Detection of Zoonotic Bartonella Pathogens in Rabbit Fleas, Colorado, USA

Shingo Sato; R. Jory Brinkerhoff; Erin Hollis; Shunta Funada; Avery B. Shannon; Soichi Maruyama


Emerging Infectious Diseases. 2020;26(4):778-781. 

In This Article

The Study

We collected fleas from live-trapped desert cottontail rabbits (Sylvilagus audubonii) in June and July 2005 from 8 sites in Boulder County, Colorado, USA.[10] We identified fleas to species by light microscopy using dichotomous keys[10] and then stored them in 96-well tissue culture plates at −20°C, except for representatives of each species that were removed and submitted to the Centers for Disease Control and Prevention (Fort Collins, CO, USA). In 2015, we extracted DNA from individual fleas using commercial DNA extraction kits (Blood and Tissue Kit; Macherey-Nagel, Inc.,, with aliquots of extracted DNA maintained at the University of Richmond (Richmond, VA, USA), and secondary aliquots sent to the Laboratory of Veterinary Public Health, Nihon University College of Bioresource Sciences (Fujisawa, Japan). Both laboratories screened samples for Bartonella infection by conventional PCR targeting part of the ssrA gene; primers used were ssrA-F (5′-GCTATGGTAATAAATGGACAATGAAATAA-3′) and ssrA-R (5′-GCTTCTGTTGCCAGGTG-3′). The targeted gene was selected because of the robustness of the PCR assay and the ability of the locus to segregate Bartonella at the species level.[11] Nihon University College of Bioresource Sciences also used real-time PCR targeting the ssrA gene to confirm Bartonella species for those samples. This PCR used a genus-specific TaqMan probe, 6-carboxyfluorescein (FAM)–labeled probe (5′-FAM-ACCCCGCTTAAACCTGCGACG-3′-BHQ1, where BHQ is black hole quencher); primers were the same as for conventional PCR. Samples that tested positive for Bartonella DNA by real-time PCR and for which unambiguous sequence data were collected in both laboratories from the target locus (ssrA) were reported as Bartonella positive. We sequenced all amplicons (301 bp) from conventional PCR and aligned them with Bartonella type strains and then subjected them to phylogenetic analysis using MEGA 7.0 (

We collected 141 fleas from 14 desert cottontail rabbits (average fleas per parasitized host 14.3, range 1–54) in the summer of 2005. Of these fleas, 105 (81 Euhoplopsyllus glacialis and 24 Cediopsylla inaequalis) collected from 7 rabbits sampled at 4 sites (Table 1; specific site locations in 10) were available for molecular screening for Bartonella. The remaining 36 fleas were processed for Yesinia pestis surveillance in a separate project (R.J. Brinkerhoff et al., unpub. data) and were not available for Bartonella testing.

We detected Bartonella DNA in 2 (8.3%) C. inaequalis fleas collected from 1 rabbit (ID no. 522) and 21 (25.9%) E. glacialis fleas collected from 5 rabbits (ID nos. 305, 522, 633, 673, and 674) (Table 1). All nucleotide sequences matched closely to 3 zoonotic Bartonella species, B. alsatica, B. vinsonii subsp. berkhoffii, and B. rochalimae (Table 2), and clustered phylogenetically with reference sequences of the type strains with high bootstrap support (Figure). The representative sequences of the 3 Bartonella species were registered in International Nucleotide Sequence Database Collaboration with accession nos. PS522-c9 (GenBank accession no. MN654366), PS674-e5 (GenBank accession no. MN654366), and PS674-e6 (GenBank accession no. MN654366). All 3 rabbits (ID nos. 522, 633, 674) from which >1 flea was PCR-positive and available for sequencing produced multiple Bartonella species (Table 2).


Phylogenetic relationships of Bartonella ssrA sequences detected in study of zoonotic Bartonella in rabbit fleas, Colorado, USA, compared with reference sequences. This tree was generated based on 253 bp by maximum likelihood and 1,000 bootstrap replicates using the Kimura 2-parameter evolutionary model with gamma-distributed rates among sites. Sample numbers are found in Table 2. GenBank accession numbers are indicated. Scale bar indicates nucleotide substitutions per site.