Article

Cite This: J. Agric. Food Chem. 2019, 67, 7755−7764 pubs.acs.org/JAFC

Utilization of Complex Pectic Polysaccharides from New Zealand ( tetragonioides and Corynocarpus laevigatus)by Gut Bacteroides Species Manuela Centanni,‡ Susan M. Carnachan,† Tracey J. Bell,† Alison M. Daines,† Simon F. R. Hinkley,† Gerald W. Tannock,‡,§,∥ and Ian M. Sims*,†,∥

† Ferrier Research Institute, Victoria University of Wellington, 69 Gracefield Road, Lower Hutt 5040, New Zealand ‡ ∥ Department of Microbiology and Immunology and Microbiome Otago, Department of Microbiology and Immunology, University of Otago, Post Office Box 56, Dunedin 9054, New Zealand § Riddet Institute Centre of Research Excellence, Palmerston North 4442, New Zealand

*S Supporting Information

ABSTRACT: Pectic polysaccharides from New Zealand (NZ) (Tetragonia tetragonioides) and karaka berries ( Corynocarpus laevigatus) were extracted and analyzed. NZ spinach polysaccharides comprised mostly homogalacturonan (64.4%) and rhamnogalacturonan I (5.8%), with side chains of arabinan (8.1%), galactan (2.2%), and type II arabinogalactan (7.1%); karaka berry polysaccharides comprised homogalacturonan (21.8%) and rhamnogalacturonan I (10.0%), with greater proportions of side chains (arabinan, 15.6%; galactan, 23.8%; and type II arabinogalactan, 19.3%). Screening of gut commensal Bacteroides showed that six were able to grow on the NZ spinach extract, while five were able to grow on the karaka berry extract. Analysis of the polysaccharides remaining after fermentation, by size-exclusion chromatography and constituent analysis, showed that the Bacteroides species that grew on these two substrates showed preferences for the different pectic polysaccharide types. Our data suggest that, to completely degrade and utilize the complex pectin structures found in plants, members of Bacteroides and other bowel bacteria work as metabolic consortia. KEYWORDS: NZ spinach, karaka berry, pectic polysaccharides, Bacteroides, gut microbiota

■ INTRODUCTION residues with side chains of arabinan, galactan, or arabinoga- The majority of dietary fiber originates from the cell walls of lactan (types I and II). HG and RG-I, together with the various , vegetables, and grains in the human diet.1 The neutral side chains, make up the majority of pectic polysaccharides (pectic polysaccharides, hemicelluloses, and polysaccharides. Rhamnogalacturonan II (RG-II) is a low- abundance pectic polysaccharide that accumulates in many cellulose) that make up these cell walls pass through the 10−12 small intestine and reach the colon largely unaltered, where processed foods as a result of its resistance to degradation. they are completely or partially fermented by the trillions of This structurally complex molecule is comprised of a branched HG backbone with conserved side chains containing a variety bacterial cells (the gut microbiota) that inhabit the colon. As 9 demonstrated by the epidemiological studies conducted by of and glycosyl linkage types.

Downloaded via VICTORIA UNIV OF WELLINGTON on June 30, 2020 at 03:00:01 (UTC). Burkitt, Cleave, Campbell, Painter, Trowell, and Walker in Members of the human gut microbiota belonging to the 2 See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. Africa decades ago (reviewed by Cummings and Engineer), genus Bacteroides are of particular interest with regards to the dietary fiber is clearly important in regulating intestinal transit hydrolysis and fermentation of pectic polysaccharides, fl although certain Firmicutes also show some capacity to utilize time and in uences the prevalence of some non-communicable 13,14 diseases of humans (bowel transit/constipation, diverticulosis, pectin. The genomes of the Bacteroides species encode a − 15 and bowel cancer).3 6 Much more research needs to be wide range of glycosyl hydrolases and polysaccharide lyases. directed to solving the question of how the gut microbiota Two recent papers have studied in detail the enzyme pathways “works” to degrade and ferment the diversity of dietary fibers involved in the degradation of pectic polysaccharides by the 16,17 18−21 present in food.7 Then, dietary interventions could provide Bacteroides species. These and other reports used fi rational and reliable remedies for gut diseases and conditions puri ed pectic polysaccharide components as substrates. Pectic where gut dysbiosis may have an etiological role.8 polysaccharides with varying degrees of methylesterification, Pectic polysaccharides are a highly diverse family of acidic including homogalacturonan, galactan, arabinan, and arabino- polysaccharides representing the major “soluble” fiber in many galactans, are available commercially, while RG-I has been and vegetables. They are comprised of backbones of homogalacturonan (HG), a linear polymer of →4]-α-D-GalpA- Received: April 17, 2019 [1→, and rhamnogalacturonan I (RG-I), which is comprised of Revised: June 18, 2019 a repeating disaccharide of →4]-α-D-GalpA-[1→2]-α-L-Rhap- Accepted: June 19, 2019 [1→.9 The RG-I backbone is branched at the rhamnosyl Published: June 19, 2019

© 2019 American Chemical Society 7755 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article purified from Arabidopsis mucilage22 and RG-II has been lyophilized fractions were resuspended in water, filtered under purified from red wine23 or from apple juice.17 However, pectic vacuum to remove any remaining insoluble material, and lyophilized. polysaccharides do not exist as distinct entities in plant cell Bacteroides Growth Assay on Pectic Polysaccharide Substrates. The growth of type strains of 15 Bacteroides species walls but form complex covalently linked macromolecular 30,31 commonly found in the human gut microbiota (>90% of humans; structures, with backbones of HG and RG domains, with ff 9,24 Table S1 of the Supporting Information) on the di erent pectic simple and/or highly complex side chains. polysaccharide substrates was determined using the procedure as We showed recently that 10 of 15 Bacteroides species tested described previously.32 Briefly, cultures of the different bacterial were able to degrade and utilize the water-soluble pectic strains were grown under anaerobic conditions for 18 h at 37 °Cin polysaccharide fraction obtained from feijoa fruit.25 We now the appropriate medium following DSMZ or ATCC protocols. To extend this work to investigate the utilization of pectic assess their ability to utilize different substrates, pure cultures of each 32,33 polysaccharides from two New Zealand (NZ) native plants. species were used to inoculate (1%, v/v) basal medium −1 fi NZ spinach (Tetragonia tetragonioides, ) is a leafy containing 2 g L of the speci c pectic polysaccharide. Three plant used by Captain James Cook R.N. (who was the first to technical replicates and two biological replicates were carried out for 26 each species studied. Optical density (A600) was measured after 24 chart the shoreline of New Zealand) as an antiscorbutic. NZ and 48 h of anaerobic incubation at 37 °C. Spent culture supernatants spinach is grown commercially as a specialty vegetable crop. were collected after centrifugation of samples at 14500g for 5 min and Karaka (Corynocarpus laevigatus) is an endemic NZ tree, which stored at −80 °C for subsequent analysis. Unpaired t bears large orange fruits in autumn. The pulpy flesh and tests were performed to compare the optical densities of bacterial kernels of the berries of karaka were eaten by NZ Maori,̅ cultures grown in the presence and absence of each carbohydrate although the kernels required extensive preparation to remove substrate, using GraphPad Prism, version 7.0b (GraphPad Software, toxic nitropropanoyl glucopyranoses.26 Polysaccharides from La Jolla, CA, U.S.A.; www.graphpad.com), considering a p value of karaka berries have not been studied before, and there is only <0.05 to be statistically significant. fi one report on the composition and structure of a water-soluble Solid-Phase Extraction (SPE) and Ultra ltration of Spent Culture Media. Prior to high-performance size-exclusion chroma- polysaccharide from NZ spinach.27 This polysaccharide was fl tography (HPSEC) and constituent sugar analysis, the supernatants reported to have anti-in ammatory properties, and other, from the spent culture media were passed through SPE cartridges to uncharacterized polysaccharide extracts from this species have remove hydrophobic compounds from the samples.32 Prior to been shown to have potential antidiabetic and antitumor constituent sugar analysis, the solid-phase-extracted samples were properties.28,29 Polysaccharides from NZ spinach and karaka passed through centrifugal ultrafiltration devices (MWCO of 30 kDa) berries have not been investigated in relation to the action of to remove lower molecular weight components. gut microbes. This paper presents new data on the Constituent Sugar Analysis. The constituent sugar composi- tions were determined by high-performance anion-exchange chroma- composition and structure of polysaccharides from native ° New Zealand plants and demonstrates that pectic poly- tography (HPAEC) after hydrolysis [3 N methanolic HCl at 80 C for 18 h, followed by 2.5 M trifluoroacetic acid (TFA) at 120 °C for 1 saccharides with different structures are differentially utilized 34 h] of the polysaccharides to their component monosaccharides. by common gut Bacteroides species. Aliquots of the hydrolysate solutions were diluted with water (20−50 μg/mL) and analyzed on a CarboPac PA-1 (4 × 250 mm) column ■ MATERIALS AND METHODS equilibrated in NaOH (25 mM) and eluted with a simultaneous −1 Materials. Ripe, bright orange karaka berries (C. laevigatus) were gradient of NaOH and sodium acetate (1 mL min ). The sugars fi collected from a local reserve (41.2172° S, 174.8790° E) in January were identi ed from their elution times relative to standard sugar 2015. Fresh NZ spinach (T. tetragonioides) was obtained from a local mixes, quantified from response calibration curves of each sugar, and farm (Wairarapa Eco Farm, Greytown, New Zealand; 41.1211° S, expressed as weight percent anhydro sugar and molar percent. 175.4097° E). Other pectic polysaccharide substrates were purchased Glycosyl Linkage Analysis. Glycosyl linkage compositions of the from Sigma-Alrich, Auckland, New Zealand (polygalacturonic acid NZ spinach and karaka berry extracts were determined by methylation and gum arabic) or Megazyme International Ireland, Bray, Co., analysis, following reduction of uronic acid residues to their 35,36 Wicklow, Ireland (potato galactan and sugar beet arabinan). dideuterio-labeled neutral sugars as described by previously. Extraction of Pectic Polysaccharides. The kernels of the karaka The two samples (10 mg) were dissolved in 500 mM imidazole−HCl − berries were removed and discarded, and pulp and skins (∼25 g of (10 mL, pH 8.0) and reduced with NaBD4, dialyzed (MWCO of 6 8 fresh weight) were extracted with distilled water (250 mL), once at kDa) for 24 h against distilled water, and freeze-dried. The free uronic 100 °C for 10 min and then at 80 °C for 2 h. After cooling to 4 °C, acids were then activated by the addition of 1-cyclohexyl-3-(2- the combined extracts were centrifuged (3000g for 15 min at 10 °C). morpholinoethyl)-carbodiimide-metho-p-toluenesulfonate (400 μL, −1 The supernatant was concentrated under reduced pressure to about 500 mg mL ) and reduced overnight with either NaBD4 or 100 mL, mixed with three volumes of ethanol, and left at 4 °C NaBH4. The carboxyl-reduced samples were dialyzed against distilled overnight. The precipitated material was collected by centrifugation, water and freeze-dried. The effectiveness of the carboxyl reductions redissolved in distilled water, dialyzed against distilled water was checked by analysis of the constituent sugar composition by [molecular weight cut-off (MWCO) of 12−14 000 Da], and freeze- HPAEC and, if needed, subjected to a further reduction, following dried (yield of 0.8 g). carbodiimide activation, until the uronic acid content was reduced to The leaves and stems of fresh NZ spinach (total of 1310 g in three <5% (w/w). batches) were blended in distilled water using a kitchen blender and Samples (0.5 mg, in duplicate) were methylated using the method then extracted with distilled water (2.5 L each batch), once at 100 °C of Ciucanu and Kerek.37 Methylated polysaccharides were then ° for 10 min and then twice at 80 °C for 3 h. After cooling to room hydrolyzed (2.5 M TFA for 1 h at 121 C), reduced (1 M NaBD4 for temperature, the extracts were centrifuged (9180g for 30 min at 15 18 h at 25 °C), and acetylated. The partially methylated alditol acetate °C). The supernatant was mixed with three volumes of ethanol and derivatives produced were separated by gas chromatography (GC) on left at 4 °C overnight. The precipitated material was collected by a BPX90 fused silica capillary column (SGE Analytical Science, centrifugation (3000g for 20 min at 8 °C), redissolved in distilled Ringwood, Victoria, Australia; 30 m × 0.25 mm inner diameter, with a water, and freeze-dried (yield of 7.9 g; 0.6% fresh weight). 0.25 μm film thickness) with the GC oven programmed from 80 °C − To remove starch and , the extracts were treated with (held for 1 min) to 130 °C at a rate of 50 °C min 1 and then to 230 − pancreatic α-amylase/amyloglucosidase and protease (subtilisin A °C at a rate of 3 °C min 1 and detected by mass spectrometry (MS) from Bacillus licheniformis, Megazyme) as described by ref 25. The using a Agilent 5973 mass selective detector (MSD). Identifications

7756 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

a Table 1. Constituent Sugar Compositions and Molecular Weights of Pectic Polysaccharide Substrates

NZ spinach karaka berry polygalacturonic acid potato galactan sugar beet arabinan gum arabic sugar (μg/mg) (μg/mg) (μg/mg) (μg/mg) (μg/mg) (μg/mg) fucose 5.3 5.8 bbb10.5 rhamnose 37.8 44.0 11.9 56.0 52.5 117.5 arabinose 80.2 95.2 3.2 11.8 432.6 221.9 xylose 2.3 2.1 bb0.8 b galactose 64.6 180.9 20.2 699.0 125.6 550.3 glucose 4.1 11.3 0.9 3.9 bb mannose 2.1 7.1 bbbb galacturonic acid 454.0 278.4 668.5 78.1 58.6 b glucuronic acid 12.5 6.2 bbb121.7 total sugar 662.9 631.0 704.8 848.9 670.1 900.1 Molecular Weight (kDa) d Mw nd nd 40.6 583 79.9 604

Mn nd nd 21.4 237 54.2 265

Mw/Mn nd nd 1.9 2.5 1.5 2.3 aValues are the averages of duplicate analyses. bNot detected. dnd = not determined. were based on peak retention times and electron impact mass spectra Table 2. Glycosyl Linkages and Polysaccharide 38 compared to partially methylated alditol acetate (PMAA) standards. Composition of NZ Spinach and Karaka Berry Water- a HPSEC. Molecular weight distributions were determined using Soluble Extracts size-exclusion chromatography coupled with multiangle laser light − scattering (SEC−MALLS). Polysaccharide substrates (2 mg mL 1) linkage composition and culture supernatants after SPE (1 mg mL−1) were dissolved in 0.1 (mol %) M NaNO3, allowed to hydrate fully by standing at room temperature polysaccharide deduced linkage NZ spinach karaka berry overnight, and centrifuged (14000g for 10 min) to clarify. The soluble μ arabinan 5-Araf 4.4 8.3 material (100 L) was injected and eluted with 0.1 M NaNO3 (0.5 −1 ° 3,5-Araf 1.6 3.2 mL min at 60 C) from three columns (TSK-Gel G5000PWXL, × 2,3,5-Araf 0.4 0.3 G4000PWXL, and G3000PWXL, 300 7.8 mm, Tosoh Corp., Tokyo, Japan) connected in series. The eluted material was detected using a terminal-5-Araf 2.4 3.8 variable wavelength detector (280 nm), a multiangle laser light 8.8 15.6 scattering detector (SLD7000, Polymer Standards Service GmbH, galactan 4-Galp 2.2 22.2 Mainz, Germany), and a refractive index monitor. The data for 3,4-Galp b 0.7 molecular weight determination were analyzed using PSS WinGPC 4,6-Galp b 0.1 Unichrom (version 8.2.1, Polymer Standards Service GmbH) using a − terminal-5-Araf b 0.8 dn/dc of 0.145 mL g 1. The system was also calibrated with a series of pullulan molecular weight standards (6−850 kDa, Shodex, Showa 2.2 23.8 Denko K.K. Tokyo, Japan). type II arabinogalactan 3-Galp 0.7 2.2 Nuclear Magnetic Resonance (NMR) Spectroscopy. The 6-Galp 0.8 0.5 spectra of the NZ spinach and karaka berry extracts (∼20 mg of 3,6-Galp 2.8 8.3 polysaccharide in 0.7 mL of D2O) were recorded on a Bruker Avance terminal-Rhap 1.2 1.6 DPX-500 spectrometer with a 5 mm probe at 500.13 (1H) and 125.77 terminal-5-Araf 1.1 5.6 13 1 13 ( C) MHz and 25 °C. The H and C chemical shifts were terminal-GlcpA 0.5 1.1 1 measured relative to an internal standard of Me2SO ( H, 2.71 ppm; 7.1 19.3 13 C, 39.5 ppm). Assignments were made from heteronuclear single rhamnogalacturonan I 2-Rhap 1.3 2.3 quantum coherence (HSQC) correlation spectroscopy (COSY) 2,4-Rhap 1.6 2.7 experiments and by comparing the spectra to published data. 3,4-GalpA 0.6 b 4-GalpA 2.3 5.0 ■ RESULTS 5.8 10.0 Composition and Structure of Pectic Polysaccharide homogalacturonan 4-GalpA 63.0 20.9 Extracts. The constituent sugar compositions of the NZ terminal-GalpA 1.4 0.9 spinach and karaka berry extracts contained mostly galactur- 64.4 21.8 onic acid, rhamnose, arabinose, and galactose, consistent with total 88.3 90.5 the presence of pectic polysaccharides (Table 1). Size- othersc 11.7 9.5 exclusion chromatography (SEC) of the water-soluble pectic aValues are the averages of duplicate analyses. bNot detected. cOther polysaccharide extracts from NZ spinach and karaka berries minor linkages. showed complex elution profiles, with material eluting across the full elution range of the columns (see Figure 3). By substrates were generally in good agreement with the comparison to pullulan molecular weight standards, the NZ specifications of the manufacturer (Table 1). spinach extract showed a minor peak at 18.3 min (∼480 kDa) Glycosyl linkage analysis of the pectic polysaccharides from and a major peak at 22 min (∼100 kDa) and the karaka berry NZ spinach and karaka berries, following reduction of uronic extract showed peaks at 18.6 min (∼550 kDa), 22 min (∼100 acids to their equivalent dideuterio neutral sugars, showed kDa), and 25 min (∼24 kDa). The constituent sugar partially methylated alditol acetate derivatives that corre- compositions and molecular weights of the commercial sponded to linkages typical of homogalacturonan (HG) and

7757 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

Figure 1. Selected regions of the HSQC spectra of pectic polysaccharide-rich extracts from (A) NZ spinach and (B) karaka berries. rhamnogalacturonan (RG) backbones with neutral pectic side linkages were grouped into different pectic polysaccharide chains (Table S2 of the Supporting Information). The various classes based on knowledge of typical plant cell wall

7758 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

Figure 2. Schematic representation of pectins from NZ spinach and karaka berries showing the HG and RG-I backbones with arabinan, type I (arabino)-galactan, and type II arabinogalactan side chains.

36 41 polysaccharide structures (Table 2). On the basis of these 1→4)-β-D-Galp-(1→. A signal at 4.50/103.5 ppm was estimations, the NZ spinach extract comprised about two- assigned to H-1/C-1→3)-β-D-Galp-(1→ and →3,6)-β-D- thirds HG; the degree of methyesterification of the uronic Galp-(1→ that were present in both extracts.43 The signals acids was ∼50%, as determined by linkage analysis (data not at 5.10/99.7 and 4.76/70.5 ppm in the spectrum on the karaka shown).35 RG-I (5.8%) with side chains of arabinan, galactan, berry extract were assigned to H-1/C-1 and H-5/C-5 of and type II arabinogalactan (AG) accounted for about a unesterified →4)-α-D-GalpA-(1→, respectively. Despite the quarter of the polysaccharides. Other minor linkages, such as higher GalpA content of NZ spinach pectic extract, anomeric 3,4-Fucp,2,3,4-Rhap,3′-apiose, 2-GalpA, and 2,4-GalpA signals for these residues (H-1/C-1, 4.95/101.0 and 5.08/ (Table S2 of the Supporting Information), indicated the 102.1 ppm) appeared at much lower intensity, possibly as a presence of small amounts of RG-II. This composition was result of the low mobility of these residues.44,45 Signals in the similar to that reported for the water-soluble extract of feijoa spectrum of this extract at 4.95−5.10/71.2, 3.80/53.4 (O- fruit determined in our previous study.25 In contrast, the methyl), and 1.99−2.17/20.7−21.2 ppm (O-acetyl) indicated karaka berry extract comprised a much lower proportion of that the →4)-α-D-GalpA-(1→ residues were present as both 6- HG (with only 5% methylesterification), with RG-I (10%) and O-methyl-esterified and 3-O-acetylated derivatives.40,43,46,47 neutral side chains accounting for more than two-thirds of the The O-methyl and O-acetyl signals were much less intense in polysaccharides. Again, minor linkages indicated the presence the karaka berry pectic extract. Signals at 1.25/17.2 ppm, of small amounts of RG-II. corresponding to H-6/C-6 of α-L-Rhap, are present in both the The composition and structure of the NZ spinach and NZ spinach and karaka berry extracts, but corresponding H-1/ karaka berry pectic extracts were also examined by NMR C-1 signals were not observed, again possibly as a result of the spectroscopy. The spectra were partially assigned on the basis low mobility of these residues.44,45 of the 1H, 13C, and HSQC experiments and by comparison to On the basis of these data, Figure 2 shows schematic published spectra of similar molecules (Figure 1). The signals representations of the major features of NZ spinach and karaka at 5.08−5.24/107.6−109.7 ppm were assigned to H-1/C-1 α- berry pectic polysaccharides. L-Araf residues, and those at 4.50−4.62/103.5−104.8 ppm Growth of Bacteroides Species and Substrate Con- 39−42 were assigned to H-1/C-1 β-D-Galp residues. The greater sumption. The growth of 15 Bacteroides species, expressed as intensity of these signals in the spectra from the karaka berry the ratio between A600 of bacterial cultures grown in the pectic extract was consistent with the higher proportion of presence of pectic polysaccharide and A600 of bacterial cultures these linkages observed (Table 2). A signal at 4.63/104.8 ppm grown in the absence of a carbohydrate source (A600 ratio), in that was present in the spectrum of the karaka berry extract but medium containing either polygalacturonic acid, potato not in that of the NZ spinach extract was assigned to H-1/C- galactan, sugar beet arabinan, or type II arabinogalactan

7759 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

Table 3. Preference of Bacteroides Species for NZ Spinach or cellulosilyticus and B. intestinalis, as evidenced by a more Karaka Berry Pectic Polysaccharide Substrates pronounced reduction in the area under the peaks (17−27 min) and, hence, the consumption of more of the 24 h 48 h polysaccharide than NZ spinach polysaccharides (Figure 3). A600 A600 a b a b Constituent sugar analysis showed that more of the total sugars ratio p value ratio p value in the karaka berry extract were consumed in comparison to NZ Spinach the NZ spinach polysaccharides (Table 4). Such differences Bacteroides cellulosilyticus DSM 1.86 <0.001 1.80 0.001 were not apparent for the three Bacteroides species, which 14838T showed a preference for the NZ spinach extract. The apparent Bacteroides intestinalis DSM 1.47 0.006 1.69 <0.001 fi 17393T changes in the SEC pro les during the growth of these species Bacteroides ovatus ATCC 8483T 1.99 0.001 2.00 <0.001 were similar for both the NZ spinach and karaka berry Bacteroides pectinophilus ATCC 10.20 0.004 15.52 0.001 polysaccharides (Figure 3). Constituent sugar analysis showed 43243T that similar amounts of the total sugars in both extracts were Bacteroides stercoris ATCC 2.00 0.001 1.42 0.061c T consumed (Table 4). 43183 Analysis of the major sugars in the NZ spinach and karaka Bacteroides finegoldii DSM 2.09 0.002 1.96 0.001 17565T berry extracts, both before and after fermentation, suggested that the different Bacteroides species consumed the various Karaka Berry ff Bacteroides cellulosilyticus DSM 2.29 0.030 2.84 <0.001 pectic polysaccharide components to di erent extents (Table 14838T 4). Thus, B. ovatus, B. stercoris, and B. pectinophilus showed Bacteroides intestinalis DSM 1.19 0.452c 2.34 0.002 similar patterns of sugar utilization after 48 h of growth on 17393T both substrates, and B. finegoldii showed a similar pattern of Bacteroides ovatus ATCC 8483T 1.44 0.033 1.32 0.111c sugar consumption to these three species on the NZ spinach Bacteroides pectinophilus ATCC 2.07 0.117c 3.76 0.001 T extract. They each consumed a high proportion of GalA 43243 present in both extracts, consistent with the consumption of Bacteroides stercoris ATCC 1.58 0.006 1.06 0.671c 43183T HG; each of these species grew on commercial polygalactur- Bacteroides finegoldii DSM dddd onic acid, and SEC showed a clear degradation of the 17565T polysaccharide (Table S3 of the Supporting Information). a A600 ratio = A600 of cultures grown in the presence of pectic These species consumed little of Gal present in the NZ spinach polysaccharides versus A600 of cultures grown in the absence of a extract but more than half in the karaka berry extract. The carbohydrate source. bp values of <0.05 were considered statistically linkage analysis data showed that Gal in the karaka berry fi c fi d signi cant. The A600 ratio was not signi cant. No growth. extract was mostly associated with galactan, whereas in the NZ spinach extract, it was mostly associated with type II AG; these (gum arabic) was, in general, similar to that reported by others four species did not grow on gum arabic, a type II AG (Table (Table S3 of the Supporting Information).16,48 However, only S3 of the Supporting Information). There was also a greater Bacteroides cellulosilyticus, Bacteroides finegoldii, Bacteroides proportion of Ara and Rha used when B. ovatus, B. stercoris, intestinalis, Bacteroides ovatus, Bacteroides pectinophilus, and and B. pectinophilus were growing on the karaka berry extract, fl ff Bacteroides stercoris showed growth on the pectic polysacchar- which may re ect the di erent pectic polysaccharide ide fractions from NZ spinach or karaka berries (Table 3). The components that these sugars are associated with or their growth of B. intestinalis and B. pectinophilus was slow and gave detailed structure. fi fi B. cellulosilyticus and B. intestinalis used much less of the insigni cant A600 ratios after 24 h but showed signi cant GalA present than the other four species, indicating that less of growth after 48 h. Conversely, the A600 ratios at 48 h for B. stercoris grown on either extract and B. ovatus grown on karaka HG present was consumed. When these two species were berry extract were not significant, indicating possible lysis of grown on polygalacturonic acid, SEC analysis of the spent the cells. Interestingly, B. finegoldii grew on the NZ spinach culture medium showed much less degradation of the substrate extract but not on the karaka berry extract, whereas the other compared to B. ovatus, B. stercoris, and B. pectinophilus, fi five species grew on both extracts. Each of these six species was although it was similar to that observed for B. negoldii (Table able to grow on polygalacturonate and galactan, and all but B. S3 of the Supporting Information). Utilization of Gal and Ara pectinophilus grew on arabinan; however, only B. cellulosilyticus in the karaka berry extract by B. cellulosilyticus and B. intestinalis − was able to grow on gum arabic. To confirm the utilization of was similar (55 82%), but B. intestinalis used much less of Gal the substrates, the cell-free supernatants were examined by (14%) from the NZ spinach extract compared to the karaka SEC for evidence of degradation of the polysaccharides (Figure berry extract, with levels comparable to the other four species. 3), and changes in the constituent sugar compositions of fractions retained by 30 kDa membranes after 48 h of growth ■ DISCUSSION were analyzed (Table 4). The water-soluble polysaccharides extracted from the leaves On the basis of our previous criteria of the A600 ratio being at and stems of NZ spinach and from the pulp and skins of karaka least 1.3 times greater on one substrate compared to the berries have not been reported previously. Both of the extracts other,25 B. cellulosilyticus and B. intestinalis showed a preference were rich in pectic polysaccharides, but these had contrasting for the karaka berry extract and B. ovatus, B. pectinophilus, and structures (Figure 2). The pectic polysaccharides from NZ B. stercoris showed a preference for the NZ spinach extract. spinach comprised about 70% backbone residues and 18% The preference for the karaka berry extract by B. cellulosilyticus side-chain residues; the backbone residues were mostly and B. intestinalis was supported by the SEC and sugar analysis homogalacturonan (>90%), with about 8% rhamnogalactur- data. The SEC chromatograms showed greater apparent onan. The side chains comprised arabinan (48.6%), type II degradation of the karaka berry polysaccharides by both B. arabinogalactan (39.2%), and galactan (12.1%). In contrast,

7760 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

Figure 3. SEC refractive index chromatograms (black, uninoculated; red, 24 h; and blue, 48 h) of pectic polysaccharide-rich extracts from NZ spinach or karaka berries from Bacteroides cultures. the karaka berry pectic polysaccharides comprised about 32% The microbiota inhabiting the colon degrades a large backbone residues and almost 60% side-chain residues; about proportion of the dietary fiber contained in foods consumed 70% of the backbone residues were homogalacturonan, and by humans. Members of the genus Bacteroides are particularly fi about 30% were rhamnogalacturonan. The side chains were well-adapted to consuming dietary ber polysaccharides, and accumulated evidence from recent studies indicates that they mostly galactan (40.5%) and type II arabinogalactan (32.9%), have varying capacities with respect to utilization of plant together with arabinan (26.6%). Our data indicate that the polysaccharides. This current study shows that, while many GalA residues of the pectic polysaccharides from NZ spinach Bacteroides species were able to grow on specific, purified fi are more methyl-esteri ed than those from karaka berry; NMR pectic polysaccharide components (arabinan, galactan, and spectroscopy also indicates that they have a greater degree of polygalacturonic acid), a more limited number grew on natural O-acetylation. pectic polysaccharide complexes that are found in fruit and

7761 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

Table 4. Monosaccharide Compositions of the Pectic Polysaccharide-Rich Substrates from NZ Spinach and Karaka Berries Following Fermentation for 48 h (Unless Stated Otherwise) with Bacteroides Species

B. ovatus B. stercoris B. pectinophilus B. finegoldii B. cellulosilyticus B. intestinalis uninoculated 48 h (μg/mL) 48 h (μg/mL) 24 h (μg/mL) 48 h (μg/mL) 48 h (μg/mL) 48 h (μg/mL) 48 h (μg/mL) (μg/mL) (% used) (% used) (% used) (% used) (% used) (% used) (% used) NZ Spinach Ara 142 138 (3) 141 (tr) 135 (5) 140 (1) 131 (8) 51 (64) 54 (62) Gal 180 149 (17) 154 (14) 165 (8) 174 (4) 151 (16) 82 (55) 154 (14) GalA 790 192 (76) 172 (78) 255 (68) 215 (73) 284 (64) 575 (27) 590 (25) Rha 59 52 (12) 61 (−)60(−)61(−) 55 (7) 58 (1) 52 (12) totala 1398 637 (54) 663 (52) 812 (42) 792 (43) 763 (54) 902 (36) 1023 (27) Karaka Berry Ara 141 122 (23) 108 (13) 128 (9) 88 (38) 52 (63) 64 (55) Gal 467 206 (56) 196 (58) 424 (9) 290 (38) 86 (82) 199 (57) GalA 373 71 (60) 85 (77) 121 (68) 95 (75) 216 (42) 205 (45) Rha 58 45 (23) 51 (13) 58 (−) 47 (20) 45 (23) 51 (13) totala 1292 567 (56) 557 (57) 944 (27) 681 (47) 514 (60) 659 (49) aIncludes other sugars not shown here. vegetables. The Bacteroides species, which grew on either the bacterial activities and perhaps specific abundances of bacterial NZ spinach or the karaka berry pectic extracts, showed species with the potential to correct intestinal dysbiosis. Thus, preferences for the pectic polysaccharide types. Genomic two closely related species, B. ovatus and B. xylanisolvens, can analysis has shown that none of these six species has the utilize complex xylan for growth but do this in different ways capacity to degrade all of the pectic polysaccharide and differently from their use of structurally simpler wheat components. Thus, the complete degradation of pectin arabinoxylans.32 Essential to these ongoing studies is sound requires the activities of consortia of Bacteroides and other knowledge of the structural chemistry of plant polysaccharides bowel bacteria.16,17,48 Many Bacteroides species are able to and determination of their utilization through chemical grow on HG, and those that grew on the NZ spinach and analysis of bacterial culture supernatants. karaka berry extracts were able to consume the HG component In conclusion, two pectic polysaccharide-rich fractions were to differing extents. In contrast, the ability to degrade RG-I is isolated from NZ spinach and karaka berries, separately. far more limited. Both B. finegoldii and B. ovatus can grow on Constituent sugar and glycosyl linkage analysis showed that pure RG-I,16 but before it can be degraded, removal of side- these extracts contained different proportions of the various chain pectic polysaccharide components (arabinan, galactan, pectic polysaccharide types, with the NZ spinach extract and type II AG) is required. Our data show that only a small containing about two times as much pectic backbone (70.2% proportion of the neutral pectic polysaccharide side chains in HG and RG-I) as the karaka berry extract (31.8%). the NZ spinach and karaka berry extracts were removed by B. Conversely, the karaka berry extract contained more than 3 finegoldii and B. ovatus, and therefore, the RG-I backbone was times as much pectic side-chain polysaccharides (58.7% probably not accessible by RG-I degrading enzymes. However, arabinan, galactan, and II AG) as NZ spinach (17.7%). Our B. cellulosilyticus and B. intestinalis, which do not have the analyses showed that members of the genus Bacteroides that are capacity to grow on RG-I, showed a greater ability to consume commonly present in the human gut microbiota displayed ff arabinan and galactan than either B. finegoldii and B. ovatus. B. di erent abilities to use these polysaccharides. As such, they cellulosilyticus was able to consume arabinan, galactan, and type are likely to be the initiators of processes that completely II AGs, whereas B. intestinalis could only consume arabinan degrade and utilize the complex pectin structures found in and galactan. Thus, cooperation between different species plants that provide energy and carbon sources for the would be required to completely degrade the pectic microbiota. polysaccharides in the NZ spinach and karaka extracts, as − indicated by others.49 51 In particular, B. cellulosilyticus would ■ ASSOCIATED CONTENT be critical in any consortium, because this was the only *S Supporting Information Bacteroides species that we tested that was able to degrade and The Supporting Information is available free of charge on the grow on type II AG. ACS Publications website at DOI: 10.1021/acs.jafc.9b02429. While the use of purified components of pectic poly- saccharide complexes in growth experiments with members of Bacteroides species tested for growth on pectic the human gut microbiota provides some information, they do polysaccharides (Table S1), full glycosyl linkage not, of course, simulate bowel conditions, where a mixture of compositions of NZ spinach and karaka berry pectic dietary fibers in varying stages of structural degradation polysaccharide extracts (Table S2), and full preference become available to the microbiota. The plant polysaccharide of Bacteroides species tested for the pectic polysaccharide extracts used in our growth experiments are probably more substrates (Table S3) (PDF) comparable to complex dietary fiber substrates, such as may be found in the colon. Already, from our work with Bacteroides AUTHOR INFORMATION species, we can see the prospect of preparing mixtures of plant ■ cell wall fractions containing polysaccharides with different Corresponding Author chemical structures and compositions that might modulate *Telephone: +64-4-6320062. E-mail: [email protected].

7762 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article

ORCID M. A.; Urbanowicz, B. R.; York, W. S.; Davies, G. J.; Abbott, D. W.; Ian M. Sims: 0000-0003-2936-5903 Ralet, M.-C.; Martens, E. C.; Henrissat, B.; Gilbert, H. J. Complex pectin metabolism by gut bacteria reveals novel catalytic functions. Funding Nature 2017, 544,65−70. The work was supported by the New Zealand Ministry of (18) Al-Tamimi, M. A. H. M.; Palframan, R. J.; Cooper, J. M.; Business, Innovation, and Employment Grant C11X1312. Gibson, G. R.; Rastall, R. A. In vitro fermentation of sugar beet Notes arabinan and arabinooligosaccharides by the human gut microflora. J. The authors declare no competing financial interest. Appl. Microbiol. 2006, 100, 407−414. (19) Benítez-Paez,́ A.; Gomeź del Pulgar, E. M.; Sanz, Y. The ■ REFERENCES glycolytic versatility of Bacteroides uniformis CECT 7771 and its genome response to oligo and polysaccharides. Front. Cell. Infect. (1) Williams, B. A.; Grant, L. J.; Gidley, M. J.; Mikkelsen, D. Gut Microbiol. 2017, 7, 383. fermentation of dietary fibres: Physico-chemistry of plant cell walls (20) Dongowski, G.; Lorenz, A.; Anger, H. Degradation of pectins and implications for health. Int. J. Mol. Sci. 2017, 18, 2203. with different degrees of esterification by Bacteroides thetaiotaomicron (2) Cummings, J. H.; Engineer, A. Denis Burkitt and the origins of isolated from human gut flora. Appl. Environ. Microbiol. 2000, 66, the dietary fibre hypothesis. Nutr. Res. Rev. 2018, 31,1−15. 1321−1327. (3) Burkitt, D. P. Epidemiology of cancer of the colon and rectum. (21) Martens, E. C.; Lowe, E. C.; Chiang, H.; Pudlo, N. A.; Wu, M.; Cancer 1971, 28,3−13. McNulty, N. P.; Abbott, D. W.; Henrissat, B.; Gilbert, H. J.; Bolam, D. (4) Cleave, T. L. Natural bran in the treatment of constipation. British Med. J. 1941, 1, 461. N.; Gordon, J. I. Recognition and degradation of plant cell wall (5) Painter, N. S. Diverticular disease of the colon. Br. Med. J. 1968, polysaccharides by two human gut symbionts. PLoS Biol. 2011, 9, 3, 475−479. e1001221. (6) Walker, A. R. P. The effect of recent changes of food habits on (22) Ralet, M.-C.; Tranquet, O.; Poulain, D.; Moïse, A.; Guillon, F. − Monoclonal antibodies to rhamnogalacturonan I backbone. Planta bowel motility. S. Afr. Med. J. 1947, 21, 590 596. − (7) Rogowski, A.; Briggs, J. A.; Mortimer, J. C.; Tryfona, T.; 2010, 231, 1373 1383. Terrapon, N.; Lowe, E. C.; Basle,́ A.; Morland, C.; Day, A. M.; Zheng, (23) Pellerin, P.; Doco, T.; Vida, S.; Williams, P.; Brillouet, J.-M.; O’Neill, M. A. Structural characterization of red wine rhamnogalactur- H.; Rogers, T. E.; Thompson, P.; Hawkins, A. R.; Yadav, M. P.; − Henrissat, B.; Martens, E. C.; Dupree, P.; Gilbert, H. J.; Bolam, D. N. onan II. Carbohydr. Res. 1996, 290, 183 197. Glycan complexity dictates microbial resource allocation in the large (24)Yapo,B.M.Pecticsubstances:Fromsimplepectic polysaccharides to complex pectins − A new hypothetical model. intestine. Nat. Commun. 2015, 6, 7481. − (8) Brüssow, H. Biome engineering-2020. Microb. Biotechnol. 2016, Carbohydr. Polym. 2011, 86, 373 385. 9, 553−563. (25) Bell, T. J.; Draper, S. L.; Carnachan, S. M.; Centanni, M.; (9) Mohnen, D. Pectin structure and biosynthesis. Curr. Opin. Plant Tannock, G. W.; Sims, I. M. Characterization of polysaccharides from Biol. 2008, 11, 266−277. feijoa fruits (Acca sellowiana Berg.) and their utilization as growth (10) Hilz, H.; Williams, P.; Doco, T.; Schols, H. A.; Voragen, A. G. J. substrates by gut commensal bacteroides species. J. Agric. Food Chem. − The pectic polysaccharide rhamnogalacturonan II is present as a 2018, 66, 13277 13284. dimer in pectic populations of bilberries and black currants in muro (26) Brooker, S. G.; Cambie, R. C.; Cooper, R. C. New Zealand and in juice. Carbohydr. Polym. 2006, 65, 521−528. Medicinal Plants; Heinemann Publishers: Auckland, New Zealand, (11) Ortiz-Basurto, R. I.; Williams, P.; Belleville, M. P.; Doco, T. 1987. Presence of rhamnogalacturonan II in the juices produced by (27) Kato, M.; Takeda, T.; Ogihara, Y.; Shimizu, M.; Nomura, T.; enzymatic liquefaction of Agave pulquero stem (Agave mapisaga). Tomita, Y. Studies on the structure of polysaccharide from Tetragonia − Carbohydr. Polym. 2009, 77, 870−875. tetragonoides.I.Chem. Pharm. Bull. 1985, 33, 3675 3680. (12) Buffetto, F.; Ropartz, D.; Zhang, X. J.; Gilbert, H. J.; Guillon, (28) Choi, H. J.; Yee, S.-T.; Kwon, G.-S.; Joo, W. H. Anti- F.; Ralet, M.-C. Recovery and fine structure variability of RGII sub- inflammatory and anti-tumor effects of Tetragonia tetragonoides & domains in wine (Vitis vinifera Merlot). Ann. Bot. 2014, 114, 1327− extracts. Han'guk Misaengmul middot;Saengmyongkong Hakhoechi − 1337. 2015, 43, 391 395. (13) Chung, W. S. F.; Meijerink, M.; Zeuner, B.; Holck, J.; Louis, P.; (29) Choi, H. J.; Yee, S.-T.; Joo, W. H. Antidiabetic activity of Meyer, A. S.; Wells, J. M.; Flint, H. J.; Duncan, S. H. Prebiotic polysaccharide extract from Tetragonia tetragonoides in streptozotocin- potential of pectin and pectic oligosaccharides to promote anti- induced diabetic mice. J. Life Sci. 2017, 27, 579−583. inflammatory commensal bacteria in the human colon. FEMS (30) Goodman, A. L.; Kallstrom, G.; Faith, J. J.; Reyes, A.; Moore, Microbiol. Ecol. 2017, DOI: 10.1093/femsec/fix127. A.; Dantas, G.; Gordon, J. I. Extensive personal human gut microbiota (14) Kim, C. C.; Healey, G. R.; Kelly, W. J.; Patchett, M. L.; Jordens, culture collections characterized and manipulated in gnotobiotic mice. Z.; Tannock, G. W.; Sims, I. M.; Bell, T. J.; Hedderley, D.; Henrissat, Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 6252−6257. B.; Rosendale, D. I. Genomic insights from Monoglobus pectinilyticus: (31) Qin, J.; Li, R.; Raes, J.; Arumugam, M.; Burgdorf, K. S.; A pectin-degrading specialist bacterium in the human colon. ISME J. Manichanh, C.; Nielsen, T.; Pons, N.; Levenez, F.; Yamada, T.; 2019, 13, 1437−1456. Mende, D. R.; Li, J.; Xu, J.; Li, S.; Li, D.; Cao, J.; Wang, B.; Liang, H.; (15) Kaoutari, A. E.; Armougom, F.; Gordon, G. I.; Raoult, D.; Zheng, H.; Xie, Y.; Tap, J.; Lepage, P.; Bertalan, M.; Batto, J. M.; Henrissat, B. The abundance and variety of carbohydrate-active Hansen, T.; Le Paslier, D.; Linneberg, A.; Nielsen, H. B.; Pelletier, E.; enzymes in the human gut microbiota. Nat. Rev. Microbiol. 2013, 11, Renault, P.; Sicheritz-Ponten, T.; Turner, K.; Zhu, H.; Yu, C.; Li, S.; 497−504. Jian, M.; Zhou, Y.; Li, Y.; Zhang, X.; Li, S.; Qin, N.; Yang, H.; Wang, (16) Luis, A. S.; Briggs, J.; Zhang, X.; Farnell, B.; Ndeh, D.; J.; Brunak, S.; Dore,́ J.; Guarner, F.; Kristiansen, K.; Pedersen, O.; Labourel, A.; Basle,́ A.; Cartmell, A.; Terrapon, N.; Stott, K.; Lowe, E. Parkhill, J.; Weissenbach, J.; MetaHIT Consortium; Bork, P.; Ehrlich, C.; McLean, R.; Shearer, K.; Schückel, J.; Venditto, I.; Ralet, M.-C.; S. D.; Wang, J. A human gut microbial gene catalogue established by Henrissat, B.; Martens, E. C.; Mosimann, S. C.; Abbott, D. W.; metagenomic sequencing. Nature 2010, 464,59−65. Gilbert, H. J. Dietary pectic glycans are degraded by coordinated (32) Centanni, M.; Hutchison, J. C.; Carnachan, S. M.; Daines, A. enzyme pathways in human colonic Bacteroides. Nature Microbiol. M.; Kelly, W. J.; Tannock, G. W.; Sims, I. M. Differential growth of 2018, 3, 210−219. bowel commensal Bacteroides species on plant xylans of differing (17) Ndeh, D.; Rogowski, A.; Cartmell, A.; Luis, A. L.; Basle,́ A.; structural complexity. Carbohydr. Polym. 2017, 157, 1374−1382. Gray, J.; Venditto, I.; Briggs, J.; Zhang, X.; Labourel, A.; Terrapon, N.; (33) Lawley, B.; Centanni, M.; Watanabe, J.; Sims, I.; Carnachan, S.; Buffetto, F.; Nepogodiev, S.; Xiao, Y.; Field, R. A.; Zhu, Y.; O’Neill, Broadbent, R.; Lee, P. S.; Wong, K. H.; Tannock, G. W. tuf Gene

7763 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764 Journal of Agricultural and Food Chemistry Article sequence variation in Bifidobacterium longum subsp. infantis detected in the fecal microbiota of Chinese infants. Appl. Environ. Microbiol. 2018, 84, e00336-18. (34) Wee, M. S. M.; Matia-Merino, L.; Carnachan, S. M.; Sims, I. M.; Goh, K. K. T. Structure of a shear-thickening polysaccharide extracted from the New Zealand black tree fern, Cyathea medullaris. Int. J. Biol. Macromol. 2014, 70,86−91. (35) Sims, I. M.; Bacic, A. Extracellular polysaccharides from suspension cultures of Nicotiana plumbaginifolia. Phytochemistry 1995, 38, 1397−1405. (36) Pettolino, F. A.; Walsh, C.; Fincher, G. B.; Bacic, A. Determining the polysaccharide composition of plant cell walls. Nat. Protoc. 2012, 7, 1590−1607. (37) Ciucanu, I.; Kerek, F. Rapid and simultaneous methylation of fatty and hydroxy fatty acids for gas−liquid chromatographic analysis. J. Chromatogr. 1984, 284, 179−185. (38) Doares, S. H.; Albersheim, P.; Darvill, A. G. An improved method for the preparation of standards for glycosyl-linkage analysis of complex . Carbohydr. Res. 1991, 210, 311−317. (39) Sims, I. M.; Furneaux, R. H. Structure of the exudate gum from Meryta sinclairii. Carbohydr. Polym. 2003, 52, 423−431. (40) Colodel, C.; Bagatin, R. M. G.; Tavares, T. M.; Petkowicz, C. L. O. Cell wall polysaccharides from pulp and peel of cubiu: A pectin- rich fruit. Carbohyr. Polym. 2017, 174, 226−234. (41) Kienteka, S. S.; Correa-Ferreira,̂ M. L.; de Oliveira Petkowicz, C. L. Characterisation of cell wall polysaccharides from Sicana odorifera fruit and structural analysis of a galactan-rich fraction pectins as side chains. Carbohydr. Polym. 2018, 197, 395−402. (42) Makarova, E. N.; Shakhmatov, E. G.; Belyy, V. A. Structural studies of water-extractable pectic polysaccharides and arabinogalac- tan protein from Picea abies greenery. Carbohydr. Polym. 2018, 195, 207−217. (43) do Nascimento, G. E.; Iacomini, M.; Cordeiro, L. M. C. New findings on green sweet pepper (Capsicum annum)pectins: Rhamnogalacturonan and type I and II arabinogalactans. Carbohydr. Polym. 2017, 171, 292−299. (44) Schols, H. A.; Posthumus, M. A.; Voragen, A. G. J. Structural features of hairy regions of pectin isolated from apple juice by liquefaction process. Carbohydr. Res. 1990, 206, 117−129. (45) Fissore, E. N.; Ponce, N. M. A.; de Escalada Pla, M.; Stortz, C. A.; Rojas, A. M.; Gerschenson, L. N. Characterization of acid- extracted pectin-enriched products obtained from red beet (Beta vulgaris L. var conditiva) and butternut (Cucurbita moscata Duch ex Poiret). J. Agric. Food Chem. 2010, 58, 3793−3800. (46) Perrone, P.; Hewage, C. M.; Thomson, A. R.; Bailey, K.; Sadler, I. H.; Fry, S. C. Patterns of methyl and O-acetyl esterification in spinach pectins: New complexity. Phytochemistry 2002, 60,67−77. (47) Patova, O. A.; Smirnov, V. V.; Golovchenko, V. V.; Vityazev, F. V.; Shashkov, A. S.; Popov, S. V. Structural, rheological and antioxidant properties of pectins from Equisetum arvense L. and Equisetum sylvaticum L. Carbohydr. Polym. 2019, 209, 239−249. (48) Cartmell, A.; Muñoz-Muñoz, J.; Briggs, J. A.; Ndeh, D. A.; Lowe, E. C.; Basle,́ A.; Terrapon, N.; Stott, K.; Heunis, T.; Gray, J.; Yu, L.; Dupree, P.; Fernandes, P. Z.; Shah, S.; Williams, S. J.; Labourel, A.; Trost, M.; Henrissat, B.; Gilbert, H. J. A surface endogalactanase in Bacteroides thetaiotaomicron confers keystone status for arabinogalactan degradation. Nat. Microbiol. 2018, 3, 1314−1326. (49) Rakoff-Nahoum, S.; Coyne, M. J.; Comstock, L. E. An ecological network of polysaccharide utilization among human intestinal symbionts. Curr. Biol. 2014, 24,40−49. (50) Rakoff-Nahoum, S.; Foster, K. R.; Comstock, L. E. The evolution of cooperation within the gut microbiota. Nature 2016, 533, 255−259. (51) Crost, E. H.; Le Gall, G.; Laverde-Gomez, J. A.; Mukhopadhya, I.; Flint, H. J.; Juge, N. Mechanistic insights into the cross-feeding of Ruminococcus gnavus and Ruminococcus bromii on host and dietary carbohydrates. Front. Microbiol. 2018, 9, 2558.

7764 DOI: 10.1021/acs.jafc.9b02429 J. Agric. Food Chem. 2019, 67, 7755−7764