Novel Detection of Coxiella spp., Theileria luwenshuni, and T. ovis Endosymbionts in Deer Keds (Lipoptena fortisetosa)

We describe for the first time the detection of Coxiella-like bacteria (CLB), Theileria luwenshuni, and T. ovis endosymbionts in blood-sucking deer keds. Eight deer keds attached to a Korean water deer were identified as Lipoptena fortisetosa (Diptera: Hippoboscidae) by morphological and genetic analyses. Among the endosymbionts assessed, CLB, Theileria luwenshuni, and T. ovis were identified in L. fortisetosa by PCR and nucleotide sequencing. Based on phylogeny, CLB 16S rRNA sequences were classified into clade B, sharing 99.4% identity with CLB from Haemaphysalis longicornis in South Korea. Although the virulence of CLB to vertebrates is still controversial, several studies have reported clinical symptoms in birds due to CLB infections. The 18S rRNA sequences of T. luwenshuni and T. ovis in this study were 98.8–100% identical to those in GenBank, and all of the obtained sequences of T. ovis and T. luwenshuni in this study were 100% identical to each other, respectively. Although further studies are required to positively confirm L. fortisetosa as a biological vector of these pathogens, strong genetic relationships among sequences from this and previous studies suggest potential transmission among mammalian hosts by ticks and keds.


Introduction
Deer keds (genus Lipoptena), also known as louse flies, are obligate, blood-feeding ectoparasites that belong to the Hippoboscidae family [1]. Deer keds typically parasitize deer, antelope, goat, and sheep [2]. After keds reach a suitable host, wings are broken off at the base, leaving behind a stump [1].
To date, ill effects by deer keds on hosts have not been well established [3]. Anemia and mechanical damage due to heavy infestation were suggested as clinical symptoms [3]. Recently, the importance of deer keds as a potential vector of various pathogens, including Anaplasma ovis [4], Bartonella spp. [5], Rickettsia spp. [4], and Trypanosoma spp. [6], was reported.
Previous studies have generally investigated pathogens in L. cervi, another deer ked species, whereas pathogens in L. fortisetosa have not been well studied.
L. fortisetosa was first identified in Japan in 1965 [7], and since been identified in only a few other countries, including the Czech Republic [8], Poland [9], and Moldavia [10]. In South Korea, studies on the distribution of Hippoboscidae have identified two species of deer keds, L. cervi and L. fortisetosa, but L. fortisetosa was found only on Jeju island (33°29'N and 126°31'E), which has a warm oceanic climate [2,11,12]. However, pathogens carried by Lipoptena have not been well characterized.
In S. Korea, reports on vector-borne diseases and its pathogens are ubiquitous, which include anaplasmosis in human [13], Borrelia burgdorferi in human [14], Bartonella spp. in Korean water deer [15], Coxiella burnetii in raw milk [16], Hepatozoon spp. in leopard cat [17], and Theileria spp. in Chinese water deer [18]. Climate change, due to global warming, has engendered a more subtropical climate, which may increase the risk of vector-borne diseases nationally [19]. Warm summer seasons, in particular, provide an ideal environment for vectors throughout the country.

Ethics statement
A wild Korean water deer in this study was road-killed and transferred to the Wildlife Treatment Center in Daegu. Ethical approval for the collection of keds and permission to conduct this study on this site were not required from any authority because the deer was dead when transferred to the center and removal of keds from deer was neither harmful nor against animal welfare. All the procedures regarding samplings and experiments were performed by veterinarians with appropriate handlings. The deer carcass was incinerated by an authorized company. While Korean water deer is designated as vulnerable species by International Union for Conservation of Nature and Natural Resources (http://www.iucnredlist.org/details/10329/0), the deer used in this study was dead. Thus, this study did not involve endangered or protected species.

DNA sequencing
Amplicons matching expected sizes were sequenced using BigDye 1 Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, New York, USA) following the manufacturer's instructions, and analyzed with ABI 3730XL DNA Analyzer (Applied Biosystems). Empirical sequences were compared with those deposited in GenBank using BLASTn.

Phylogenetic analysis
Phylogenetic relationships of L. fortisetosa, Coxiella-like bacteria (CLB), T. luwenshuni, and T. ovis investigated in this study were assessed using MEGA 6.06 based on a maximum likelihood method [21]. To estimate the reliability of constructed trees, bootstrap analysis was performed

Identification of ked species
Morphologically, keds collected in this study were flattened dorso-ventrally, with a depressed head and compound eyes (Fig 1A and 1B). The thorax and abdomen gives these insects their louse-like appearance; however, wing vestiges were observed at the margin of the thorax from the dorsal view (Fig 1B, yellow arrows). In addition, the lengths of the head and thorax regions ranged between 1.3-1.5 mm, and body lengths reached~3.0 mm (Fig 1C and 1D). These results are consistent with characteristics of L. fortisetosa, but differ from those of L. cervi. Head and thorax region lengths of L. cervi are~2.0 mm, and body lengths are 4-7 mm [1,11,20]. Moreover, universal primers for the cox-1 gene were used to amplify 472-bp fragments from ked samples. The sequences (accession nos. KU356895 and KU356896) of the cox-1 gene in two ked samples exhibited 95.4% and 96.0% identity, respectively, with the cox-1 gene of L. fortisetosa (accession no. AB632572) deposited in GenBank (Fig 2). Based on morphological characteristics and cox-1 sequence identity, keds collected from a Korean water deer were identified as L. fortisetosa.

PCR and phylogenetic analysis
In all, three species of endosymbionts in L. fortisetosa were identified by PCR. Coxiella spp. were detected in five keds, and T. luwenshuni and T. ovis were detected in six keds. None of the rickettsias, Babesia spp., Bartonella spp., Borrelia spp., and Hepatozoon spp. was detected ( Table 2). Individual keds carried from zero to three endosymbionts.
For Coxiella spp., 719-bp fragments of 16S rRNA were amplified. All five of the obtained sequences exhibited 100% identity to each other, and 99.4% identity with a Coxiella endosymbiont sequence from Haemaphysalis longicornis in S. Korea (AY342035) deposited in the Gen-Bank database. Based on a previous study [22], CLB detected in this study were classified into clade B (Fig 3).

Discussion
Coxiella genus includes C. burnetii, C. cheraxi, and unclassified CLB [22,23]. Among them, C. burnetii is a zoonotic pathogen that causes acute or chronic illness and flu-like symptoms in humans, and abortion in animals [23]. C. burnetii is shed in milk, feces, and urine from infected animals and can be transmitted by inhalation of aerosolized microorganisms [24]. Due to its environmental resistance, route of transmission, and difficulty to diagnose, C. burnetii is designated as a category B potential biological weapon by the United States [25].
According to a recent study, Coxiella spp. could be classified into four different clades (A-D), according to their genetic characteristics [22], with C. burnetii belonging to clade A. In this study, CLB detected from L. fortisetosa was classified into clade B through genetic analysis of 16S rRNA. In addition, CLB detected in our study exhibited 100% identity with Coxiella endosymbiont (AY342035) from H. longicornis in S. Korea. The perfect sequence identity of 16S rRNA between CLB in this study and AY342035 in a previous study suggests that CLB is potentially transmitted between mammals by ticks and keds.
In previous studies, C. burnetii and CLB were detected from various sources including ticks, flies, dairy cattle, raw milk, and aborted fetuses [23,24,26]. In 1958, C. burnetii was detected from Melophagus ovinus (Family: Hippoboscidae), also known as sheep ked, and confirmed as a bona fide vector of C. burnetii through animal experiments, clinical manifestations, and antigen-antibody tests [27]. Based on these results, we suggest that L. fortisetosa could act as a vector not only for CLB but also for C. burnetii, though further studies are required to confirm this hypothesis. Although the virulence of CLB to vertebrates is still controversial [23], several studies have reported clinical symptoms (systematic and even fatal) in birds due to CLB infections [28][29][30]. Therefore, further studies are required to assess the virulence of CLB to vertebrate.  Till now, different Theileria spp. including T. lestoquardi, T. ovis, T. uilenbergi, T. luwenshuni, T. separate, and T. recondite have been found in small ruminants [31]. Of these species, T. lestoquardi, T. luwenshuni, and T. uilenbergi are known to be highly pathogenic to sheep and goats [31]. In S. Korea, T. luwenshuni and T. ovis were reported in Chinese water deer [18], and T. luwenshuni was found in roe deer and H. longicornis [32]. In the past, T. luwenshuni was considered to be a Theileria sp. indistinct from T. ovis and T. lestoquardi that caused ovine and caprine theileriosis [33]; however, it can now be distinguished from T. ovis and T. lestoquardi based on their biological characteristics and 18S rRNA sequence [33]. This is particularly important as the pathogenicities of T. ovis and T. luwenshuni to ruminants vary. In this study, 259-bp of Theileria 18S rRNA was amplified by PCR but amplicon was insufficient to clearly differentiate the species, owing to the high sequence similarity among Theileria spp. Using species-specific primer sets, we revealed the presence of mixed infection of T. ovis and T. luwenshuni in L. fortisetosa.
The sequences of T. luwenshuni in this study showed 100% identity to a Theileria sp. (FJ668369) that was detected from a Chinese water deer in S. Korea [18], and 98.7% identity to T. luwenshuni (KC735157) from a goat in China. While the sequence FJ668369 was originally submitted simply as Theileria sp., we now propose that species is T. luwenshuni based on our phylogenetic analysis. The sequences of T. ovis in this study showed 98.8% identity to T. ovis (FJ668373) that was detected from a Chinese water deer in S. Korea [18]. The high sequence similarity between these sequences and sequences obtained in previous studies suggest the potential transmission of Theileria among mammals, ticks, and keds.
Deer, antelope, goats, and sheep are the main hosts of deer keds [2]. However, incidental infestation of deer keds in other animals, including dogs, horses, and other ruminants, has also been reported [34]. Moreover, human dermatitis caused by ked bites has been often reported in Finland [35].
Recently, due to global warming and an increasing number of wild animals in S. Korea, caution to vector-borne diseases has been raised. Vectors generally transmit pathogens mechanically or biologically [36]. The results of this study suggest that L. fortisetosa is a potential biological vector of CLB endosymbionts. However, ecological aspects, such as the life cycle and reproduction of deer keds, are yet to be determined.
In this study, we report the first identification of CLB, T. ovis and T. luwenshuni in L. fortisetosa. Further investigations are required to confirm L. fortisetosa as a biological vector of these pathogens. Moreover, owing to the possible transmission of vector-borne pathogens, members of the medical and veterinary field need to be cautious of potential contact with deer keds. luwenshuni detected in this study showed 100% identity with one another and with the Theileria sp. (FJ668369) identified from Korean water deer. All 6 T. ovis detected in this study showed 100% identity with one another and with T. ovis (GU726904) identified from a sheep in Iran. The sequences of T. luwenshuni and T. ovis are marked with an arrow and arrowhead, respectively. Phylogenetic trees were constructed based on the maximum likelihood method with 1,000 replicates. Scale bar represents the phylogenetic distance between sequences. Species, host, region of isolation, and GenBank accession numbers are included in the figure. doi:10.1371/journal.pone.0156727.g004