Electronic Supplementary Material
Immune Priming in Ant Larvae: Social Immunity Does Not Undermine Individual Immunity
Rebeca B. Rosengaus1*, Tanya Malak, and Christopher MacKintosh
1Department of Biology, Northeastern University, Boston, MA 02115, USA.
E-mail: , ,
*Author for correspondence ()
Methods
Colony maintenance
Five colonies of Camponotus pennsylvanicus, collected from Newton MA, were maintained in the laboratory inside covered plastic tubs with soil, original wood, and corrugated cardboard for nesting. The colonies were fed honey agar [26], supplemented with cut mealworms and water tubes as needed.
Bacterial cultures
Serratia marcescens, a gram-negative bacterium, was chosen as the microorganism to elicit an immune response in these immature ants. Following an overnight culture of bacteria in tryptic soy broth [27], the bacterial pellets were washed twice with sterile Ringer solution and then separated into 1 ml aliquots that were boiled for 15 minutes in order to produce the heat-killed vaccine suspension. We verified that the vaccine had no live bacteria by plating 100μl into tryptic soy agar plates which were incubated at 25°C for 24 hours. No colonies were observed from any of the prepared vaccine aliquots. Active bacteria to be used for the immune challenge treatment were identically grown with the exception of the boiling procedure. In their case, pink CFUs were recovered after similar incubation.
Larvae injections and confirmation of S. marcescens infection
Depending on their size, second and third instar larvae were injected under a dissecting scope at 40-60X magnification. The larvae, while on ice, were injected (dorsally or ventrally) with a pulled glass capillary tube (~2 microns in diameter) through the intersegmental membrane, usually between segments IV-V or V-VI (Figure S1a). To visualize and confirm that the respective solutions were successfully injected into the insect’s hemocoel, both the Ringer and vaccine solutions were colored by adding green sterile food coloring. A successful injection typically resulted in larvae attaining a green coloration (Figure S1b). The food coloring was metabolized within the first 24 hours post-injection, as injected larvae returned to their typical white/translucent tint.
Larvae were considered dead when they remained unresponsive while prodding under a dissecting microscope and when they attained a yellow hue and appeared somewhat “shriveled” relative to the white, “plump” and responsive live larvae. Dead larvae were removed daily and moved to the “morgue” for confirmation purposes. Cadavers confirming positive for S. marcescens infection attained a pink coloration due to the presence of prodigiosin (Figure S1c), a pigment typically expressed in this bacterium. Dead larvae that turned brown or black were not positively confirmed for S. marcescens infection.
Statistics
Overall susceptibility to bacterial infection was determined through the following parameters: the time course of survival (survival distributions), median survival time (LT50) and percent survival at the end of the census period (Kaplan-Meir Survival Analyses, [28]). Larvae that disappeared from the subnests were assumed cannibalized and hence, were censored on the last day they were seen alive. Relative hazard ratios of death were calculated using Cox Regressions and the Wald Statistic (WS, [28]). Factors in the regression model included colony of origin and treatment (Naïve/Challenge, Ringer/Challenged and Vaccinated/Challenge). Average percent cannibalism and confirmations as a function of treatment were compared with an ANOVA. Any post-hoc pairwise comparisons included Tukey’s HSD test.
Figure Legends Electronic Supplementary Material
Figure S1. Larva undergoing injection with a pulled glass capillary tube (a). To visually ensure that injections were successful, the solutions were colored with green food coloring (b). In (c) dead larva positively confirmed S. marcescens infection after challenge while live vaccinated larva (d) underwent injection with heat-killed bacteria three days prior. Larvae are capable of cuticle healing as evidenced by the brown melanotic scars at the point of injection (e and f, arrows).
Figure S2. Time course of survival as a function of colony of origin. Each line represents the survival of each of the five colonies after controlling for the effect of treatment. Colony of origin was significant and independent predictor of larvae survival (Wald Statistic=54.0, df=4, p≤0.0001, Cox proportional regression, [28]), even after controlling for the effect of treatment.
Figure S3. Time course of survival of C. pennsylvanicus larvae as a function of treatment. Each graph represents one colony. These graphs resulted from Kaplan-Meir Survival analyses and pairwise significant differences between distributions are indicated with different letters next to the curves. Due to multiple pairwise comparisons, the significance level was adjusted to a more conservative p value of 0.01 (Bonferroni correction). After such adjustment, Colony # 3 was the only one with no significant pairwise comparisons between treatments. Cannibalized larvae were considered censored on the last day they were seen alive. Hence, dashed lines represent the time course of survival when their censored data was included. With the exception of Colony # 5, each distribution within each colony is based on 20 larvae. For Colony # 5, the sample size for each distribution is based on 30 larvae.
.
Electronic Supplementary Material Figure. S1
Electronic Supplementary Material Figure. S2
Electronic Supplementary Material Figure. S3
References
26. Bhatkar, A.P., Whitcomb, W.H. 1970 Artificial diet for rearing various species of ants. Florida Entomol. 53, 229–232
27. Rosengaus, R.B., Traniello, J.F.A., Chen, T., Brown, J.J., Karp, R.D. 1999 Immunity in a social insect. Naturwissenschaften 86, 588-591
28. SPSS 1990. SPSS/PC+4.0 Advanced Statistic Manual. SPSS, Chicago
4