Abstract
Urban Norway rats (Rattus norvegicus) are a reservoir for Bartonella spp. - a genus of zoonotic bacteria transmitted by hematophagous vectors, particularly fleas. Rats and fleas may be infected with more than one Bartonella species; however, mixed infections may be difficult to detect using culture and/or mono-locus PCR. We set out to characterize Bartonella spp. using gltA PCR and Sanger sequencing on blood (n = 480) and Nosopsyllus fasciatus flea pools (n = 200) obtained from a population of urban Norways rats from Vancouver, Canada. However, when contamination of a subset of flea pools necessitated the use of a second target (ssrA) and the results of gltA and ssrA were discordant, a metagenomic approach was used to better characterize the Bartonella spp. present in these samples and our objective transitioned to comparing data obtained via metagenomics to those from PCR/sequencing. Among the Bartonella spp.-positive rats (n = 95), 52 (55.3%), and 41 (43.6%) had Sanger sequences consistent with Bartonella tribocorum and Bartonella vinsonii, respectively. One rat had a mixed infection. All sequences from Bartonella spp.-positive flea pools (n = 85), were consistent with B. tribocorum, and re-analysis of 34 bloods of varying Bartonella spp. infection status (based gltA PCR and sequencing) using ssrA PCR showed that the assay was capable of identifying B. tribocorum but not B. vinsonii. Metagenomics analysis of a subset of PCR-positive blood samples (n = 70) and flea pools (n = 24) revealed that both B. tribocorum and B. vinsonii were circulating widely in the study population with 31/70 (44.3%) rats and 5/24 (2.1%) flea pools infected with both species. B. vinsonii, however, made up a smaller relative proportion of the reads for samples with mixed infections, which may be why it was generally not detected by genus-specific PCR and Sanger sequencing. Further analysis of 16S-23S ITS sequences amplified from a subset of samples identified the B. vinsonii strain as B. vinsonii subsp. berkhoffii type II. This demonstrates the value of a metagenomic approach for better characterizing the ecology and health risks associated with this bacterium, particularly given that the less dominant species, B. vinsonii is associated with greater pathogenicity in people.