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

Learn More →

Variable responses of human microbiomes to dietary supplementation with resistant starch

Variable responses of human microbiomes to dietary supplementation with resistant starch Background: The fermentation of dietary fiber to various organic acids is a beneficial function provided by the microbiota in the human large intestine. In particular, butyric acid contributes to host health by facilitating maintenance of epithelial integrity, regulating inflammation, and influencing gene expression in colonocytes. We sought to increase the concentration of butyrate in 20 healthy young adults through dietary supplementation with resistant starch (unmodified potato starch—resistant starch (RS) type 2). Methods: Fecal samples were collected from individuals to characterize butyrate concentration via liquid chromatography and composition of the microbiota via surveys of 16S rRNA-encoding gene sequences from the Illumina MiSeq platform. Random Forest and LEfSe analyses were used to associate responses in butyrate production to features of the microbiota. Results: RS supplementation increased fecal butyrate concentrations in this cohort from 8 to 12 mmol/kg wet feces, but responses varied widely between individuals. Individuals could be categorized into three groups based upon butyrate concentrations before and during RS: enhanced, high, and low (n = 11, 3, and 6, respectively). Fecal butyrate increased by 67 % in the enhanced group (from 9 to 15 mmol/kg), while it remained ≥11 mmol/kg in the high group and ≤8 mmol/kg in the low group. Microbiota analyses revealed that the relative abundance of RS-degrading organisms—Bifidobacterium adolescentis or Ruminococcus bromii—increased from ~2 to 9 % in the enhanced and high groups, but remained at ~1.5 % in the low group. The lack of increase in RS-degrading bacteria in the low group may explain why there was no increase in fecal butyrate in response to RS. The microbiota of individuals in the high group were characterized by an elevated abundance of the butyrogenic microbe Eubacterium rectale (~6 % in high vs. 3 % in enhanced and low groups) throughout the study. Conclusions: We document the heterogeneous responses in butyrate concentrations upon RS supplementation and identify characteristic of the microbiota that appear to underlie this variation. This study complements and extends other studies that call for personalized approaches to manage beneficial functions provided by gut microbiomes. Background functions could promote health and reduce the incidence The microbiota in the large intestine provides several of preventable diseases including obesity and type 2 dia- functions that are beneficial to human health such as pro- betes [3], colon cancer [4], chronic and acute undernutri- ducing short-chain fatty acids, modifying primary to sec- tion [5], and infections by Clostridium difficile [6]. ondary bile acids, and providing colonization resistance to One of the beneficial functions derived from the co- some enteric pathogens [1, 2]. Managing this community lonic microbiota is the production of butyric acid that is of microbes to maintain and improve these beneficial generated from the fermentation of dietary fiber. The conjugate base of the acid—butyrate—is the preferred * Correspondence: [email protected] energy source for colonocytes [7]. Butyrate improves the Department of Internal Medicine, University of Michigan, Ann Arbor, MI intestinal barrier by facilitating tight-junction assembly 48105, USA [8], suppresses inflammatory and allergic responses by Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Venkataraman et al. Microbiome (2016) 4:33 Page 2 of 9 inducing differentiation of colonic regulatory T cells [9], 4 fecal samples during RS regulates cell apoptosis [10], and stimulates production of anorectic hormones [11]. Indeed, reduced concentra- Acclimatization period tions of butyrate have been associated with the incidence 4 baseline fecal samples No fecal samples before consuming RS of graft-versus-host disease [12, 13], kwashiorkor [14], 4 8 11 17 colon cancer [15], and obesity [11]. Especially in these Day circumstances, increased butyrate production in the Fig. 1 Experimental design. Participants supplemented their habitual large intestine may be beneficial to human health. diets with resistant starch. Filled symbols represent points at which One approach to increasing butyrate production is to fecal samples were collected deliver more fermentable carbohydrates to gut micro- biomes. Here, we report on the impact of dietary supple- mentation with resistant starch (unmodified potato 48 g of potato starch—24-g doses twice per day—for starch; RS type 2) on fecal butyrate concentrations and seven more days. Participants were provided a 1- composition of the gut microbiota in 20 healthy young tablespoon scoop, which they used to measure out the adults. RS consists of starch that is resistant to hydroly- appropriate amount of potato starch (each tablespoon sis by human enzymes and passes through the small in- ≈12 g) and consumed it after mixing the starch with cold testine unabsorbed. In the large intestine, RS can be water. metabolized and then fermented by microbes to a variety of products, including butyric acid [16]. In this study, we Fecal collection find that the response to RS supplementation varies be- Participants were provided waxed tissue paper (Epitope tween individuals in ways that can be explained, at least Diagnostics, San Diego, CA), which was laid down on in part, by variation in the composition of their micro- the water in the toilet bowl prior to defecation. This biota. Recognizing inter-individual variability in re- paper sticks to the sides of the toilet bowl so that the sponses to fiber supplementation and determining the fecal sample is not readily contaminated by water in the microbiota characteristics that underlie it are essential toilet. Fecal samples were collected using the deoxyri- first steps towards personalized plans for managing gut bose nucleic acid (DNA) Genotek Omnigene Gut microbiomes for desirable functionality, including the Collection kits (DNA Genotek, Ontario, Canada) follow- production of butyrate. ing the manufacturer’s instructions. The DNA Genotek tube contains 2 mL of a stabilization buffer and a steel Methods ball to facilitate mixing of the fecal material with the Participants buffer. Approximately, 0.67 g of fecal material was col- Prospective volunteers were students in research-based lected per tube. The kits were returned to a −20 °C sections of Introductory Biology 173 at the University of freezer within 24 h of collection. Frozen samples were Michigan. Individuals with a self-reported history of thawed and aliquots were withdrawn for DNA extrac- bowel disorders such as irritable bowel syndrome, in- tion, measurement of fermentation products, and to cre- flammatory bowel disease, or colorectal cancer were ex- ate sample archives. cluded from the study. Twenty individuals (10 males, 10 females) participated in the 3-week study. The age range DNA extraction, PCR, sequencing, and sequence was 19–20 years, and the BMI range was from 19–63. processing Each participant gave his or her written, informed For DNA extraction, 0.25 mL of the fecal suspension in consent before participating in the study. This research the DNA Genotek tube was deposited into a MoBio was carried out in compliance with the Helsinki Declar- PowerMag Soil DNA Isolation Bead Plate. DNA was ex- ation and was approved by the Institutional Review tracted following MoBio’s instructions on an epMotion Board of the University of Michigan Medical School 5075 liquid handling workstation (Eppendorf, Hauppauge, (HUM00094242). NY). The V4 region of the 16S-ribosomal RNA (rRNA) encoding genes was amplified from each sample using Study design the dual-indexing sequencing strategy [17]. Sequen- Participants consumed their habitual diet throughout cing was performed on the Illumina MiSeq platform, the study period. During the intervention phase, raw un- using a MiSeq Reagent Kit V2 500 cycles (Cat# MS- modified potato starch (Bob’s Red Mill, Milwaukie, OR) 102-2003), according to the manufacturer’sinstruc- was gradually added to their diet (day 1—12 g, day tions. Sequences were curated using mothur v1.31.2 2—24 g, day 3—48 g; Fig. 1). This potato starch contains [18] and clustered into operational taxonomic units at approximately 50 % resistant starch (type 2) by weight. ≥ 97 % sequence identity using the average neighbor After the acclimatization period, the subjects consumed algorithm. Total umodified potato starch consumed (g) Venkataraman et al. Microbiome (2016) 4:33 Page 3 of 9 Measurement of fermentation products composition of gut microbiota in the study population One millimeter of fecal suspension was aliquoted from (p < 0.05). For further analyses, we used the median each DNA Genotek OmniGut tube and centrifuged at abundance of each operational taxonomic unit (OTU) in 4 °C for 10 min (4500×g). The supernatant fraction was each individual before and during RS consumption. To withdrawn and passed sequentially through 1.2, 0.65, identify specific OTUs that had changed with RS con- and 0.22-μm hydrophilic low protein-binding Durapore sumption, LEfSe and Random Forest analyses were per- membrane filters (EMD Millipore, Darmstadt, Germany). formed for each individual separately (comparing their Filtered samples were stored at −20 °C until analysis. four samples before and four samples during RS con- Once thawed, samples were maintained at 4 °C in an sumption). LEfSe [21] was implemented within mothur autosampler. A Shimadzu HPLC (Shimadzu Scientific using the correction for multiple comparisons. PERMA- Instruments, Columbia, MD) equipped with a dual UV NOVA, analysis of similarity (ANOSIM), and permuta- (214 nm for short-chain fatty acids (SCFAs))/refractive tional analyses of dispersions (PERMDISP) analyses were index detector (for ethanol) and an Aminex HPX-87H performed with the package vegan in R. Random Forest column (Bio-Rad Laboratories, Hercules, CA) heated to was implemented using the package randomForest in R. 50 °C was used to measure acetic, propionic, and butyric acids and ethanol in the filtrate. The mobile phase was Results 0.01 N H SO at a flow rate of 0.6 ml/min, and the injec- The impact of RS on the composition of the gut micro- 2 4 tion volume was 10 μl. Samples were randomized in re- biota and the concentration of three short-chain fatty gard to injection order, and the average value from two acids (acetic, propionic, and butyric) in feces was deter- technical replicates was used in all subsequent analyses. mined in 20 healthy young adults. RS was gradually intro- Ethanol, which was a component of the stabilization buf- duced into their regular diets during an acclimatization fer, served to monitor injection volumes. Samples in which period of 3 days (Fig. 1). Thereafter, study participants in- the peak height for ethanol was less than 70 % of the aver- cluded 48 g of unmodified potato starch in their daily diet age peak height for ethanol in all samples were excluded (~24 g as RS) for 7 days. Four fecal samples were collected from further analysis. Eight external standards (0.1– before the introduction of RS and another four during the 20 mM) for acetic, propionic, and butyric acids were used period of maximum RS supplementation (Fig. 1). to generate the standard curve. These standards were run There was considerable intra-individual variability in after every 100 samples. LC Solutions Software (Shimadzu the concentration of acetate, propionate, and butyrate Scientific Instruments, Columbia, MD) was used to curate (coefficient of variation = 20–90 %). This variability is the data and calculate concentrations based on the stand- not unexpected because fecal SCFA concentrations are ard curve generated during the run. The concentrations influenced by host absorption, transit time through the were then normalized by the average weight of the fecal GI tract, host diet, and time since last meal. Due to this samples (0.67 g) collected in the DNA Genotek tubes. variability and the number of samples, Student’s t tests did not reveal measurable differences in the concentra- Statistical analyses tion of SCFA in response to RS supplementation for an All statistical analyses were performed using RStudio individual (four samples before vs. four during RS for version 0.99.489 [19] and the software PAST [20]. To each individual; p ≥ 0.10 for all individuals). However, a evaluate the impact of RS supplementation on acetate, repeated measures ANOVA showed that RS supplemen- propionate, and butyrate concentrations in the study tation increased the fecal concentration of butyrate by population, a nested repeated measures analysis of vari- 50 % and acetate by 26 % in the study population as a ance (ANOVA) was used. This revealed that fecal butyr- whole (p = 0.02 and 0.03, respectively; Table 1). The ate and acetate concentrations increased with RS intake concentrations of propionate did not change signifi- in the study population (p < 0.05). For all subsequent cantly in this cohort (p =0.85; Table 1). analyses, the median value of butyrate for each person before and during RS consumption was used so that the Table 1 Effect of dietary supplementation with RS on number of samples compared was not artificially in- concentration of select fermentation products (mmol/kg flated. Paired or unpaired t tests were employed as ap- wet feces) propriate. To determine if RS intake altered microbial Fermentation Before RS During RS Change p value (repeated community composition, a permutational analysis of product measures ANOVA) Median ± IQR Median ± IQR variance (PERMANOVA) analysis was conducted with Butyric acid 8 ± 6 12 ± 7 50 % 0.03 the Bray-Curtis similarity index. Each individual was Acetic acid 27 ± 6 34 ± 10 26 % 0.02 used as the blocking factor to account for repeated mea- Propionic acid 13 ± 6 12 ± 5 −8 % 0.85 sures of microbiota composition from each individual. This analysis revealed that RS intake altered the IQR interquartile range Venkataraman et al. Microbiome (2016) 4:33 Page 4 of 9 Like the fecal SCFA concentrations, there was also some intra-individual variability in the composition of microbiota. However, at least 75 % of OTUs were con- sistently detected in all samples from an individual with 10 the coefficient of variation in their relative abundances ranging from 39 to 65 %. Unlike the SCFAs, where stat- Individuals arranged in ascending order of butyrate before RS istical tests did not detect differences in response to RS Fig. 2 Median butyrate concentrations for each individual before for an individual, ANOSIM tests showed that RS supple- (triangles) and during consumption of RS (circles). Dotted and dashed mentation altered the composition of microbiota in most lines denote the median values for butyrate before and during RS, individuals in the study (ANOSIM with Bray-Curtis respectively, for the entire study population similarity; four samples before vs. four during RS for each individual; p < 0.10 for 16 out of 20 individuals). This conclusion was corroborated by results from a Baseline butyrate concentrations were not predictive of PERMANOVA analysis to determine if RS supplementa- butyrate concentrations during RS supplementation tion altered the composition of microbiota in the overall (R = 0.08, p = 0.20). study population. In order to incorporate intra- To identify characteristics of the microbiota that may individual variability, each individual was considered as underlie the variable responses to RS, we first used Ran- the blocking factor in the PERMANOVA. This analysis dom Forest regression to identify relationships between revealed that the composition of the gut microbiota was the abundances of OTUs and butyrate concentrations altered with RS supplementation in our study population before and during consumption of RS. No OTUs were (PERMANOVA using Bray-Curtis similarity blocking for particularly strong predictors of butyrate concentrations each individual; p = 0.001). A PERMDISP analysis further either before or during RS consumption. Butyrate con- revealed that the PERMANOVA was not affected by dif- centrations before RS were weakly related to baseline ferences in the dispersion of communities before and abundances of OTU 4 (Eubacterium rectale)(R = 0.14; during RS (PERMDISP p = 0.44). p = 0.10). Unexpectedly, this relationship was not detect- The relative abundance of OTUs belonging to the able during RS supplementation. phylum Actinobacteria increased with RS, and there was Population-wide relationships between OTUs and bu- a small decrease in the abundance of Firmicutes.No tyrate concentrations could be masked by the heterogen- changes were detected in the relative abundances of eity of both variables between individuals. We therefore Bacteroidetes or Proteobacteria (Table 2). Finally, RS looked for correlations between features of the micro- supplementation did not change the overall richness or biota and butyrate concentrations in subsets of partici- evenness of the microbial community in the study popu- pants that had similar responses in fecal butyrate lation (repeated measures ANOVA Chao1 index before following RS supplementation. The study population vs. during p = 0.49; Simpson’s index before vs. during was separated into three groups using k-means cluster- p = 0.96). ing based on butyrate concentrations before and during Therefore, despite the intra-individual variability in RS. An elbow plot [22] revealed that there were three fecal SCFA concentrations and composition of micro- “clusters”. The categories identified were enhanced, high, biota, repeated measures ANOVA and PERMANOVA and low (Fig. 3a, b). The concentration of butyrate in reveal that RS consumption led to increases in fecal bu- the “enhanced” group (n = 11) increased significantly tyrate concentrations and altered the composition of the following consumption of RS (from 9 to 15 mmol/kg microbiota in the study population. However, the start- wet feces, paired t test p = 0.0003). Individuals in the ing concentrations of butyrate and the changes in butyr- “high” group (n = 3) maintained butyrate concentrations ate during RS varied widely between individuals (Fig. 2). ≥11 mmol/kg wet feces during the course of the study. Individuals in the “low” group (n = 6) had less than or equal to 8 mmol butyrate/kg wet feces both before and Table 2 Effect of dietary supplementation with RS on relative during RS (paired t test p = 0.14; Fig. 3c). abundance of four dominant bacterial phyla The OTUs that distinguished these three clusters were Phylum Before RS During RS (%) Change p value identified using Random Forest analysis and LEfSe Median ± IQR Median ± IQR paired t test (Table 3). Random Forest revealed OTU #7 as the most Actinobacteria 1.3 ± 0.6 6.2 ± 0.6 377 % 0.02 prominent feature of the microbiota distinguishing the Firmicutes 37.1 ± 10.7 33.2 ± 5.2 -11 % 0.04 low from the enhanced group (Table 3). Sequences Bacteroidetes 53.3 ± 13.2 51.5 ± 9 -3 % 0.82 within this OTU are identical to those from Bifidobac- Proteobacteria 4.7 ± 4.1 4.8 ± 4.1 2 % 0.82 terium adolescentis. The relative abundance of this OTU IQR interquartile range before RS was similar in all three groups (~0.7–1.4 %; Butyrate (mmol/kg wet feces) Venkataraman et al. Microbiome (2016) 4:33 Page 5 of 9 ab c During RS Before RS U005 p < 0.001 U011 U031 U014 U034 U029 U032 U020 U025 U035 U007 U033 U004 8 U006 U015 U022 U026 U008 Before During Before During Before During U001 5 101520 U024 Enhanced High Low Butyrate before RS (mmol/kg wet feces) Butyrate 5 10 15 20 (mmol/kg wet feces) Fig. 3 Clustering individuals based upon their butyrate response to RS. a Three groups generated by k-means clustering. b Median butyrate concentrations before and during RS supplementation. c Average butyrate concentrations before and during RS in the three groups t test p > 0.27; Fig. 4a). However, the enhanced and was detected in 14 out of 20 total individuals and in- high groups had dramatically higher abundances of creased in abundance in 12 individuals. This OTU did this OTU during RS (average 8.9 % in enhanced and not increase in individuals U026 and U024 (LEfSe 7.8 % in high; p < 0.05; Fig. 4a). The abundance of p > 0.05 correcting for multiple comparisons). However, this OTU did not change in the low group (average OTU #7 [B. adolescentis] was not detected in six individ- before RS = 1.5 %, average during = 3.7 %, p = 0.13). uals. In these individuals, LEfSe was used to find other This finding was further corroborated with a LEfSe OTUs that increased in abundance during RS supple- analyses comparing four samples before to four dur- mentation, since these could be RS-degrading bacteria. ing RS consumption for each individual. As one In three individuals in whom B. adolescentis sequences would expect, the abundance of OTU 7 increased in did not increase, OTU 19 increased in abundance (LEfSe 8 out of 11 individuals in the enhanced group, 2 out p < 0.05 correcting for multiple comparisons). Sequences of 3 individuals in the high group, but in only 2 out in that OTU are identical to Ruminococcus bromii, an- of 6 individuals in the low group (LEfSe p <0.05 cor- other group of RS-degrading bacteria [23]. In individual recting for multiple comparisons). U005, neither B. adolescentis nor R. bromii increased in Cultivars of B. adolescentis are capable of breaking abundance with RS. A potential candidate for a RS- down RS [23]. OTU #7 in which sequences were identi- degrading organism in this individual is OTU 50, whose cal to the 16S rRNA encoding gene of B. adolescentis average abundance increased from 2.6 to 7.1 % albeit not statistically significant (p = 0.29). Sequences in this OTU are most closely related to Ruminiclostridium Table 3 Results of Random Forest regression and LEfSe to [Eubacterium] siraeum. This organism has not been identify OTUs that distinguish the three response groups reported to degrade RS, but it is in the same taxonomic Comparison Distinguishing microbiota features family as R. bromii. LEfSe Random Forest Since B. adolescentis and R. bromii are the strongest Low vs. enhanced None identified [OTU 7] candidates for RS-degrading organisms in this study, we during RS summed the abundance of these two organisms for each High vs. low [OTU 4] [OTU 4] before and during RS before and during RS ‡ individual (Fig. 4b). Individuals in all three groups start [OTU 3] before RS with a similar abundance of these RS-degrading mi- High vs. enhanced [OTU 4] [OTU 4] before and during RS before and during RS crobes. RS elicits dramatic increases in their abundance A post-hoc ANOVA analysis revealed that the abundance of OTU #3 was not in high and enhanced groups, but not in the low group significantly different between the groups being compared (p > 0.10). Hence, this OTU was not considered further (Fig. 4b). Butyrate during RS (mmol/kg wet feces) Enhanced Low High Butyrate (mmol/kg wet feces) Venkataraman et al. Microbiome (2016) 4:33 Page 6 of 9 OTU 7 OTU 19 Summed B. adolescentis R. bromi i abundance (%) p = 0.11 p = 0.003 p = 0.02 Before During Before During Before During U015 U022 U026 U008 U001 U024 U005 U011 U031 U014 U034 U029 U032 U020 U025 U035 U007 U033 U004 Before During Before During Before During U006 Low Enhanced High Abundance (%) 5 10 15 20 Fig. 4 Identifying microbiota features that distinguish the low from other groups. a Median abundance of OTUs taxonomically related to B. adolescentis and R. bromii before and during RS for each individual. b Summed abundance of the two proposed RS-degrading OTUs before and during RS in the three groups Random Forest and LEfSe also revealed that OTU 4 supplementation (Fig. 5a, LEfSe p < 0.05 correcting for distinguished the high group from the enhanced and low multiple comparisons), but in general, its abundance did groups (Table 3). Sequences within this OTU are identi- not change as a result of RS addition to diet in any of cal to E. rectale—a known and prominent butyrogenic the three groups (Fig. 5b). We looked specifically at microbe in human guts [24]. E. rectale was more abun- other butyrogenic organisms known to be present in the dant in individuals in the high group than in the other human colon such as Faecalibacterium prausnitzii [25]. two groups—both before and during RS (Fig. 5b). While An OTU with the same V4 16S-rRNA encoding gene se- individuals in the high group fall into a very distinct quence as F. prausnitzii was present at about 3–6% cluster in the k-means clustering (Fig. 3), there are only relative abundance in all three groups, but did not three individuals in this group. This limited sample size change with RS in any of the groups (t test p > 0.15). does constrain the strength of the finding that E. rectale Since B. adolescentis—the prominent RS-degrading mi- is more abundant in the high group vs. the other two crobe in our cohort—produces lactate as the primary groups. Some individuals in the enhanced group exhib- fermentation product, we also investigated if organisms ited increases in abundances of E. rectale during RS known to produce butyrate from lactate increased in a b p = 0.02 OTU 4 E. rectale p = 0.02 Before During U033 U004 U006 7.5 U005 U011 U031* U014 U034 5.0 U029* U032 U020* U025* U035* 2.5 U007 U015 U022 U026 0.0 U008 BeforeDuring Before During Before During U001 High Low Enhanced U024 Abundance (%) 510 15 20 Fig. 5 The relative abundance of OTU 4 Eubacterium rectale distinguishes the high from other groups. a Median abundance of OTU 4 before and during RS for each individual. Individuals in whom the abundance of this OTU increased with RS are marked with an asterisk (LEfSe p < 0.05). b Abundance of OTU 4 before and during RS in the three groups High Enhanced Low Low Enhanced High OTU 4 [E. rectale] (%) Proposed RS−degrading OTUs (%) Venkataraman et al. Microbiome (2016) 4:33 Page 7 of 9 abundance. OTUs in which sequences were most closely clustering using butyrate concentrations before and dur- related to Eubacterium halii and Anaerostipes caccae ing RS identified three groups: enhanced, high, and low. were present at less than 0.4 % abundance in our cohort Butyrate increased on average from 9 to 15 mmol/kg wet and did not change with RS (paired t test p = 0.4). feces in the enhanced group, whereas butyrate concentra- tions remained consistently high or low in the other two Discussion groups (≥11 and ≤8 mmol/kg wet feces, respectively; The impact of dietary supplementation with resistant Fig. 3). With these clusters, we were able to identify fea- starch on fecal butyrate concentrations and the compos- tures of the microbial communities that differentiated the ition of the microbiota was examined in 20 healthy three groups. young adults. In this cohort, the average ratios of aceta- The relative abundance of proposed RS-degrading te:propionate:butyrate before and during RS supplemen- bacteria—B. adolescentis or R. bromii—increased with tation were 58:25:15 and 58:22:20, respectively. These RS supplementation in the enhanced and high groups ratios agree well with previously documented ratios of (from 2 to 9 %), but not in the low group (~1.5 % ~60:20:20 in human feces [26–29]. In response to RS throughout; Fig. 4). This finding complements and ex- supplementation, the concentration of fecal butyrate in- tends previously published reports on the effects of RS creased from 8 to 12 mmol/kg wet feces in the overall supplementation on the composition of the human gut study population (repeated measures ANOVA p = 0.02). microbiota. In a study with overweight adult males, only The inter-individual variation in butyrate concentrations individuals with detectable abundances of R. bromii in before and during RS was striking (Fig. 2). Most previ- their gut microbiota were able to degrade RS (type ous studies with RS have documented only population- 3—Novelose 330; Walker et al. [32]). In another study wide response in butyrate concentrations [16], while one with ten human subjects (five males and females; 28– other study reported considerable heterogeneity between 38 years old), dietary supplementation with either type 4 individuals in regard to fecal butyrate concentrations RS (FiberSym RW) or type 2 RS (HiMaize260) increased [30]. This pronounced inter-individual variability sug- the relative abundance of B. adolescentis and R. bromii, gests that a single approach to improving beneficial respectively (Martinez et al. [33]). These in vivo studies functions from the microbiome is unlikely to be univer- are nicely complemented by in vitro microcosm studies, sally successful. Rather, personalized approaches may be which showed that isolates of R. bromii and B. adoles- needed to manage microbiomes for health. In line with centis are capable of degrading several forms of RS [23]. this endeavor, we attempted to identify features in the Thus, different types of resistant starches are likely to microbiota that could explain different responses to diet- promote the growth of B. adolescentis or R. bromii, with ary supplementation with RS. both organisms exhibiting substrate specificity depending Random Forest was used to identify relationships be- upon the type and source of RS. tween butyrate concentrations and OTUs in the study The relative abundance of RS-degrading organisms population. Weak relationships were identified between and butyrate concentrations did not increase in the low butyrate concentrations before RS and the abundances group. This suggests that their microbiota did not break of OTU 4 (E. rectale). Random Forest failed to reveal down RS and so could not lead to increased butyrate any relationships between OTU abundances and butyr- production. This suggestion is supported by the fact that ate concentrations during RS supplementation. It is pos- the concentrations of neither acetate nor propionate in- sible that the number of individuals in this study (n = 20) creased in the low group. So why does the abundance of may have been too small to detect robust correlations RS-degrading organisms remain at 1.5 % in the low between butyrate and the abundance of OTUs at the group? Possible explanations include limitation by antag- population level. Alternatively, it is not necessary that onistic microbes or lack of synergistic microbes. Our the abundance of a single organism be correlated with attempts to identify antagonistic and synergistic interac- butyrate concentrations. After all, the potential for pro- tions by constructing correlation networks [34] between ducing butyrate is fairly widespread within the phylum RS-degrading organisms and other OTUs did not yield Firmicutes [31]. Rather, we expect that the abundance of any compelling relationships. genes encoding butyryl-CoA:acetate CoA transferase In addition to RS-degrading organisms, another OTU (but) and butyrate kinase (buk) would be correlated to (OTU #4) was identified as distinguishing the high butyrate concentrations. group from the enhanced and low groups (Table 3, Our next step was to create groupings in the data Fig. 5). Its V4 sequence is identical to that of E. rectale, based upon similar responses to RS. This clustering ap- well established as a prominent butyrogenic bacterium proach should constrain the variability within each group in the human gut microbiome [24, 35]. Its relative abun- and increase the probability of identifying microbiota fea- dance in the high group was consistently about 6 %, tures that characterize each type of response. K-means compared to 3 % in the other two groups. Surprisingly, Venkataraman et al. Microbiome (2016) 4:33 Page 8 of 9 the abundance of E. rectale did not change in any of the changes in the microbiota in the enhanced group groups in response to RS—even though butyrate con- (ANOSIM R = 0.64). It is tempting to suggest that the centrations increased appreciably in the enhanced group microbiota of these individuals are performing well with (from 9 to 15 mmol/kg wet feces). E. rectale generates regard to butyrate production, and they do not benefit butyrate from acetate, and there is a net gain of ATP in from additional dietary input of fermentable carbohy- that process [35]. So if E. rectale was responsible for in- drates. A third subset of our study population (6 of 20) creased butyrate production in the enhanced group with had consistently low concentrations of butyrate even RS, one might expect an increase in its relative abun- when consuming RS (≤8 mmol/kg wet feces). In this dance. We offer two possible explanations for the in- group, RS-degrading organisms did not increase in creased butyrate production in the enhanced group abundance, suggesting that their microbiota did not without a measurable increase in the abundance of E. break down RS. Based upon this result, we propose that rectale. First, changes in the relative abundance of E. rec- increasing butyrate in these individuals will require tale may be subtle and masked by the dramatic increase either (i) testing another form of dietary fiber such as in the abundance of RS-degrading organisms (~from 2 inulin or arabinoxylan that their microbiota might de- to 10 % in the enhanced group). It is also possible that grade, (ii) a synbiotic approach that combines a dietary populations of E. rectale take longer to respond to RS fiber with the appropriate fiber-degrading bacteria, or supplementation. We used an acclimatization period of (iii) targeted removal of microbes if any are antagonistic only 3 days during which the amounts of RS in diet were to RS-degrading organisms. The findings of this study il- gradually increased, and after this period, four fecal sam- lustrate the importance of studying individual responses ples were collected. It remains to be seen whether there to dietary modifications. This will uncover the mecha- is a detectable increase in the abundance of E. rectale nisms that underlie these responses and in time provide following a longer duration of RS consumption. Also, as actionable insights towards precision management of mentioned earlier, it is well documented that butyrate microbiomes. production is widespread within the phylum Firmicutes in human gut microbiomes. Therefore, rather than a sin- Abbreviations ANOVA, analysis of variance; DNA, deoxyribose nucleic acid; g, gram(s); gle organism, if more acetate is being converted to bu- h, hour(s); HPLC, high-pressure liquid chromatography; LC, liquid tyrate, the genes encoding butyryl-CoA:acetate CoA chromatography; ml, milliliter(s); mM, millimolar; °C, degrees centigrade; transferase (but) and butyrate kinase (buk) should in- OTU, operational taxonomic unit; rRNA, ribosomal RNA; RS, resistant starch; μl, microliters crease in abundance with RS supplementation. Acknowledgements Conclusions The authors thank Drs. Frank Gerberick and JoseCarlos Garcia-Garcia (Procter & Gamble Company, Cincinnati, OH) for providing feedback on Our data show that dietary supplementation with RS data analysis and interpretation. type 2 as unmodified potato starch increases fecal butyr- ate concentration, but with remarkable inter-individual Funding variation. We were able to infer potential explanations This research was supported by grants from the Procter and Gamble Company and the Howard Hughes Medical Institute (52008119). for some of these differential effects by investigating the composition of microbiota. Fecal butyrate concentra- Availability of data and materials tions increased by an average of 67 % in a subset of the The sequence and metadata have been deposited in the GenBank database [accession number: SRP067761]. study population (n = 11 of 20; 9 to 15 mmol/kg wet feces). Most individuals in this group showed a dramatic Authors’ contributions increase in the relative abundance of the RS-degrading AV, JRS, AS, CW, KRT, and TMS conceived of the study, participated in its design and coordination, and helped draft the manuscript. AV, AS, and CW organisms—B. adolescentis or R. bromii. In five of these conducted the microbiota analyses. JRS conducted the analyses of individuals, the prominent butyrogenic microbe E. rec- fermentation products. KRT participated and advised in statistical analyses. All tale also increased in abundance. Another subset of the authors read and approved the final manuscript. population (3 of 20) consistently maintained high butyr- Competing interests ate concentrations both before and with RS (≥12 mmol/ The authors declare that they have no competing interests. kg wet feces). In this subset, RS-degrading organisms in- creased in abundance, suggesting that RS is being de- Author details Department of Internal Medicine, University of Michigan, Ann Arbor, MI graded. But there was no concomitant increase in fecal 48105, USA. Present address: Department of Biology, University of butyrate concentrations. These individuals may be ex- 3 Minnesota, Duluth, MN 55812, USA. Present address: Department of periencing a plateau effect [36] in butyrate production Immunology and Microbiology, Wayne State University School of Medicine, Detroit, MI 48201, USA. even before RS is administered. In fact, the microbiota structures of these individuals also did not change as Received: 16 February 2016 Accepted: 10 June 2016 much with RS (ANOSIM R = 0.17) as compared to Venkataraman et al. Microbiome (2016) 4:33 Page 9 of 9 References 24. Duncan SH, Flint HJ. Proposal of a neotype strain (A1-86) for Eubacterium 1. Tremaroli V, Backhed F. Functional interactions between the gut microbiota rectale. Request for an opinion. Int J Syst Evol Microbiol. 2008;58(Pt 7):1735–6. and host metabolism. Nature. 2012;489(7415):242–9. doi:10.1099/ijs.0.2008/004580-0. 2. Flint HJ, Duncan SH, Scott KP, Louis P. Links between diet, gut microbiota 25. Khan MT, Duncan SH, Stams AJM, van Dijl JM, Flint HJ, Harmsen HJM. The gut anaerobe Faecalibacterium prausnitzii uses an extracellular electron shuttle to composition and gut metabolism. Proc Nutr Soc. 2015;74(1):13–22. grow at oxic-anoxic interphases. ISME J. 2012:doi:10.1038/ismej.2012.5. doi:10.1017/s0029665114001463. 26. den Besten G, van Eunen K, Groen AK, Venema K, Reijngoud D-J, Bakker BM. 3. Hartstra AV, Bouter KE, Backhed F, Nieuwdorp M. Insights into the role of The role of short-chain fatty acids in the interplay between diet, gut the microbiome in obesity and type 2 diabetes. Diabetes Care. 2015;38(1): microbiota, and host energy metabolism. J Lipid Res. 2013;54(9):2325–40. 159–65. doi:10.2337/dc14-0769. doi:10.1194/jlr.R036012. 4. Zackular JP, Baxter NT, Iverson KD, Sadler WD, Petrosino JF, Chen GY, et al. 27. Hijova E, Chmelarova A. Short chain fatty acids and colonic health. Bratisl The gut microbiome modulates colon tumorigenesis. mBio. 2013;4(6). doi: Lek Listy. 2007;108(8):354–8. 10.1128/mBio.00692-13. 28. Binder HJ. Role of colonic short-chain fatty acid transport in diarrhea. Annu 5. Blanton LV, Charbonneau MR, Salih T, Barratt MJ, Venkatesh S, Ilkaveya O, Rev Physiol. 2010;72:297–313. doi:10.1146/annurev-physiol-021909-135817. et al. Gut bacteria that prevent growth impairments transmitted by 29. Cummings JH, Pomare EW, Branch WJ, Naylor CP, Macfarlane GT. Short microbiota from malnourished children. Science. 2016;351(6275):3311-1–7. chain fatty acids in human large intestine, portal, hepatic and venous blood. 6. Buffie CG, Bucci V, Stein RR, McKenney PT, Ling L, Gobourne A, et al. Gut. 1987;28(10):1221–7. Precision microbiome reconstitution restores bile acid mediated resistance 30. McOrist AL, Miller RB, Bird AR, Keogh JB, Noakes M, Topping DL, et al. Fecal to Clostridium difficile. Nature. 2015;517(7533):205–8. doi:10.1038/nature13828. butyrate levels vary widely among individuals but are usually increased 7. Donohoe DR, Garge N, Zhang X, Sun W, O’Connell TM, Bunger MK, et al. by a diet high in resistant starch. J Nutr. 2011;141(5):883–9. doi:10.3945/ The microbiome and butyrate regulate energy metabolism and autophagy jn.110.128504. in the mammalian colon. Cell Metab. 2011;13(5):517–26. doi:10.1016/j.cmet. 31. Vital M, Howe AC, Tiedje JM. Revealing the bacterial butyrate synthesis 2011.02.018. pathways by analyzing (meta)genomic data. mBio. 2014;5(2). doi:10.1128/ 8. Peng L, Li ZR, Green RS, Holzman IR, Lin J. Butyrate enhances the intestinal mBio.00889-14. barrier by facilitating tight junction assembly via activation of AMP-activated 32. Walker AW, Ince J, Duncan SH, Webster LM, Holtrop G, Ze X, et al. protein kinase in Caco-2 cell monolayers. J Nutr. 2009;139(9):1619–25. Dominant and diet-responsive groups of bacteria within the human colonic doi:10.3945/jn.109.104638. microbiota. ISME J. 2011;5(2):220–30. doi:10.1038/ismej.2010.118. PubMed 9. Furusawa Y, Obata Y, Fukuda S, Endo TA, Nakato G, Takahashi D, et al. PMID: PMC3105703. Commensal microbe-derived butyrate induces the differentiation of colonic 33. MartÌnez I, Kim J, Duffy PR, Schlegel VL, Walter J. Resistant starches types 2 regulatory T cells. Nature. 2013;504(7480):446–50. doi:10.1038/nature12721. and 4 have differential effects on the composition of the fecal microbiota in 10. Ruemmele FM, Schwartz S, Seidman EG, Dionne S, Levy E, Lentze MJ. human subjects. PLoS One. 2010;5(11):e15046. doi:10.1371/journal.pone. Butyrate induced Caco-2 cell apoptosis is mediated via the mitochondrial pathway. Gut. 2003;52(1):94–100. 34. Friedman J, Alm EJ. Inferring correlation networks from genomic survey data. 11. Mikkelsen KH, Allin KH, Knop FK. Effect of antibiotics on gut microbiota, PLoS Comput Biol. 2012;8(9):e1002687. doi:10.1371/journal.pcbi.1002687. glucose metabolism and body weight regulation: a review of the literature. 35. Rivière A, Gagnon M, Weckx S, Roy D, De Vuyst L. Mutual cross-feeding Diabetes Obes Metab. 2016;18(5):444–53. doi:10.1111/dom.12637. interactions between Bifidobacterium longum subsp. longum NCC2705 12. Mathewson ND, Jenq R, Mathew AV, Koenigsknecht M, Hanash A, Toubai T, and Eubacterium rectale ATCC 33656 explain the bifidogenic and et al. Gut microbiome-derived metabolites modulate intestinal epithelial butyrogenic effects of arabinoxylan oligosaccharides. Appl Environ cell damage and mitigate graft-versus-host disease. Nat Immunol. Microbiol. 2015;81(22):7767–81. doi:10.1128/aem.02089-15. 2016;17(5):505–13. doi:10.1038/ni.3400. 36. Guggenheim KY. Rudolf Schoenheimer and the concept of the dynamic 13. Mathewson N, Mathew A, Oravecz-Wilson K, Wu J, Toubai T, Cummings E, state of body constituents. J Nutr. 1991;121(11):1701–4. et al. Microbial metabolites modulate GI mucosal damage from graft versus host disease (GHVD). (MUC4P.850). J Immunol. 2014;192(1 Supplement):133.26. 14. Smith MI, Yatsunenko T, Manary MJ, Trehan I, Mkakosya R, Cheng J, et al. Gut microbiomes of Malawian twin pairs discordant for kwashiorkor. Science. 2013;339(6119):548–54. doi:10.1126/science.1229000. 15. Ohland Christina L, Jobin C. Bugs and food: a recipe for cancer? Cell Metab. 2014;20(6):937–8. 16. Birt DF, Boylston T, Hendrich S, Jane J-L, Hollis J, Li L, et al. Resistant starch: promise for improving human health. Adv Nutr Int Review J. 2013;4(6):587–601. doi:10.3945/an.113.004325. 17. Kozich JJ, Westcott SL, Baxter NT, Highlander SK, Schloss PD. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol. 2013. doi:10.1128/aem.01043-13. 18. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: open-source, platform-independent, community- supported software for describing and comparing microbial communities. Submit your next manuscript to BioMed Central Appl Environ Microbiol. 2009;75(23):7537–41. 19. R Development Core Team. R: a language and environment for statistical and we will help you at every step: computing. In: Computing RFfS, editor. Vienna, Austria: R Foundation for • We accept pre-submission inquiries Statistical Computing; 2015. 20. Hammer O, Harper DAT, Ryan PD. PAST: paleontological statistics software � Our selector tool helps you to find the most relevant journal package for education and data analysis. In: Electronica P, editor.2001. � We provide round the clock customer support 21. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, et al. � Convenient online submission Metagenomic biomarker discovery and explanation. Genome Biol. 2011;12(6):R60. doi:10.1186/gb-2011-12-6-r60. � Thorough peer review 22. Thorndike RL. Who belongs in the family? Psychometrika.18(4):267-76. � Inclusion in PubMed and all major indexing services doi:10.1007/bf02289263. � Maximum visibility for your research 23. Ze X, Duncan SH, Louis P, Flint HJ. Ruminococcus bromii is a keystone species for the degradation of resistant starch in the human colon. ISME J. Submit your manuscript at 2012;6(8):1535–43. doi:10.1038/ismej.2012.4. www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Microbiome Springer Journals

Variable responses of human microbiomes to dietary supplementation with resistant starch

Download PDF
9 pages

Loading next page...
 
/lp/springer-journals/variable-responses-of-human-microbiomes-to-dietary-supplementation-LaEFJc0wRk

References (28)

Publisher
Springer Journals
Copyright
Copyright © 2016 by The Author(s).
Subject
Biomedicine; Medical Microbiology; Bioinformatics; Microbial Ecology; Microbiology; Microbial Genetics and Genomics; Virology
eISSN
2049-2618
DOI
10.1186/s40168-016-0178-x
pmid
27357127
Publisher site
See Article on Publisher Site

Abstract

Background: The fermentation of dietary fiber to various organic acids is a beneficial function provided by the microbiota in the human large intestine. In particular, butyric acid contributes to host health by facilitating maintenance of epithelial integrity, regulating inflammation, and influencing gene expression in colonocytes. We sought to increase the concentration of butyrate in 20 healthy young adults through dietary supplementation with resistant starch (unmodified potato starch—resistant starch (RS) type 2). Methods: Fecal samples were collected from individuals to characterize butyrate concentration via liquid chromatography and composition of the microbiota via surveys of 16S rRNA-encoding gene sequences from the Illumina MiSeq platform. Random Forest and LEfSe analyses were used to associate responses in butyrate production to features of the microbiota. Results: RS supplementation increased fecal butyrate concentrations in this cohort from 8 to 12 mmol/kg wet feces, but responses varied widely between individuals. Individuals could be categorized into three groups based upon butyrate concentrations before and during RS: enhanced, high, and low (n = 11, 3, and 6, respectively). Fecal butyrate increased by 67 % in the enhanced group (from 9 to 15 mmol/kg), while it remained ≥11 mmol/kg in the high group and ≤8 mmol/kg in the low group. Microbiota analyses revealed that the relative abundance of RS-degrading organisms—Bifidobacterium adolescentis or Ruminococcus bromii—increased from ~2 to 9 % in the enhanced and high groups, but remained at ~1.5 % in the low group. The lack of increase in RS-degrading bacteria in the low group may explain why there was no increase in fecal butyrate in response to RS. The microbiota of individuals in the high group were characterized by an elevated abundance of the butyrogenic microbe Eubacterium rectale (~6 % in high vs. 3 % in enhanced and low groups) throughout the study. Conclusions: We document the heterogeneous responses in butyrate concentrations upon RS supplementation and identify characteristic of the microbiota that appear to underlie this variation. This study complements and extends other studies that call for personalized approaches to manage beneficial functions provided by gut microbiomes. Background functions could promote health and reduce the incidence The microbiota in the large intestine provides several of preventable diseases including obesity and type 2 dia- functions that are beneficial to human health such as pro- betes [3], colon cancer [4], chronic and acute undernutri- ducing short-chain fatty acids, modifying primary to sec- tion [5], and infections by Clostridium difficile [6]. ondary bile acids, and providing colonization resistance to One of the beneficial functions derived from the co- some enteric pathogens [1, 2]. Managing this community lonic microbiota is the production of butyric acid that is of microbes to maintain and improve these beneficial generated from the fermentation of dietary fiber. The conjugate base of the acid—butyrate—is the preferred * Correspondence: [email protected] energy source for colonocytes [7]. Butyrate improves the Department of Internal Medicine, University of Michigan, Ann Arbor, MI intestinal barrier by facilitating tight-junction assembly 48105, USA [8], suppresses inflammatory and allergic responses by Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Venkataraman et al. Microbiome (2016) 4:33 Page 2 of 9 inducing differentiation of colonic regulatory T cells [9], 4 fecal samples during RS regulates cell apoptosis [10], and stimulates production of anorectic hormones [11]. Indeed, reduced concentra- Acclimatization period tions of butyrate have been associated with the incidence 4 baseline fecal samples No fecal samples before consuming RS of graft-versus-host disease [12, 13], kwashiorkor [14], 4 8 11 17 colon cancer [15], and obesity [11]. Especially in these Day circumstances, increased butyrate production in the Fig. 1 Experimental design. Participants supplemented their habitual large intestine may be beneficial to human health. diets with resistant starch. Filled symbols represent points at which One approach to increasing butyrate production is to fecal samples were collected deliver more fermentable carbohydrates to gut micro- biomes. Here, we report on the impact of dietary supple- mentation with resistant starch (unmodified potato 48 g of potato starch—24-g doses twice per day—for starch; RS type 2) on fecal butyrate concentrations and seven more days. Participants were provided a 1- composition of the gut microbiota in 20 healthy young tablespoon scoop, which they used to measure out the adults. RS consists of starch that is resistant to hydroly- appropriate amount of potato starch (each tablespoon sis by human enzymes and passes through the small in- ≈12 g) and consumed it after mixing the starch with cold testine unabsorbed. In the large intestine, RS can be water. metabolized and then fermented by microbes to a variety of products, including butyric acid [16]. In this study, we Fecal collection find that the response to RS supplementation varies be- Participants were provided waxed tissue paper (Epitope tween individuals in ways that can be explained, at least Diagnostics, San Diego, CA), which was laid down on in part, by variation in the composition of their micro- the water in the toilet bowl prior to defecation. This biota. Recognizing inter-individual variability in re- paper sticks to the sides of the toilet bowl so that the sponses to fiber supplementation and determining the fecal sample is not readily contaminated by water in the microbiota characteristics that underlie it are essential toilet. Fecal samples were collected using the deoxyri- first steps towards personalized plans for managing gut bose nucleic acid (DNA) Genotek Omnigene Gut microbiomes for desirable functionality, including the Collection kits (DNA Genotek, Ontario, Canada) follow- production of butyrate. ing the manufacturer’s instructions. The DNA Genotek tube contains 2 mL of a stabilization buffer and a steel Methods ball to facilitate mixing of the fecal material with the Participants buffer. Approximately, 0.67 g of fecal material was col- Prospective volunteers were students in research-based lected per tube. The kits were returned to a −20 °C sections of Introductory Biology 173 at the University of freezer within 24 h of collection. Frozen samples were Michigan. Individuals with a self-reported history of thawed and aliquots were withdrawn for DNA extrac- bowel disorders such as irritable bowel syndrome, in- tion, measurement of fermentation products, and to cre- flammatory bowel disease, or colorectal cancer were ex- ate sample archives. cluded from the study. Twenty individuals (10 males, 10 females) participated in the 3-week study. The age range DNA extraction, PCR, sequencing, and sequence was 19–20 years, and the BMI range was from 19–63. processing Each participant gave his or her written, informed For DNA extraction, 0.25 mL of the fecal suspension in consent before participating in the study. This research the DNA Genotek tube was deposited into a MoBio was carried out in compliance with the Helsinki Declar- PowerMag Soil DNA Isolation Bead Plate. DNA was ex- ation and was approved by the Institutional Review tracted following MoBio’s instructions on an epMotion Board of the University of Michigan Medical School 5075 liquid handling workstation (Eppendorf, Hauppauge, (HUM00094242). NY). The V4 region of the 16S-ribosomal RNA (rRNA) encoding genes was amplified from each sample using Study design the dual-indexing sequencing strategy [17]. Sequen- Participants consumed their habitual diet throughout cing was performed on the Illumina MiSeq platform, the study period. During the intervention phase, raw un- using a MiSeq Reagent Kit V2 500 cycles (Cat# MS- modified potato starch (Bob’s Red Mill, Milwaukie, OR) 102-2003), according to the manufacturer’sinstruc- was gradually added to their diet (day 1—12 g, day tions. Sequences were curated using mothur v1.31.2 2—24 g, day 3—48 g; Fig. 1). This potato starch contains [18] and clustered into operational taxonomic units at approximately 50 % resistant starch (type 2) by weight. ≥ 97 % sequence identity using the average neighbor After the acclimatization period, the subjects consumed algorithm. Total umodified potato starch consumed (g) Venkataraman et al. Microbiome (2016) 4:33 Page 3 of 9 Measurement of fermentation products composition of gut microbiota in the study population One millimeter of fecal suspension was aliquoted from (p < 0.05). For further analyses, we used the median each DNA Genotek OmniGut tube and centrifuged at abundance of each operational taxonomic unit (OTU) in 4 °C for 10 min (4500×g). The supernatant fraction was each individual before and during RS consumption. To withdrawn and passed sequentially through 1.2, 0.65, identify specific OTUs that had changed with RS con- and 0.22-μm hydrophilic low protein-binding Durapore sumption, LEfSe and Random Forest analyses were per- membrane filters (EMD Millipore, Darmstadt, Germany). formed for each individual separately (comparing their Filtered samples were stored at −20 °C until analysis. four samples before and four samples during RS con- Once thawed, samples were maintained at 4 °C in an sumption). LEfSe [21] was implemented within mothur autosampler. A Shimadzu HPLC (Shimadzu Scientific using the correction for multiple comparisons. PERMA- Instruments, Columbia, MD) equipped with a dual UV NOVA, analysis of similarity (ANOSIM), and permuta- (214 nm for short-chain fatty acids (SCFAs))/refractive tional analyses of dispersions (PERMDISP) analyses were index detector (for ethanol) and an Aminex HPX-87H performed with the package vegan in R. Random Forest column (Bio-Rad Laboratories, Hercules, CA) heated to was implemented using the package randomForest in R. 50 °C was used to measure acetic, propionic, and butyric acids and ethanol in the filtrate. The mobile phase was Results 0.01 N H SO at a flow rate of 0.6 ml/min, and the injec- The impact of RS on the composition of the gut micro- 2 4 tion volume was 10 μl. Samples were randomized in re- biota and the concentration of three short-chain fatty gard to injection order, and the average value from two acids (acetic, propionic, and butyric) in feces was deter- technical replicates was used in all subsequent analyses. mined in 20 healthy young adults. RS was gradually intro- Ethanol, which was a component of the stabilization buf- duced into their regular diets during an acclimatization fer, served to monitor injection volumes. Samples in which period of 3 days (Fig. 1). Thereafter, study participants in- the peak height for ethanol was less than 70 % of the aver- cluded 48 g of unmodified potato starch in their daily diet age peak height for ethanol in all samples were excluded (~24 g as RS) for 7 days. Four fecal samples were collected from further analysis. Eight external standards (0.1– before the introduction of RS and another four during the 20 mM) for acetic, propionic, and butyric acids were used period of maximum RS supplementation (Fig. 1). to generate the standard curve. These standards were run There was considerable intra-individual variability in after every 100 samples. LC Solutions Software (Shimadzu the concentration of acetate, propionate, and butyrate Scientific Instruments, Columbia, MD) was used to curate (coefficient of variation = 20–90 %). This variability is the data and calculate concentrations based on the stand- not unexpected because fecal SCFA concentrations are ard curve generated during the run. The concentrations influenced by host absorption, transit time through the were then normalized by the average weight of the fecal GI tract, host diet, and time since last meal. Due to this samples (0.67 g) collected in the DNA Genotek tubes. variability and the number of samples, Student’s t tests did not reveal measurable differences in the concentra- Statistical analyses tion of SCFA in response to RS supplementation for an All statistical analyses were performed using RStudio individual (four samples before vs. four during RS for version 0.99.489 [19] and the software PAST [20]. To each individual; p ≥ 0.10 for all individuals). However, a evaluate the impact of RS supplementation on acetate, repeated measures ANOVA showed that RS supplemen- propionate, and butyrate concentrations in the study tation increased the fecal concentration of butyrate by population, a nested repeated measures analysis of vari- 50 % and acetate by 26 % in the study population as a ance (ANOVA) was used. This revealed that fecal butyr- whole (p = 0.02 and 0.03, respectively; Table 1). The ate and acetate concentrations increased with RS intake concentrations of propionate did not change signifi- in the study population (p < 0.05). For all subsequent cantly in this cohort (p =0.85; Table 1). analyses, the median value of butyrate for each person before and during RS consumption was used so that the Table 1 Effect of dietary supplementation with RS on number of samples compared was not artificially in- concentration of select fermentation products (mmol/kg flated. Paired or unpaired t tests were employed as ap- wet feces) propriate. To determine if RS intake altered microbial Fermentation Before RS During RS Change p value (repeated community composition, a permutational analysis of product measures ANOVA) Median ± IQR Median ± IQR variance (PERMANOVA) analysis was conducted with Butyric acid 8 ± 6 12 ± 7 50 % 0.03 the Bray-Curtis similarity index. Each individual was Acetic acid 27 ± 6 34 ± 10 26 % 0.02 used as the blocking factor to account for repeated mea- Propionic acid 13 ± 6 12 ± 5 −8 % 0.85 sures of microbiota composition from each individual. This analysis revealed that RS intake altered the IQR interquartile range Venkataraman et al. Microbiome (2016) 4:33 Page 4 of 9 Like the fecal SCFA concentrations, there was also some intra-individual variability in the composition of microbiota. However, at least 75 % of OTUs were con- sistently detected in all samples from an individual with 10 the coefficient of variation in their relative abundances ranging from 39 to 65 %. Unlike the SCFAs, where stat- Individuals arranged in ascending order of butyrate before RS istical tests did not detect differences in response to RS Fig. 2 Median butyrate concentrations for each individual before for an individual, ANOSIM tests showed that RS supple- (triangles) and during consumption of RS (circles). Dotted and dashed mentation altered the composition of microbiota in most lines denote the median values for butyrate before and during RS, individuals in the study (ANOSIM with Bray-Curtis respectively, for the entire study population similarity; four samples before vs. four during RS for each individual; p < 0.10 for 16 out of 20 individuals). This conclusion was corroborated by results from a Baseline butyrate concentrations were not predictive of PERMANOVA analysis to determine if RS supplementa- butyrate concentrations during RS supplementation tion altered the composition of microbiota in the overall (R = 0.08, p = 0.20). study population. In order to incorporate intra- To identify characteristics of the microbiota that may individual variability, each individual was considered as underlie the variable responses to RS, we first used Ran- the blocking factor in the PERMANOVA. This analysis dom Forest regression to identify relationships between revealed that the composition of the gut microbiota was the abundances of OTUs and butyrate concentrations altered with RS supplementation in our study population before and during consumption of RS. No OTUs were (PERMANOVA using Bray-Curtis similarity blocking for particularly strong predictors of butyrate concentrations each individual; p = 0.001). A PERMDISP analysis further either before or during RS consumption. Butyrate con- revealed that the PERMANOVA was not affected by dif- centrations before RS were weakly related to baseline ferences in the dispersion of communities before and abundances of OTU 4 (Eubacterium rectale)(R = 0.14; during RS (PERMDISP p = 0.44). p = 0.10). Unexpectedly, this relationship was not detect- The relative abundance of OTUs belonging to the able during RS supplementation. phylum Actinobacteria increased with RS, and there was Population-wide relationships between OTUs and bu- a small decrease in the abundance of Firmicutes.No tyrate concentrations could be masked by the heterogen- changes were detected in the relative abundances of eity of both variables between individuals. We therefore Bacteroidetes or Proteobacteria (Table 2). Finally, RS looked for correlations between features of the micro- supplementation did not change the overall richness or biota and butyrate concentrations in subsets of partici- evenness of the microbial community in the study popu- pants that had similar responses in fecal butyrate lation (repeated measures ANOVA Chao1 index before following RS supplementation. The study population vs. during p = 0.49; Simpson’s index before vs. during was separated into three groups using k-means cluster- p = 0.96). ing based on butyrate concentrations before and during Therefore, despite the intra-individual variability in RS. An elbow plot [22] revealed that there were three fecal SCFA concentrations and composition of micro- “clusters”. The categories identified were enhanced, high, biota, repeated measures ANOVA and PERMANOVA and low (Fig. 3a, b). The concentration of butyrate in reveal that RS consumption led to increases in fecal bu- the “enhanced” group (n = 11) increased significantly tyrate concentrations and altered the composition of the following consumption of RS (from 9 to 15 mmol/kg microbiota in the study population. However, the start- wet feces, paired t test p = 0.0003). Individuals in the ing concentrations of butyrate and the changes in butyr- “high” group (n = 3) maintained butyrate concentrations ate during RS varied widely between individuals (Fig. 2). ≥11 mmol/kg wet feces during the course of the study. Individuals in the “low” group (n = 6) had less than or equal to 8 mmol butyrate/kg wet feces both before and Table 2 Effect of dietary supplementation with RS on relative during RS (paired t test p = 0.14; Fig. 3c). abundance of four dominant bacterial phyla The OTUs that distinguished these three clusters were Phylum Before RS During RS (%) Change p value identified using Random Forest analysis and LEfSe Median ± IQR Median ± IQR paired t test (Table 3). Random Forest revealed OTU #7 as the most Actinobacteria 1.3 ± 0.6 6.2 ± 0.6 377 % 0.02 prominent feature of the microbiota distinguishing the Firmicutes 37.1 ± 10.7 33.2 ± 5.2 -11 % 0.04 low from the enhanced group (Table 3). Sequences Bacteroidetes 53.3 ± 13.2 51.5 ± 9 -3 % 0.82 within this OTU are identical to those from Bifidobac- Proteobacteria 4.7 ± 4.1 4.8 ± 4.1 2 % 0.82 terium adolescentis. The relative abundance of this OTU IQR interquartile range before RS was similar in all three groups (~0.7–1.4 %; Butyrate (mmol/kg wet feces) Venkataraman et al. Microbiome (2016) 4:33 Page 5 of 9 ab c During RS Before RS U005 p < 0.001 U011 U031 U014 U034 U029 U032 U020 U025 U035 U007 U033 U004 8 U006 U015 U022 U026 U008 Before During Before During Before During U001 5 101520 U024 Enhanced High Low Butyrate before RS (mmol/kg wet feces) Butyrate 5 10 15 20 (mmol/kg wet feces) Fig. 3 Clustering individuals based upon their butyrate response to RS. a Three groups generated by k-means clustering. b Median butyrate concentrations before and during RS supplementation. c Average butyrate concentrations before and during RS in the three groups t test p > 0.27; Fig. 4a). However, the enhanced and was detected in 14 out of 20 total individuals and in- high groups had dramatically higher abundances of creased in abundance in 12 individuals. This OTU did this OTU during RS (average 8.9 % in enhanced and not increase in individuals U026 and U024 (LEfSe 7.8 % in high; p < 0.05; Fig. 4a). The abundance of p > 0.05 correcting for multiple comparisons). However, this OTU did not change in the low group (average OTU #7 [B. adolescentis] was not detected in six individ- before RS = 1.5 %, average during = 3.7 %, p = 0.13). uals. In these individuals, LEfSe was used to find other This finding was further corroborated with a LEfSe OTUs that increased in abundance during RS supple- analyses comparing four samples before to four dur- mentation, since these could be RS-degrading bacteria. ing RS consumption for each individual. As one In three individuals in whom B. adolescentis sequences would expect, the abundance of OTU 7 increased in did not increase, OTU 19 increased in abundance (LEfSe 8 out of 11 individuals in the enhanced group, 2 out p < 0.05 correcting for multiple comparisons). Sequences of 3 individuals in the high group, but in only 2 out in that OTU are identical to Ruminococcus bromii, an- of 6 individuals in the low group (LEfSe p <0.05 cor- other group of RS-degrading bacteria [23]. In individual recting for multiple comparisons). U005, neither B. adolescentis nor R. bromii increased in Cultivars of B. adolescentis are capable of breaking abundance with RS. A potential candidate for a RS- down RS [23]. OTU #7 in which sequences were identi- degrading organism in this individual is OTU 50, whose cal to the 16S rRNA encoding gene of B. adolescentis average abundance increased from 2.6 to 7.1 % albeit not statistically significant (p = 0.29). Sequences in this OTU are most closely related to Ruminiclostridium Table 3 Results of Random Forest regression and LEfSe to [Eubacterium] siraeum. This organism has not been identify OTUs that distinguish the three response groups reported to degrade RS, but it is in the same taxonomic Comparison Distinguishing microbiota features family as R. bromii. LEfSe Random Forest Since B. adolescentis and R. bromii are the strongest Low vs. enhanced None identified [OTU 7] candidates for RS-degrading organisms in this study, we during RS summed the abundance of these two organisms for each High vs. low [OTU 4] [OTU 4] before and during RS before and during RS ‡ individual (Fig. 4b). Individuals in all three groups start [OTU 3] before RS with a similar abundance of these RS-degrading mi- High vs. enhanced [OTU 4] [OTU 4] before and during RS before and during RS crobes. RS elicits dramatic increases in their abundance A post-hoc ANOVA analysis revealed that the abundance of OTU #3 was not in high and enhanced groups, but not in the low group significantly different between the groups being compared (p > 0.10). Hence, this OTU was not considered further (Fig. 4b). Butyrate during RS (mmol/kg wet feces) Enhanced Low High Butyrate (mmol/kg wet feces) Venkataraman et al. Microbiome (2016) 4:33 Page 6 of 9 OTU 7 OTU 19 Summed B. adolescentis R. bromi i abundance (%) p = 0.11 p = 0.003 p = 0.02 Before During Before During Before During U015 U022 U026 U008 U001 U024 U005 U011 U031 U014 U034 U029 U032 U020 U025 U035 U007 U033 U004 Before During Before During Before During U006 Low Enhanced High Abundance (%) 5 10 15 20 Fig. 4 Identifying microbiota features that distinguish the low from other groups. a Median abundance of OTUs taxonomically related to B. adolescentis and R. bromii before and during RS for each individual. b Summed abundance of the two proposed RS-degrading OTUs before and during RS in the three groups Random Forest and LEfSe also revealed that OTU 4 supplementation (Fig. 5a, LEfSe p < 0.05 correcting for distinguished the high group from the enhanced and low multiple comparisons), but in general, its abundance did groups (Table 3). Sequences within this OTU are identi- not change as a result of RS addition to diet in any of cal to E. rectale—a known and prominent butyrogenic the three groups (Fig. 5b). We looked specifically at microbe in human guts [24]. E. rectale was more abun- other butyrogenic organisms known to be present in the dant in individuals in the high group than in the other human colon such as Faecalibacterium prausnitzii [25]. two groups—both before and during RS (Fig. 5b). While An OTU with the same V4 16S-rRNA encoding gene se- individuals in the high group fall into a very distinct quence as F. prausnitzii was present at about 3–6% cluster in the k-means clustering (Fig. 3), there are only relative abundance in all three groups, but did not three individuals in this group. This limited sample size change with RS in any of the groups (t test p > 0.15). does constrain the strength of the finding that E. rectale Since B. adolescentis—the prominent RS-degrading mi- is more abundant in the high group vs. the other two crobe in our cohort—produces lactate as the primary groups. Some individuals in the enhanced group exhib- fermentation product, we also investigated if organisms ited increases in abundances of E. rectale during RS known to produce butyrate from lactate increased in a b p = 0.02 OTU 4 E. rectale p = 0.02 Before During U033 U004 U006 7.5 U005 U011 U031* U014 U034 5.0 U029* U032 U020* U025* U035* 2.5 U007 U015 U022 U026 0.0 U008 BeforeDuring Before During Before During U001 High Low Enhanced U024 Abundance (%) 510 15 20 Fig. 5 The relative abundance of OTU 4 Eubacterium rectale distinguishes the high from other groups. a Median abundance of OTU 4 before and during RS for each individual. Individuals in whom the abundance of this OTU increased with RS are marked with an asterisk (LEfSe p < 0.05). b Abundance of OTU 4 before and during RS in the three groups High Enhanced Low Low Enhanced High OTU 4 [E. rectale] (%) Proposed RS−degrading OTUs (%) Venkataraman et al. Microbiome (2016) 4:33 Page 7 of 9 abundance. OTUs in which sequences were most closely clustering using butyrate concentrations before and dur- related to Eubacterium halii and Anaerostipes caccae ing RS identified three groups: enhanced, high, and low. were present at less than 0.4 % abundance in our cohort Butyrate increased on average from 9 to 15 mmol/kg wet and did not change with RS (paired t test p = 0.4). feces in the enhanced group, whereas butyrate concentra- tions remained consistently high or low in the other two Discussion groups (≥11 and ≤8 mmol/kg wet feces, respectively; The impact of dietary supplementation with resistant Fig. 3). With these clusters, we were able to identify fea- starch on fecal butyrate concentrations and the compos- tures of the microbial communities that differentiated the ition of the microbiota was examined in 20 healthy three groups. young adults. In this cohort, the average ratios of aceta- The relative abundance of proposed RS-degrading te:propionate:butyrate before and during RS supplemen- bacteria—B. adolescentis or R. bromii—increased with tation were 58:25:15 and 58:22:20, respectively. These RS supplementation in the enhanced and high groups ratios agree well with previously documented ratios of (from 2 to 9 %), but not in the low group (~1.5 % ~60:20:20 in human feces [26–29]. In response to RS throughout; Fig. 4). This finding complements and ex- supplementation, the concentration of fecal butyrate in- tends previously published reports on the effects of RS creased from 8 to 12 mmol/kg wet feces in the overall supplementation on the composition of the human gut study population (repeated measures ANOVA p = 0.02). microbiota. In a study with overweight adult males, only The inter-individual variation in butyrate concentrations individuals with detectable abundances of R. bromii in before and during RS was striking (Fig. 2). Most previ- their gut microbiota were able to degrade RS (type ous studies with RS have documented only population- 3—Novelose 330; Walker et al. [32]). In another study wide response in butyrate concentrations [16], while one with ten human subjects (five males and females; 28– other study reported considerable heterogeneity between 38 years old), dietary supplementation with either type 4 individuals in regard to fecal butyrate concentrations RS (FiberSym RW) or type 2 RS (HiMaize260) increased [30]. This pronounced inter-individual variability sug- the relative abundance of B. adolescentis and R. bromii, gests that a single approach to improving beneficial respectively (Martinez et al. [33]). These in vivo studies functions from the microbiome is unlikely to be univer- are nicely complemented by in vitro microcosm studies, sally successful. Rather, personalized approaches may be which showed that isolates of R. bromii and B. adoles- needed to manage microbiomes for health. In line with centis are capable of degrading several forms of RS [23]. this endeavor, we attempted to identify features in the Thus, different types of resistant starches are likely to microbiota that could explain different responses to diet- promote the growth of B. adolescentis or R. bromii, with ary supplementation with RS. both organisms exhibiting substrate specificity depending Random Forest was used to identify relationships be- upon the type and source of RS. tween butyrate concentrations and OTUs in the study The relative abundance of RS-degrading organisms population. Weak relationships were identified between and butyrate concentrations did not increase in the low butyrate concentrations before RS and the abundances group. This suggests that their microbiota did not break of OTU 4 (E. rectale). Random Forest failed to reveal down RS and so could not lead to increased butyrate any relationships between OTU abundances and butyr- production. This suggestion is supported by the fact that ate concentrations during RS supplementation. It is pos- the concentrations of neither acetate nor propionate in- sible that the number of individuals in this study (n = 20) creased in the low group. So why does the abundance of may have been too small to detect robust correlations RS-degrading organisms remain at 1.5 % in the low between butyrate and the abundance of OTUs at the group? Possible explanations include limitation by antag- population level. Alternatively, it is not necessary that onistic microbes or lack of synergistic microbes. Our the abundance of a single organism be correlated with attempts to identify antagonistic and synergistic interac- butyrate concentrations. After all, the potential for pro- tions by constructing correlation networks [34] between ducing butyrate is fairly widespread within the phylum RS-degrading organisms and other OTUs did not yield Firmicutes [31]. Rather, we expect that the abundance of any compelling relationships. genes encoding butyryl-CoA:acetate CoA transferase In addition to RS-degrading organisms, another OTU (but) and butyrate kinase (buk) would be correlated to (OTU #4) was identified as distinguishing the high butyrate concentrations. group from the enhanced and low groups (Table 3, Our next step was to create groupings in the data Fig. 5). Its V4 sequence is identical to that of E. rectale, based upon similar responses to RS. This clustering ap- well established as a prominent butyrogenic bacterium proach should constrain the variability within each group in the human gut microbiome [24, 35]. Its relative abun- and increase the probability of identifying microbiota fea- dance in the high group was consistently about 6 %, tures that characterize each type of response. K-means compared to 3 % in the other two groups. Surprisingly, Venkataraman et al. Microbiome (2016) 4:33 Page 8 of 9 the abundance of E. rectale did not change in any of the changes in the microbiota in the enhanced group groups in response to RS—even though butyrate con- (ANOSIM R = 0.64). It is tempting to suggest that the centrations increased appreciably in the enhanced group microbiota of these individuals are performing well with (from 9 to 15 mmol/kg wet feces). E. rectale generates regard to butyrate production, and they do not benefit butyrate from acetate, and there is a net gain of ATP in from additional dietary input of fermentable carbohy- that process [35]. So if E. rectale was responsible for in- drates. A third subset of our study population (6 of 20) creased butyrate production in the enhanced group with had consistently low concentrations of butyrate even RS, one might expect an increase in its relative abun- when consuming RS (≤8 mmol/kg wet feces). In this dance. We offer two possible explanations for the in- group, RS-degrading organisms did not increase in creased butyrate production in the enhanced group abundance, suggesting that their microbiota did not without a measurable increase in the abundance of E. break down RS. Based upon this result, we propose that rectale. First, changes in the relative abundance of E. rec- increasing butyrate in these individuals will require tale may be subtle and masked by the dramatic increase either (i) testing another form of dietary fiber such as in the abundance of RS-degrading organisms (~from 2 inulin or arabinoxylan that their microbiota might de- to 10 % in the enhanced group). It is also possible that grade, (ii) a synbiotic approach that combines a dietary populations of E. rectale take longer to respond to RS fiber with the appropriate fiber-degrading bacteria, or supplementation. We used an acclimatization period of (iii) targeted removal of microbes if any are antagonistic only 3 days during which the amounts of RS in diet were to RS-degrading organisms. The findings of this study il- gradually increased, and after this period, four fecal sam- lustrate the importance of studying individual responses ples were collected. It remains to be seen whether there to dietary modifications. This will uncover the mecha- is a detectable increase in the abundance of E. rectale nisms that underlie these responses and in time provide following a longer duration of RS consumption. Also, as actionable insights towards precision management of mentioned earlier, it is well documented that butyrate microbiomes. production is widespread within the phylum Firmicutes in human gut microbiomes. Therefore, rather than a sin- Abbreviations ANOVA, analysis of variance; DNA, deoxyribose nucleic acid; g, gram(s); gle organism, if more acetate is being converted to bu- h, hour(s); HPLC, high-pressure liquid chromatography; LC, liquid tyrate, the genes encoding butyryl-CoA:acetate CoA chromatography; ml, milliliter(s); mM, millimolar; °C, degrees centigrade; transferase (but) and butyrate kinase (buk) should in- OTU, operational taxonomic unit; rRNA, ribosomal RNA; RS, resistant starch; μl, microliters crease in abundance with RS supplementation. Acknowledgements Conclusions The authors thank Drs. Frank Gerberick and JoseCarlos Garcia-Garcia (Procter & Gamble Company, Cincinnati, OH) for providing feedback on Our data show that dietary supplementation with RS data analysis and interpretation. type 2 as unmodified potato starch increases fecal butyr- ate concentration, but with remarkable inter-individual Funding variation. We were able to infer potential explanations This research was supported by grants from the Procter and Gamble Company and the Howard Hughes Medical Institute (52008119). for some of these differential effects by investigating the composition of microbiota. Fecal butyrate concentra- Availability of data and materials tions increased by an average of 67 % in a subset of the The sequence and metadata have been deposited in the GenBank database [accession number: SRP067761]. study population (n = 11 of 20; 9 to 15 mmol/kg wet feces). Most individuals in this group showed a dramatic Authors’ contributions increase in the relative abundance of the RS-degrading AV, JRS, AS, CW, KRT, and TMS conceived of the study, participated in its design and coordination, and helped draft the manuscript. AV, AS, and CW organisms—B. adolescentis or R. bromii. In five of these conducted the microbiota analyses. JRS conducted the analyses of individuals, the prominent butyrogenic microbe E. rec- fermentation products. KRT participated and advised in statistical analyses. All tale also increased in abundance. Another subset of the authors read and approved the final manuscript. population (3 of 20) consistently maintained high butyr- Competing interests ate concentrations both before and with RS (≥12 mmol/ The authors declare that they have no competing interests. kg wet feces). In this subset, RS-degrading organisms in- creased in abundance, suggesting that RS is being de- Author details Department of Internal Medicine, University of Michigan, Ann Arbor, MI graded. But there was no concomitant increase in fecal 48105, USA. Present address: Department of Biology, University of butyrate concentrations. These individuals may be ex- 3 Minnesota, Duluth, MN 55812, USA. Present address: Department of periencing a plateau effect [36] in butyrate production Immunology and Microbiology, Wayne State University School of Medicine, Detroit, MI 48201, USA. even before RS is administered. In fact, the microbiota structures of these individuals also did not change as Received: 16 February 2016 Accepted: 10 June 2016 much with RS (ANOSIM R = 0.17) as compared to Venkataraman et al. Microbiome (2016) 4:33 Page 9 of 9 References 24. Duncan SH, Flint HJ. Proposal of a neotype strain (A1-86) for Eubacterium 1. Tremaroli V, Backhed F. Functional interactions between the gut microbiota rectale. Request for an opinion. Int J Syst Evol Microbiol. 2008;58(Pt 7):1735–6. and host metabolism. Nature. 2012;489(7415):242–9. doi:10.1099/ijs.0.2008/004580-0. 2. Flint HJ, Duncan SH, Scott KP, Louis P. Links between diet, gut microbiota 25. Khan MT, Duncan SH, Stams AJM, van Dijl JM, Flint HJ, Harmsen HJM. The gut anaerobe Faecalibacterium prausnitzii uses an extracellular electron shuttle to composition and gut metabolism. Proc Nutr Soc. 2015;74(1):13–22. grow at oxic-anoxic interphases. ISME J. 2012:doi:10.1038/ismej.2012.5. doi:10.1017/s0029665114001463. 26. den Besten G, van Eunen K, Groen AK, Venema K, Reijngoud D-J, Bakker BM. 3. Hartstra AV, Bouter KE, Backhed F, Nieuwdorp M. Insights into the role of The role of short-chain fatty acids in the interplay between diet, gut the microbiome in obesity and type 2 diabetes. Diabetes Care. 2015;38(1): microbiota, and host energy metabolism. J Lipid Res. 2013;54(9):2325–40. 159–65. doi:10.2337/dc14-0769. doi:10.1194/jlr.R036012. 4. Zackular JP, Baxter NT, Iverson KD, Sadler WD, Petrosino JF, Chen GY, et al. 27. Hijova E, Chmelarova A. Short chain fatty acids and colonic health. Bratisl The gut microbiome modulates colon tumorigenesis. mBio. 2013;4(6). doi: Lek Listy. 2007;108(8):354–8. 10.1128/mBio.00692-13. 28. Binder HJ. Role of colonic short-chain fatty acid transport in diarrhea. Annu 5. Blanton LV, Charbonneau MR, Salih T, Barratt MJ, Venkatesh S, Ilkaveya O, Rev Physiol. 2010;72:297–313. doi:10.1146/annurev-physiol-021909-135817. et al. Gut bacteria that prevent growth impairments transmitted by 29. Cummings JH, Pomare EW, Branch WJ, Naylor CP, Macfarlane GT. Short microbiota from malnourished children. Science. 2016;351(6275):3311-1–7. chain fatty acids in human large intestine, portal, hepatic and venous blood. 6. Buffie CG, Bucci V, Stein RR, McKenney PT, Ling L, Gobourne A, et al. Gut. 1987;28(10):1221–7. Precision microbiome reconstitution restores bile acid mediated resistance 30. McOrist AL, Miller RB, Bird AR, Keogh JB, Noakes M, Topping DL, et al. Fecal to Clostridium difficile. Nature. 2015;517(7533):205–8. doi:10.1038/nature13828. butyrate levels vary widely among individuals but are usually increased 7. Donohoe DR, Garge N, Zhang X, Sun W, O’Connell TM, Bunger MK, et al. by a diet high in resistant starch. J Nutr. 2011;141(5):883–9. doi:10.3945/ The microbiome and butyrate regulate energy metabolism and autophagy jn.110.128504. in the mammalian colon. Cell Metab. 2011;13(5):517–26. doi:10.1016/j.cmet. 31. Vital M, Howe AC, Tiedje JM. Revealing the bacterial butyrate synthesis 2011.02.018. pathways by analyzing (meta)genomic data. mBio. 2014;5(2). doi:10.1128/ 8. Peng L, Li ZR, Green RS, Holzman IR, Lin J. Butyrate enhances the intestinal mBio.00889-14. barrier by facilitating tight junction assembly via activation of AMP-activated 32. Walker AW, Ince J, Duncan SH, Webster LM, Holtrop G, Ze X, et al. protein kinase in Caco-2 cell monolayers. J Nutr. 2009;139(9):1619–25. Dominant and diet-responsive groups of bacteria within the human colonic doi:10.3945/jn.109.104638. microbiota. ISME J. 2011;5(2):220–30. doi:10.1038/ismej.2010.118. PubMed 9. Furusawa Y, Obata Y, Fukuda S, Endo TA, Nakato G, Takahashi D, et al. PMID: PMC3105703. Commensal microbe-derived butyrate induces the differentiation of colonic 33. MartÌnez I, Kim J, Duffy PR, Schlegel VL, Walter J. Resistant starches types 2 regulatory T cells. Nature. 2013;504(7480):446–50. doi:10.1038/nature12721. and 4 have differential effects on the composition of the fecal microbiota in 10. Ruemmele FM, Schwartz S, Seidman EG, Dionne S, Levy E, Lentze MJ. human subjects. PLoS One. 2010;5(11):e15046. doi:10.1371/journal.pone. Butyrate induced Caco-2 cell apoptosis is mediated via the mitochondrial pathway. Gut. 2003;52(1):94–100. 34. Friedman J, Alm EJ. Inferring correlation networks from genomic survey data. 11. Mikkelsen KH, Allin KH, Knop FK. Effect of antibiotics on gut microbiota, PLoS Comput Biol. 2012;8(9):e1002687. doi:10.1371/journal.pcbi.1002687. glucose metabolism and body weight regulation: a review of the literature. 35. Rivière A, Gagnon M, Weckx S, Roy D, De Vuyst L. Mutual cross-feeding Diabetes Obes Metab. 2016;18(5):444–53. doi:10.1111/dom.12637. interactions between Bifidobacterium longum subsp. longum NCC2705 12. Mathewson ND, Jenq R, Mathew AV, Koenigsknecht M, Hanash A, Toubai T, and Eubacterium rectale ATCC 33656 explain the bifidogenic and et al. Gut microbiome-derived metabolites modulate intestinal epithelial butyrogenic effects of arabinoxylan oligosaccharides. Appl Environ cell damage and mitigate graft-versus-host disease. Nat Immunol. Microbiol. 2015;81(22):7767–81. doi:10.1128/aem.02089-15. 2016;17(5):505–13. doi:10.1038/ni.3400. 36. Guggenheim KY. Rudolf Schoenheimer and the concept of the dynamic 13. Mathewson N, Mathew A, Oravecz-Wilson K, Wu J, Toubai T, Cummings E, state of body constituents. J Nutr. 1991;121(11):1701–4. et al. Microbial metabolites modulate GI mucosal damage from graft versus host disease (GHVD). (MUC4P.850). J Immunol. 2014;192(1 Supplement):133.26. 14. Smith MI, Yatsunenko T, Manary MJ, Trehan I, Mkakosya R, Cheng J, et al. Gut microbiomes of Malawian twin pairs discordant for kwashiorkor. Science. 2013;339(6119):548–54. doi:10.1126/science.1229000. 15. Ohland Christina L, Jobin C. Bugs and food: a recipe for cancer? Cell Metab. 2014;20(6):937–8. 16. Birt DF, Boylston T, Hendrich S, Jane J-L, Hollis J, Li L, et al. Resistant starch: promise for improving human health. Adv Nutr Int Review J. 2013;4(6):587–601. doi:10.3945/an.113.004325. 17. Kozich JJ, Westcott SL, Baxter NT, Highlander SK, Schloss PD. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol. 2013. doi:10.1128/aem.01043-13. 18. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: open-source, platform-independent, community- supported software for describing and comparing microbial communities. Submit your next manuscript to BioMed Central Appl Environ Microbiol. 2009;75(23):7537–41. 19. R Development Core Team. R: a language and environment for statistical and we will help you at every step: computing. In: Computing RFfS, editor. Vienna, Austria: R Foundation for • We accept pre-submission inquiries Statistical Computing; 2015. 20. Hammer O, Harper DAT, Ryan PD. PAST: paleontological statistics software � Our selector tool helps you to find the most relevant journal package for education and data analysis. In: Electronica P, editor.2001. � We provide round the clock customer support 21. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, et al. � Convenient online submission Metagenomic biomarker discovery and explanation. Genome Biol. 2011;12(6):R60. doi:10.1186/gb-2011-12-6-r60. � Thorough peer review 22. Thorndike RL. Who belongs in the family? Psychometrika.18(4):267-76. � Inclusion in PubMed and all major indexing services doi:10.1007/bf02289263. � Maximum visibility for your research 23. Ze X, Duncan SH, Louis P, Flint HJ. Ruminococcus bromii is a keystone species for the degradation of resistant starch in the human colon. ISME J. Submit your manuscript at 2012;6(8):1535–43. doi:10.1038/ismej.2012.4. www.biomedcentral.com/submit

Journal

Microbiome – Springer Journals

Published: Jun 29, 2016

There are no references for this article.