX for 24 h increased the Bacteroides-Prevotella (probe BAC303) populations significantly (by 1.3 Ϯ 0.2 log 10 cells/ml), while FOS increased the population by 0.9 Ϯ 0.1 log 10 cells/ml. These results are in agreement with those obtained for low-molecular-weight XOS from oil palm empty fruit bunches and corn cobs (13,15). Similar results were also reported with wood-derived ␤-MOS (21). Previous studies demonstrated that bifidobacteria are equipped with efficient machinery for utilization of XOS (40); thus, the increase in the population could be attributed to the fermentation of the low-molecular-weight undecorated and acetyl-decorated XOS which have been shown to be present in the AcAGX mixture ( Fig. 2e and f). Alternatively, it is likely that the Bacteroides serve as primary degraders of the high-molecular-weight carbohydrates into shorter molecules; bifidobacteria can eventually utilize these degradation products thanks to their crossfeeding behavior (39).
While the populations of butyrogenic bacteria of the Faecalibacterium prausnitzii (probe FPRAU655) cluster and of propionate-producing bacteria belonging to clostridial cluster IX (probe PROP853) declined or remained stable throughout FOS fermentation, overall significant increases were detected with AcGGM (0.3 Ϯ 0.01 log 10 cells/ml for FPRAU655 and 0.6 Ϯ 0.2 log 10 cells/ml for PROP853) and AcAGX (0.8 Ϯ 0.3 log 10 cells/ml for FPRAU655 and 0.7 Ϯ 0.1 log 10 cells/ml for PROP853) between 0 and 24 h. In pure culture experiments, F. prausnitzii DSM 17677 was not able to degrade CGM, KGM, or arabinoxylans from oat and wheat to any extent (27). However, the increased numbers of F. prausnitzii cells in the batch culture experiment performed with AcGGM and AcAGX can be explained by the presence of strains equipped with machinery for utilization of these hemicelluloses. Indeed, we identified a cluster encoding enzymes involved in ␤-mannan metabolism in two F. prausnitzii strains whose genome were sequenced (Fig. S5). The cluster lacks a GH26 endomannanase, as would be required to hydrolyze the ␤-mannan backbone into smaller oligosaccharides, but it may allow growth of these F. prausnitzii strains through cross-feeding on ␤-MOS released by other colonic microorganisms. F. prausnitzii strain 2789STDY5834930, isolated in the United Kingdom from a fecal sample, possesses two ␤-xylosidases/ arabinosidases (ERS852542_01151/01152) that might support the growth on AXOS. It is likely that similar strains present in the fecal samples that we employed in this study are able to hydrolyze and ferment the XOS present in the AcAGX preparation. Aside from being a ubiquitous primary source of butyrate in the colon, F. prausnitzii is thought to regulate gut homeostasis and to have anti-inflammatory properties (49). Thus, stimulation of F. prausnitzii growth through AcGGM and AcAGX treatments could result in an overall beneficial effect on host health.
The counts of another member of the Actinobacteria group, i.e., the Atopobium cluster (probe ATO291), increased after 24 h of growth with AcGGM (0.8 Ϯ 0.3 log 10 cells/ml) and FOS (0.9 Ϯ 0.1 log 10 cells/ml), whereas AcAGX did not enhance its growth. These findings are in line with previous studies reporting that FOS promoted the growth of Atopobium, which, on the other hand, was unaffected by high-molecularweight XOS or AXOS (15,50). To date, no study has investigated the ␤-mannan fermentation capabilities of the Atopobium cluster; thus, it remains unclear whether the increase in this population seen with the AcGGM mixture is supported by the activity of mannolytic Atopobium strains.
No significant changes were detected at any time point in other bacterial populations, namely, Lactobacillus spp. (probe LAB158), Clostridium coccoides-Eubacterium rectale genus (probe EREC482), and the Clostridium histolyticum group (probe CHIS150). AcGGM and AcAGX supported a stable Roseburia sp. population throughout the fermentation, whereas a significant decrease was seen in the cultures supplemented with FOS after 5 h. Also, an overall decrease in numbers of the sulfate-reducing Desulfovibrio spp. was observed with all substrates tested.
Although the total number of bacteria decreased during fermentation in the absence of any carbon source, we observed an increase in the populations of clostridium cluster IX and Bacteroides after 10 h of fermentation and constant numbers of bifidobacteria and the Atopobium cluster bacteria over time (Fig. 4). Growth of the Bacteroides members is consistent with their ability to utilize peptide and yeast ␣-mannan/␤glucans, deriving from the peptone and yeast extract components of the basal medium as previously shown (28,51).
Organic acid analysis. Table 2 summarizes the amounts of the SCFAs acetate, propionate, and butyrate as measured by the use of a gas chromatograph analyzer equipped with a flame ionization detector (GC-FID) in the fermentation supernatants of the batch cultures over time. Total organic acid concentrations increased throughout the fermentation, and the increases were more pronounced in the cultures supplemented with the carbohydrates than in the negative-control samples (P Ͻ 0.05). Acetate was the dominant SCFA produced (65% to 75% of the total) for all substrates tested, with significant differences seen at all time points. The highest concentration was achieved in the cultures containing AcAGX, followed by those containing FOS and AcGGM. The increase in acetate levels is consistent with the higher abundance of bifidobacteria detected by FISH-FCM. Previous studies (14,15) have shown that fermentation of XOS resulted in a higher concentration of acetate, which is in agreement with the present findings. Acetate is considered the primary SCFA in colonic fermentation and is used for de novo lipogenesis once it enters the systemic circulation. In addition, the increase in acetate could result in a more acidic colonic pH, thus inhibiting the proliferation of pathogenic bacteria, which are generally sensitive to low pH values (52).
The production profile of propionate was similar to that of acetate, although it reached lower levels. This is in accordance with the significant enhancement in the Bacteroides-Prevotella and clostridial cluster IX populations, as these bacteria are known to be propionate producers (53). The propionate levels of AcGGM at 24 h were comparable to those measured for FOS; however, fermentation of AcAGX generated more propionate (1:0.31:0.095 mM [acetate/propionate/butyrate] ratio) than fermentation of AcGGM and FOS (Table 2). Propionic acid production has been reported to result in beneficial health effects on the host, including appetite regulation, lower levels of glucose-induced insulin secretion, and antiproliferative effects on liver cancer cells (54,55). Moreover, while acetate acts as a precursor for biosynthesis of blood lipids, propionate is able to inhibit this process. Thus, a low acetate/propionate ratio helps in regulating levels of blood lipids and potentially reducing cardiovascular disease risk (54).
The wood-derived substrates tested generated a major change in the level of butyrate production (P Ͻ 0.05). This is in good agreement with the results showing an increased abundance of F. prausnitzii, which is a major butyrate producer, in the samples supplemented with AcGGM and AcAGX. Interestingly, fermentation of AcGGM generated more butyrate (1:0.19:0.34 mM [acetate/propionate/butyrate] ratio) than fermentation of AcAGX and FOS at 24 h, indicating that that substrate promotes butyrogenic fermentation in particular ( Table 2). The ability of FOS to promote butyrate synthesis has been reported previously, and it has been linked to the conversion of acetate into butyrate by the gut microbiota (15,56). The production of butyrate by commensal bacteria has received much attention for its health-promoting functions (57). It serves as the main energy source for colonocytes (58) and exhibits anticarcinogenic, anti-inflammatory, and barrier-protective properties in the distal gut (59)(60)(61).
Fermentation of the negative control resulted in a limited but still significant increase of acetate, propionate and butyrate levels after 10 h of fermentation ( Table 2). Production of acetate and propionate could be linked to the increases in the levels of the Bacteroides and clostridial cluster IX populations (53). The gut microbiota is capable of converting some of the acetate into butyrate that thus does not originate from the butyrate-producing capacity in the fecal slurry (62).
Conclusions. Our results prove that AcGGM and AcAGX, which consist of oligosaccharides and polysaccharides with different DP and structure, can influence the com-  position of the gut microbiota. Both preparations displayed in vitro prebiotic activity, in terms of increasing the number of bifidobacteria, besides stimulating the growth of other beneficial bacteria, including the Bacteroides-Prevotella, clostridial cluster IX, and F. prausnitzii groups. Metabolite production correlated well with changes in bacterial populations; acetate was the prevalent SCFA released, followed by butyrate for AcGGM and propionate for AcAGX. The total levels of SCFAs produced by fermentation of AcGGM and AcAGX were comparable with those seen with the established prebiotic FOS.
Although further in vivo investigations should be conducted to ultimately confirm the beneficial health effect for the host, AcGGM and AcAGX from lignocellulosic biomass can be considered new potential prebiotics for human consumption.
Sugar composition analysis. The neutral sugar composition of AcGGM and AcAGX was analyzed using HPAEC (Dionex ICS-3000) (CarboPac PA1 columns [2 ϫ 50 mm and 2 ϫ 250 mm]; 30°C) coupled with a PAD detector (HPAEC-PAD). Samples (5 mg/ml) were heated at 121°C for 1 h with 4% H 2 SO 4 to induce hydrolysis. Hydrolysates were diluted with 15 mM NaOH to obtain an optimal range of concentrations for the quantification. Neutral sugars were eluted with an isocratic concentration of 1 mM NaOH at a flow rate of 0.25 ml/min for 35 min. The uronic acid quantification was adapted from reference 63. Hydrolysis was initiated in silicate tubes with a mixture of 0.3 ml of polysaccharides (1 mg/ml) and 0.3 mg sodium boric acid solution (2% boric acid and 3% NaCl dissolved in water). Samples were incubated at 70°C for 40 min with 5 ml of concentrated sulfuric acid. The hydrolysates were cooled for 20 min. The colorimetric reaction was performed with the addition of 0.2 ml dimethylphenol (0.1% [wt/vol] in glacial acetic acid) at room temperature (RT) for 15 min and kept on ice until the absorbance was read. The total uronic concentration was proportional to the absorbance difference at wavelengths of 450 nm and 400 nm using glucuronic acid as a standard. The quantification of acetic acid was initiated with the saponification of the acetyl groups; this was done by incubating the carbohydrates (10 mg/ml) overnight at 4°C with 0.1 M NaOH to release bound acetyl groups. The analysis was performed using a Dionex Ultimate 3000 high-performance liquid chromatography (HPLC) system with 30 min of isocratic elution (5 mM H 2 SO 4 ) on a Rezex ROA-organic acid-Hϩ column (7.8 ϫ 300 mm) coupled to a Carbo-H4 security guard cartridge (3.0-mm internal diameter). Acetic acid was detected at the wavelength of 210 nm.
Analysis of sugar monosaccharides and oligosaccharides by HPAEC-PAD. Samples were analyzed by HPAEC-PAD using a Dionex ICS-3000 chromatographic system operated using Chromeleon software version 7 (Dionex, Thermo Scientific). The system was equipped with a CarboPac PA1 analytical column (Dionex, Thermo Scientific) (2 ϫ 250 mm) in combination with a CarboPac PA1 guard column (2 ϫ 50 mm), and it was run at a flow rate of 0.25 ml/min. The elution conditions for the analysis of manno-oligosaccharides and manno-polysaccharides were 0 to 9 min 0.1 M NaOH; 9 to 35 min 0.1 M NaOH with a 0 to 0.3 M sodium acetate (NaOAc) gradient; 35 to 40 min 0.1 M NaOH with 0.3 M NaOAc; and 40 to 50 min 0.1 M NaOH. The elution conditions for the xylo-oligosaccharides and xylopolysaccharides were 0 to 10 min 0.1 M NaOH with a 0 to 0.1 M NaOAc gradient; 10 to 35 min 0.1 M NaOH with a 0.1 to 0.3 M NaOAc gradient; 35 to 40 min 0.1 M NaOH with a 0.3 to 1 M NaOAc gradient; and 40 to 50 min 0.1 M NaOH. Commercial manno-oligosaccharides and xylo-oligosaccharides (DP, 2 to 6) from Megazyme were used as standards.
MALDI-TOF. MALDI-TOF analyses were conducted using an Ultraflex MALDI-TOF/TOF instrument (Bruker Daltonics, Germany) equipped with a 337-nm-wavelength nitrogen laser. All measurements were performed in positive mode. Data were collected from 100 laser shots using the lowest energy level necessary to obtain sufficient spectral intensity. The mass spectrometer was calibrated with a mixture of manno-oligosaccharides or xylo-oligosaccharides (DP, 2 to 6) obtained from Megazyme. For sample preparation, 1 l of sample solution was mixed with 2 l of matrix (9% 2,5-dihydroxybenzoic acid [DHB]-30% acetonitrile [vol/vol]), directly applied on a MTP 384 target plate (Bruker Daltonics, Germany), and dried under a stream of warm air.
Organic acid analysis. Acetic, propionic, and butyric acid concentrations were quantified using a gas chromatograph analyzer equipped with a flame ionization detector (GC-FID) following the method described by Richardson et al. (67) with ethyl butyric acid used as an internal standard. A GC analyzer (Agilent/HP 6890) equipped with an HP-5MS column (30 m ϫ 0.25 mm) with a 0.25-m-pore-size coating (cross-linked 5% phenyl methylpolysiloxane; Hewlett Packard, United Kingdom) was employed. Helium was used as the carrier gas at a flow rate of 1.7 ml/min (head pressure, 133 KPa). The oven initial temperature was set at 63°C followed by a temperature ramp of 15°C/min to 190°C and was then held constant for 3 min. A split ratio of 100:1 was used. Quantification of SCFA in the chromatograms was confirmed based on the retention times of the respective commercial SCFA standards (Sigma-Aldrich, United Kingdom) at concentrations ranging between 2 to 10 mM.
Statistical analysis. A paired Student's t test was used to evaluate significant differences among bacterial numbers and organic acid concentrations at the time of inoculation and the different sampling points. Data are presented as means Ϯ standard deviations. Differences were considered statistically significant at P values of Ͻ0.05.