Tetranychus urticae mites do not mount an induced immune response against bacteria

Abstract

The genome of the spider mite Tetranychus urticae, a herbivore, is missing important elements of the canonical Drosophila immune pathways necessary to fight bacterial infections. However, it is not known whether spider mites can mount an immune response and survive bacterial infection. In other chelicerates, bacterial infection elicits a response mediated by immune effectors leading to the survival of infected organisms. In T. urticae, infection by either Escherichia coli or Bacillus megaterium did not elicit a response as assessed through genome-wide transcriptomic analysis. In line with this, spider mites died within days even upon injection with low doses of bacteria that are non-pathogenic to Drosophila Moreover, bacterial populations grew exponentially inside the infected spider mites. By contrast, Sancassania berlesei, a litter-dwelling mite, controlled bacterial proliferation and resisted infections with both Gram-negative and Gram-positive bacteria lethal to T. urticae This differential mortality between mite species was absent when mites were infected with heat-killed bacteria. Also, we found that spider mites harbour in their gut 1000-fold less bacteria than S. berlesei We show that T. urticae has lost the capacity to mount an induced immune response against bacteria, in contrast to other mites and chelicerates but similarly to the phloem feeding aphid Acyrthosiphon pisum Hence, our results reinforce the putative evolutionary link between ecological conditions regarding exposure to bacteria and the architecture of the immune response.

Keywords: Sancassania berlesei; Tetranychus urticae; host–parasite interactions; immunity; microbiota.

Figure 1.

Bacterial infection severely affects survival of T. urticae. (a) T. urticae adult females were infected with E. coli at ODs 0.1, 1 and 10 and with LB as control. There is a clear reduction in survival of T. urticae after infection, independently of the bacteria concentration tested. (b) T. urticae adult females were infected with B. megaterium at ODs 0.1, 1 and 10, with LB and E. coli at OD 0.1 as controls. A reduction in the survival of T. urticae was observed after infection with B. megaterium with the two highest concentrations tested, but not for OD 0.1. In (a,b), vertical bars correspond to the standard errors of survival estimates, obtained from the Cox proportional hazards models. (c) Hazard ratios between T. urticae adults infected with LB or with bacteria (light grey, B. megaterium; dark grey, E. coli). Vertical bars correspond to the 95% CIs of the estimated hazard ratios. ***p < 0.001.

Figure 2.

Spider mites do not trigger an induced immune response against bacterial infection. T. urticae adult females were infected with E. coli, B. megaterium and LB buffer by an injection method and collected 3, 6 or 12 h post-infection (pi) to analyse the transcriptomic responses. Using non-injected mites as a reference, the relative transcription levels of the gene sets that showed significant differential expression in any time point of each bacterial and LB-control treatment were subjected to hierarchical clustering based on the distance calculated by dynamic time warping alignments. Resulting clusters were grouped, of which the means (solid line) and confidence interval (α: 0.05) (shaded regions) are shown for infections with E. coli (in red) (a) or B. megaterium (in green) (b) together with their respective LB controls (in black).

Figure 3.

Sancassania berlesei can resist bacterial infection. (a,b) S. berlesei adult females were infected with E. coli (a) or B. megaterium (b) at three concentrations or with LB as control and their survival was followed daily over 4 days. Bacterial infection decreased survival but to a lesser degree than that observed in T. urticae. In (a,b), vertical bars correspond to the standard errors of survival estimates, obtained from the Cox proportional hazards models. (c) Hazard ratios of S. berlesei adults injected with bacteria relative to LB-injected controls. Vertical bars correspond to the 95% CIs of the estimated hazard ratios. *p < 0.05; **p < 0.01 and ***p < 0.001.

Figure 4.

Escherichia coli proliferation in different species correlates with host survival patterns and gut microbiota size. T. urticae, D. melanogaster and S. berlesei were infected with 5–100 CFUs of E. coli per individual and bacterial dynamics within the host was quantified daily for up to 4 days after infection. (a) In T. urticae, an increase in the number of CFUs in living individuals was observed over the course of 2 days. At 3 and 4 days (asterisk), only dead individuals were plated as there were no survivors, but the CFUs did not increase significantly. (b) For S. berlesei injection, the dynamics of E. coli growth was similar to that of D. melanogaster. (c) Gut microbiota is virtually absent in T. urticae as opposed to S. berlesei. Untreated mites (control) or treated with different protocols to remove bacteria from their surface (sterilized), from their gut (rifampicin) and from both their surface and gut (sterilized and rifampicin) were homogenized and plated on a Petri dish containing LB agar. T. urticae is colonized by two to three orders of magnitude less bacteria than S. berlesei. ***p < 0.001.