Zoonotic Bacteria in Fleas Parasitizing Common Voles, Northwestern Spain

We detected Francisella tularensis and Bartonella spp. in fleas parasitizing common voles (Microtus arvalis) from northwestern Spain; mean prevalence was 6.1% for F. tularensis and 51% for Bartonella spp. Contrasted vector–host associations in the prevalence of these bacteria suggest that fleas have distinct roles in the transmission cycle of each pathogen in nature.

careful visual inspection and by gently blowing the vole's fur while holding the animal over a white plastic tray (520  420  95 mm) filled with water. We counted, collected and preserved in labeled tubes with 70% ethanol all the fleas collected from 225 individual voles. Each flea was later identified using a binocular microscope based on morphologic criteria following Gómez et al. (2). Although in a previous study, 240 common voles were screened for the occurrence of Francisella tularensis (3), for this current study, we used fleas collected from these same voles, but we only considered those animals that arrived alive to the lab to have a reliable estimation of flea infestation in the voles. Individual fleas often abandon carcasses of hosts that die in traps or during transport. From the 225 voles (141 voles were infested with fleas and 84 were not infested; Appendix Table 1), we selected voles that were infested with only 1 flea species, reducing the sample size to 90 individual voles (Appendix Table 1 Table 1). We did not analyze pools containing a mix of the 2 flea species, i.e., voles that simultaneously had fleas of both species were not considered in this study.
Flea pools were selected based on an a priori knowledge of F. tularensis prevalence in the voles that hosted them (3). In particular, from the initial 225 voles, 48 were F.tularensis PCR-positive and 177 were F. tularensis PCR-negative (Appendix Table 1). The proportion of voles infested with fleas and F. tularensis PCR-positive was 70.8% (34/48); while the proportion of voles infested with fleas and F. tularensis PCR-positive was 60.5% (107/177).
We found that F. tularensis infection did not affect flea infestation. There were no significant difference between the proportions ( 2 1.74, g.l. = 1, p>0.05). The selected 90 monospecific flea pools were made up of 27 flea pools from F. tularensis PCR-positive voles and 63 pools from F. tularensis PCR-negative voles. Since in a previous study we used a multiplex PCR method to analyze the DNA of the common voles that hosted the fleas studied here (see [4]; and PCR methods sub-section below), we also knew the prevalence of other zoonotic pathogens in these common voles, specifically Anaplasma phagocytophilum, Bartonella spp., Borrelia spp., Coxiella burnetii, F. tularensis, and Rickettsia spp.

DNA Extraction
DNA from each flea pool was extracted using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the standard procedures of the manufacturer.

PCR Methods
Pathogen detection in the DNA extracted from fleas was carried out using a multiplex PCR that simultaneously detected 6 vectorborne pathogens (A. phagocytophilum, Bartonella spp., Borrelia spp., C. burnetii, F. tularensis, and Rickettsia spp.), combined with a reverse line blotting (RLB), as previously described (5,6). Sensitivity of the multiplex PCR was between 10 and 100 GE (Genome Equivalents), and specificity with unrelated bacteria, mammals and arthropods was 100% (5). The same methodology was used to detect these same pathogens in common voles (4), including those hosting the fleas analyzed here. All the samples that tested positive to any given pathogen were further tested separately using specific probes with an individual PCR and subsequent RLB.
For detection of F. tularensis DNA in a flea pool, a phylogenetically informative region of lpnA (231 bp) was amplified by conventional PCR and further hybridization with specific probes by RLB, as previously described in Escudero et al. (7). Positive samples were tested for confirmation of the results using a real-time multitarget TaqMan PCR, targeting tul4 and ISFtu2 assays (8). A negative PCR control, as well as a negative control for DNA extraction, was included in each group of samples tested.

Identification of Bartonella Species
Bartonella-positive samples were further analyzed using a multiplex PCR targeting the 16S rRNA and the intergenic transcribed spacer (ITS) 16S-23S rRNA. Subsequently, amplicons were analyzed with a RLB that included 36 probes for the identification of the different genotypes and species of Bartonella (9,10). Results are shown in Appendix Table 2.
Page 4 of 5

Statistical Analyses
As the number of fleas per pool ranged from 1 to 9, and all the fleas in each pool were screened together, we estimated an average pathogen prevalence per pool as the mean prevalence between the minimum and maximum prevalence. We assumed that either only 1 of the fleas was positive (minimum prevalence estimate) or that all the fleas from the pool were positive (maximum prevalence estimate). Average pathogen prevalence was estimated for all the fleas and for each flea species separately.
We used an analysis of variance (ANOVA) to test whether the pathogen prevalence in voles had an effect on the average pathogen prevalence in fleas. We also tested whether the average prevalence of a pathogen in fleas was related to the average prevalence of other pathogens in fleas. A p<0.05 was considered significant. Analyses were done with R v3.5.1 (11).