Genomic analysis of the population structure of Paenibacillus larvae in New Zealand

Abstract

New Zealand is a remote country in the South Pacific Ocean. The isolation and relatively late arrival of humans into New Zealand has meant there is a recorded history of the introduction of domestic species. Honey bees (Apis mellifera) were introduced to New Zealand in 1839, and the disease American foulbrood was subsequently found in the 1870s. Paenibacillus larvae, the causative agent of American foulbrood, has been genome sequenced in other countries. We sequenced the genomes of P. larvae obtained from 164 New Zealand apiaries where American foulbrood was identified in symptomatic hives during visual inspection. Multi-locus sequencing typing (MLST) revealed the dominant sequence type to be ST18, with this clonal cluster accounting for 90.2% of isolates. Only two other sequence types (with variants) were identified, ST5 and ST23. ST23 was only observed in the Otago area, whereas ST5 was limited to two geographically non-contiguous areas. The sequence types are all from the enterobacterial repetitive intergenic consensus I (ERIC I) genogroup. The ST18 and ST5 from New Zealand and international P. larvae all clustered by sequence type. Based on core genome MLST and SNP analysis, localized regional clusters were observed within New Zealand, but some closely related genomes were also geographically dispersed, presumably due to hive movements by beekeepers.

Keywords: American foulbrood (AFB); genome; honey bee (Apis mellifera L.); multilocus sequence typing (MLST); sequencing.

FIGURE 1

Geographic distribution of Paenibacillus larvae multi-locus sequence types (MLST) across New Zealand. Pie charts show the proportions of sequence types present in each region. The sequence types are grouped as clonal clusters. Each clonal cluster consists of the main sequence type (ST) and the single nucleotide polymorphism (SNP) variants of the main sequence type. An untypable genome, referred to in the figure as “unknown”, shared six loci with ST23 but the seventh loci was not found.

FIGURE 2

Minimum-spanning of the core genome multilocus sequence types (cgMLST) for n = 163 Paenibacillus larvae in New Zealand, by region. Regions are color coded, with the tally of isolates shown in square brackets. The dashed lines delimit three clusters, each representing one of the 7-gene MLST we found in this study (viz., ST5, ST18, and ST23). The total number of allele differences between each node are shown on the branch lines. Scale bar shows 60 allele differences.

FIGURE 3

Maximum-likelihood phylogenetic analysis comparing n = 163 New Zealand and n = 20 international genomes of Paenibacillus larvae. The scale bar shows the genetic distance. The multilocus sequence types are delineated in gray, alongside the phylogenetic tree.

FIGURE 4

Population structure analysis using hierarchical clustering of core genomes for Paenibacillus larvae from New Zealand. Clustering was based on analysis of single nucleotide polymorphisms. (A) Cladogram shown with subtypes assigned by hierarchical clustering analysis using RhierBAPS. Sequence types (ST) are coded by color for each isolate, and annotated with the unique genome identifier. (B) Cladogram by hierarchical clustering analysis using RhierBAPS without subtype assignment, and with the region of New Zealand annotated for each isolate.