A Bacterial Symbiont Protects Honey Bees from Fungal Disease

Abstract

Fungal pathogens, among other stressors, negatively impact the productivity and population size of honey bees, one of our most important pollinators (1, 2), in particular their brood (larvae and pupae) (3, 4). Understanding the factors that influence disease incidence and prevalence in brood may help us improve colony health and productivity. Here, we examined the capacity of a honey bee-associated bacterium, Bombella apis, to suppress the growth of fungal pathogens and ultimately protect bee brood from infection. Our results showed that strains of B. apis inhibit the growth of two insect fungal pathogens, Beauveria bassiana and Aspergillus flavus, in vitro. This phenotype was recapitulated in vivo; bee broods supplemented with B. apis were significantly less likely to be infected by A. flavus. Additionally, the presence of B. apis reduced sporulation of A. flavus in the few bees that were infected. Analyses of biosynthetic gene clusters across B. apis strains suggest antifungal candidates, including a type 1 polyketide, terpene, and aryl polyene. Secreted metabolites from B. apis alone were sufficient to suppress fungal growth, supporting the hypothesis that fungal inhibition is mediated by an antifungal metabolite. Together, these data suggest that B. apis can suppress fungal infections in bee brood via secretion of an antifungal metabolite. IMPORTANCE Fungi can play critical roles in host microbiomes (5-7), yet bacterial-fungal interactions are understudied. For insects, fungi are the leading cause of disease (5, 8). In particular, populations of the European honey bee (Apis mellifera), an agriculturally and economically critical species, have declined in part due to fungal pathogens. The presence and prevalence of fungal pathogens in honey bees have far-reaching consequences, endangering other species and threatening food security (1, 2, 9). Our research highlights how a bacterial symbiont protects bee brood from fungal infection. Further mechanistic work could lead to the development of new antifungal treatments.

Keywords: Apis mellifera; Bombella; Bombella apis; Nosema; Parasaccharibacter apium; acetic acid microbes; symbiosis.

FIG 1

B. apis outcompetes fungal pathogens in vitro. (a) The ability of each fungal isolate to grow in the presence of B. apis was qualitatively assayed by plating 10³ spores of each isolate across a lawn of B. apis. (b) Compared to controls of 10³ spores plated on fresh media, the presence of B. apis completely inhibited fungal growth. (c) When cocultured in liquid media, the presence of B. apis significantly reduced the number of spores produced by B. bassiana (Kruskal-Wallis; χ² = 11.7, df = 4, P = 0.01973; pairwise comparisons to control A29, t = 13.114, df = 2.0996, P = 0.019056; B8, t = 11.147, df = 2.9658, P = 0.00652; C6, t = 10.121, df = 2.7744, P = 0.011404; SME1, t = 12.352, df = 2.0277, P = 0.024652). (d) B. apis also significantly reduced the number of spores produced by A. flavus (Kruskal-Wallis; χ²  = 9.9, df = 4, P  = 0.04215); however, pairwise comparisons between control and SM from each strain were not significantly different due to the variation in the control samples. To control for nutritional effects, in panels c and d, experimental wells contained the same volume of fresh media as the controls in addition to B. apis culture. Each experimental group consists of three biological replicates. Sporulation was quantified for each well via hemocytometer.

FIG 2

Bee broods supplemented with B. apis are less susceptible to infection with A. flavus. (a) First-instar larvae (n = 45) collected from the apiary were reared on sterile larval diet with or without B. apis (AJP2). Five days after pupation, each pupa was inoculated with 10³ spores of A. flavus with or without B. apis or carrier (0.01% Triton X-100) as a control (without A. flavus/without B. apis, n = 8; without A. flavus/with B. apis, n = 11; with A. flavus/without B. apis, n = 11; without A. flavus/with B. apis, n = 15). (b) Of the pupae inoculated with A. flavus, those without B. apis all showed signs of infection by 48 h, whereas 66% of those with B. apis never developed infections (χ² = 14.8, df  = 1, P < 0.001). (c and d) Representative photos of individual pupae 48 h postinfection show the prominent difference in infection between the two groups. (e) Pupae with B. apis that did become infected (4 out of 15) had lower intensity infections, producing significantly (t = 5.5052, df = 5.5751, P = 0.0019) fewer spores than those without B. apis. The same outcome was replicated in three separate experiments, including with a different B. apis strain (Fig. S2).

FIG 3

Fungal inhibition is mediated by B. apis-secreted metabolites. (a) Spores of fungal isolates were incubated in spent medium (SM) from B. apis cultures in addition to fresh medium or in fresh medium alone. (b) The growth of both B. bassiana (A29, t = −15.315, df = 119, P < 0.001; B8, t = −13.925, df = 119, P < 0.001; C6, t = −13.202, df = 119, P < 0.001; SME1, t = −11.963, df = 119, P < 0.001) and A. flavus (A29, t = −11.398, df = 59, P < 0.001; B8, t = −13.022, df = 59, P < 0.001; C6, t = −13.282, df = 59, P < 0.001; SME1, t = −11.261, df = 59, P < 0.001) in SM was strongly reduced compared to the control in fresh media, suggesting secreted metabolites from B. apis mediate fungal inhibition. The same volume of fresh media was in control and experimental wells. Significant inhibition via SM alone was observed in upwards of three independent experiments. (c) Genomic architecture of the terpene, type 1 polyketide synthase, and arylpolyene secondary metabolite gene clusters identified by antiSMASH; gene models are colored based on putative function within the cluster and are oriented to show direction of transcription.