|Home | About | Journals | Submit | Contact Us | Français|
Deer keds (Lipoptena cervi) are thought to have been introduced into New England from Europe during the 1800s. We sought to determine whether L. cervi from Massachusetts deer contained evidence of infection by Bartonella schoenbuchensis, which appears to be maintained by L. cervi in Europe. Five of 6 keds were found to contain B. schoenbuchensis DNA, and 2 deer ticks cofeeding on deer with such keds did as well. The detection of Bartonella DNA in deer ticks probably represents contamination by infected deer blood.
The gram-negative bacterial genus Bartonella contains 18 species and 3 subspecies that are associated with fleas, lice, and sand-flies. The diversity of Bartonella spp. and that of their hosts is increasingly being described using molecular analyses; previously, Bartonella spp. were seemingly restricted to rodents or humans. Ungulates and phocids are now known to be infected, and putative Bartonella DNA has been detected even in sea turtles (Valentine et al. 2007). B. schoenbuchensis was first isolated from wild roe deer (Capreolus capreolus) in Germany (Dehio et al. 2001), as well as from French cattle (Rolain et al. 2003) and from deer keds (Lipoptena cervi) collected from roe deer and red deer in Germany (Dehio et al. 2004). Halos et al. (2004) detected Bartonella spp. in hippoboscid flies, including B. schoenbuchensis in L. cervi, and suggested that these flies may be vectors of Bartonella. In the southern United States, evidence of bacteria closely related to B. schoenbuchensis was detected by polymerase chain reaction (PCR) in L. mazamae (southern deer keds) collected from white-tailed deer (Odocoileus virginianus) (Reeves et al. 2006). L. cervi was first identified on American deer in 1907 (independently in New Hampshire and Pennsylvania), perhaps introduced into the northeastern United States during the late 1800s by importation of European fallow or roe deer (Bequaert 1942). It may be that B. schoenbuchensis was simultaneously introduced and American L. cervi may frequently be infected. Accordingly, we determined whether keds infesting deer in Massachusetts were infected by B. schoenbuchensis. We also sought molecular confirmation of the identity of L. cervi in our sites by sequencing its 18S rDNA. In addition, because ticks have been suggested as vectors for Bartonella spp. (Chang et al. 2001), we determined whether deer ticks (Ixodes dammini, known also as I. scapularis) that concurrently infested deer with keds were infected by B. schoenbuchensis.
Deer keds and ticks were collected from white-tailed deer (O. virginianus) shot in Massachusetts in November 2006. Ticks were collected for ongoing surveys of tick-borne infections in Massachusetts, the results of which will be reported elsewhere. To corroborate our morphological identification of the keds, we amplified and sequenced the 18S rRNA gene, using primers 18S 1.2F, 18S a0.7, 18S a3.5, 18S b7.0, 18S b2.5, 18S 7R, and 18S 9R as described elsewhere (Whiting 2002).
DNA was extracted using the QIAamp DNA Mini Kit (QIAGEN, GmbH, Germany) according to the manufacturer's recommendations. To determine whether keds were infected by Bartonella spp. we used primer pairs BhCS.781p and BhCS.1137n, which amplify a 379-bp fragment of the gltA gene of Bartonella spp. (Norman et al. 1995). PCR products were visualized by electrophoresis. DNA samples positive for this small target were also tested by PCR, targeting a different gene as well as a larger portion of the gltA, using primer pairs 1400F/2300R and BvCS.205p/BhCS.1137n, which amplify a 893-bp fragment of the rpoB gene and a 958-bp fragment of the gltA gene of Bartonella spp., respectively (Renesto et al. 2001, Winoto et al. 2005). Negative controls for PCR consisted of a blank DNA extraction and distilled water added to the PCR mix instead of DNA. One sample of DNA from B. vinsonii arupensis (strain 2198) was included as a positive control.
Individual ticks (I. dammini) were homogenized, aliquots of homogenized samples were pooled (1–6 ticks per pool), and DNA was extracted using the Isoquick kit (Orca Research) or QIAamp Viral RNA Mini Kit (QIAGEN). (Because ticks were analyzed for viral as well as bacterial or protozoal agents for an ongoing survey of tick-borne infections, different methods were used to extract tick DNA.) DNA template was tested by PCR using primers BhCS.781p and BhCS.1137n. Pools that screened positive were reanalyzed by extracted aliquots of homogenates from individual ticks using the Isoquick method.
To identify the bacterial amplicons, phylogenetic analysis was performed. A 958-bp fragment of gltA was sequenced using the primers BvCS.205p, BhCS.781p, Bv322R (5′-GAG ATG AGG CGA ACA GAA GC-3′), and BhCS.1137n (Norman et al. 1995, Winoto et al. 2005); a 893-bp fragment of rpoB using primers 1400F, 2028F, 1596R, and 2300R (Renesto et al. 2001) was sequenced as well. The sequences were aligned with sequences of other Bartonella species registered in GenBank using ClustalW (Thompson et al. 1994). Neighbor-Joining analysis using alignment data was performed using MEGA3 software (Kumar et al. 2004).
Six deer keds were collected from 4 deer in 3 different locations (Table 1). No other deer were found infested. Our samples were identified as L. cervi by morphology using the available keys (Bequaert 1942, Maa 1965). The 1923-bp fragment of 18S rRNA gene from these keds had 99.7% (1788/1794) similarity to the 18S rRNA gene of L. cervi from Europe (GenBank Accession Number AF322426) (Fig. 1). A total of 203 ticks (54 females and 149 males) were collected from 27 deer, and all of them were I. dammini. Of these, 21 ticks (7 females and 14 males) coinfested deer with deer keds (Table 1).
Five (DF11-1, DF35-1, DF41-1, DF41-2, and DF41-3) of 6 deer ked samples were PCR positive for Bartonella species (Table 1). These 5 samples were also tested by PCR for rpoB and gltA genes, and all samples tested positive. In each PCR, negative controls and positive controls were negative and positive, respectively.
Of 38 pools tested, 2 tick pools which contained 12 ticks were positive. PCR for individual ticks from positive pools found 4 ticks positive by PCR (2.0%, 4/203). One and 3 positive ticks came from deer on which Bartonella-positive deer keds DF11-1 and DF41-1 had attached, respectively (Table 1). Tick samples from deer on which ked samples DF17-1 and DF35-1 had attached tested negative by PCR for Bartonella gltA gene.
All PCR products of deer ked samples were sequenced, and 915-bp fragments of gltA gene and 852-bp fragments of rpoB gene were obtained. A Neighbor-Joining analysis based on gltA (Fig. 2A) and rpoB (Fig. 2B) genes demonstrated that the sequences from keds clustered with B. schoenbuchensis.
In Europe, B. schoenbuchensis has been isolated from L. cervi, and L. cervi appears to be a natural reservoir supporting the replication of B. schoenbuchensis (Dehio et al. 2004). In our study, keds morphologically identified as L. cervi were collected from deer. To corroborate our identification of the keds, we generated a 1923-bp portion of the 18S rRNA gene from representative keds, and these proved to be 99.7% similar to those of L. cervi in Europe. Whether the 0.3% divergence between the sequences comprises a difference attributable to allopatry or is typical of intrapopulation variation is not clear. We detected B. schoenbuchen-sis in L. cervi keds collected from white-tailed deer in Massachusetts, as expected if this ked was indeed introduced to the United States from Europe within recent times (Bequaert 1942). The alternative hypothesis is that L. cervi and B. schoenbuchensis may have been widely distributed by elk or caribou given its known range from Britain to eastern Siberia (Bequaert 1942), and that American L. cervi are part of a long-standing Holarctic distribution.
Among 3 deer with deer keds containing Bartonella DNA, 2 were infested with ticks containing the same DNA. The presence of deer keds with bacterial DNA does not necessarily imply that cofeeding ticks become infected: several ticks from 1 deer infested with keds with Bartonella DNA tested negative by PCR. Although it might be argued that deer keds become infected by cofeeding with Bartonella-infected ticks, all tick samples on deer without Bartonella-positive deer keds tested negative by PCR. Bartonella spp. have been detected in ticks including I. pacificus (Chang et al. 2001) and I. ricinus (Halos et al. 2005), but the role of ticks as vector of Bartonella species is unknown. The most parsimonious explanation for our finding Bartonella DNA in ticks is that the deer that they infested were infected themselves, although we did not test any of their blood samples.
The public health implications of B. schoenbuchensis in deer and their ectoparasites remain unclear. L. cervi will readily feed on humans under experimental conditions and there is anecdotal evidence that they do so naturally (Bequaert 1942). Human biting may imply risk for occasional transmission of B. schoenbuchensis to hunters, forestry workers, and crosscountry runners as suggested previously (Dehio et al. 2004), although L. cervi has not yet been definitively demonstrated to transmit B. schoenbuchensis by bite. It is also possible that deer hunters exposed to deer blood are at risk for infection by B. schoenbuchensis, as they may or may not be due to Anaplasma phagocytophilum (Telford 1997). Whether I. dammini is a competent vector for B. schoenbuchensis remains speculative.
We are supported, in part, by grants from the National Institutes of Health (R01 AI 064218, N01 AI 30050). We thank the Massachusetts Division of Fisheries and Wildlife for their cooperation in collecting ticks from deer check stations.