Antibacterial immune competence of honey bees (Apis mellifera) is adapted to different life stages and environmental risks

Abstract

The development of all honey bee castes proceeds through three different life stages all of which encounter microbial infections to a various extent. We have examined the immune strength of honey bees across all developmental stages with emphasis on the temporal expression of cellular and humoral immune responses upon artificial challenge with viable Escherichia coli bacteria. We employed a broad array of methods to investigate defence strategies of infected individuals: (a) fate of bacteria in the haemocoel; (b) nodule formation and (c) induction of antimicrobial peptides (AMPs). Newly emerged adult worker bees and drones were able to activate efficiently all examined immune reactions. The number of viable bacteria circulating in the haemocoel of infected bees declined rapidly by more than two orders of magnitude within the first 4-6 h post-injection (p.i.), coinciding with the occurrence of melanised nodules. Antimicrobial activity, on the other hand, became detectable only after the initial bacterial clearance. These two temporal patterns of defence reactions very likely represent the constitutive cellular and the induced humoral immune response. A unique feature of honey bees is that a fraction of worker bees survives the winter season in a cluster mostly engaged in thermoregulation. We show here that the overall immune strength of winter bees matches that of young summer bees although nodulation reactions are not initiated at all. As expected, high doses of injected viable E.coli bacteria caused no mortality in larvae or adults of each age. However, drone and worker pupae succumbed to challenge with E.coli even at low doses, accompanied by a premature darkening of the pupal body. In contrast to larvae and adults, we observed no fast clearance of viable bacteria and no induction of AMPs but a rapid proliferation of E.coli bacteria in the haemocoel of bee pupae ultimately leading to their death.

Figure 1. Fate of injected viable Escherichia coli bacteria in honey bee larvae.

(A, C) Number of colony-forming units (CFU) recovered from the haemolymph of infected larvae. Four-day-old worker larvae (n = 4) (A) and six-day-old drone larvae (n = 3) (C) were challenged with ∼10⁵ E. coli cells and haemolymph was collected at the indicated times post-injection (p.i.). Aliquots of the haemolymph were spread onto agar plates to estimate the number of surviving bacteria. Each point represents the mean number of CFUs ± standard deviation. Separate analyses of covariance testing for CFUs recovered at different times post-injection revealed significant effects for worker larvae at 0.5 h, 6 h and 24 h p.i. (p<0.0001) as well as for drone larvae at 0.5 h (p = 0.0108), 24 h, 48 h and 72 h (p = 0.0106). (B, D) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of infected larvae. Haemolymph aliquots derived from larvae examined in (A, C) were applied onto agar plates together with Micrococcus flavus as indicator bacteria. The original area of application is encircled.

Figure 2. Fate of injected viable E. coli bacteria in honey bee adult workers.

(A) Number of colony-forming units (CFU) recovered from the haemolymph (left y-axis, •) and nodule formation in the haemocoel of infected workers (right y-axis, ▪). Newly emerged worker bees (1d) were challenged with ∼10⁵ E. coli cells and were subsequently divided into two groups: haemolymph was recovered from one batch of individual workers (n = 4–5) at the indicated times (p.i.) and aliquots were spread on nutrient agar plates in order to determine the number of CFUs and the number of melanised nodules per individual worker bee (n = 5–12) was assessed of the second group. The error bars represent the standard deviation. (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of adult worker bees after infection with E. coli at the indicated times.

Figure 3. Fate of injected viable E. coli bacteria in winter bees.

Worker bees were removed from the hive during the winter season (January/February, 2010) and kept in small cages in an incubator at 26°C. They were artificially infected with ∼10⁵ viable E. coli cells and divided into two groups. (A) Number of CFUs recovered from the haemolymph of winter bees (n = 4) at the indicated times p.i. (left y-axis, •) and number of melanised nodules per individual winter bee (n = 15) (right y-axis ▪). (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of winter bees after infection with E. coli at the indicated times. Aliquots of the samples examined in (A) were applied onto agar plates together with M. flavus as indicator bacteria. Inhibition-zones looked brownish in the area of sample application possibly because haemolymph from older worker bees contains small amounts of melanin . (C) Two –dimensional gel electrophoretic analyses of soluble proteins present in the haemolymph of winter bees at 24 h post-infection with 10⁵ E. coli bacteria. For comparison, the protein pattern of non-infected (n.i.) individuals is shown to the left. Proteins were separated by isoelectric focusing (IEF) in the first dimension and by SDS-PAGE in the second dimension. Encircled spots were excised and the proteins were identified by subsequent MS/MS analyses as described . Differentially expressed immune-responsive proteins are marked by red circles. Vg, vitellogenin (gi|58585104); ApoLpII, apolipophorin (gi|66513966); OBP14, odorant binding protein 14 (gi|94158822); IRp30 (gi|66507096); H, hymenoptaecin (gi|58585174), D, defensin 1 (gi|37703274). Gels were stained with Coomassie Brilliant Blue G250.

Figure 4. Fate of injected viable E. coli bacteria in honey bee adult drones.

(A) Number of CFUs recovered from the haemolymph (left y-axis •) and nodule formation in the haemocoel of infected drones (right y-axis ▪). Newly emerged drones (1 d) were challenged with ∼10⁵ E. coli bacteria and were subsequently divided into two groups: haemolymph was recovered from one batch of individual drones (n = 6) at the indicated times (p.i.) and aliquots were spread on nutrient agar plates in order to determine the number of CFUs and the number of melanised nodules per individual drone (n = 5–11) was assessed of the second group. The error bars represent the standard deviation. (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of adult drones after infection with E. coli at the indicated times. (C) Gel electrophoretic analyses of proteins present in the haemolymph of adult drones at the indicated times post-injection of ∼10⁵ E. coli bacteria. Aliquots of haemolymph samples were applied onto a 15% polyacrylamide/0.1% SDS gel . Gels were stained with Coomassie Brilliant Blue G250. The induced antimicrobial peptides and immune-responsive proteins are indicated in red.

Figure 5. Immune reactions of honey bee queens.

(A) Characterization of induced antimicrobial peptides (AMPs) in honey bee queens by gel electrophoretic analyses of haemolymph proteins. Two young queens (2 days old) were left untreated and two queens were challenged with ∼10⁵ viable E. coli bacteria. Single queens were kept in small cages with five worker bees in an incubator at 35°C and 70% relative humidity. They were supplied with Apifonda ad libitum. The haemolymph was collected 24 h post-injection of each individual. Aliquots of these samples were applied onto a 15% polyacrylamide/0.1% SDS gel . Gels were stained with Coomassie Brilliant Blue G250. Induced AMPs are indicated in red. (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of queens challenged with E. coli. Aliquots of the same haemolymph samples as in (A) were applied onto agar plates together with Micrococcus flavus as indicator bacteria. (C) Nodulation reaction of queens. A total of 8 young queens (7 days old) were challenged with 10⁵ viable E. coli bacteria. Upon 24 h p.i., abdominal tergits were removed and the body cavity of queens was exposed followed by inspection for the presence of melanised nodules (indicated by red arrows). Photomicrographs were taken with an Olympus SZX7 stereomicroscope equipped with an Olympus UC30 camera. Magnification was x20 for the complete abdomen and x32 for the detailed view.

Figure 6. Comparison of in vitro cultured worker pupae after aseptic and septic wounding.

White-eyed worker pupae were left untreated (n.i.), challenged with PBS, 10⁵ E. coli bacteria or with 10³ ABPV particles. The pupae were kept in tissue culture plates in an incubator at 35°C. Their development was recorded until the emergence of worker bees in the control group (n.i.).

Figure 7. Fate of injected viable E. coli bacteria in honey bee worker pupae.

White-eyed pupae were collected from sealed combs, transferred into empty tissue culture plates and subsequently kept in an incubator at 35°C and 70% relative humidity. They were challenged with E. coli bacteria as demonstrated in Figure 6. (A) Number of CFUs recovered from the haemolymph of worker pupae (n = 3) post-infection with ∼10⁵ E. coli cells at the indicated times. The error bars represent the standard deviation. (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of worker pupae infected with E. coli. Aliquots of the samples were applied onto agar plates together with M. flavus as indicator bacteria. The area which is covered with newly grown bacteria colonies is indicated by a dotted ring.

Figure 8. Fate of injected viable E. coli bacteria in honey bee drone pupae challenged with a high dose.

White-eyed drone pupae were collected from sealed combs, transferred into empty tissue culture plates and subsequently kept in an incubator at 35°C and 70% relative humidity. They were left either untreated (n.i.) or challenged with E. coli. (A) Number of CFUs recovered from the haemolymph of drone pupae (n = 3–4) at the indicated times post-infection with ∼10⁵ E. coli bacteria. The error bars represent the standard deviation. (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of drone pupae infected with E. coli. Aliquots of the samples were applied onto agar plates together with M. flavus as indicator bacteria. The area which is covered with newly grown bacteria colonies is indicated by a dotted ring. (C) Development of in vitro cultured white-eyed drone pupae within 7 days post-injection of ∼10⁵ E. coli bacteria. For comparison, the development of non-infected control pupae to the imago is shown to the left.

Figure 9. Fate of injected viable E. coli bacteria in honey bee drone pupae challenged with a low dose.

White-eyed drone pupae were collected from sealed combs, transferred into empty tissue culture plates and subsequently kept in an incubator at 35°C and 70% relative humidity. They were left either untreated (n.i.) or challenged with E. coli. (A) Number of CFUs recovered from the haemolymph of drone pupae (n = 3–4) at the indicated times post-infection with ∼10² E. coli bacteria. The error bars represent the standard deviation. (B) Inhibition-zone assay for the detection of antimicrobial activities in the haemolymph of drone pupae infected with E. coli. Aliquots of the samples were applied onto agar plates together with M. flavus as indicator bacteria. The area which is covered with newly grown bacteria colonies is indicated by a dotted ring. (C) Development of in vitro cultured white-eyed drone pupae within 24 h post-injection of ∼10² E. coli bacteria. For comparison, the development of non-infected control pupae is shown to the left.