Ecto- and endoparasite induce similar chemical and brain neurogenomic responses in the honey bee (Apis mellifera)

Abstract

Background: Exclusion from a social group is an effective way to avoid parasite transmission. This type of social removal has also been proposed as a form of collective defense, or social immunity, in eusocial insect groups. If parasitic modification of host behavior is widespread in social insects, the underlying physiological and neuronal mechanisms remain to be investigated. We studied this phenomenon in honey bees parasitized by the mite Varroa destructor or microsporidia Nosema ceranae, which make bees leave the hive precociously. We characterized the chemical, behavioral and neurogenomic changes in parasitized bees, and compared the effects of both parasites.

Results: Analysis of cuticular hydrocarbon (CHC) profiles by gas chromatography coupled with mass spectrophotometry (GC-MS) showed changes in honey bees parasitized by either Nosema ceranae or Varroa destructor after 5 days of infestation. Levels of 10-HDA, an antiseptic important for social immunity, did not change in response to parasitism. Behavioral analysis of N. ceranae- or V. destructor- parasitized bees revealed no significant differences in their behavioral acts or social interactions with nestmates. Digital gene expression (DGE) analysis of parasitized honey bee brains demonstrated that, despite the difference in developmental stage at which the bee is parasitized, Nosema and Varroa-infested bees shared more gene changes with each other than with honey bee brain expression gene sets for forager or nurse castes.

Conclusions: Parasitism by Nosema or Varroa induces changes to both the CHC profiles on the surface of the bee and transcriptomic profiles in the brain, but within the social context of the hive, does not result in observable effects on her behavior or behavior towards her. While parasitized bees are reported to leave the hive as foragers, their brain transcription profiles suggest that their behavior is not driven by the same molecular pathways that induce foraging behavior.

Figure 1

Nosema-infected bees developed different cuticular hydrocarbons profiles. Discriminant analysis based on the cuticular hydrocarbons profiles of Nosema-infected and control bees at day 5 and 10. The analysis was repeated on bees from 3 different colonies (N = 11–12 bees/colony and treatment). Young (5 days) parasitized and control bees did not display different chemical profiles, but 5 days later both profiles were distinct (see Table 2).

Figure 2

Varroa-infested bees developed different cuticular hydrocarbons profiles. Discriminant analysis based on the cuticular hydrocarbons profiles of Varroa-infested and control bees. The analysis was repeated on bees from 3 different colonies (N = 12 bees/colony and treatment).

Figure 3

Parasitism did not affect the levels of 10-HDA levels in (A) Nosema- or (B) Varroa-parasitized bees. The 10-HDA levels did not differ between Nosema-infected and control bees at day 5 and 10 but increase with age in colony 120 and 231 (except for control) (Kruskall-Wallis tests: Colony 98: H = 4.53, P = 0.2; Colony 120: H = 19.32, P < 0.001; Colony 231: H = 8.57, P = 0.035; * denotes significant differences after Conover-Iman post-hoc tests, P < 0.05). Varroa did not modify the 10-HDA levels (Mann–Whitney tests: Colony 98: P = 0.38, Colony 120: P = 0.68; Colony 231: P = 0.058).

Figure 4

Parasitism did not induce different behavioral treatment by colony nestmates. The rate of behavioral acts or interactions did not differ between Nosema-infected and control bees (A and B), and Varroa-infested and control bees (C and D) (Mann–Whitney tests, P > 0.05 for each behavior). The experiment was performed on two colonies (19–20 bees/state in the Nosema experiment and 12 bees/state in the Varroa experiment).

Figure 5

Expression levels show similar directionality for 20 brain genes commonly affected by Nosema and Varroa parasites. For the total number of genes changed by parasitism by Varroa (N = 455 genes) or Nosema (N = 57 genes), a statistically significant number of genes occur in both lists (N = 21) with 20 genes expressed in the same direction (Exact hypergeometric probability test: P < 0.0001). Color scale for the heatmap (red to green) indicates log2 transcription ratios where red color indicates underexpression of the gene in the parasitized bee and green color indicates overexpression. For each gene, the accession number and annotation are indicated.

Figure 6

Deformed wing virus titers increased in the honey bee brain with parasitism. Viral titers are expressed as the total number of tags in a sample for each colony individually (Colony 1 and Colony 2). Deformed wing virus (DWV) levels were significantly higher in Varroa-infested bees compared to Nosema- and control bees (Generalized Linear Model with Cox-Reid method for estimating dispersion; Varroa vs. control: P < 0.05, Varroa vs. Nosema:P < 0.01), while increased DWV levels in Nosema-infested bees were marginally significant (P < 0.10). Varroa destructor virus (VDV) titers did not differ between parasitized and control bees (For P-values see Additional file 5: Table S6.)