INTRODUCTION
Bartonellae are gram-negative, facultative intracellular, vector-borne bacteria, widely distributed among animal reservoirs, worldwide (Chomel et al. Reference Chomel, Boulouis, Breitschwerdt, Kasten, Vayssier-Taussat, Birtles, Koehler and Dehio2009). Most Bartonella species establish long-term and subclinical infections in their associated reservoir host (Chomel et al. Reference Chomel, Boulouis, Breitschwerdt, Kasten, Vayssier-Taussat, Birtles, Koehler and Dehio2009). Bartonella species have been associated with a wide range of domesticated and wild animals, and to date more than 33 known Bartonella species and subspecies have been described, and many other Candidatus species and uncharacterized genotypes have been documented (Kosoy et al. Reference Kosoy, Hayman and Chan2012). Among their host reservoirs predominate mammals, including rodents, carnivores, lagomorphs, insectivorous, marine mammals, bats and primates (Vayssier-Taussat et al. Reference Vayssier-Taussat, Le Rhun, Bonnet and Cotte2009). Only a single report of a non-mammalian host (sea turtles) has been reported (Valentine et al. Reference Valentine, Harms, Cadenas, Birkenheuer, Marr, Braun-McNeill, Maggi and Breitschwerdt2007). Notably, several Bartonella species have been recognized as emerging pathogens for incidental hosts, such as humans and other animals (Chomel and Kasten, Reference Chomel and Kasten2010). Among them, Bartonella rochalimae was isolated from a bacteremic patient, who presented fever and splenomegaly (Eremeeva et al. Reference Eremeeva, Gerns, Lydy, Goo, Ryan, Mathew, Ferraro, Holden, Nicholson, Dasch and Koehler2007). In addition, B. rochalimae was later associated with a fatal case of endocarditis in a dog (Henn et al. Reference Henn, Gabriel, Kasten, Brown, Koehler, MacDonald, Kittleson, Thomas and Chomel2009b ), revealing its interspecies pathogenic potential. Bartonella rochalimae has been associated with wild carnivores, including coyotes (Canis latrans), grey foxes (Urocyon cinereoargenteus), red foxes (Vulpes vulpes) and raccoons (Procyon lotor), which are considered to be reservoirs of this species (Henn et al. Reference Henn, Chomel, Boulouis, Kasten, Murray, Bar-Gal, King, Courreau and Baneth2009a ).
Rodents have been reported as reservoirs of several Bartonella spp., including the zoonotic species Bartonella elizabethae (Daly et al. Reference Daly, Worthington, Brenner, Moss, Hollis, Weyant, Steigerwalt, Weaver, Daneshvar and O'Connor1993), Bartonella grahamii (Birtles et al. Reference Birtles, Harrison, Saunders and Molyneux1995) and Bartonella vinsonii subsp. arupensis (Welch et al. Reference Welch, Carroll, Hofmeister, Persing, Robison, Steigerwalt and Brenner1999). Interestingly, many novel and uncharacterized Bartonella strains and genotypes are continuously detected in many rodent species (Inoue et al. Reference Inoue, Maruyama, Kabeya, Hagiya, Izumi, Une and Yoshikawa2009). Furthermore, rodent hosts are easily found to be co-infected with more than one Bartonella sp. or variant (Buffet et al. Reference Buffet, Marsot, Vaumourin, Gasqui, Masseglia, Marcheteau, Huet, Chapuis, Pisanu, Ferquel, Halos, Vourc'h and Vayssier-Taussat2012; Gutiérrez et al. Reference Gutiérrez, Morick, Cohen, Hawlena and Harrus2014b ).
In Israel, a diverse range of Bartonella species has been detected in various animals and their associated ectoparasites. These include stray and domestic cats (Avidor et al. Reference Avidor, Graidy, Efrat, Leibowitz, Shapira, Schattner, Zimhony and Giladi2004; Gutiérrez et al. Reference Gutiérrez, Morick, Gross, Winkler, Abdeen and Harrus2013) and cat fleas (Gutiérrez et al. Reference Gutiérrez, Nachum-Biala and Harrus2015b ), dogs (Ohad et al. Reference Ohad, Morick, Avidor and Harrus2010) and dog fleas (Sofer et al. Reference Sofer, Gutiérrez, Morick, Mumcuoglu and Harrus2015), cattle and their lice (Gutiérrez et al. Reference Gutiérrez, Cohen, Morick, Mumcuoglu, Harrus and Gottlieb2014a ; Rudoler et al. Reference Rudoler, Rasis, Sharir, Novikov, Shapira and Giladi2014), domestic camels (Camelus dromedarius) (Rasis et al. Reference Rasis, Rudoler, Schwartz and Giladi2014), rodents, including black rats (Rattus rattus), Cairo spiny mice (Acomys cahirinus), Sundevall's jirds (Meriones crassus), Balochistan gerbils (Gerbillus nanus), Anderson's gerbils (Gerbillus andersoni) and their associated-fleas (Harrus et al. Reference Harrus, Bar-Gal, Golan, Elazari-Volcani, Kosoy, Morick, Avidor and Baneth2009; Morick et al. Reference Morick, Baneth, Avidor, Kosoy, Mumcuoglu, Mintz, Eyal, Goethe, Mietze, Shpigel and Harrus2009, Reference Morick, Krasnov, Khokhlova, Gottlieb and Harrus2011; Gutiérrez et al. Reference Gutiérrez, Morick, Cohen, Hawlena and Harrus2014b ) and red foxes (Henn et al. Reference Henn, Chomel, Boulouis, Kasten, Murray, Bar-Gal, King, Courreau and Baneth2009a ). Hence, as the ongoing interaction between pets, farm animals and humans as well as the gradual geographic extension of urban areas to the wild, the risk of transmission of emerging zoonoses is highlighted. The aim of this study was to detect and characterize the Bartonella spp. circulating in wild animals in Israel.
MATERIALS AND METHODS
Animal blood and tissue sampling
Spleen samples, EDTA blood, blood spotted onto filter paper and heart clot samples were collected from wild animals during 2008–2011, in 82 locations in Israel, from Rosh Hanikra in the north to Nitzana in the south. A total of 365 samples were collected from 275 animals belonging to the following orders: Rodentia (156 animals), Carnivora (81 animals), Hyracoidea (35 animals) and Erinaceomorpha (three animals). Accordingly, 73 samples from 46 Social voles (Microtus socialis), 62 from 43 Tristram's jirds (Meriones tristrami), 62 from 57 Cairo spiny mice (A. cahirinus), ten from six house mice (M. musculus), four from four Indian crested porcupines (Hystrix indica), 76 from 70 Golden jackals (Canis aureus), 15 from 11 red foxes (V. vulpes), 58 from 35 rock hyraxes (Procavia capensis) and five samples from three southern white-breasted hedgehogs (Erinaceus concolor) were screened. The animals were captured and sampled by the Nature and Parks Authorities of Israel as part of epidemiological and diagnostic studies on leishmaniasis, conducted by the Ministry of Health and the ministry of Environment of Israel. Maps, indicating the collection sites were constructed using their coordinates in AcrMap 10.0 software (Esri, Redlands, CA, USA).
DNA extraction
DNA was extracted from the tissue samples by guanidine thiocyanate technique (Hoss and Paabo, Reference Hoss and Paabo1993), with the following modifications. Each sample was cut and a portion of 0·5–1 g was placed in a 1·5 mL sterile Eppendorf tube containing 500 µL solution of 4 m Guanidinium thiocyanate (GuSCN), 0·1 m Tris–HCl (pH 6·4), EDTA 0·02 m (pH 6·4) and 1·3% Triton X-100. All tubes, including a control (with all the reagents except a sample), were incubated overnight at 56 °C with constant agitation (50 rpm), followed by a second incubation at 94 °C for 10 min. Then, samples were centrifuged at 14 000 rpm for 5 min. The supernatants were recovered into new sterile 1·5 mL tubes, and 900 µL sodium iodide (NaI, Sigma-Aldrich, MO, USA), 15 µL silica beads (Sigma-Aldrich, MO, USA) and 15 µL of linear acrylamide were added and placed on ice for 1 h, stirring frequently by vortex. The extraction solutions were centrifuged at 5000 rpm for 30 s. The supernatants were discarded and the silica pellets were washed with 500 µL of washing buffer (10 m GuSCN and 0·1 m Tris–HCI, pH 6·4), and were centrifuged at 5000 rpm for 30 s. The supernatants were discarded and the beads were washed with 200 µL of ethanol absolute and centrifuged at 5000 rpm. The ethanol was removed by pipetting, and the tubes were set to dry at room temperature for 2 h. Silica beads were then treated with 90 µL of ultra-pure water (UPW), 10 µL of TE buffer (10 mm Tris pH 7·5, EDTA pH 7·5–8·0, UPW). Purified DNA was obtained in 100 µL of elution buffer. The samples were incubated at 56 °C for 1 h, and finally stored at 20 °C.
DNA from EDTA blood samples (200 µL) was extracted using the Illustra Tissue and Cells genomicPrep Mini Spin kit (GE Healthcare, Buckinghamshire, UK), following the manufacturer's recommendations.
DNA from blood samples spotted onto a filter paper was extracted using the phenol–chloroform–isoamyl alcohol method following the modifications described elsewhere (Strauss-Ayali et al. Reference Strauss-Ayali, Jaffe, Burshtain, Gonen and Baneth2004).
HRM real-time PCR analysis
Screening for Bartonella spp. DNA was performed by HRM real-time PCR analysis targeting the 16S–23S internal transcribed spacer (ITS), following procedures and protocols described earlier (Gutiérrez et al. Reference Gutiérrez, Morick, Gross, Winkler, Abdeen and Harrus2013). In brief, an approximately 190 bp fragment was amplified using primers 321 s and H493 as, described elsewhere (Maggi and Breitschwerdt, Reference Maggi and Breitschwerdt2005). The ITS real-time PCR reactions were carried out in a 20 µL final volumes containing 1 µL of 10 µ m solution of each primer, 0·6 µL of 50 µ m Syto9 solution (Invitrogen, CA, USA), 5·4 µL of UPW, 10 µL of MAXIMA Hot-Start PCR Master Mix 2X (Thermo Scientific, Surrey, UK) and 2 µL of each genomic DNA. A Bartonella-positive DNA (Bartonella henselae and Bartonella sp. FG 4-1 strains), a Bartonella-negative DNA and a non-template DNA (NTC) were used as controls in each run. All reactions carried out using the rotor gene 6000 cycler (Corbett Research, Sydney, Australia).
All samples positive for the ITS Bartonella-DNA were later screened for other loci by targeting partial fragments of the transfer-mRNA (ssrA), RNA polymerase β-subunit (rpoB) and citrate synthase (gltA) genes by HRM real-time PCR assays. Accordingly, an approximately 300 bp fragment of the ssrA locus was amplified using ssrAF and ssrAR primers, as previously described (Diaz et al. Reference Diaz, Bai, Malania, Winchell and Kosoy2012), and following reaction protocols described earlier (Gutiérrez et al. Reference Gutiérrez, Cohen, Morick, Mumcuoglu, Harrus and Gottlieb2014a ). An approximately 200 bp rpoB gene fragment was amplified using primers 600f and 800r, according to previously published conditions and reagent volumes (Morick et al. Reference Morick, Baneth, Avidor, Kosoy, Mumcuoglu, Mintz, Eyal, Goethe, Mietze, Shpigel and Harrus2009). Finally, an approximately 340 bp gltA gene fragment was amplified using primers 443F (Birtles and Raoult, Reference Birtles and Raoult1996) and 781R, according to previously published conditions (Sofer et al. Reference Sofer, Gutiérrez, Morick, Mumcuoglu and Harrus2015).
Bartonella isolation from DNA-positive EDTA blood samples
Culture isolation from EDTA blood samples of hosts with Bartonella-positive DNA was attempted (when adequate blood samples were available). Accordingly, the samples were diluted 1:2 in Schneider's Insect supplemented with 10% fetal bovine serum (Biological Industries, Israel), 5% sucrose (Riess et al. Reference Riess, Dietrich, Schmidt, Kaiser, Schwarz, Schafer and Kempf2008), and 2·0 µg mL−1 amphotericin B, to enhance the Bartonella isolation and reduce fungal contamination, respectively, as previously recommended (Kosoy et al. Reference Kosoy, Regnery, Tzianabos, Marston, Jones, Green, Maupin, Olson and Childs1997). Briefly, 100 µL of each diluted sample was directly plated on chocolate agar and incubated at 37 °C with 5% CO2 atmosphere for up to 8 weeks. In addition, a pre-enrichment of the samples in liquid medium was attempted to increase the chances of Bartonella isolation. Thus, 200 µL of the sample solution described above was inoculated in 5 mL fresh Schneider's Insect Medium with additives (as mentioned above) and incubated at 37 °C with constant agitation (100 rpm) for 10 days. After the incubation, the tubes were centrifuged at 3000 rpm for 10 min and the supernatant was discarded. The cellular pellet was washed with 1 mL sterile PBS, twice. Finally, a bacteriological loop was used to inoculate the pellet in new chocolate agar and incubated at 37 °C with 5% CO2 atmosphere for 8 weeks. Any small, round and Bartonella-like colony was re-isolated in a new chocolate agar plate. DNA was extracted from the colonies in 200 µL of PBS by a thermal protocol (i.e. 95 °C for 12 min). DNA was collected from the supernatant after centrifugation at 4 °C at 8500 rpm for 5 min.
Sequencing
All positive PCR products were purified and cleaned by NEB Exo-SAP PCR purification kit (New England Biolabs, Inc., Ipswich, MA, USA) and subsequently sequenced with sense and antisense primers using BigDye Terminator cycle sequencing chemistry from Applied Biosystems ABI PRISM 3730xl DNA Analyser and the ABI's Data collection and Sequence Analysis software (ABI, Carlsbad, CA, USA). Further analyses were done by MEGA 5 (Tamura et al. Reference Tamura, Peterson, Peterson, Stecher, Nei and Kumar2011). Then, the clean sequences were identified using BLASTn against the GenBank database (http://www.ncbi.nlm.nih.gov). All sequences with a length ⩾200 bp were deposited in the GenBank database. All sequences detected were classified in clones according to the host source and/or to similar characterized Bartonella spp. A numeric code was assigned to distinguished clone sequences with identities <99%. In addition, clones with similarities between 99 and 100% were identified with the same number and differentiated with additional alphabetic characters. Co-infected hosts were detected when different sequences were obtained from the different tissues of the same animal (e.g. blood and spleen samples) or by the isolation of different Bartonella strains from the same host.
RESULTS
Bartonella-DNA detection
Bartonella-DNA was detected in a total of 58 samples (spleen and/or blood) obtained from 46 animals. The infection rates and the geographical distribution of the positive animal hosts are shown in Table 1 and Fig. 1, respectively. Table 2 shows the characterization of the Bartonella sequences detected and obtained from the animals. Further information on the positive animals and the details of the sequence clones detected in this study are included as online supplementary material (Tables S1 and S2).
Fig. 1. Maps of Israel and the Palestinian Authority indicating the geographical localization of the collected samples according their Bartonella-infection status and host species. (A) Total distribution of Bartonella-positive animals in all the collected sites; (B) Distribution of B. rochalimae and other Bartonella spp. among positive hosts; (C–E) Distribution Bartonella-infection according to the different host species collected in the study.
Table 1. Bartonella infection rates of wild animals from Israel
Table 2. Characterization of Bartonella species and strains detected in wild animals from Israel
a Characterization of the loci obtained from the samples is found in the online supplemental material.
b Co-infection with different variants detected in the blood and spleen samples.
Canis aureus (Golden jackal): Nine golden jackals were found positive for Bartonella-DNA. Bartonella rochalimae-DNA sequences (1–3 loci detected per sample, 100% identical to B. rochalimae sequences) were detected in 55% (5/9) of the positive jackals (Table 2, Fig. 2). Three other animals (33·3%; 3/9) contained DNA sequences closely related to Bartonella sp. HMD strains, recently re-named as Candidatus Bartonella merieuxii (Chomel et al. Reference Chomel, McMillan-Cole, Kasten, Stuckey, Sato, Maruyama, Diniz and Breitschwerdt2012), which were clustered phylogenetically with ruminant-associated Bartonella spp. (Fig. 2). The ITS amplicons from these samples were 100% identical to the Bartonella HMD clones (accession number FJ177635.5 and EF614393). In addition, the ssrA sequences were closely related to the ruminant bartonellae, B. bovis and B. chomelii (97 and 96% identity; KF218228 and KM215712 GenBank accession numbers, respectively). Lastly, one jackal harboured sequences of mixed-origin, with an ITS sequence 100% identical to B. vinsonii subsp. berkhoffii (HQ185695.2) and an ssrA sequence identical to the ruminant-like clone described above.
Fig. 2. Maximum-likelihood phylogenetic tree based on the partial ITS locus sequences (~200 bp). Phylogenetic tree was constructed using the MEGA software version 5. Bootstrap replicates were performed to estimate the node reliability, and values were obtained from 1000 randomly selected samples of the aligned sequence data. Bootstrap values higher than 50% are indicated. The host sources of the sequences obtained in this study are indicated in parentheses, highlighted with light grey and marked with an asterisk (*). GenBank reference sequences were used with their accession numbers in parentheses.
Vulpes vulpes (Red fox): Two red fox samples were detected positive for Bartonella-DNA, one harbored DNA sequences (ITS and ssrA loci) 100% identical to B. rochalimae and the other was positive for a Bartonella sp. closely related to the HMD strains/Candidatus B. merieuxii (Table 2, Fig. 2). The latter sample contained ITS sequences closely related to the Candidatus B. merieuxii detected in the golden jackals, but with a deletion-gap of 22 nucleotides, and the ssrA amplicon was 100% identical to the clone detected in the golden jackals (Table 2).
Procavia capensis (Rock hyrax): Three hyraxes were positive for at least two Bartonella-DNA loci. All sequences obtained (ITS, rpoB and ssrA loci) were 100% identical to B. rochalimae (Table 2, Fig. 2). All positive animals were collected from the same geographical area (Table 1) and were positive only in their spleen samples.
Erinaceus concolor (Southern white-breasted hedgehog): A single positive blood sample was detected in a hedgehog. The ITS sequence obtained was distant from any previously described Bartonella species or strain, with 91% identity to Bartonella strain JB-15 (GenBank accession number AB674235.1). This sequence clustered in one ITS clade together with Bartonella clarridgeiae and B. rochalimae (Fig. 2). No other locus could be amplified from this blood sample.
Acomys cahirinus (Cairo spiny mouse): Five Cairo spiny mice were found positive for Bartonella-ITS. All ITS sequences obtained were closely related to Bartonella strains previously detected in this host species. Two mice carried ITS sequences 100% identical to Bartonella acomydis (GenBank accession number AB602564.1; Fig. 2). One of these mice was found positive for a gltA fragment (Table 2; deposited in GenBank under accession number KU316220). However, no gltA sequences from B. acomydis have been deposited in the GenBank database to date, thus the association of the obtained gltA sequence with B. acomydis could not be confirmed. The other three positive animals carried Bartonella genotypes closely related to an uncultured clone previously detected in A. cahirinus mice (GenBank accession number FJ686049.1).
Meriones tristrami (Tristram's jird): Tristram's jird was the animal species with the highest Bartonella prevalence in this study (Table 1). The sequenced loci revealed a great diversity of infecting genotypes within the two M. tristrami populations (Fig. 1), including three clones of ITS, six of rpoB, five of gltA and five of the ssrA fragments. From the 25 positive Tristram's jirds detected, sequence clones were closely related to bartonellae previously detected from rodents and/or their fleas, as well as B. rochalimae-like sequences (Table 2, Fig. 2). Moreover, three jirds were confirmed to be co-infected with different Bartonella genotypes, since different Bartonella sequences were detected in their blood and spleens. Four additional jirds contained loci from different Bartonella origin, also suggesting co-infections (online Table S1).
Mus musculus (house mouse): One house mouse was found positive for Bartonella-DNA. The ITS amplicon obtained was only distantly related to all known Bartonella species or strains, with Bartonella coopersplainsensis (GenBank accession number HQ444157.1) and Bartonella japonica (GenBank accession number AB498007.2) being the closest matches, with 91 and 90% sequence identities, respectively (Fig. 2).
Microtus socialis (social vole): No Bartonella-DNA was detected in any of the social voles screened.
Hystrix indica (Indian crested porcupine): No Bartonella-DNA was detected in any of the Indian crested porcupines screened.
Bartonella culture isolation
Bartonella isolation was attempted from the blood of 18 of 46 PCR-positive animals, including seven golden jackals, one red fox, three positive hyraxes, six Tristram's jirds and one southern white-breasted hedgehog. Six individual colonies per positive sample were characterized by conventional PCRs and sequencing targeting the same loci used for the molecular detection of all samples. Three Bartonella strains were successfully isolated from only two Tristram's jirds (Mt-2290 and Mt-2286, Fig. 3). All isolated colonies from rodent Mt-2290 were genetically identical (e.g. same ITS, ssrA, gltA and rpoB sequences). Nevertheless, the gltA sequence detected in the blood extracted DNA (B. rochalimae-like clone, online Table S1) was different from the gltA fragment detected in the isolated colonies, confirming the circulation of various genotypes in this jird. Similarly, two different Bartonella strains were isolated from rodent Mt-2286 (Mt-2286.1 and Mt-2286.3). The phylogenetic analyses of those isolates demonstrated that the two strains represent distantly related spp., with an identity of 84·5% (Fig. 3). Moreover, the gltA and rpoB DNA sequences detected from the blood were different from those detected in the isolated strains, demonstrating the presence of at least a third co-infecting Bartonella genotype in this host (online Tables S1). The isolated strains Mt-2286.1 and Mt-2290 were closely related with 97·7% sequence identity, demonstrating that similar but not identical genotypes circulate among the same host species (Fig. 3). No Bartonella isolates were obtained by the pre-enrichment liquid medium. On the other hand, the overgrowth of other co-infecting bacterial genera in the sub-cultures was a common finding.
Fig. 3. Maximum-likelihood phylogenetic tree based on the concatenation of four loci (ITS, gltA, rpoB and ssrA) representing ~2400 bp. Phylogenetic tree was constructed using the MEGA software version 4. Bootstrap replicates were performed to estimate the node reliability, and values were obtained from 1000 randomly selected samples of the aligned sequence data. Bootstrap values higher than 50% are indicated. Strains isolated from M. tristrami jirds in this study are highlight with light grey. Reference sequences were obtained from whole-sequenced strains deposited in GenBank database.
Nucleotide sequence accession numbers
Newly identified sequences, longer than 200 bp, obtained directly from the samples were deposited in GenBank database under ITS accession numbers: KU316206-KU316209; ssrA: KU316210-KU316219; and gltA: KU316220- KU316225.
Sequences (ITS, rpoB, gltA and ssrA) from isolated strains from M. tristrami jirds were deposited under the accession numbers: Mt-2290 strain: KU316226- KU316229; Mt-2286.1 strain: KU316230, KU316232, KU316234, KU316236; and Mt-2286.3 strain: KU316231, KU316233, KU316235 and KU316237.
DISCUSSION
This study reports the detection of Bartonella species in wild animals from Israel. DNA sequences of several recognized Bartonella spp., such as B. rochalimae, B. vinsonii subsp. berkhoffii and B. acomydis, Candidatus Bartonella species and several uncharacterized genotypes were detected among seven different animal species. Notably, the widespread distribution of the zoonotic B. rochalimae, and closely related strains, among different host species and across all the sampling areas is highlighted. This study reports infection with Bartonella for the first time in three animal species: the rock hyrax, Tristam's jird and the southern hedgehog. Moreover, the great diversity of Bartonella genotypes in Tristram's jirds is emphasized.
Infections with Bartonella spp. in wild canids have been reported worldwide (Henn et al. Reference Henn, Gabriel, Kasten, Brown, Theis, Foley and Chomel2007, Reference Henn, Chomel, Boulouis, Kasten, Murray, Bar-Gal, King, Courreau and Baneth2009a , Schaefer et al. Reference Schaefer, Kasten, Coonan, Clifford and Chomel2011; Chomel et al. Reference Chomel, McMillan-Cole, Kasten, Stuckey, Sato, Maruyama, Diniz and Breitschwerdt2012). In this study, DNA sequences from three Bartonella spp. were detected in golden jackals (C. aureus) and in red foxes (V. vulpes), including the zoonotic B. rochalimae (in both canids), B. vinsonii subsp. berkhoffii-like organism (in a golden jackal) and Bartonella clones closely related to the proposed Candidatus B. merieuxii (in both canids) (Chomel et al. Reference Chomel, McMillan-Cole, Kasten, Stuckey, Sato, Maruyama, Diniz and Breitschwerdt2012). To the best of our knowledge, this study represents the first description of B. rochalimae in golden jackals. This zoonotic Bartonella species has been isolated from domestic dogs (Canis lupus familiaris), coyotes (C. latrans), wolfs (Canis lupus), island foxes (Urocyon littoralis), grey foxes (U. cinereoargenteus), red foxes (V. vulpes) and raccoons (P. lotor) (Henn et al. Reference Henn, Gabriel, Kasten, Brown, Theis, Foley and Chomel2007, Reference Henn, Chomel, Boulouis, Kasten, Murray, Bar-Gal, King, Courreau and Baneth2009a , Schaefer et al. Reference Schaefer, Kasten, Coonan, Clifford and Chomel2011; Gerrikagoitia et al. Reference Gerrikagoitia, Gil, Garcia-Esteban, Anda, Juste and Barral2012). Thus, our results expand the list of potential reservoirs of B. rochalimae in wild carnivores. Moreover, the detection of clones closely related to the Bartonella sp. HMD strains (Candidatus B. merieuxii) in both canids is notable. These strains were first detected from dogs and Rhipicephalus sanguineus ticks collected in Italy and Greece (Diniz et al. Reference Diniz, Billeter, Otranto, De Caprariis, Petanides, Mylonakis, Koutinas and Breitschwerdt2009). In the latter study, the sequences obtained (i.e. ITS locus, the 16S and the rpoB genes) were closely related to ruminant-associated bartonellae. Interestingly, the ssrA fragments detected from the canids in our study were also closely related to ruminant bartonellae, suggesting that this Bartonella locus belongs to the same species. In addition, as reported previously, we detected differences between the ITS amplicons from the golden jackal (100% identical to those reported from HMD strains) and the red fox (92% identical due to a deletion-gap of 22 nucleotides). Chomel and others (2012), compared the Bartonella strain detected from Iraqi dogs and jackals with the original HMD strains, and concluded that they were the same Bartonella species, and therefore renamed them as Candidatus B. merieuxii. Furthermore, another study reported the infection of dogs from Sri Lanka with the HMD strain (Brenner et al. Reference Brenner, Chomel, Singhasivanon, Namekata, Kasten, Kass, Cortes-Vecino, Gennari, Rajapakse, Huong and Dubey2013). In summary, the detection of closely related strains from the jackals and a fox in this study reflects the widespread distribution of this newly canid Candidatus Bartonella species. Finally, one golden jackal was found to carry a genotype closely related to two different Bartonella species. The ITS amplicon was 100% similar to B. vinsonii subsp. berkhoffii and the ssrA fragment was 100% identical to the one detected in the other jackals, suggesting a co-infection with two different Bartonella species in this animal. The former finding is surprising, since B. vinsonii subsp. berkhoffii has not been found in Israel, neither in a dog or human or any other animal, to date. However, the short length of the ITS fragment sequenced (213 bp) and the different Bartonella-origin of the ssrA fragment, prevent confirming that this animal was infected with B. vinsonii subsp. berkhoffii and not with a closely related genotype. Interestingly, Chomel et al. (Reference Chomel, McMillan-Cole, Kasten, Stuckey, Sato, Maruyama, Diniz and Breitschwerdt2012) detected this Bartonella sp. in jackals from Iraq. Hence, the capability of golden jackals to be reservoirs of B. vinsonii subsp. berkhoffii or closely related strains need to be further evaluated in future studies.
In the present study B. rochalimae DNA was detected from rock hyraxes (P. capensis). All samples were positive for at least two genomic loci. To the best of our knowledge, this represents the first report of Bartonella infection in rock hyraxes and adds one non-carnivore animal species as a potential reservoir for B. rochalimae. The positive animals were trapped in the same geographical area and only their spleen samples were positive, suggesting that these bacteria are harboured in the spleen in greater concentrations than the blood, allowing their molecular detection. It should be noted that in a recent study conducted in the Palestinian Authority, the authors screened Bartonella-DNA from fleas collected from various animals including hyraxes, and no Bartonella-DNA was detected in the hyraxes associated-fleas (Nasereddin et al. Reference Nasereddin, Risheq, Harrus, Azmi, Ereqat, Baneth, Salant, Mumcuoglu and Abdeen2014), suggesting a low exposure to Bartonella from these hosts in these regions. Hyraxes serve as a reservoir host for Leishmania tropica, transmitted by sand flies Phlebotomus sergenti (Jaffe et al. Reference Jaffe, Baneth, Abdeen, Schlein and Warburg2004) and Phlebotomus arabicus (Svobodova et al. Reference Svobodova, Votypka, Peckova, Dvorak, Nasereddin, Baneth, Sztern, Kravchenko, Orr, Meir, Schnur, Volf and Warburg2006). Our samples were derived from epidemiological and diagnostic studies on leishmaniasis, and none of the Bartonella-infected hyraxes were infected with Leishmania (data not shown). Thus, the potential role of rock hyraxes as reservoirs for Bartonella spp. and the potential vectors involved in their transmission require further investigation.
Rodents have been considered to be important reservoirs of Bartonella spp. (Buffet et al. Reference Buffet, Kosoy and Vayssier-Taussat2013; Gutiérrez et al. Reference Gutiérrez, Krasnov, Morick, Gottlieb, Khokhlova and Harrus2015a ). Wild rodents and their associated fleas, collected from suburban areas and the Negev desert of Israel, have been widely reported as Bartonella carriers (Morick et al. Reference Morick, Baneth, Avidor, Kosoy, Mumcuoglu, Mintz, Eyal, Goethe, Mietze, Shpigel and Harrus2009, Reference Morick, Krasnov, Khokhlova, Shenbrot, Kosoy and Harrus2010, Reference Morick, Krasnov, Khokhlova, Gottlieb and Harrus2011; Gutiérrez et al. Reference Gutiérrez, Morick, Cohen, Hawlena and Harrus2014b ). In this study, five wild rodent species were screened for Bartonella infection, including two species whose Bartonella infection status was unexplored previously in Israel. The highest prevalence of Bartonella infection was detected in Tristram's jirds (M. tristrami) (58% of the total animals sampled). Tristram's jirds carried Bartonella genotypes closely related to previously reported bartonellae genotypes and to B. rochalimae-like organisms. Bartonella rochalimae-like loci amplified from these jirds were 98–100% similar to B. rochalimae ATCC BAA-1498 and to the Bartonella sp. 1-1C, a B. rochalimae-like organism isolated from Rattus norvegicus from Taiwan (Lin et al. Reference Lin, Chen, Chen, Chomel and Chang2008). Isolation of Bartonella spp. was obtained from two out of six animals tested. These isolations confirmed a co-infection status of the samples with different Bartonella spp. and genotypes. Additionally, the detection of up to three different genotypes, including two distantly related Bartonella spp., emphasizes the complex infection composition that these jirds harboured. A previous study performed by our group, detected Bartonella-DNA in flea pools collected from this jird species, and identified several genotypes based on the gltA and rpoB genes (Morick et al. Reference Morick, Krasnov, Khokhlova, Shenbrot, Kosoy and Harrus2010). However, in the latter study, no mammalian samples were collected. In the present study, two jird populations from two geographically distant regions were screened. Interestingly, both populations contained multiple Bartonella genotypes, showing a remarkable Bartonella diversity among these jirds. Thus, the present study confirmed the role of these jirds as carriers and potential reservoirs of Bartonella strains.
Cairo spiny mice (A. cahirinus) were positive for B. acomydis-DNA and two other uncharacterized Bartonella genotypes. This is the first detection of B. acomydis in this mouse species. This Bartonella species was previously isolated and described from the golden spiny mouse, Acomys russatus (Sato et al. Reference Sato, Kabeya, Fujinaga, Inoue, Une, Yoshikawa and Maruyama2013). The other two Bartonella genotypes were closely related to a previously identified clone detected in A. cahirinus collected from Israel (Morick et al. Reference Morick, Baneth, Avidor, Kosoy, Mumcuoglu, Mintz, Eyal, Goethe, Mietze, Shpigel and Harrus2009).
One house mouse (M. musculus) was detected positive for an uncharacterized Bartonella genotype. The genotype was distantly related (~91% sequence identity) to B. coopersplainsensis and B. japonica, which were first isolated from Australian rats (Gundi et al. Reference Gundi, Taylor, Raoult and La Scola2009) and Apodemus argenteus from Japan (Inoue et al. Reference Inoue, Kabeya, Shiratori, Ueda, Kosoy, Chomel, Boulouis and Maruyama2010), respectively. A previous study detected Bartonella genotypes in fleas collected from M. musculus from Israel (Morick et al. Reference Morick, Krasnov, Khokhlova, Shenbrot, Kosoy and Harrus2010). The role of M. musculus as Bartonella-reservoir seems apparently minor, since epidemiological studies have reported none or very low infection rates in this rodent species (Holmberg et al. Reference Holmberg, Mills, McGill, Benjamin and Ellis2003; Morick et al. Reference Morick, Baneth, Avidor, Kosoy, Mumcuoglu, Mintz, Eyal, Goethe, Mietze, Shpigel and Harrus2009).
Another interesting finding in this study was the first detection of Bartonella-DNA in a Southern white-breasted hedgehog. The sequence detected from this host showed low similarity with other Bartonella spp., being the closest match to a genotype detected from a badger (Meles anakuma) in Japan (Sato et al. Reference Sato, Kabeya, Miura, Suzuki, Bai, Kosoy, Sentsui, Kariwa and Maruyama2012). Both genotypes appear to be different from other known Bartonella species and cluster together. However, only the ITS fragment was amplified from this sample. This could be attributed to the possibility that as a newly uncharacterized genotype, the primers used were unable to amplify other genetic loci. Another possibility is that the bacterial loads were below the detection level of the assays used, however could be amplified by the ITS only, which is known to be a double-copy locus and thus more sensitive than the other targets. This argument applies for other samples that could be amplified only when this locus was targeted.
No Bartonella infection was identified in M. socialis (social vole) or H. indica (Indian crested porcupine), in this study. In a previous study, social voles trapped from suburban areas from Israel also tested negative for Bartonella-DNA (Morick et al. Reference Morick, Baneth, Avidor, Kosoy, Mumcuoglu, Mintz, Eyal, Goethe, Mietze, Shpigel and Harrus2009). Thus, the ecological factors that may limit Bartonella infection of M. socialis, in areas where other reservoirs co-habit, need to be further explored.
In conclusion, this study reports the identification and genetic characterization of several Bartonella species and genotypes in wildlife from Israel. Several zoonotic Bartonella spp., especially the widespread distribution of B. rochalimae and B. rochalimae-like bacteria across wildlife hosts, deserve special attention. Our study indicates that infection with zoonotic and other Bartonella species is likely to be widely prevalent among wild animals and stresses their potential threat to public health. It also suggests further exploration of Bartonella transmission and its relationship to infection of humans, domestic and wild animals.
SUPPLEMENTARY MATERIAL
The supplementary material for this article can be found at http://dx.doi.org/10.1017/S0031182016000603
ACKNOWLEDGEMENTS
We thank Alicia Rojas and Lia Prokopiev for their assistance in the processing of animal samples.
FINANCIAL SUPPORT
This study was supported by the Israel Science Foundation (grant number 30/11 to Shimon Harrus).
CONFLICT OF INTEREST
None.
