Gut microbiome remodeling provides protection from an environmental toxin

Abstract

Gut microbiomes contribute to animal health and fitness. The immense biochemical diversity of bacteria holds particular potential for neutralizing environmental toxins and thus helping hosts deal with new toxic challenges. To explore this potential, we used Caenorhabditis elegans harboring a defined microbiome, and the antibiotic neomycin as a model toxin, differentially affecting microbiome strains, and also toxic to worms. Worms exposed to neomycin showed delayed development and reduced survival but were protected when colonized with neomycin-resistant Stenotrophomonas. 16S rRNA sequencing, bacterial load quantification, genetic manipulation, and behavioral assays showed that protection was linked to enrichment of Stenotrophomonas carrying a neomycin-modifying enzyme. Enrichment was facilitated by altered bacterial competition in the gut, as well as by KGB-1/JNK-dependent behavioral changes. While microbiome remodeling conferred toxin resistance, it was associated with reduced infection resistance and metabolic changes. These findings suggest that microbiome adaptation can help animals cope with stressors but may have long-term consequences that add to effects of direct intoxication.

Keywords: Microbiology; Microbiome.

Graphical abstract

Figure 1

The gut microbiome protects host from neomycin toxicity (A) Sensitivity to neomycin, quantified in CeMbio community members by disk diffusion growth inhibition as demonstrated in inset (red intensity signifies extent of inhibition), or by growth in liquid media (green intensity signifies growth rate). Resistant strains are boxed. (B) Effects of neomycin (or kanamycin) on development of worms raised on E. coli or on CeMbio. Proportion of worms reaching L4 following 30 h of growth from L1 at 20°C. Shown are averages ±SD from two independent experiments, each with two plate replicates (total n = 4); 17–91 worms/plate; ∗, p < 7.64E-4, t test. (C) Survival of wild type worms raised on designated bacterial cultures and shifted as first day gravids (t = 0) to plates with neomycin and E. coli as food. Shown are averages ±SD from three plates (60–73 worms/group); ∗, p = 1E-10, log-rank test.

Figure 2

Neomycin exposure enriches for Neo-resistant gut Stenotrophomonas (A) Survival assays of wild type worms raised on monocultures of Neo-resistant CeMbio strains and shifted as gravids to plates with neomycin and E. coli as food. Each curve shows averages ±SD from three plates (31–110 worms/group); ∗, p < 1E-4, log-rank test. Inset. Persistence of designated strains in worms 48 h after shift to neomycin plates (10 worms/group, averages ±SD for three replicates); ∗, p < 0.011, t test compared to E. coli. (B) A representative experiment showing microbiome composition (analyzed by 16S sequencing, color-coded by strains) in wild type worms raised on CeMbio with or without neomycin (W) or in their lawn environment (Env). Each bar represents a swab of the bacterial lawn or a population of ∼70 worms. (C) Gut enrichment of Ochrobactrum and Stenotrophomonas in worms raised with or without neomycin (left, 3 independent experiments, each with 2–3 separate populations as in B (total n = 7), each population marked with a dot), or worms raised on CeMbio and shifted as gravids for 12 h to neomycin with E. coli food (right, two independent experiments, 2-3 separate populations each, total n = 5). Shown are averages ±SD, ∗, p < 0.05, t test. (D) Bacterial load in gravid worms raised on CeMbio with or without neomycin. CFU counts of gut bacteria cultured on LB with neomycin (left) or without (right); averages ±SD for two independent experiments, each with two replicates (n = 4, 15 worms/experiment); ∗, p = 0.0018 (left) and p = 0.047 (right), t test. (E and F) Survival assays of wild type worms raised on designated bacteria (live or dead) and transferred as gravid to plates with neomycin and E. coli food. St.i., S. indicatrix (JUb19, APH(3′)+), St.r., S. rhizophila (BIGb262, APH(3′)-), St.r. APH(3′)+, APH(3')-transformed BIGb262. Each curve shows averages ±SD for three plates; 27–45 worms/plate (E), N = 44–89 worms/plate (F); ∗, p < 7E-7, log-rank test compared to live E. coli.

Figure 3

kgb-1-dependent toxin avoidance enables preference and acquisition of Neo-protective Stenotrophomonas (A) Avoidance of neomycin by wild type or kgb-1 gravid worms within an hour of application. Shown is one experiment, with averages ±SD for distances of individual worms from point of neomycin application (red dot on bacterial lawn in scheme); 24–92 worms/group; ∗, p < 2E-5, t test. (B) Preference assays (direction marked in schemes by arrows) with wild type or kgb-1 worms between CeMbio and E. coli, with or without neomycin. Shown are median (lines) and interquartile ranges (boxes) for five assay plates from two independent experiments (n = 5), 46–134 worms/plate; ∗, p = 8.5E-5, t test. (C) Preference assay of wild type worms in a multi-choice assay on spatially segregated monocultures of CeMbio members, with or without neomycin. Shown are averages ±SD from two independent experiments each with two plate replicates (n = 4), 46–185 worms/plate; ∗, p = 0.01, t test. (D and E) Preference assays with wild type or kgb-1 worms between Stenotrophomonas strains or E. coli, live or dead, and with or without neomycin. D. Shown are medians and interquartile ranges for three independent experiments, each with 2–3 assay plates (n = 7, 40–157 worms/plate); ∗, p = 4.5E-7, t test. St.i., JUb19; St. r., BIGb262; St.r. APH(3′)+, APH(3′)-transformed BIGb262. E. Each graph shows medians and interquartile ranges for three assay plates, 37–135 worms/plate; ∗, p < 0.009, t test. (F) Gut microbiome composition and the derived Stenotrophomonas abundance in worms raised on a mixed CeMbio culture, or on spatially segregated monocultures, in plates with or without neomycin. Bar graphs as in Figure 2B, 80 worms/population; Stenotrophomonas abundance shows averages ±SD of two populations, ∗, p < 0.05, t test.

Figure 4

Consequences of neomycin-induced microbiome remodeling (A) CFU counts of gut-colonizing Neo-resistant CeMbio members from worms raised from L1 to mid gravid stage (4.5 days) on CeMbio with or without neomycin, or first with neomycin (two days) and then shifted for 2.5 days more to plates without neomycin (Post-Neo). Averages ±SD of two technical repeats per group (13 worms/group); ∗∗∗∗, p < 0.0001, two-way ANOVA followed by a Bonferroni’s test. (B) Left, representative Images of lipid staining (proportional to lipid content) in wild type gravid worms raised on CeMbio, on CeMbio with neomycin, on reconfigured CeMbio (Re.) or on Stenotrophomonas alone; scale bar, 200 μm. Right, quantification of staining in individual worms from three independent experiments. Averages ±SD for 16–21 worms/group; p = 2.4E-8, ANOVA; ∗, p < 3.7E-05, Tukey’s HSD post hoc. (C) Infection resistance in worms raised on designated communities and shifted in adulthood to Pseudomonas aeruginosa. Each curve shows averages ±SD for three plates (67–83 worms/group); p = 1E-7 (CeMbio vs. E. coli), p = 0.003 (CeMbio vs. Re.), log-rank test. Inset. Persistence of CeMbio members in worms 24 h after shift to P. aeruginosa. Averages ±SD for three technical replicates, 10 worms/group; p = 8E-7, ANOVA; ∗, p < 3.4E-05 Tukey’s HSD post hoc. (D) Infection resistance as in C with worms raised on CeMbio including Pseudomonas-protective strains, without them, or without three non-protective strains (low diversity). Curves show averages ±SD for three plates (62–87 worms/group), ∗, p < 3.4–05, log-rank test.