The diversity, evolution and ecology of Salmonella in venomous snakes

Abstract

Background: Reptile-associated Salmonella bacteria are a major, but often neglected cause of both gastrointestinal and bloodstream infection in humans globally. The diversity of Salmonella enterica has not yet been determined in venomous snakes, however other ectothermic animals have been reported to carry a broad range of Salmonella bacteria. We investigated the prevalence and diversity of Salmonella in a collection of venomous snakes and non-venomous reptiles.

Methodology/principle findings: We used a combination of selective enrichment techniques to establish a unique dataset of reptilian isolates to study Salmonella enterica species-level evolution and ecology and used whole-genome sequencing to investigate the relatedness of phylogenetic groups. We observed that 91% of venomous snakes carried Salmonella, and found that a diverse range of serovars (n = 58) were carried by reptiles. The Salmonella serovars belonged to four of the six Salmonella enterica subspecies: diarizonae, enterica, houtanae and salamae. Subspecies enterica isolates were distributed among two distinct phylogenetic clusters, previously described as clade A (52%) and clade B (48%). We identified metabolic differences between S. diarizonae, S. enterica clade A and clade B involving growth on lactose, tartaric acid, dulcitol, myo-inositol and allantoin.

Significance: We present the first whole genome-based comparative study of the Salmonella bacteria that colonise venomous and non-venomous reptiles and shed new light on Salmonella evolution. Venomous snakes examined in this study carried a broad range of Salmonella, including serovars which have been associated with disease in humans such as S. Enteritidis. The findings raise the possibility that venomous snakes could be a reservoir for Salmonella serovars associated with human salmonellosis.

Fig 1. The distribution of the 58 Salmonella enterica serovars isolated from venomous and non-venomous reptiles.

Each bar represents the total number of isolates which belonged to each serovar. Serovars containing isolates that had multiple serovar designations from SISTR are indicated with asterisks. Human pictographs are displayed on serovars which are amongst the top 20 isolated from humans in Africa. Data is based on the global monitoring of Salmonella serovar distribution from the WHO global foodborne infections network data [69].

Fig 2. The diversity of Salmonella isolated from a collection of venomous snakes and non-venomous reptiles.

Core genome maximum likelihood phylogenetic tree. The tree was rooted using S. bongori (S1 Fig). 25 contextual reference genomes representing previously sequenced isolates from each Salmonella subgroup are indicated in red. A cluster of S. Souhanina isolates which demonstrate a high level of genetic similarity are indicated. Colour strips showing metadata are as follows; Subspecies–Subspecies of isolate, Origin—The country of origin of the snake from which the isolate was taken, V or NV—Depicts whether the reptile host was venomous (V) or non-venomous (NV), AMR phenotype—Isolates shown in red were resistant to one or more antimicrobial agent. Metadata for the contextual reference genomes appear as white. Tree was visualised using ITOL (https://itol.embl.de).

Fig 3. Phylogenetic context of carbon-source utilisation by reptile-derived-Salmonella isolates.

Carbon source utilisation data were mapped against the core genome phylogenetic tree. The maximum likelihood tree includes all reptile-derived Salmonella isolates which were assessed for carbon utilisation and had high quality genome sequences (see S4 Table, S5 Table and S1 Text). Reference sequences for the majority of carbon utilisation and acquisition genes were taken from S. Typhimurium strain 4/74, and the lac gene sequences were from E. coli MG1655. Carbon sources which required anaerobic conditions are indicated with asterisk.