Genomic Epidemiology and Multilevel Genome Typing of Australian Salmonella enterica Serovar Enteritidis

Abstract

Salmonella enterica serovar Enteritidis is one of the leading causes of salmonellosis in Australia. In this study, a total of 568 S. Enteritidis isolates from two Australian states across two consecutive years were analyzed and compared to international strains, using the S. Enteritidis multilevel genome typing (MGT) database, which contained 40,390 publicly available genomes from 99 countries. The Australian S. Enteritidis isolates were divided into three phylogenetic clades (A, B, and C). Clades A and C represented 16.4% and 3.5% of the total isolates, respectively, and were of local origin. Clade B accounted for 80.1% of the isolates which belonged to seven previously defined lineages but was dominated by the global epidemic lineage. At the MGT5 level, three out of five top sequence types (STs) in Australia were also top STs in Asia, suggesting that a fair proportion of Australian S. Enteritidis cases may be epidemiologically linked with Asian strains. In 2018, a large egg-associated local outbreak was caused by a recently defined clade B lineage prevalent in Europe and was closely related, but not directly linked, to three European isolates. Additionally, over half (54.8%) of predicted multidrug resistance (MDR) isolates belonged to 10 MDR-associated MGT-STs, which were also frequent in Asian S. Enteritidis . Overall, this study investigated the genomic epidemiology of S. Enteritidis in Australia, including the first large local outbreak, using MGT. The open MGT platform enables a standardized and sharable nomenclature that can be effectively applied to public health for unified surveillance of S. Enteritidis nationally and globally. IMPORTANCE Salmonella enterica serovar Enteritidis is a leading cause of foodborne infections. We previously developed a genomic typing database (MGTdb) for S. Enteritidis to facilitate global surveillance of this pathogen. In this study, we examined the genomic features of Australian S. Enteritidis using the MGTdb and found that Australian S. Enteritidis is mainly epidemiologically linked with Asian strains (especially strains carrying antimicrobial resistance genes), followed by European strains. The first large-scale egg-associated local outbreak in Australia was caused by a recently defined lineage prevalent in Europe, and three European isolates in the MGTdb were closely related but not directly linked to this outbreak. In summary, the S. Enteritidis MGTdb open platform is shown to be a potentially powerful tool for national and global public health surveillance of this pathogen.

Keywords: Salmonella Enteritidis; foodborne outbreak; genomic epidemiology; genomic typing database; multilevel genome typing; standardized genomic typing.

FIG 1

Schematics of MGT and definition of single-linkage clusters. (a) MGT scheme of 9 levels with an increasing number of loci from MGT1 (the classic 7-gene MLST scheme) to MGT9 (the cgMLST scheme of S. Enteritidis with 4,986 loci). MGT1 to MGT7 were exclusive from each other with no shared loci. MGT8 and MGT9 included loci of the lower levels. (Adapted from reference .) (b) ST calling at different MGT levels. For each strain, an ST was assigned based on exact comparison and a unique combination of alleles at each MGT level. (c) Single-linkage clustering of STs using different cutoffs. For all MGT levels, clonal complex (CC) is defined as STs with a one-allele difference. At MGT9, outbreak detection clusters (ODC) or genomic clusters (GC) are defined as follows: ODC1 (GC1, also MGT9 CC), ODC2 (GC2), ODC5 (GC5), and ODC10 (GC10), allowing for pairwise allele differences of 1, 2, 5, and 10, respectively.

FIG 2

The collection dates of the newly sequenced genomes from NSW and QLD in Australia. The x axis shows the month and the year of collection, while y axis shows the number of isolates for each month.

FIG 3

MGT typing of the newly sequenced genomes. (a) Sequence types (STs) assigned at each MGT level. (b) Single-linkage clustering of STs into clonal complexes (CCs) (assigned based on STs from MGT2 to MGT9) and genomic clusters (GCs) (assigned based on MGT9-STs). STs/CCs/GCs with ≥2 isolates each were separated into different colors, and singleton types (one isolate each) were grouped in gray.

FIG 4

The population structure of the Australian S. Enteritidis using existing nomenclatures. NA, North America. The population structure of all Australian isolates (896), including the publicly available genomes. Seven previously defined lineages of clade B were observed, with the majority of isolates belonging to the global lineage MGT4-CC1. The two African lineages were observed with three isolates. Note that the geographic prevalence of different clades and lineages was based on the metadata in MGTdb, which may have sampling bias. The total number of isolates by state was marked with colour. Total with unknown origin (Null) was marked gray.

FIG 5

Genomic epidemiological characteristics of the MGT5-STs in Australia. (a) The main MGT5-STs (≥5 isolates each) of the newly sequenced Australian S. Enteritidis genomes, and their distribution by states (NSW and QLD). (b) Distribution of the main MGT5-STs in different months. STs with ≥5 isolates each are represented with different colors. Some STs occurred almost every month, whereas several STs appeared and persisted over several months.

FIG 6

Identification of putative outbreak clusters using a 4-week window. Clusters with at least two isolates occurring within 4 weeks in clade A and B were identified at different-resolution genomic cluster (GC) levels. Clusters at each GC level were categorized into four groups. Singletons were clusters with only one isolate each, which were collapsed into one column in gray. For the nonsingleton clusters: 4 weeks, clusters including at least two isolates collected within 4 weeks; not 4 weeks, clusters that did not include two isolates collected within 4 weeks; unknown, clusters without collection date information.

FIG 7

Phylodynamic analysis of the NSW outbreak-related isolates. A total of 29 isolates were sampled with the exponential growth strict clock model as the optimal model. Country and collection year-month information of the sampled isolates is represented in different colors. Isolates belonging to the Australian outbreaks are highlighted in green. The most recent common ancestor (MRCA) of the Australian outbreak isolates was around 2015 (95% CI, 2013 to 2017). Six isolates from the United Kingdom were phylogenetically close to the Australian outbreak isolates, and the MRCA was around 2011 (95% CI, 2008 to 2012). Note that one Australian isolate from 2018 that was unrelated to the large outbreak was identified and included in the tree.

FIG 8

Visualization of an international cluster (GC10-C89) in Microreact. (a) Visualization of the whole cluster. The dendrogram was generated based on the cluster types from different resolution levels (MGT9 STs to GC10). The root is the GC10 cluster, and each leaf is an MGT9 ST. Each internal node is a GC between GC1 and GC9 as labeled. At each GC level, the branches that are linked with a horizontal line belong to the same cluster type. Different colors represent different countries. Isolates that caused possible invasive infections are labeled in green. The size of the circles in the map corresponds to the number of isolates. Note that international isolates with collection year and month information but no date were assumed to be collected on the 15th of each month. Subclusters at different GC levels can be selected and visualized. (b) Visualization of a GC4 subcluster (marked by the red arrow in panel a) as a separate tree. The two Australian S. Enteritidis genomes collected in 2018 were of the same GC2 type as three isolates from the United Kingdom, United States, and Ireland. The UK and Irish isolates were also observed in 2018.