The evolution of cooperation within the gut microbiota - PubMed (original) (raw)

. 2016 May 12;533(7602):255-9.

doi: 10.1038/nature17626. Epub 2016 Apr 25.

Affiliations

The evolution of cooperation within the gut microbiota

Seth Rakoff-Nahoum et al. Nature. 2016.

Abstract

Cooperative phenotypes are considered central to the functioning of microbial communities in many contexts, including communication via quorum sensing, biofilm formation, antibiotic resistance, and pathogenesis. The human intestine houses a dense and diverse microbial community critical to health, yet we know little about cooperation within this important ecosystem. Here we test experimentally for evolved cooperation within the Bacteroidales, the dominant Gram-negative bacteria of the human intestine. We show that during growth on certain dietary polysaccharides, the model member Bacteroides thetaiotaomicron exhibits only limited cooperation. Although this organism digests these polysaccharides extracellularly, mutants lacking this ability are outcompeted. In contrast, we discovered a dedicated cross-feeding enzyme system in the prominent gut symbiont Bacteroides ovatus, which digests polysaccharide at a cost to itself but at a benefit to another species. Using in vitro systems and gnotobiotic mouse colonization models, we find that extracellular digestion of inulin increases the fitness of B. ovatus owing to reciprocal benefits when it feeds other gut species such as Bacteroides vulgatus. This is a rare example of naturally-evolved cooperation between microbial species. Our study reveals both the complexity and importance of cooperative phenotypes within the mammalian intestinal microbiota.

PubMed Disclaimer

Figures

Extended Data Figure 1

Extended Data Figure 1. Limited cooperation during polysaccharide utilization by B. thetaiotaomicron

a-d, Independent experiments for Fig. 1b-e. (a,b) Upper panels n=2 biological replicates, lower panels n=1; (c,d) n=2 biological replicates. e, TLC analysis of conditioned media from Bt grown in amylopectin (left panel) or levan (right panel) minimal media. EL = early log, ML = late log, LL = late log, Stat = stationary phase. Glu = Glucose, Fru = fructose, Suc = sucrose. See SI Figure 1 for uncropped scanned images. In all panels, error bars represent standard error_; p_ values derived from two-tailed Student's t test.

Extended Data Figure 2

Extended Data Figure 2. Bo 04502 and 04503 mutants grow equivalently to wild type on limiting concentrations of inulin and do not require surface digestion for utilization of inulin

a, Growth of Bo WT, Δ04502, Δ04503 and Δ04502/3 mutants in varying concentrations of inulin as indicated. Biological replicates of each condition are plotted as individual lines (n=2 cell culture biological replicates). Upper and lower panels are independent experiments. b, Independent experiments for Fig. 2d. EL = early log, ML = mid log, LL = late log, St = stationary phase. See SI Figure 1 for uncropped scanned images. c, Complementation of Bo Δ04502 and Δ04502/3 mutants with the respective genes in trans. TLC analysis of conditioned media from Bo Δ04502/3 (left panel) complemented in trans with BACOVA_04502, BACOVA_04503, BACOVA_04502/3 or vector alone (pFD340) and Bo Δ04502 (right panel) with BACOVA_04502 or vector alone grown in defined inulin media. See SI Figure 1 for uncropped scanned images. S = stationary phase.

Extended Data Figure 3

Extended Data Figure 3. SusC and SusD homologs BACOVA_04505 and BACOVA_04505 are required for inulin utilization

a, Independent experiments for Fig. 2e. Left panel n=2 biological replicates, right panel each line represents n=1 sample per condition. b, Complementation of Bo Δ04504, Bo Δ04505 and Δ04504/5 mutants with the genes in trans. Growth of Bo Δ04504, Bo Δ04505, Bo Δ04504/5 with BACOVA_04504, BACOVA_04505, BACOVA_04504/5 or vector alone (pFD340) in trans in defined inulin media. Each line represents n=1 sample per condition. In all panels, error bars represent standard error_; p_ values derived from two-tailed Student's t test.

Extended Data Figure 4

Extended Data Figure 4. Costs of extracellular inulin digestion by B. ovatus

a,b, Independent experiments for Fig. 3a,c. n=3 biological replicates at day 1,2; n=2, biological replicates at day 4 (a). In upper and upped middle panels biological replicates of each condition are plotted as individual lines (n=2 cell culture biological replicates); in lower middle and lower panels each line represents n=1 sample per condition (b). c, TLC analysis of conditioned media from Bo WT and Bo Δ04502/3 cultured in 0.5% inulin with trace (0.06%) amounts of fructose. St = stationary phase. See SI Figure 1 for uncropped scanned images. In all panels, error bars represent standard error_; p_ values derived from two-tailed Student's t test

Extended Data Figure 5

Extended Data Figure 5. Preferential utilization of undigested inulin by B. ovatus and costs of inulin digestion by 04502/3

a, Time to mid-log (estimated at 50% maximal OD (OD50) by linear regression analysis) of Fig. 3d. b, Additional independent experiments for Fig. 3d. Each line represents n=1 sample per condition. c, Competition of Bo WT and Bo Δ04502/3 co-cultured in 0.5% inulin with trace (0.06%) amounts of fructose. n=2 cell culture biological replicates; d,e, Growth (d) and TLC analysis of conditioned media (e) of Bo WT and Bo_Δ04502/3 cultured in 0.1% inulin with trace (0.06%) amounts of fructose. See SI Figure 1 for uncropped scanned images. For d, each panel is an independent experiment. Each condition is plotted as individual lines in each panel (n=2 cell culture biological replicates). In all panels, error bars represent standard error; p_ values derived from two-tailed Student's t test.

Extended Data Figure 6

Extended Data Figure 6. A complex mouse microbiota differentially affects B. ovatus WT and Δ04502/3 pre-colonized gnotobiotic mice and analysis of cross feeding of the predominant murine Bacteroidales of the murine gut microbiota

a, Germ-free mice were monocolonized with Bo WT or Δ04502/3 and maintained on a diet supplemented with inulin as the sole polysaccharide and housed under gnotobiotic conditions for 2 weeks. Bacteria were enumerated from feces prior to gavage with intestinal microbiota of conventionally raised mice (n=5 mice, cell culture biological replicates). b, Growth of two dominant mouse microbiota Bacteroidales strains (Parabacteroides goldsteinii and a strain with 96% 16S rRNA gene identity to B. thetaiotaomicron) with inulin breakdown products derived from the conditioned media of B. ovatus grown in inulin (all inulin had been digested) or undigested inulin minimal media. Each data point is a different isolate of the indicated species from the ceca of the conventionally raised mice used for gavage (n=11 isolates for upper panel, n=6 for lower panel. In all panels, error bars represent standard error.

Extended Data Figure 7

Extended Data Figure 7. B. ovatus, but not B. fragilis benefits from B. vulgatus in co-culture in inulin

a, b, Independent experimentsfor Fig. 4b,d. The left panel corresponds to starting culture with ~107 CFU Bo corresponding to starting culture of ~106 CFU Bo in Fig. 4d. The two right panels are a duplicate pair of experiments of starting CFU of 106 and 107. In each panel n=2 biological replicates. c, Enumeration of B. fragilis (c) in monoculture or co-culture with Bv on defined inulin plates, n=3 cell culture biological replicates.. Letters in parentheses refer to values correlating to color of line used for statistical analysis. In (a), for example, p (g,r) refers to comparison of values of green (g) and red (r) values at the time-point indicated. In (b) color of p value indicates comparison of monoculture and co-culture for the given condition at the time-point indicated. For all panels, error bars represent standard error_; p_ values derived from two-tailed Student's t test.

Extended Data Figure 8

Extended Data Figure 8. Secreted factors from Bo and isogenic mutants, but not B. fragilis (Bf), support Bv survival

a, Growth of Bv in conditioned media from Bo WT, Δ04502, Δ04502/3 or Bf grown in defined media with inulin as the sole carbohydrate and in inulin media. End time-point corresponded to peak growth of Bv in conditioned media derived from Bo WT. B. fragilis, which utilizes inulin, but similar to Bo Δ04502/3 does not liberate inulin breakdown products, does not support the survival of Bv during co-culture. Left and right panels are independent experiments. Left panel; start, n=4 biological replicates; end n=2 biological replicates. Right panel; t0, n=4 cell culture biological replicates; t13 and 24, n=2 biological replicates. b, Growth of Bv in dialyzed (2 kD MW membrane) or undialyzed conditioned media from Bo Δ04502/3 grown in inulin, n=2 cell culture biological replicates. c, Gas chromatographic analysis of acetate, propionate and succinate in conditioned media during growth of Bo WT and Δ04502/3 in defined media with inulin as the sole carbohydrate. Other volatile and non-volatile substances (as listed in Methods) were undetectable, n=2 cell culture biological replicates, except Δ04502/3 stationary phase, n=1. d, Growth of Bv in defined medium with inulin as the sole carbohydrate with or without addition with 15 mM of acetate, propionate or succinate, t0, n=4 biological replicates; t13 and 24, n=2 biological replicates. CM = conditioned media, UD = undetected. For all panels, error bars represent standard error_; p_ values derived from two-tailed Student's t test

Extended Data Figure 9

Extended Data Figure 9. Spatial aspects of mutualism between Bo and Bv via cross feeding and interspecies cooperation in vivo via 04502/3

a, Independent experiments for Fig. 4c, n=4 biological replicates. b, Independent experiments for Fig. 4e, n=3 cell culture biological replicates. c, Scatter plot of experiment in Fig. 4f, n=3 cell culture biological replicates. d, Ratios of WT and Δ04502/3 in feces 21 days after co-colonization of three germ-free mice (n=3 mice biological replicates) on a diet of inulin as the sole dietary polysaccharide (pre-Bv) and then four days after introduction of Bv (post-Bv). Each panel shows the ratio pre- and post- Bv of an individual mouse. p values are Fisher exact test comparing the frequency of Bo WT and Δ04502/3 pre- and post- colonization with Bv for each individual mouse. At day 21, all mice were colonized with a higher ratio of the mutant (ranging to 86% in the mouse depicted in the lowest panel), with each mouse showing a statistically significant increase in the proportion of the WT after introduction of Bv. e, Growth of B. ovatus in 0.5% inulin (upper panel) or fructose (lower panel) in minimal media to which filter sterilized conditioned media from early log, OD600 matched growth of Bv or Bo in 0.125% fructose minimal media or fresh 0.125% fructose minimal media control was added at 1:1 ratio. In (e) numbers refer to p values (* <0.05, **<0.01, ***<0.001) of comparison of values of green (g), red (r) or black (b) values by unpaired, two-tailed student t test at the time-point indicated, n=2 cell culture biological replicates. For all panels, error bars represent standard error, for all panels except (d) p values derived from two-tailed Student's t test.

Extended Data Figure 10

Extended Data Figure 10. Schematic of forms of cooperativity via polysaccharide digestion among Bacteroidales

a, Limited cooperation. Privatization of extracellularly digested public goods by the individual performing the digestion leads to greater individual (dark arrow) than shared benefits (light arrow) as seen in Bt during growth on levan and amylopectin. b, Cooperation between species is seen between Bo and Bv during growth on inulin. Surface digestion of inulin by Bo creates breakdown products that it does not need to grow on inulin. Rather, inulin breakdown represents a dedicated cross-feeding system that provides benefits to Bv, with reciprocal fitness benefits to Bo.

Figure 1

Figure 1. Direct and cooperative benefits of polysaccharide digestion by surface glycoside hydrolases (GH)

a, Polysaccharide utilization loci of B. thetaiotaomicron (Bt) for amylopectin and levan with the products or properties each gene encodes listed above and color coded. SPI or SPII = signal peptidase I or II cleavage site. b, c, Growth of Bt WT and surface GH mutants in media with amylopectin (n=2, cell culture biological replicates) (b) or levan (n=2, cell culture biological replicates) (c). See Extended Data Fig.1 for additional independent experiments. d,e, Growth of Bt WT and surface GH mutants in mono- and co -culture in media with amylopectin (n=2, cell culture biological replicates) (d) or levan (n=2, cell culture biological replicates) (e). See Extended Data Fig.1 for additional independent experiments. In all panels error bars represent standard error_; p_ values derived from two-tailed Student's t test.

Figure 2

Figure 2. B. ovatus does not require surface digestion for utilization of inulin

a, Predicted inulin utilization locus of B. ovatus (Bo). Gene designations shown are preceded by BACOVA_. The color coding of gene products is as in Fig.1. b, Thin layer chromatography (TLC) analysis of inulin defined medium incubated for the indicated times with purified His-tagged 04502, 04503 or 04502 and 04503, or vector control. See SI Figure 1 for uncropped scanned images. c, Growth of Bo WT and mutants in inulin defined medium. d, TLC analysis of conditioned media during the growth of Bo WT and mutants in 0.5% and 0.1% inulin. EL = early log, ML = mid log, LL = late log, St = stationary. e, Growth of Bo WT, Δ04502/3, Δ04504 (susD ortholog) and Δ04505 (susC ortholog) mutants in inulin defined medium. For c,e, each line represents n=1 sample per condition. See Extended Data Fig.2 for additional independent experiments. In all panels error bars represent standard error_; p_ values derived from two-tailed Student's t test.

Figure 3

Figure 3. Cost of inulin digestion by surface glycoside hydrolases

a, Growth yield of Bo WT and mutants on inulin agarose plates. Each line represents n=1 sample per condition. See Extended Data Fig. 4 for additional independent experiments. b, Bacteria per gram feces of gnotobiotic mice seven days after monocolonization with Bo WT or Δ04502/3 on a polysaccharide free diet or supplemented with inulin (n=5 biological replicate mice per condition). c, Growth curves (upper and middle panels) and maximal OD600 (bottom panel) of Bo WT or Δ04502/3 in minimal media with carbon sources as indicated. Each line represents n=1 sample per condition. See Extended Data Fig. 4 for additional independent experiments. In the bottom panel, p values are displayed as paired (matched WT and Δ04502/3 within same experiment) followed by unpaired (all experiments) two-tailed Student's t test (n=7, biological replicates). d, Growth of Bo WT or mutants in inulin or a stoichiometric equivalent amount of inulin breakdown products after digestion with purified 04502 and 04503 enzymes (n=2, cell culture biological replicates); *, p < 0.05; ** p < 0.01; ***, p < 0.001; See Extended Data Fig. 5 for additional independent experiments. In all panels error bars represent standard error_; p_ values derived from two-tailed Student's t test.

Figure 4

Figure 4. Interspecies cooperation mediated by surface digestion of inulin is stabilized by reciprocal benefits

a, Germ-free mice were monocolonized with Bo WT or Δ04502/3 and maintained on a polysaccharide free diet supplemented with inulin and housed for 2 weeks. Mice were then gavaged with the cecal contents of conventionally raised mice. At day 5 post gavage, cecal contents were plated for enumeration of Bo (n=5 biological replicate mice per group). b, Enumeration of Bv in monoculture or co-culture with Bo WT or Δ04502 on defined inulin plates. α, p = 0.001 for number of Bv in Bv/Bo WT co-culture vs. Bv alone; ψ, p = 0.001 for number of Bv in Bv/Bo WT co-culture vs. Bv/Δ04502 co-culture; α, p = 0.003 for number of Bv in Bv/Δ04502 co-culture vs. Bv alone (n=2, cell culture biological replicates). See Extended Data Figs. 7 and 9 for additional independent experiments. c, (left) Photos of patches of Bv plated at varying distances around Bo WT or Bo Δ04502/3 or alone on inulin agarose plates. (right) Enumeration of Bv after five days of culture plated at the same distance to Bo WT or Bo Δ04502/3 or alone on inulin plates. See Extended Data Fig. 9 for additional independent experiments. d, Enumeration of Bo WT or Δ04502 in monoculture or co-culture with Bv on defined inulin plates. p values correlate to color of line/genotype used for statistical analysis. p value indicates comparison of monoculture and co-culture for the given condition at the time-point indicated. The benefit Bo receives from Bv is most robust when starting with fewer Bo. Depicted are starting CFU of ~106 Bo. (n=3, cell culture biological replicates at day 1,2; n=2, biological replicates at day 4). See Extended Data Fig. 7 for additional independent experiments and Extended Data Fig.7b for starting CFU of ~107 Bo. e, Ratios of WT and Δ04502/3 at the start and day 2 of culture when co-plated with or without Bv on inulin plates (n=3, cell culture biological replicates). See Extended Data Fig. 9 for additional independent experiments. f, Ratios of WT and Δ04502/3 in the inoculum (day 0) and in feces at various time points (days 4, 7, 11, 14, 18) post co-colonization of gnotobiotic mice (n=3 cell culture biological replicates) with Bo WT and Δ04502/3. Polysaccharide free diet was supplemented with inulin at day 7. Bv was introduced at day 14. A Fisher exact test comparing the frequency of Bo WT and Δ04502/3 pre- (day 14) and post- (day 18) colonization with Bv was significant with a p value of 0.0001 for each individual mouse. Bo CFU in feces were maximal after the switch to inulin diet and addition of Bv changed the abundance of Bo WT compared to Δ04502/3 but not total CFU of Bo. See Extended Data Fig. 9 for additional independent experiments. In all panels error bars represent standard error_; p_ values displayed are derived from two-tailed Student's t test.

References

    1. Sonnenburg ED, et al. Specificity of Polysaccharide Use in Intestinal Bacteroides Species Determines Diet-Induced Microbiota Alterations. Cell. 2010;141:1241–1252. - PMC - PubMed
    1. West SA, Griffin AS, Gardner A, Diggle SP. Social evolution theory for microorganisms. Nat Rev Microbiol. 2006;4:597–607. - PubMed
    1. Rakoff-Nahoum S, Coyne MJ, Comstock LE. An ecological network of polysaccharide utilization among human intestinal symbionts. Curr Biol. 2014;24:40–49. - PMC - PubMed
    1. Drescher K, Nadell CD, Stone HA, Wingreen NS, Bassler BL. Solutions to the public goods dilemma in bacterial biofilms. Curr Biol. 2014;24:50–55. - PMC - PubMed
    1. Frank SA. A general model of the public goods dilemma. Journal of Evolutionary Biology. 2010;23:1245–1250. - PMC - PubMed

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources