Caenorhabditis elegans employs innate and learned aversion in response to bacterial toxic metabolites tambjamine and violacein

Abstract

Bacteriovorus eukaryotes such as nematodes are one of the major natural predators of bacteria. In their defense bacteria have evolved a number of strategies to avoid predation, including the production of deterrent or toxic metabolites, however little is known regarding the response of predators towards such bacterial defenses. Here we use the nematode C. elegans as a model to study a predators' behavioral response towards two toxic bacterial metabolites, tambjamine YP1 and violacein. We found that C. elegans displays an innate avoidance behavior towards tambjamine YP1, however requires previous exposure to violacein before learning to avoid this metabolite. The learned avoidance of violacein is specific, reversible, is mediated via the nematode olfactory apparatus (aversive olfactory learning) and is reduced in the absence of the neurotransmitter serotonin. These multiple strategies to evade bacterial toxic metabolites represent a valuable behavioral adaptation allowing bacteriovorus predators to distinguish between good and bad food sources, thus contributing to the understanding of microbial predator-prey interactions.

Figure 1

(A)C. elegans assessed for aversion behavior using tambjamine-producing clone AA11 and crude extracts in ring assay. The ability to induce avoidance behavior was observed for AA11 clone and AA11 crude extract while AA11tamG⁻ and M13: Methanol (1:1) did not induce avoidance in the nematode. Glycerol was used as positive control. Each data point represents means ± the standard error of three replicates. * denotes a 0.001 < p < 0.05. Statistical significance was calculated comparing single treatments. (B) C. elegans assessed for aversive olfactory learning behavior in the food choice assay using tambjamine producing clone AA11. A choice index of −1.0 represents complete preference for the control bacterium, a choice index of 1.0 represents complete preference for the test bacterium, and an index of zero represents an equal distribution. The choice indexes of nematodes trained with AA11 clone (black bars) and naïve nematodes grown on AA11tamG⁻(white bars) were similar (p-value = 0.959) and no aversive learning was observed. Each data point represents means ± the standard error of three replicates.

Figure 2

(A) Aversion behavior using violacein producing clone 20G8 and pure violacein. Innate aversion was not observed in nematodes exposed to either 20G8 clone, 20G8vioA⁻, or pure violacein 10mM. Each data point represents means ± the standard error of three replicates. ***denotes a p < 0.001 compared to Glycerol used as a positive control. Statistical significance was calculated using a one-way ANOVA with a Tukeys HSD post hoc test.

Figure 3

(A) Aversive olfactory learning in the food choice assay using violacein producing clone 20G8 (test) plus 20G8vioA⁻ (control). A choice index of −1.0 represents preference for the control bacterium, a choice index of 1.0 represents preference for the test bacterium, and an index of zero represents equal distribution. Animals trained with 20G8 (black bars) in the food choice assay avoid 20G8 more than the animals exposed to 20G8vioA⁻ (white bar) do. (B) C. elegans trained using 20G8 (black bar) and 20G8vioA⁻ (white bar) assessed for aversive olfactory learning behavior in the food choice assay tested using 20G8vioC⁻ (test) plus 20G8vioA⁻ (control). Nematodes grown in presence of 20G8 do not learn to avoid 20G8vioC⁻. (C) Food preferences assessed at different time points in the food choice assay using 20G8 (test) plus 20G8vioA⁻ (control). In order to keep track of the movements of nematodes during time anaesthetic was not used. At all time points nematodes trained using 20G8 (black bars) preferred the control bacteria over the test bacteria (CI always negative). Naïve nematodes, grown on 20G8vioA⁻ (white bars) changed their preferences from the test bacteria at 1 hr (CI = 0.15) to the control bacteria at 6 hrs (CI = −0.4). (D) Aversive olfactory learning behavior in the memory assay. Nematodes were trained with 20G8 (black bars) or 20G8vioA⁻ (white bars), moved to control bacteria for different lengths of time and then the nematode preference for 20G8 or 20G8vioA⁻ was tested using 20G8 (test) plus 20G8vioA⁻ (control). N2 wild type and tph-1 mutants GR1321 (E) and MT15434 (F) assessed for aversive learning in the food choice assay tested using 20G8 (test bacterial spot) plus 20G8vioA⁻ (control bacterial spot). The two serotonin mutants, GR1321 and MT15434 displayed a lower learning index compared to N2 animals (p = 0.05 and p = 0.09 respectively). For panels A, B, C, and D each data point represents means ± the standard error of three replicates. In all panels each data point represents means ± the standard error of three replicates. * denotes a 0.001 < p < 0.05.

Figure 4

C. elegans wild type and mutant osm-6 assessed for avoidance behavior in the food choice assay trained using 20G8-20G8vioA⁻ and tested using 20G8-20G8vioA⁻. After 20 minutes most of the osm-6 nematodes remained in the centre of the test plate, therefore the results of this food choice were not represented as choice index. Instead the plate was divided in three sections (A) and the results were plotted as percentage of the nematodes’ distribution in the three sections after 30 minutes (B,D), 3 hours (C,E). No wild-type animals remained in the center of the plate at all times tested. Three replicates of each experiment were performed and statistical significance was calculated comparing single sections ***denotes a p < 0.0001.