Genome Evolution of a Symbiont Population for Pathogen Defense in Honeybees

Abstract

The honeybee gut microbiome is thought to be important for bee health, but the role of the individual members is poorly understood. Here, we present closed genomes and associated mobilomes of 102 Apilactobacillus kunkeei isolates obtained from the honey crop (foregut) of honeybees sampled from beehives in Helsingborg in the south of Sweden and from the islands Gotland and Åland in the Baltic Sea. Each beehive contained a unique composition of isolates and repeated sampling of similar isolates from two beehives in Helsingborg suggests that the bacterial community is stably maintained across bee generations during the summer months. The sampled bacterial population contained an open pan-genome structure with a high genomic density of transposons. A subset of strains affiliated with phylogroup A inhibited growth of the bee pathogen Melissococcus plutonius, all of which contained a 19.5 kb plasmid for the synthesis of the antimicrobial compound kunkecin A, while a subset of phylogroups B and C strains contained a 32.9 kb plasmid for the synthesis of a putative polyketide antibiotic. This study suggests that the mobile gene pool of A. kunkeei plays a key role in pathogen defense in honeybees, providing new insights into the evolutionary dynamics of defensive symbiont populations.

Keywords: Apilactobacillus kunkeei; defensive symbionts; evolution; mobile elements; plasmids; transposons.

Fig. 1.

Origin of samples and A. kunkeei strain code nomenclature. Schematic illustration of the origins of the cultivated A. kunkeei strains obtained from honeybee hives located on the islands Åland and Gotland in the Baltic Sea and from hives sampled between May and August in Helsingborg, Sweden. Numbers below honeybees represent the number of A. kunkeei strains sampled from each bee. The bacterial strains from Åland and Gotland were named indicating the island (e.g., G0104 or A0901), the host bee (e.g., G0104) and the bacterial isolate (e.g., G0104). The bacterial strains obtained from Helsingborg were named following an alphanumerical system indicating the site (e.g., H3B2-04M), the hive number (e.g., H3B2-04M), the host bee (e.g., H3B2-04M), the bacterial isolate (e.g., H3B2-04M) and the month in which the sample was collected (e.g., H3B2-04M). Abbreviations for months; M = May, J = June, X = July, A = August.

Fig. 2.

Phylogenetic representation of A. kunkeei strains. The strains have been color-coded to represent the site from which they were obtained; blue = Gotland; green = Helsingborg; brown = Åland; black = previously published genomes (Djukic et al. 2015; Porcellato et al. 2015; Sun et al. 2015; Tamarit et al. 2015; Asenjo et al. 2016). Letters A–F indicate phylogroup designations. The phylogeny was obtained using IQ-Tree with the LG + F + R5 model based on 682 single-copy orthologous proteins in 119 A. kunkeei strains and A. apinorum Fhon13. The tree is displayed such that branch lengths are proportional to substitution frequencies. The statistical support for the branching pattern was estimated from 1,000 ultrafast bootstraps and SH-like pseudoreplicates. Nodes supported by 100% by both methods are highlighted with filled circles. For ease of visualization, the branch to A. apinorum Fhon13 is not displayed. The full tree, including A. apinorum Fhon13 and the statistical support values for all nodes are shown in supplementary figure S2, Supplementary Material online.

Fig. 3.

Gene contents in the A. kunkeei population. (A) The pie chart shows the division of 2,656 protein families in the A. kunkeei population into core families and three categories of accessory protein families (shell, soft-core and cloud). The families were obtained from chromosomally encoded proteins in 104 A. kunkeei isolates. (B) The number of protein families plotted against the number of isolates for the conserved core families and the total set of families. Isolates were sampled randomly 10,000 times without replacement. Full and dashed lines show mean number of protein families, and the gray areas show upper and lower bounds. (C) Functional analyses of core and variably present protein families. The number of core and variable protein families is shown as a function of COG category for the proteome of the A. kunkeei population. One-letter abbreviations and full names of COG categories: Amino acid transport and metabolism (E), carbohydrate transport and metabolism (G), cell cycle control, cell division, chromosome partitioning (D), cell motility (N), cell wall/membrane/envelope biogenesis (M), chromatin structure and dynamics (B), coenzyme transport and metabolism (H), defense mechanisms (V), energy production and conversion (C), extracellular structures (W), function unknown (S), general function prediction only (R), inorganic ion transport and metabolism (P), intracellular trafficking, secretion, and vesicular transport (U), lipid transport and metabolism (I), mobilome: prophages, transposons (X), nucleotide transport and metabolism (F), posttranslational modification, protein turnover, chaperones (O), RNA processing and modification (A), replication, recombination and repair (L), secondary metabolites biosynthesis, transport and catabolism (Q), signal transduction mechanisms (T), transcription (K), translation, ribosomal structure and biogenesis (J).

Fig. 4.

Distribution of transposons and prophages. (A) Density plot representing the distribution of transposase genes across the genome in 34 representative strains of A. kunkeei. (B) The location of transposases and prophages in genomes from 34 isolates, including only one representative genome among those that are more than 99.9% identical. The tree topology is schematic and taken from figure 2. Vertical red lines above or below chromosome lines indicate transposase genes located in the Watson or Crick strands, respectively. Blue boxes on the chromosomes represent prophage genes predicted by PHASTER (score more than 30). Gray, connecting lines represent best-reciprocal Blastn hits with e-values lower than 1e−5. An equivalent plot with all complete genomes included in this study is shown in supplementary figure S3, Supplementary Material online.

Fig. 5.

Genes located on plasmids in A. kunkeei strains. Synteny plots for (A) the 19.5 kb pKUN plasmid carrying genes for the biosynthesis of kunkecin A, (B) the 33.5 kb pPKS plasmid carrying genes for the biosynthesis of putative novel polyketide antibiotic, and (C) the 33 kb plasmid carrying a gene for the cell surface protein with the LPxTG motif. Detailed information about the annotation of plasmid genes is provided in supplementary table S6, Supplementary Material online.

Fig. 6.

Phyletic distribution patterns and gene order structures for A. kunkeei proteins with LPxTG-domains. (A) Phyletic distribution patterns of proteins containing the LPxTG domain as predicted by InterProScan. The phylogeny has been taken from figure 2 and the colors of the isolates correspond to the phylogroups. Protein families containing surface proteins with LPxTG motifs, indicated by numbers 1–11 on the right side and arbitrarily named LPxTG-1 to LPxTG-11, are color-coded to represent different orthogroups. Proteins clustered into these orthogroups but without an identified LPxTG domain are shown as boxes without a black border. Gene synteny is shown for a few representative isolates for protein families classified as (B) LPxTG-1, LPxTG-6 and LPxTG-7 (C) LPxTG-3, LPxTG-4 and LPxTG-5.

Fig. 7.

Spot-on-lawn assay for assessing inhibition of M. plutonius by A. kunkeei. Cell-free supernatants of the A. kunkeei strains with predicted pKUN plasmids were added on DSM1582 agar plates with pre-streaked lawns of M. plutonius. The figure shows the presence or absence of inhibition zones for the three biological replicates from each strain. The number in parenthesis indicates how many of the biological replicates demonstrated inhibitory potency against M. plutonius. Raw images are shown in supplementary figure S6, Supplementary Material online and the data is summarized in supplementary table S7, Supplementary Material online.

Fig. 8.

Phyletic distribution patterns of plasmids and phage–plasmids in comparison to generation times. Detailed information about the gene content of plasmids and phage–plasmids is shown in figure 5 and supplementary table S6, Supplementary Material online. Generation times are based on data from multiple experiments per isolate as presented in supplementary table S2, Supplementary Material online. One strain marked with an asterisk had multiple isolates sequenced.