A high-throughput sequencing survey characterizing European foulbrood disease and Varroosis in honey bees

Abstract

As essential pollinators of ecosystems and agriculture, honey bees (Apis mellifera) are host to a variety of pathogens that result in colony loss. Two highly prevalent larval diseases are European foulbrood (EFB) attributed to the bacterium Melissococcus plutonius, and Varroosis wherein larvae can be afflicted by one or more paralytic viruses. Here we used high-throughput sequencing and qPCR to detail microbial succession of larval development from six diseased, and one disease-free apiary. The disease-free larval microbiome revealed a variety of disease-associated bacteria in early larval instars, but later developmental stages were dominated by beneficial symbionts. Microbial succession associated with EFB pathology differed by apiary, characterized by associations with various gram-positive bacteria. At one apiary, diseased larvae were uniquely described as "melting and deflated", symptoms associated with Varroosis. We found that Acute Bee Paralysis Virus (ABPV) levels were significantly associated with these symptoms, and various gram-negative bacteria became opportunistic in the guts of ABPV afflicted larvae. Perhaps contributing to disease progression, the ABPV associated microbiome was significantly depleted of gram-positive bacteria, a likely result of recent antibiotic application. Our results contribute to the understanding of brood disease diagnosis and treatment, a growing problem for beekeeping and agriculture worldwide.

Figure 1

Experimental design and sample size used to infer disease progression. From a disease free “Healthy” apiary in Tucson AZ, we sampled first through fifth instar larvae from three different colonies, five repetitions per colony. From each diseased apiary in Illinois, we sampled one “Healthy” hive with no disease symptoms, and one “Diseased” hive. From the diseased hive, we sampled both asymptomatic and symptomatic larvae. We sampled 3rd, 4th, and 5th instar larvae and also advanced disease wherein larvae instar was indeterminate.

Figure 2

Bacterial load (log scale) of asymptomatic larval instars differs by apiary. Boxplots contain 25–75% of the data, whiskers are 10th and 90th percentiles, and dots represent the range. The red and black horizontal lines represent the mean and median respectively. Light grey box-plots represent larval instars 0–4 from an apiary in Arizona with no history or recent evidence of EFB disease (n = 15 for each larval instar, see Fig. 3). The dark grey box-plots* represent samples of asymptomatic larvae from asymptomatic colonies in apiaries in Illinois with active EFB disease (n = 23 or 24 for each larval instar). As indicated by brackets above the x-axis, asymptomatic larval instars differ in bacterial load according to t-tests of log-transformed normalized abundance (3rd instar; T36 = − 1.8 p < 0.07, 4th instar; T36 = − 3.9, p < 0.0004).

Figure 3

Microbial succession associated with healthy larval development sampling each 24 h to approximate zero to fourth instar. Results from an apiary in Tucson, AZ. with no history or recent evidence of European foulbrood disease. Data is from three colonies, five reps per day per colony, yielding n = 15 for each larval age (x-axis). Estimated average hours after egg hatch (range =  ± 8 h) is listed across the x-axis. Individual larval microbiotas are displayed as relative abundance bar charts (y-axis), and BactQuant (BQ) values estimating microbiome size are displayed as a log notation heatmap vertically across the x-axis.

Figure 4

Panels (A–F) illustrate disease progression of the larval microbiota from six different apiaries; five with European foulbrood disease (A–E) and one with Acute bee paralysis virus-Parasitic mite syndrome (panel F). Each vertical bar represents a single larval microbiota displayed as relative abundance (y-axis). Larval instar and symptomology are displayed across the x-axis; from left to right are four distinct sample groups selected to capture disease progression: Asymptomatic larvae from an asymptomatic colony, asymptomatic larvae from a symptomatic colony, symptomatic larvae from a symptomatic colony, and advanced disease from a symptomatic colony. BactQuant (BQ) results estimating microbiome size are listed as log notation vertically across each x-axis. On panel (F), an RT-qPCR estimate of ABPV viral load is shown at log10 scale below BQ values.

Figure 4

Panels (A–F) illustrate disease progression of the larval microbiota from six different apiaries; five with European foulbrood disease (A–E) and one with Acute bee paralysis virus-Parasitic mite syndrome (panel F). Each vertical bar represents a single larval microbiota displayed as relative abundance (y-axis). Larval instar and symptomology are displayed across the x-axis; from left to right are four distinct sample groups selected to capture disease progression: Asymptomatic larvae from an asymptomatic colony, asymptomatic larvae from a symptomatic colony, symptomatic larvae from a symptomatic colony, and advanced disease from a symptomatic colony. BactQuant (BQ) results estimating microbiome size are listed as log notation vertically across each x-axis. On panel (F), an RT-qPCR estimate of ABPV viral load is shown at log10 scale below BQ values.

Figure 4

Panels (A–F) illustrate disease progression of the larval microbiota from six different apiaries; five with European foulbrood disease (A–E) and one with Acute bee paralysis virus-Parasitic mite syndrome (panel F). Each vertical bar represents a single larval microbiota displayed as relative abundance (y-axis). Larval instar and symptomology are displayed across the x-axis; from left to right are four distinct sample groups selected to capture disease progression: Asymptomatic larvae from an asymptomatic colony, asymptomatic larvae from a symptomatic colony, symptomatic larvae from a symptomatic colony, and advanced disease from a symptomatic colony. BactQuant (BQ) results estimating microbiome size are listed as log notation vertically across each x-axis. On panel (F), an RT-qPCR estimate of ABPV viral load is shown at log10 scale below BQ values.

Figure 5

Estimated copy number of acute bee paralysis virus (ABPV) in symptomatic (Sym.) and asymptomatic (Asy.) larvae diagnosed with either European foulbrood (EFB) disease or parasitic mite syndrome (PMS). All displayed values are relative to the geometric mean viral load of asymptomatic larvae from the asymptomatic Spring Valley colony. Each boxplot is comprised of 20–24 individual larvae (see File S4 for details). The grey box contains 25–75% of the data, whiskers are 10th and 90th percentiles, and dots represent the range. The red and black horizontal lines represent the arithmetic mean and median respectively.