Larval defense against attack from parasitoid wasps requires nociceptive neurons

Abstract

Parasitoid wasps are a fierce predator of Drosophila larvae. Female Leptopilina boulardi (LB) wasps use a sharp ovipositor to inject eggs into the bodies of Drosophila melanogaster larvae. The wasp then eats the Drosophila larva alive from the inside, and an adult wasp ecloses from the Drosophila pupal case instead of a fly. However, the Drosophila larvae are not defenseless as they may resist the attack of the wasps through somatosensory-triggered behavioral responses. Here we describe the full range of behaviors performed by the larval prey in immediate response to attacks by the wasps. Our results suggest that Drosophila larvae primarily sense the wasps using their mechanosensory systems. The range of behavioral responses included both "gentle touch" like responses as well as nociceptive responses. We found that the precise larval response depended on both the somatotopic location of the attack, and whether or not the larval cuticle was successfully penetrated during the course of the attack. Interestingly, nociceptive responses are more likely to be triggered by attacks in which the cuticle had been successfully penetrated by the wasp. Finally, we found that the class IV neurons, which are necessary for mechanical nociception, were also necessary for a nociceptive response to wasp attacks. Thus, the class IV neurons allow for a nociceptive behavioral response to a naturally occurring predator of Drosophila.

Figure 1. Behavioral responses of Drosophila melanogaster larvae to attack by LB.

(A) Classification of behavioral responses to attacks by parasitoid wasps. The cartoons depict peristaltic locomotion, turning, writhing and nocifensive escape locomotion ((NEL) also see supplemental movies). (B) Ethogram of behaviors shown by third instar larvae (based on observations of 124 attacks). The size of the arrow is weighted according to the observed frequency of the behavior. Primary behaviors are indicated by large cartoons, and secondary behaviors are indicated by the smaller cartoons. Tertiary behaviors are not shown. (C) Attack position along the larval body wall influences behavioral response. Fisher’s Exact Test with Holm-Bonferroni correction. Data are presented as percentages ± 95% confidence intervals. P< .05 =*, P<.01=**, P< .001 =***. N=54(Anterior), N=37(Medial), N=33(Posterior).

Figure 2. Cuticle penetration and mortality is more frequent in attacks with nociceptive behaviors.

(A) Larvae that showed nocifensive escape locomotion showed frequent penetration of the cuticle (N=14, 71% (+17/-26)). Gentle touch-like behaviors (turning and locomotion) rarely showed penetration to the cuticle (N=25, 24% (+19/-13)). Writhing behaviors were associated with an intermediate level of penetration (N=9, 56%, (+26/-29)). Fisher’s Exact Test with Holm-Bonferroni correction. Data are presented as percentages ±95% confidence intervals. (B) Representative photomicrograph of a wasp ovipositor (scale bar=20 µm). The arrowhead indicates the location of the ovipositor clip. (C) The melanotic spot was similar to the diameter of the ovipositor clip (N=12, 18µm, ± .4) when attacks triggered either writhing (N=5, 17µm, ±2.6) or nocifensive escape locomotion (N=9, 23µm, ±2.2). The size of the melanotic spot was smaller than the diameter of the ovipositor clip when larvae that showed non-nociceptive behaviors (N=6, 12µm, ±.9). T-test with Holm-Bonferroni correction. Error bars denote standard error of the mean. (D) Mortality was high in larvae that displayed writhing (N=30, 47%, (+17/-16)) or nocifensive escape locomotion (N=24, 50%, (+19/-19)) relative to larvae that displayed locomotion and turning (N=76, 7%, (+8/-4)) . Fisher’s Exact Test with Holm-Bonferroni correction. Data are presented as percentages ±95% confidence intervals. P<.05=*, P<.01=**, P< .001 =***.

Figure 3. Role of the class IV neurons and nocifensive escape locomotion in response to wasp attack.

(A) Confocal micrograph of the dendritic field of the dorsal (ddaC) class IV neuron taken from a larva (ppk-GAL4 UAS-mCD8::GFP/+) that displayed nocifensive escape locomotion following wasp attack. Scale bar is 20 μm. The location of ovipositor penetration is denoted by an arrowhead. See also Figure S1. (B) Expression of dTRPA1-A in 5-10 class IV neurons is sufficient to cause nocifensive escape locomotion. (0 neurons in the no heat shock control (N=65, 15%, (+11/-7)), 0 neurons following heat shock (N=35, 14% (+15/-8)), 1 neuron (N=24, 21%, (+20/-12)), 2 neurons (N=24, 13%, (+19/-8)), 3 neurons (N=11, 9%, (+29/-7)), 4 neurons (N=12, 33%, (+28/-20)), 5-10 neurons (N=20, 50% (+20/-20)), 11-40 neurons (N=14, 71%, (+17/-26)), positive controls (N=112, 88%, (+5/-7)). The genotype used was w;hs-flp/pickpocket1.9-GAL4, UAS-dTRPA1-A;tub>GAL80>/UAS-mCD8::GFP. For positive controls the genotype was pickpocket1.9-GAL4, UAS-dTRPA1-A/+; UAS-mCD8::GFP/+. (C) Larvae with class IV neurons silenced by UAS-TNT-E (N=125) show no nocifensive escape locomotion (0%, (+3/-0)) and increased locomotion (68%, (+8/-9)) compared to larvae expressing impotent TNT in the class IV neurons (N=164, nocifensive escape locomotion 13%, (+6/-4), locomotion 47%, (+8/-7)). The genotypes used were w;ppk-GAL4/UAS-TNT, w;ppk-GAL4/UAS-IMP TNT. Fisher’s Exact Test with Holm-Bonferroni correction. Data are presented as percentages ±95% confidence intervals. P<.05=*, P<.01=**, P< .001 =***.