American Dog Ticks (Dermacentor variabilis) as Biological Indicators of an Association between the Enteric Bacterium Moellerella wisconsensis and Striped Skunks (Mephitis mephitis) in Southwestern Manitoba, Canada Open Access

Several studies have highlighted the use of hematophagous arthropods for xenosurveillance, a noninvasive approach for the detection of zoonotic pathogens in the blood of vertebrate hosts (Stadler et al. 2011; Calvignac-Spencer et al. 2013; Grubaugh et al. 2015). Sometimes referred to as vertebrate samplers or living syringes, hematophagous arthropods are particularly valuable surveillance tools when blood sampling of wildlife is difficult or needs to be minimized (Stadler et al. 2011). Ixodid ticks are ideal for xenosurveillance because they spend several days feeding on the blood and lymph of their vertebrate hosts (Balashov 1967).

The American dog tick (Dermacentor variabilis) is an important vector of pathogens to humans, domestic animals, and wildlife in North America (Lindquist et al. 2016). The microbiome of D. variabilis also contains nonpathogenic bacterial endosymbionts (Hawlena et al. 2013) that influence tick physiology or behavior. For example, species in the genus Arsenophonus (order Enterobacteriales) reduce the motility of questing D. variabilis larvae (Kageman and Clay 2013). Previously, we determined the prevalence of Arsenophonus within seven Canadian populations of D. variabilis using PCR and DNA sequencing (Dergousoff and Chilton 2010). Ticks in that study were host-seeking adults collected by flagging, except for 100 adults collected from vertebrate hosts near Minnedosa in Manitoba (Table 1). Although Arsenophonus was not detected by PCR in any tick from Minnedosa (Dergousoff and Chilton 2010), some amplicons derived from these ticks contained the DNA of other bacteria in the order Enterobacteriales (possibly Moellerella and Serratia spp.) and the order Pseudomonadales (Pseudomonas spp.; Dergousoff 2011). Given that Moellerella wisconsensis is a potential enteric human pathogen (Cardentey-Reyes et al. 2009) and has been associated with wildlife in the US (Sandfort et al. 2002; Wilson and Kurz 2017), we examined whether ticks from Minnedosa were infected with Moellerella.

A PCR assay was designed to amplify about 780 base pairs (bp) of the bacterial 16S rRNA gene using primers Moellerella-F (5′-TGGGAATGGCATCTAAAACTGGTC-3′) and NC-1387-mod-R (5′-GGGCGGTGTGTA CAAGGC-3′) and the reaction conditions described in Dergousoff and Chilton (2010). The total genomic (g)DNA of the 100 ticks from Minnedosa (Table 1) was tested using this PCR assay. Twelve samples (M28–M30, M32–39, and M44) were PCR positive; however, of the four amplicons (M29, M35, M36, and M38) selected at random for sequencing, only M35 and M38 had 16S sequences (739 bp) that were >99% similar to those of Moellerella sequences on GenBank.

To confirm the identity of the putative Moellerella, 1,304 bp of the 16S rRNA gene were amplified by PCR from the gDNA of one partially engorged female tick (M38). We performed PCR using primers 63F (5′-CAGGCCTAACACATGCAAGTC-3′) following Marchesi et al. (1998) and NC-MS-R (5′-GTTCGCTTCTCTTTGTATACG-3′) under the conditions described previously (Dergousoff and Chilton 2010). The purified amplicon was sequenced using primers 63F and NC-MS-R in separate reactions. The sequence obtained (GenBank accession no. LT986671) differed at two to four nucleotide positions compared with the 16S rDNA sequences of different strains or isolates of M. wisconsensis. A Bayesian inference analysis was conducted using the Monte Carlo Markov chain method in MrBayes v.3.2.2 (Ronquist and Huelsenbeck 2003). Likelihood parameters were based on the Akaike information criterion test in jModeltest v.2.1.5 (Posada 2008). Posterior probability values were calculated by running 2,000,000 generations with four simultaneous tree-building chains. The results revealed strong statistical support (i.e., P=0.989) for M. wisconsensis forming a distinct clade to the exclusion of other related bacteria (Providencia and Morganella spp.), and for the Moellerella in the gDNA of D. variabilis belonging to a clade (with a posterior probability value of 0.881) that included M. wisconsensis isolated from human stool samples in the US and France (Fig. 1). Our report of M. wisconsensis is unique in Canada.

Since the PCR assay using primers Moellerella-F and NC-1387-mod-R was not 100% specific for Moellerella, the total gDNA of the 100 D. variabilis adults was retested for the presence of M. wisconsensis using a new PCR assay designed to amplify about 350 bp of the 16S rRNA gene of Moellerella. Primers Mwiscon-16S-88F (5′-ATGGGGATCTGCCTGA CA-3′) and Mwiscon-16S-419R (5′-GATAGT ATTAATATCAACG-3′) were based on the 16S sequence of a type strain (DSM 5676T) of M. wisconsensis (GenBank accession no. AM040754). We performed PCRs using 25-µL reaction volumes containing 5 µL of 5× Phusion Green HF buffer and 0.5 U of Phusion Hot Start II DNA polymerase^™ (Thermo Fisher Scientific, Toronto, Ontario, Canada), 200 µM each deoxyribonucleoside triphosphate, 3 mM magnesium chloride, 25 pmol of each primer, and 1.5 µL of gDNA or water (as negative controls) under the following conditions: 95 C for 5 min, 30 cycles of 95 C for 60 s, 52 C for 60 s, and 72 C for 60 s, and then 74 C for 5 min. Amplicons were produced from the gDNA of one male (M35) and one female (M38) tick collected from two striped skunks.

The external surface of each D. variabilis adult collected from Minnedosa had been washed and vortexed several times in sterile water before DNA extraction. By chance, a wash sample for ticks M35 and M38 and two other individuals (i.e., M34 and M39) attached to the same skunks had been stored at –80 C. The gDNA extracted from each of the four wash samples was PCR positive using primers Mwiscon-16S-88F and Mwiscon-16S-419R. The DNA sequences (313 bp) of the four purified amplicons were >99% similar to the sequences of M. wisconsensis in GenBank.

Detection of M. wisconsensis DNA in the gDNA of D. variabilis does not imply that these ticks are vectors of the bacterium, or that they represent a health risk to humans. After feeding, female ticks detach from hosts to lay eggs and then die, whereas most males remain on hosts to mate with other females. Tick-borne bacteria transmitted transovarially (i.e., from female ticks to their offspring) usually occur at a high prevalence in tick populations; hence this is unlikely for M. wisconsensis given that this bacterium has rarely been detected in nature. Furthermore, immature stages (larvae and nymphs) of D. variabilis are unlikely to become infected with M. wisconsensis because in Manitoba, they feed mainly on small rodents (e.g., shrews, voles, and mice) and not skunks (Dergousoff et al. 2013). Given the detection of M. wisconsensis DNA on the external surface of the four ticks, the most plausible explanation is that D. variabilis is not a vector of M. wisconsensis and that the presence of this bacterium in the tick gDNA occurred as a consequence of the adult ticks coming into physical contact with, or feeding on, contaminated or infected striped skunks. However, it does not exclude the possibility that M. wisconsensis may have been present in host blood and lymph consumed by the ticks since this bacterium has been found in the lymphatic tissues of 23% of elk (Cervus elaphus) in Utah (Wilson and Kurz 2017). This needs to be verified by direct testing of Mephitis mephitis. Unfortunately, no tissue or blood was available from the skunks that were hosts to the ticks used in this study. Nonetheless, the results of the present study show that the four ticks represent biological indicators of the presence of Moellerella wisconsensis, a rare facultative human pathogen (Cardentey-Reyes et al. 2009).

Although M. wisconsensis has been primarily found in human clinical samples (feces, peritoneal fluid, sputum, and blood) in the US (Wisconsin, Virginia, and New York), France, Germany, Belgium, Switzerland, and the Czech Republic (Hickman-Brenner et al. 1984; Stock et al. 2003; Aller et al. 2009; Cardentey-Reyes et al. 2009), it has also been isolated from unchlorinated drinking water in South Dakota (Hickman-Brenner et al. 1984) and the lung of a goat in Italy (Casalinuovo and Musarella 2009). The bacterium has also been isolated from the oral secretions of a raccoon in northern California, raising the question as to whether raccoons are reservoir hosts and a possible source of zoonotic infection given their close association with human populations (Sandfort et al. 2002). However, in our study, M. wisconsensis was not detected in the gDNA of ticks feeding on raccoons, whereas it was detected in the total gDNA of ticks feeding on two skunks collected near Minnedosa. Striped skunks are a source of zoonotic pathogens (Berrada et al. 2006), and an association with M. wisconsensis may represent another potential health risk for researchers, veterinarians, and the public that come into contact with contaminated or infected animals. Our study highlights the use of blood-feeding arthropods, such as the American dog tick, as biological indicators of microorganisms associated with wildlife.

We thank Travis Quirk for kindly providing ticks used in this study and Alexander Halpin for some technical assistance. This research was supported in part by grants from the Natural Sciences and Engineering Research Council of Canada and the Canadian Foundation for Innovation (to N.B.C.).