Parasitism causes changes in caterpillar odours and associated bacterial communities with consequences for host-location by a hyperparasitoid

Abstract

Microorganisms living in and on macroorganisms may produce microbial volatile compounds (mVOCs) that characterise organismal odours. The mVOCs might thereby provide a reliable cue to carnivorous enemies in locating their host or prey. Parasitism by parasitoid wasps might alter the microbiome of their caterpillar host, affecting organismal odours and interactions with insects of higher trophic levels such as hyperparasitoids. Hyperparasitoids parasitise larvae or pupae of parasitoids, which are often concealed or inconspicuous. Odours of parasitised caterpillars aid them to locate their host, but the origin of these odours and its relationship to the caterpillar microbiome are unknown. Here, we analysed the odours and microbiome of the large cabbage white caterpillar Pieris brassicae in relation to parasitism by its endoparasitoid Cotesia glomerata. We identified how bacterial presence in and on the caterpillars is correlated with caterpillar odours and tested the attractiveness of parasitised and unparasitised caterpillars to the hyperparasitoid Baryscapus galactopus. We manipulated the presence of the external microbiome and the transient internal microbiome of caterpillars to identify the microbial origin of odours. We found that parasitism by C. glomerata led to the production of five characteristic volatile products and significantly affected the internal and external microbiome of the caterpillar, which were both found to have a significant correlation with caterpillar odours. The preference of the hyperparasitoid was correlated with the presence of the external microbiome. Likely, the changes in external microbiome and body odour after parasitism were driven by the resident internal microbiome of caterpillars, where the bacterium Wolbachia sp. was only present after parasitism. Micro-injection of Wolbachia in unparasitised caterpillars increased hyperparasitoid attraction to the caterpillars compared to untreated caterpillars, while no differences were found compared to parasitised caterpillars. In conclusion, our results indicate that host-parasite interactions can affect multi-trophic interactions and hyperparasitoid olfaction through alterations of the microbiome.

Fig 1. Experimental setup for the caterpillar odour and microbiome collection.

Groups of eight C. glomerata parasitised or unparasitised caterpillars were placed in a glass jar with a restriction device. A dynamic air current was led through the setup and VOCs were trapped in a Tenax trap. After headspace (odour) trapping, the same caterpillars were subjected to the collection of their external and internal microbiome in PBS-Tween80 solutions.

Fig 2. Volatile compounds (tentative identification) detected in the headspace of untreated, starved or starved then external microbiome disrupted unparasitised or C. glomerata parasitised caterpillars (P. brassicae) with a Variable Importance in the Projection (VIP) values > 1 for the OPLS-DA.

Amounts of individual compounds are given as the average of peak height /10⁴ (±SE). Statistical differences among treatments for compounds with VIP score > 1 are indicated with different letters based on Kruskal-Wallis tests with Dunn’s test for multiple comparisons including a Bonferroni correction. a) Compounds in yellow background are correlated with parasitism status. b) Compounds in green background are correlated with the presence/absence of frass. c) Compound in blue background is not correlated to either of these. Abbreviations used: Pb = Untreated unparasitised caterpillars; Cg = Untreated parasitised caterpillars; Pb-ST = starved unparasitised caterpillars; Cg-ST = starved parasitised caterpillars; Pb-ST+EMD = starved then external microbiome disrupted unparasitised caterpillars. Cg-ST+EMD = starved then external microbiome disrupted parasitised caterpillars.

Fig 3. Overview of caterpillar volatile profiles of untreated, starved or starved then external microbiome disrupted C. glomerata parasitised or unparasitised caterpillars (P. brassicae).

a) OPLS-DA (Orthogonal Projection to Latent Structures Discriminant Analysis) plot for the volatile blends of different groups of caterpillars. The Hotelling’s T2 ellipse confines the confidence region (95%) of the score plot. b) Loading plot defining the contribution of each of the volatile compounds to the separation of treatment groups. Volatile compounds closer to a treatment in the plot correlate stronger with the treatment. For compound identity see Table 1. Abbreviations used: Cg = Untreated parasitised caterpillars; Cg-ST = starved parasitised caterpillar; Cg-ST+EMD = starved then external microbiome disrupted parasitised caterpillars. Pb = Untreated unparasitised caterpillars; Pb-ST = starved unparasitised caterpillars; Pb-ST+EMD = starved then external microbiome disrupted unparasitised caterpillars.

Fig 4. Overview of caterpillar-associated bacterial communities of untreated, starved or starved then external microbiome disrupted C. glomerata parasitised or unparasitised caterpillars (P. brassicae).

a) Non-metric multidimensional scaling (NMDS) ordination plots based on Bray–Curtis distances of Hellinger-transformed relative abundance data of the external bacterial communities. b) NMDS ordination for the internal bacterial communities. c) External bacterial community profiles of the different caterpillar samples. d) Internal bacterial community profiles of the different caterpillars samples. For each zOTU, the average relative abundance (%) for each group is given in the cell as a percentage, whereas the colour indicates prevalence (white is absent). Only bacteria with an overall relative abundance >1.5% are shown in (c) and (d). The full overview is provided as supporting information (Table B in S1 Supporting Information). zOTUs are identified by a BLAST search against type materials in GenBank. When no significant similarity was found with type materials, the BLAST analysis was performed against entire GenBank (indicated with an asterisk). Identifications were performed at genus level; when identical scores were obtained for different genera, identifications were performed at family level. Abbreviations used: Cg = untreated parasitised caterpillars; Cg-ST = starved parasitised caterpillar; Cg-ST+EMD = starved then external microbiome disrupted parasitised caterpillars. Pb = untreated unparasitised caterpillars; Pb-ST = starved unparasitised caterpillars; Pb-ST+EMD = starved then external microbiome disrupted unparasitised caterpillars.

Fig 5. Forward selection redundancy analyses (RDA) on volatiles of untreated, starved or starved then external microbiome disrupted C. glomerata parasitised or unparasitised caterpillars (P. brassicae) and their bacterial communities.

The triplots indicate the volatile composition of the caterpillars (dots), the volatile compounds (numbers as in Table 1) and forward selected zOTUs after 9999 permuations (vectors) for the external (a) and internal (b) bacterial communities. Further, relative abundance-prevalence data of the selected zOTUs are presented for the external (c) and internal (d) bacterial community. The average relative (%) abundance for each group is given in the cell as a percentage, whereas the colour indicates prevalence (white is absent). zOTUs were identified by a BLAST search against type materials in GenBank. When no significant similarity was found with type materials, the BLAST analysis was performed against entire GenBank (indicated with an asterisk). Identifications were performed at genus level. When identity percentages were lower than 99%, the percentage of sequence identity with the GenBank entry is given between brackets. Abbreviations used: Cg-ST = starved parasitised caterpillar; Cg-ST+EMD = starved then external microbiome disrupted parasitised caterpillars; Pb-ST = starved unparasitised caterpillars; Pb-ST+EMD = starved then external microbiome disrupted unparasitised caterpillars.

Fig 6. Preference of the hyperparasitoid B. galactopus for caterpillar body odours of C. glomerata parasitised and unparasitised P. brassicae caterpillars.

This was tested with Y-tube olfactometer tests. Tested combinations were selected according to the hypothesis that hyperparasitoids can use caterpillar body odours, which are (at least partially) determined by the external microbiome. Numbers between brackets indicate the number of wasps that made a choice within 10 min from the start of the experiment versus the total number of wasps tested. *** P <0.001; two-sided binomial test. Abbreviations used: Air = clean air (black); Cg-ST = starved parasitised caterpillar (solid blue); Cg-ST+EMD = starved and external microbiome disrupted parasitised caterpillars (solid red). Pb-ST = starved unparasitised caterpillars (blue border); Pb-ST+EMD = starved then external microbiome disrupted unparasitised caterpillars (red border).

Fig 7. Results of the no-choice assay to assess the role of Wolbachia in hyperparasitoid host location.

Bar plots showing (a) the average time (±SE) until first contact of B. galactopus with the caterpillar host P. brassicae, and (b) total mounting time (±SE) of B. galactopus on the caterpillars within 1 h after first contact was made. Hyperparasitoids were subjected to four treatments, including unparasitised caterpillars (Unparasitised-PBS; n = 39), caterpillars parasitised by C. glomerata (Parasitised-PBS; n = 36), caterpillars injected once with Wolbachia (Wolb-PBS; n = 50), and caterpillars injected twice with Wolbachia (Wolb-Wolb; n = 46). Statistical differences among treatments are indicated with different letters based on Kruskal-Wallis tests with Dunn’s test for multiple comparisons including a Hochberg correction. Pie charts show the percentage of responding (black) and non-responding (grey) hyperparasitoids.

Fig 8. Y-tube olfactometer setup.

a) The setup consisted of a Y-tube with an 8.5 cm-long stem, two 5-cm-long arms (angle between both arms of 65°) and a diameter of 0.9 cm. A fine gauze mesh was placed over each end of the Y-tube to create a physical barrier, then a 5 mL transparent pipette tip containing caterpillars was mounted on each end of the Y-tube. A charcoal filtered airflow of 500 mL min^-1 was split, continuously measured with flowmeters and led through both arms, sealed with Teflon tape. Hyperparasitoid females were released at the basis of the Y-tube and tests concluded once they passed the finish line for at least 15 seconds. b) The setup was mounted on a wooden board and tilted upward with an angle of 40° towards a single light source.