The assassin bug Pristhesancus plagipennis produces two distinct venoms in separate gland lumens

Abstract

The assassin bug venom system plays diverse roles in prey capture, defence and extra-oral digestion, but it is poorly characterised, partly due to its anatomical complexity. Here we demonstrate that this complexity results from numerous adaptations that enable assassin bugs to modulate the composition of their venom in a context-dependent manner. Gland reconstructions from multimodal imaging reveal three distinct venom gland lumens: the anterior main gland (AMG); posterior main gland (PMG); and accessory gland (AG). Transcriptomic and proteomic experiments demonstrate that the AMG and PMG produce and accumulate distinct sets of venom proteins and peptides. PMG venom, which can be elicited by electrostimulation, potently paralyses and kills prey insects. In contrast, AMG venom elicited by harassment does not paralyse prey insects, suggesting a defensive role. Our data suggest that assassin bugs produce offensive and defensive venoms in anatomically distinct glands, an evolutionary adaptation that, to our knowledge, has not been described for any other venomous animal.

Fig. 1

Assassin bug venom apparatus. a–d Gland reconstructions from MRI. a Sagittal view and b dorsal view, showing position of glands in relation to the surface of the insect (green). VP, venom pump (pink); AMG, anterior main gland (dark blue); PMG, posterior main gland (red); AG, accessory gland (light blue); G, gut (yellow). c Dorsal view of main gland. VD, venom ducts connecting the hilus (H, green) to VP (pink). d Dorsal view of AG and accessory gland ducts (AGDs). e–g 2D MRI slices with compartments in upper panels coloured dark blue (AMG), red (PMG) and green (hilus). Insets show location of each slice. e Parasagittal slice showing separate AMG and PMG lumens. f Dorsal plane slice showing inner (HI) and outer (HO) chambers of hilus. g Parasagittal slice (more medial than that in e), showing hilus connecting AMG and PMG. h MRI reconstruction of hilus showing connectivity to AMG, PMG, VD and AGD. i Confocal laser scanning microscopy (CLSM) images of the hilus and surrounding structures. Left, nuclear stain showing arrangement of hilus, AMG, PMG, VD and AGD. Right, nuclear stain (grey) with actin stain (red) superimposed. Concentrated actin in muscle fibres (M) reveals a sphincter valve (S) at the junction of AMG and HI. j CLSM and differential interference contrast (DIC) images of hilus region. Left, DIC image. Right, antibody against acetylated tubulin (green) highlights nerves (N) and neuromuscular junctions (black arrows) as well as fibrous-appearing staining near the start of VD (white arrows) and gland cells. Red: actin staining of muscle fibres. Blue: nuclear stain. Insets: enlargements of neuromuscular junctions marked by black arrows. k Schematic of hilus with arrows indicating putative direction of flow of gland contents. Red bands represent muscle fibres (M). l–m µCT scans showing arrangement of VP and VD in relation to gut (yellow). l Sagittal view of the head. VC, venom channel; FC, food channel; P, proboscis; A, antenna. m Oblique posterolateral view showing strong concavity on the posterior surface of the VP (left) connected to muscle bundle (MB, orange)

Fig. 2

Predicted secreted proteins in each compartment of the venom glands. a Secretory activity, as measured by the proportion of total transcription encoding predicted secreted proteins (normalised to fragments per kilobase million, FPKM). PMG, main gland posterior lobe; AMG, main gland anterior lobe; AG, accessory gland. b Overlap of transcriptional activity between gland compartments. Each area represents transcript abundance (FPKM) of all secretome entries that are separate or shared between the three compartments

Fig. 3

Different classes of protein are secreted in each gland compartment. a Transcript abundance of major classes of secreted protein classes in each gland compartment. b Proportion of precursor ion counts observed by LC-MS/MS originating from each major protein class (putative housekeeping proteins excluded). PMG, posterior lobe of main gland; AMG, anterior lobe of main gland; AG, accessory gland

Fig. 4

Electrostimulation yields PMG venom and harassment yields AMG venom. a, b Expression levels of proteins unique to venoms obtained either by electrostimulation (a) or harassment (b). c, d Comparison of protein abundance (calculated from precursor ion counts) and expression level in each gland compartment. c Venom obtained by electrostimulation. Values of r and p indicate the regression to PMG expression values. d Venom obtained by harassment. Values of r and p are for regression to AMG expression values

Fig. 5

MALDI spectra of venoms and venom gland extracts. Each spectrum was obtained using MALDI-TOF of samples obtained from the same adult male bug. a MS spectra acquired over the range m/z 2500–4500 Da. Expansions of the regions around 3632 Da (b) and 3796 Da (c) reveal close correspondence of detected masses in these regions between venom samples and gland extracts. d Schematic showing the source of samples analysed. Samples PMG1–PMG4, posterior lobe extracts; AMG, anterior lobe extract

Fig. 6

Venom obtained by electrostimulation but not by harassment paralyses insects. a Effect of injecting venom obtained by electrostimulation or harassment, or water, on cricket escape. For each venom condition, 0.17 µl venom equivalent was injected into the abdomen and the time to escape an upturned Petri dish lid (in s, up to 300 s, mean ± SD) was scored. b Dose–response curve for inhibition of escape success by venom obtained by electrostimulation