C. elegans responses to predatory threat can be organized into three defense modes.

(A) Predatory imminence model of C. elegans defensive responses to a predator-inhabited bacterial food patch. Upon being bitten by a predator, C. elegans executes an escape response, exits the predator-inhabited patch, and then ultimately reenters the patch (circa-strike mode). After extended exposure, C. elegans adopts an outstretched feeding posture to minimize predator contact (post-encounter mode). When confronted with a new patch, the predator-exposed C. elegans cautiously explores for potential predators in the patch (pre-encounter mode). (B) Arena setup for assessing circa-strike and post-encounter behaviors includes a 9.5 mm circular arena with a 2 mm bacterial patch, housing one C. elegans and four RS5194 P. pacificus predators (or none). (C) Number of exits and (D) total time that C. elegans spent off the patch during 1-hour exposure to predator and predator-free conditions (Wilcoxon rank sum test, nC.elegans = 13). (E) Percentage of C. elegans animals adopting outstretched feeding posture across different exposure durations to both predator and predator-free conditions (Fisher’s exact test, nC.elegans = 44-45). (F) Arena setup for studying pre-encounter behaviors involved placing one C. elegans and four RS5194 P. pacificus predators (or none) in either a wide exit arena (open space around the food patch) or a narrow exit arena (narrow corridor to/from the food patch) with a 2 mm bacterial patch. After 2- or 4-hour predator exposure, C. elegans was transferred to a predator-free arena for a 15-minute exploration period. (G) Latency to enter a new patch (Wilcoxon rank sum test with Benjamini-Hochberg adjustment, nC.elegans = 21-24) and (H) number of bins explored by C. elegans following either 2- or 4-hour exposure to predator or predator-free conditions (Welch’s t-test with Benjamini-Hochberg adjustment, nC.elegans = 13-15). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Images of arena setups.

(A) Wide exit arena. A 9.5 mm wide arena contains a 2 mm wide bacterial food patch as well as one adult C. elegans (arrow) and four adult Pristionchus sp. predators. (B) Representative example of C. elegans adopting outstretched feeding posture (arrow), with only its head in contact with bacteria. The typical sinusoidal waveform observed in (A) is distorted in the outstretched feeding posture. (C) Narrow exit arena. A dumbbell-shaped arena consisting of a pair of 2 mm-wide circular cutouts separated by a 3 mm x 0.7 mm corridor. One circular cutout is completely filled with a 2 mm wide bacterial food patch. (D). Example image of a 10×10 square grid overlaid on top of the patch for counting bins explored by C. elegans.

Progression of post-encounter behavior acquisition.

Percentage of C. elegans animals adopting outstretched feeding posture at different hours of exposure to RS5194 P. pacificus (Fisher’s exact test with Benjamini-Hochberg adjustment, nC.elegans = 45).

Pre-encounter behavior is reversible and not explained by injury-induced changes to locomotor speed.

(A-C) The pre-patch period is defined as starting from the initial placement of C. elegans in the empty circular cutout of a narrow exit arena and lasting until its mouth touches the new patch. (A) Distance traveled in the pre-patch period and (B) duration of the pre-patch period for C. elegans exposured either to RS5194 P. pacificus predators or no predators (Wilcoxon’s ranked sum test, nC.elegans = 38-41). (C) Association between pre-patch distance and pre-patch duration. Bold lines represent linear regression lines, with shaded regions representing 95% confidence intervals from linear regression models. Pre-patch speed was estimated as the slopes of regression lines. (Kruskal-Wallis test on residuals of linear regression models, nC.elegans = 38-41). (D) Bins explored immediately or 6 hours after 4-hour exposure to either RS5194 P. pacificus predators or no predators (Student’s t-test and paired t-test with Benjamini-Hochberg adjustment, nC.elegans = 8-9). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Post-encounter and pre-encounter behaviors are not explained by food deprivation.

(A) Percentage of C. elegans animals adopting outstretched feeding posture at different hours of exposure to RS5194 P. pacificus or food deprivation (Fisher’s exact test, nC.elegans = 26-31). (B) Latency to enter a new patch following 4-hour exposure to RS5194 predators, food deprivation, or neither (Dunn’s test with Benjamini-Hochberg adjustment, nC.elegans = 23-31). (C) Number of bins explored by C. elegans following 4-hour exposure to RS5194 predators, food deprivation, or neither (Dunns’s test with Benjamini-Hochberg adjustment, nC.elegans = 7-20). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Defensive response intensity increases with predation risk.

(A) Schematic of predatory harm potential of various Pristionchus spp. predator strains, based on studies previously conducted by this lab. (B) Number of bite-induced escape responses during 1-hour exposure to Pristionchus spp. predators (Dunn’s test with Benjamini-Hochberg adjustment, nC.elegans = 13-25). (C) Probability of exit following an escape response (binomial logistic regression followed by Wald test with single-step adjustment for Tukey contrasts, nC.elegans = 10-25). (D) Latency to reenter the patch following an exit, averaged across escape-induced exits, for various predator strains (Kruskal-Wallis test, nC.elegans= 10-25). (E) Percentage of C. elegans animals adopting outstretched feeding posture across different exposure durations to various predator conditions (Fisher’s exact test with Benjamini-Hochberg adjustment, nC.elegans = 36-45). Pairwise comparisons between predator strains are displayed on the right. (F) Latency to enter a new patch (Dunn’s test with Benjamini-Hochberg adjustment, nc.elegans = 22-40) and (G) bins explored (Dunn’s test with Benjamini-Hochberg adjustment, nC.elegans = 11-28) following 4-hour exposure to various predator conditions. Error bars in (C) are predicted P(exit | escape) and 95% CIs from binomial logistic regression model of data. All other error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Patch exit latency is unaffected by predator strain.

(A) Total number of exits, with or without preceding bite, in various predator and predator-free conditions (Dunn’s test with Benjamini-Hochberg adjustment, nC.elegans = 13-25). (B) Latency between bite-induced escape response and subsequenty exiting the patch, averaged across escape-induced exits, for various predator strains (Kruskal-Wallis test, nC.elegans = 10-25). (C) Probability of an aborted escape given that the escape did not result in an exit (binomial logistic regression followed by Wald test with single-step adjustment for Tukey contrasts, nC.elegans = 9-18). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Effect of extended exposure to various predations on pre-encounter behavior.

(A-B) Number of bins explored by C. elegans following (A) 6-hour (Tukey’s test, nC.elegans = 15-21)) and (B) 8-hour (Dunn’s test with Benjamini-Hochberg adjustment, C.elegans = 15-20) exposure to various predator conditions. Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

SEB-3 and NLP-49 peptides differentially regulate defense modes

(A-B) Probability of exit following an escape response by (A) seb-3 strains (nC.elegans = 17-34) and (B) nlp-49 strains (nC.elegans = 15-18) (binomial logistic regression followed by Wald test with single-step adjustment for Tukey contrasts). (C-D) Latency to reenter the patch following an exit, averaged across escape-induced exits, for (C) seb-3 strains (nC.elegans = 17-34) and (D) nlp-49 strains (nC.elegans = 14-18) (Dunn’s test with Benjamini-Hochberg adjustment). (E-F) Percentage of animals adopting outstretched feeding posture in (E) seb-3 strains (nC.elegans = 27-63) and (F) nlp-49 strains (nC.elegans = 31-33) (Fisher’s exact test with Benjamini-Hochberg adjustment). (G-H) Bins explored following 4-hour exposure to predator or predator-free conditions, by (G) seb-3 strains (Wilcoxon’s ranked sum test with Benjamini-Hochberg adjustment, nC.elegans = 16-18) and (H) nlp-49 strains (Welch’s t-test with Benjamini-Hochberg adjustment, nC.elegans = 9-12). Error bars in (A-B) are predicted P(exit | escape) and 95% CIs from binomial logistic regression model of data. All other error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Baseline speeds of seb-3 and nlp-49 strains.

(A) Image of arena setup for assaying baseline on-bacteria escape speed. A 2 mm wide arena is completely filled with a 2 mm wide bacterial food patch. One adult C. elegans (arrow) and four adult RS5194 P. pacificus were placed in the arena and recorded for two hours. (B-C) Histogram of pooled mean latencies between escape response and exit from the patch for (B) seb-3 strains and their controls (nC.elegans = 80) and (C) nlp-49 strains and their controls (nC.elegans = 49). Based on theses histograms, escape speed was evaluated within the 15-second period following a bite. (D-E) Bite escape speed across consecutive bites in (D) seb-3 strains (nC.elegans = 20-27) and (E) nlp-49 strains (nC.elegans = 11-13). Bold lines indicate mean speed over a 5-bite sliding window, with shaded areas denoting bootstrap 95% confidence intervals. Significance (*p<0.05) determined by non-overlapping confidence intervals between wildtype and other strains. (F-G) Baseline speed in bacteria-free, predator-free conditions for (F) seb-3 strains (Tukey’s test, nC.elegans = 10) and (G) nlp-49 strains (one-way ANOVA, nC.elegans = 9-10). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Escapes and exit latencies by seb-3 and nlp-49 strains.

(A-B) Number of bite-induced escape responses executed by (A) seb-3 strains (Kruskal-Wallis test, nC.elegans = 17-34) and (B) nlp-49 strains (one-way ANOVA, nC.elegans = 15-18) during 1-hour predator. (C-D) Latency to exit the patch following a bite-induced escape response, averaged across escape-induced exits, for (C) seb-3 strains (Kruskal-Wallis test, nC.elegans = 17-34) and (D) nlp-49 strains (Kruskal-Wallis test, nC.elegans = 14-18). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Outstretched feeding on high-density bacteria by nlp-49 strains.

Percentage of animals adopting outstretched feeding posture in nlp-49 deletion and overexpression strains (Fisher’s exact test with Benjamini-Hochberg adjustment, C.elegans = 35-54).

New patch entry latency by seb-3 and nlp-49 strains.

(A-B) Latency to enter a new patch for (A) seb-3 strains (nC.elegans = 16-35) and (B) nlp-49 strains (nC.elegans = 14-22) (Wilcoxon’s ranked sum test with Benjamini-Hochberg adjustment). (C-D) Comparison of latency to enter a new patch, within predator or predator-free conditions for (C) seb-3 strains (Kruskal-Wallis test, nC.elegans = 16-35) and (D) nlp-49 strains (Dunn’s test with Benjamini-Hochberg adjustment, nC.elegans = 14-22). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Interdependence between SEB-3 and NLP-49 peptides shifts across defense modes

(A) Probability of exit following an escape response by seb-3 deletion strains with nlp-49 deletion or overexpression (binomial logistic regression followed by Wald test with single-step adjustment for Tukey contrasts, nC.elegans = 11-24). (B) Latency to reenter the patch following an exit, averaged across escape-induced exits, for strains with seb-3 deletion by itself or with nlp-49 deletion or overexpression (Games-Howell test, nC.elegans = 11-21). (C) Percentage of animals adopting outstretched feeding posture in seb-3 deletion strains with nlp-49 deletion or overexpression (Fisher’s exact test with Benjamini-Hochberg adjustment, nC.elegans = 67-78). Significance asterisks represent comparisons with seb-3(-);nlp-49(-) (blue). (D) Percentage of animals adopting outstretched feeding posture in seb-3 deletion strains with nlp-49 deletion or overexpression (Fisher’s exact test with Benjamini-Hochberg adjustment, nC.elegans = 64-67). Significance asterisks represent comparisons with seb-3(-);nlp-49(o/e) (yellow). (E) Bins explored following 4-hour exposure to predator or predator-free conditions, by seb-3 deletion strains with nlp-49 deletion or overexpression (Student’s t-test with Benjamini-Hochberg adjustment, nC.elegans = 9-10). (F) Bins explored following 4-hour exposure to predator or predator-free conditions, by seb-3 deletions strains with or without nlp-49 overexpression (Student’s t-test with Benjamini-Hochberg adjustment, nC.elegans = 13-16). Error bars in (A) are predicted P(exit | escape) and 95% CIs from binomial logistic regression model of data. All other error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Circa-strike behavior of seb-3 deletion strains with nlp-49 deletion or overexpression.

(A) Number of bite-induced escape responses (one-way ANOVA, nC.elegans = 11-24). (B) Latency to exit the patch following a bite-induced escape response, averaged across escape-induced exits (Tukey’s test, nC.elegans = 11-24). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Baseline locomotion of seb-3 deletion strains with nlp-49 deletion or overexpression..

(A) Baseline speed in bacteria-free, predator-free conditions (one-way ANOVA, nC.elegans = 10). (B) Latency to enter a new patch for mock-exposed double deletion mutants and wildtype animals (Wilcoxon’s ranked sum test, nC.elegans = 9-10). Error bars are 95% bootstrap CIs containing the mean. n.s.=p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

C. elegans and Pristionchus spp. strains.