Sylvatic Plague in a Canadian Black-Tailed Prairie Dog (Cynomys ludovicianus) Open Access

In July 2010, a black-tailed prairie dog (Cynomys ludovicianus) was found deceased on the Larson Prairie Dog Colony in Grasslands National Park, Saskatchewan, Canada (49°6′0″, 107°24′0″). Concurrently, researchers from Canadian Wildlife Service observed that prairie dogs at a separate colony in the park had disappeared, and anecdotal evidence indicated that active prairie dogs were last observed on the 4.6-ha colony in October 2009. Although prairie dog population abundance is highly variable among years, the population during summer 2010 was 50–70% below the long-term average (Parks Canada unpubl. data). Hypothesized causes for the population decline included climate (drought) or predation; it was also hypothesized that plague may have been acting in a supporting role. The prairie dog was submitted for necropsy with suspected sylvatic plague.

On gross necropsy, the prairie dog was in excellent condition. The spleen was moderately enlarged and there was generalized mild to moderate congestion and edema of all tissues. Histologic examination of tissues revealed thrombi containing small gram-negative bacilli and moderate numbers of neutrophils in numerous blood vessels of the brain, liver, kidney, heart, and adrenal gland. Pulmonary alveoli were filled with protein-rich edema fluid and the spleen contained a marked neutrophil and fibrin exudate with a myriad of small bacteria. Based on gross and histologic findings, a preliminary diagnosis of bacterial septicemia, likely due to Yersinia pestis, was made.

Identification of this isolate, named “EBD10-058” was confirmed using an National Microbiology Laboratory in-house (unpublished) real-time PCR assay targeting three regions of the plasmid and chromosome in addition to conventional bacteriologic tests (identification of the capsular F1 antigen and CDC A1122 bacteriophage). In addition, two single-nucleotide polymorphism (SNP) assays were undertaken to rapidly determine the origin of the strain (Vogler et al. 2008). These assays demonstrated the presence of a biovar Orientalis, North America–specific SNP, and lack of the Y. pestis CO92–specific SNP.

Whole genomic sequencing was performed using the genome analyzer IIx (Illumina Inc, San Diego, California, USA) following the manufacturer's instructions, generating single end reads (150-bp length). De novo sequence assemblies were completed using Edena v3 (Hernandez et al. 2008), followed by reference mapping of the EBD10-058 de novo contigs to another characterized North American Y. pestis (CO92) using SMALT map version 0.6.4 (Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, UK). Unmapped regions of CO92 (<4 mapped reads) were utilized to identify regions missing within EBD10-058. The reference mapping analyses revealed that the genetic material for EBD10-058 was encoded on one chromosome and three plasmids of similar size to that of Y. pestis CO92 (Parkhill et al. 2001). There did not appear to be any large regions of divergence.

We identified SNPs from reference-mapped pileups using FreeBayes version 0.9.6 (Garrison and Marth 2012). Regions were visualized using Tablet version 1.12.09.03 (Milne et al. 2010). For SNP analysis only high-quality SNPs that had >5× coverage, >85% consensus among reads, and minimum of 250 bp of consecutively mapped sequence flanking on either side were used. Once these SNPs were identified in EBD10-058, alignments to 12 publically available North American Y. pestis strains were conducted using progressive Mauve aligner (Darling et al. 2010). Thirty-three SNPs between EBD10-058 and CO92 (Table 1) were found. This included 14 unique SNPs (10 coding and four noncoding) not previously reported. The remaining 19 SNPs were previously reported between CO92 and other sequenced North American strains (Auerbach et al. 2007; Touchman et al. 2007).

Comparison of gene content within EBD10-058 to other North American Y. pestis strains revealed that there were few to no differences, confirming the genetic monomorphic nature of 1.ORI North American Y. pestis strains (Morelli et al. 2010). A SNP matrix table (Table 2) illustrates relatedness and provides context among publicly available isolates that would otherwise yield a linear tree structure. The sequence has been deposited at DDBJ/EMBL/GenBank under accession number ATWT00000000.

In Canada, there have been very few documented outbreaks of sylvatic plague. Sporadic outbreaks were reported in ground squirrels (species not reported) in Alberta and Saskatchewan in the 1930s and a 1939 report describes an outbreak in farmed mink (Mustela vison), presumably the result of using ground squirrels as a food source; the mink rancher also died (Wobeser et al. 2009). A more recent report described two cases of plague among bushy-tailed woodrats (Neotoma cinerea) in southern British Columbia (Lewis 1989). Although confirmed animal cases of plague are rare, the disease is present in the western provinces as evidenced by detection of antibodies to Y. pestis in domestic dogs (Canis lupus familiaris) and cats (Felis catus) in southern Saskatchewan and Alberta (Leighton et al. 2001). In contrast, plague occurs and is reported frequently in US Gunnison's prairie dog (Cynomys gunnisoni) populations throughout the midwestern states (Friggens et al. 2010; Busch et al. 2011). Wobeser et al (2009) theorized that plague in Canada is less visible than in the US because colonial rodents that are highly susceptible to the disease, such as prairie dogs, are confined to a small area of southern Saskatchewan, in Grasslands National Park.

This is the first report of sylvatic plague in Canadian black-tailed prairie dogs and the first Canadian isolate to be subject to molecular analysis. As expected, this isolate belongs to the biovar Orientalis group (1.ORI branch, specific to North America), further confirming that all Y. pestis in North America seem to have arisen from one introduction event (Morelli et al. 2010). In addition, 14 unique SNPs were found. This parallels previous studies that have shown a range of two to 18 unique SNPs for each North American Y. pestis strain compared to CO92 (Auerbach et al. 2007; Touchman et al. 2007). Hopefully, with enhanced surveillance and molecular characterization of future isolates we can define Canadian-specific SNPs and contribute further to the understanding of the radiation of plague in Canada.

We thank B. Balcewich and K. Hayden for technical expertise. This work was supported by the Public Health Agency of Canada, the Canadian Cooperative Wildlife Health Centre, and Parks Canada. The views and opinions expressed herein are those of the authors only and do not necessarily represent the views and opinions of the Public Health Agency of Canada, the Canadian Food Inspection Agency, or the Government of Canada.