P0, F1 and F2 generation responses to P0 PA14 exposure.

Three representative experimental results for trained and inherited aversion behavior (A-C) and induced and inherited elevated daf-7p::gfp in ASI neurons (D-I). A-C, Quartile box plots for each generation and training condition showing the distribution of choice index values ([number of animals choosing OP50 - number of animals choosing PA14] / total number of choices) for each OP50 vs PA14 choice assay plate following the published protocols. The experimental and assay conditions are indicated in the panel titles; [Exp #] [growth temperature (20°C or 25°C)] [PA14 isolate (B, Balskus or M, Murphy)] [light/dark assay condition (L or D)] [azide/cold paralytic (A or C)]. D, F, H, Quartile box plots displaying average ASI neuron daf-7p::gfp expression levels per worm and E, G, I, show the 95% confidence intervals for the difference in absolute mean between the conditions. For these experiments the FK181 strain (integrated multi-copy [MC] daf-7p::gfp reporter) was cultured at 20°C at all generations and was exposed to one of three different PA14 isolates (H, Hunter; B, Balskus; M, Murphy labs). AFU arbitrary fluorescence units normalized to mean OP50 levels. Red dots in choice assay results indicate outlier data points that were included in all statistical tests. Statical significance **** P < 0.0001, *** < 0.001, ** < 0.01, * < 0.05, ns > 0.05. See Methods section for statistical methods.

Experimental conditions and results for daf-7p::gfp expression levels in ASI neurons of wild-type animals.

Experimental conditions and results for PA14 avoidance (Choice) assay.

Effect of experimental design changes on multigenerational ASI daf-7p::gfp expression levels after P0 PA14 exposure.

Box plots of results for 21 independent experiments are normalized to the average OP50 value by generation within each experiment for summary presentation (see Table 1 and Table S1 for experimental conditions). This figure includes the experiments shown in Figure 1. FK181 contains an integrated multi-copy (MC) tandem array composed of the daf-7p::gfp reporter and the co-injection marker rol-6(su1006) (Murakami et al., 2001). QL296 is a single copy (SC) insert of daf-7p::gfp with unc-119(+) as the co-selection marker (Zhan et al., 2015). Worms were cultured at either 20°C or 25°C and exposed to one of three different PA14 isolates (B, Balskus; H, Hunter; M, Murphy labs). In some experiments, worms were grown at 15°C for at least three generations prior to the P0 generation (indicated with parentheses, i.e. (15)20). The 95% confidence interval for the predicted absolute difference in means between conditions, normalized to the predicted P0 difference for each experiment (when applicable), is presented in the lower portion of the figure. Figure S2 shows the same data for individual neurons, rather than the mean of the two neurons. Statistical significance **** P < 0.0001, *** < 0.001, ** < 0.01, * < 0.05. Non-significant labels (p > 0.05) are omitted for clarity. † Indicate statistical significance with control higher than experiment. See Methods section for statistical methods.

Effects of experimental design changes on multigenerational PA14 avoidance after PA14 exposure.

Growth and assay conditions for each experiment are described in Table 2. This figure includes the experiments shown in Figure 1. Choice index calculated as described in Figure 1. A-I, The experimental and assay conditions are indicated in the panel titles; [Exp #] [growth temperature (20°C or 25°C)] [PA14 isolate (B, Balskus or M, Murphy)] [light/dark assay condition (L or D)] [azide/cold paralytic (A or C)]. J, Summary panel showing learning indexes (choice index for trained minus choice index for control) for all nine results (eight for F1 animals). See Figure S3 for Fisher’s exact test analysis of choice assay results. Red dots in choice assay results indicate outlier data points that were included in all statistical tests. Statical significance **** P < 0.0001, *** < 0.001,** < 0.01, * < 0.05, ns > 0.05. See Methods section for statistical methods.

Effect of OP50 growth conditions on OP50 aversion.

N2 worms grown to adulthood on HG plates were “trained” on either HG OP50 or NG OP50 plates for 24 hours and then assayed on OP50 vs PA14 choice plates. The choice index was calculated as described in Figure 1. The learning index, difference between mean HG choice index and mean NG choice index, for all eight experiments is plotted in the last panel. See Figure S5 for Fisher’s exact test analysis of choice assay results. Statical significance **** P < 0.0001, *** < 0.001, ** < 0.01, * < 0.05, ns > 0.05. See Methods section for statistical methods. Data for this figure is presented in Table S6.

Effects of RNAi pathway mutants on intergenerational (F1) inheritance of avoidance behavior and ASI daf-7p::gfp expression levels.

A-D, sid-1(qt9) and sid-2(qt42) choice assays. Growth and assay conditions for each experiment are described in Table 2. The choice index was calculated as described in Figure 1. See Figure S6 for Fisher’s exact test analysis of sid-1 and sid-2 choice assay results. N2 choice assay results performed in parallel with the sid-1 and sid-2 experiments are presented in Figure S7. E-H, ASI expression experiments showing average ASI neuron daf-7p::gfp expression levels per worm for each of the four genotypes are shown. ASI daf-7p::gfp expression levels are normalized to the average OP50 value by generation within each experiment for ease of presentation. The experiments shown in panels E-H were performed with animals cultured and trained at 20°C. Statical significance **** P < 0.0001, *** < 0.001, ** < 0.01, * < 0.05, ns > 0.05. See Methods section for statistical methods.

Re-analysis of small RNA and mRNA sequence data from PA14 exposed and control P0 animals.

RNA sequence data (PRJNA509938) was downloaded and re-analyzed as described in Methods. A, B, Scatter plots comparing sense and antisense strand small RNAs (17-29 nucleotides) from worms grown on PA14 and OP50. Red dots represent ≥ 2-fold change and Benjamini-Hochberg corrected P values (P adj. ≤ 0.05) and grey dots correspond to less than a 2-fold change or insignificant difference (P adj. > 0.05). Volcano plots of the same data are shown in Figure S8. C, Black dots represent genes associated with small RNAs that show a significant 2-fold or greater change in abundance (Y axis) and map to an mRNA that also shows a significant 2-fold or greater change in abundance (X axis). Blue dots correspond to the subset of target mRNAs associated by GO analysis (panel D) with anti-bacterial or innate immune responses. The red dot represents the maco-1 gene. D, Gene Ontology analysis of differentially expressed mRNAs targeted by differentially expressed small RNAs. Eight of the 116 small RNA targeted differentially expressed mRNAs are also associated with antibacterial or innate immune responses, while 257 of all 14,194 detectably expressed genes share similar GO annotations representing a 3.8-fold (P < 0.0001, hypergeometric distribution) enrichment over neutral expectations.