ORIGINAL ARTICLES

Bartonella dromedarii sp. nov. Isolated from DomesticatedCamels (Camelus dromedarius) in Israel

Michal Rasis,1 Nir Rudoler,2 David Schwartz,3 and Michael Giladi1,4,5

Abstract

Bartonella spp. are fastidious, Gram-negative bacilli that cause a wide spectrum of diseases in humans. MostBartonella spp. have adapted to a specific host, generally a domestic or wild mammal. Dromedary camels(Camelus dromedarius) have become a focus of growing public-health interest because they have been iden-tified as a reservoir host for the Middle East respiratory syndrome coronavirus. Nevertheless, data on camelzoonoses are limited. We aimed to study the occurrence of Bartonella bacteremia among dromedaries in Israel.Nine of 51 (17.6%) camels were found to be bacteremic with Bartonella spp.; bacteremia levels ranged fromfive to > 1000 colony-forming units/mL. Phylogenetic reconstruction based on the concatenated sequences ofgltA and rpoB genes demonstrated that the dromedary Bartonella isolates are closely related to other ruminant-derived Bartonella spp., with B. bovis being the nearest relative. Using electron microscopy, the novel isolateswere shown to be flagellated, whereas B. bovis is nonflagellated. Sequence comparisons analysis of thehousekeeping genes ftsZ, ribC, and groEL showed the highest homology to B. chomelii, B. capreoli, andB. birtlesii, respectively. Sequence analysis of the gltA and rpoB revealed *96% identity to B. bovis, a previouslysuggested cutoff value for sequence-based differentiation of Bartonella spp., suggesting that this approach does nothave sufficient discriminatory power for differentiating ruminant-related Bartonella spp. A comprehensive mul-tilocus sequence typing (MLST) analysis based on nine genetic loci (gltA, rpoB, ftsZ, internal transcribed spacer(ITS), 16S rRNA, ribC, groEL, nuoG, and SsrA) identified seven sequence types of the new dromedary isolates.This is the first description of a Bartonella sp. from camelids. On the basis of a distinct reservoir and ecologicalniche, sequence analyses, and expression of flagella, we designate these isolates as a novel Bartonella sp. namedBartonella dromedarii sp. nov. Further studies are required to explore its zoonotic potential.

Key Words: Bartonella infection—Bartonella dromedarii—Dromedary camels.

Introduction

Bartonella are fastidious, Gram-negative bacteriathat infect and persist in mammalian erythrocytes and

endothelial cells and can be found in a broad range of do-mestic and wild mammalian hosts (Breitschwerdt and Kor-dick 2000, Greub and Raoult 2002, Jacomo et al. 2002). Todate, the genus Bartonella constitutes over 30 different spe-cies and subspecies, the majority of which exhibit restrictedhost specificity. An increasing number of Bartonella speciesare being recognized as important zoonotic pathogens, re-

sponsible for a growing spectrum of emerging diseases ( Ja-como et al. 2002, Saisongkorh et al. 2009, Vayssier-Taussatet al. 2009, Giladi and Ephros 2011). Hematophagous ar-thropod vectors (e.g., fleas, biting flies, lice, mites, and ticks)have been found to be naturally infected and are frequentlyimplicated in Bartonella spp. transmission (Higgins et al.1996, Jacomo et al. 2002, Kim et al. 2005, Boutellis et al.2012, Kumsa et al. 2014).

Several Bartonella spp. have been described in cattle andwild ruminants, such as roe deer, mule deer, elk, and waterbuffalo (Chang et al. 2000, Dehio et al. 2001, Bermond et al.

1The Bernard Pridan Laboratory for Molecular Biology of Infectious Diseases, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel.2Koret School of Veterinary Medicine, Robert H. Smith Faculty of Agriculture, Food and Environment, The Hebrew University of

Jerusalem, Rehovot, Israel.3Clinical Microbiology Laboratory, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel.4The Infectious Disease Unit, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel.5Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.

VECTOR-BORNE AND ZOONOTIC DISEASESVolume 14, Number 11, 2014ª Mary Ann Liebert, Inc.DOI: 10.1089/vbz.2014.1663

775

2002, Bai et al. 2011, Bai et al. 2013). Cattle comprise themain reservoir of B. bovis and B. chomelii (Bermond et al.2002, Maillard et al. 2004), whereas B. schoenbuchensis andB. capreoli were first described in roe deer (Dehio et al. 2001,Bermond et al. 2002). Additionally, B. melophagi was firstisolated from commercial sources of sheep blood, and itsDNA was detected in sheep ked (Bemis and Kania 2007,Kumsa et al. 2014). The clinical relevance of these species tohumans is unclear, although Candidatus Bartonella melo-phagi was isolated from the blood of two female patients(Maggi et al. 2009). Bartonella infection in ruminants isusually subclinical. However, in a few cases, B. bovis wasidentified as a cause of prolonged bacteremia and bovineendocarditis (Maillard et al. 2007, Erol et al. 2013).

Dromedary camels (Camelus dromedarius) have a closeassociation with humans; in addition to being employed fortransportation of people and goods in some camel-rearingcountries and a valuable resource for milk, meat, and wool,camels are also recreational animals used for camel racingand tourism. Camels have been uncommonly implicated as asource for zoonoses. In fact, apart from camel brucellosis,camel-to-human transmission of zoonotic agents is either rareor unsubstantiated. Nonetheless, camels have been proposedas the reservoir for the newly described Middle East respi-ratory syndrome coronavirus (Alagaili et al. 2014, Haagmanset al. 2014). Given the apparent growing scientific and publichealth interest in camels as potential zoonotic hosts under a‘‘One Health’’ approach, we decided to examine the preva-lence of Bartonella spp. in domestic dromedary camels fromIsrael. In this study, we describe the first known Bartonellasp. from camelids. Their potential role as reservoir for humanbartonellosis warrants further investigation.

Materials and Methods

Animals and specimen collection

During May of 2013, a total of 51 dromedary camels fromthree camel farms located in central and southern Israel werestudied (herd A, n = 9; herd B, n = 15 and herd C, n = 27).Dromedary blood samples (2–3 mL of whole blood inEDTA-coated tubes) were obtained by jugular vein punctureaccording to standard veterinary procedures by trained per-sonnel. Samples were kept at - 80�C until processed. Data ongender, age group ( < 10 years, 10–20 years, > 20 years old),and the presence of ectoparasites were recorded for eachanimal. This study was approved ethically by the Tel AvivSourasky Medical Center Institutional Animal Care and UseCommittee (study no. 18-6-13).

Culture conditions and Bartonella spp. isolation

Dromedary blood samples were thawed, plated in dupli-cates on fresh chocolate agar plates (Novamed, Jerusalem,Israel), and incubated at 37�C in a humid 5% CO2 atmospherefor 8 weeks or until growth. Plates were inspected twiceweekly for growth. Putative identification of Bartonella sp.was based on Gram stain and morphological examination ofthe colonies. Quantitative evaluation of the number of colo-nies and their morphology were recorded. Bacterial stockswere prepared in brain heart infusion broth (Novamed, Jer-usalem, Israel) and 15% glycerol vol/vol (Sigma Aldrich,Rehovot, Israel) and kept at - 80�C.

DNA extraction

Bacterial colonies were scraped off chocolate agar plates,washed once, and resuspended in 200 lL of sterile 1 ·phosphate-buffered saline (PBS; Sigma Aldrich, Rehovot,Israel). Genomic DNA was extracted using the QIAampDNA Mini Kit (QIAGEN, Valencia, CA) according to themanufacturer’s instructions.

PCR and sequencing

Conventional PCR was performed with Q5� or Phusion�

High-Fidelity DNA Polymerase (New England Biolabs,Hitchin, UK) according to the manufacturer’s protocol. Theprimers (Sigma Aldrich, Rehovot, Israel) used in this studyare listed in Table 1. One or 2 lL of genomic DNA prepa-rations of Bartonella isolates were used as template. PCRproducts were analyzed by gel electrophoresis and purifiedwith Illustra� ExoProStar 1-Step Kit (GE Healthcare,Buckinghamshire, UK). Sequencing was performed in theDNA Sequencing Unit at the G.S. Wise Faculty of LifeSciences, Tel Aviv University. Nucleotide sequences weresubmitted to a BLAST search in the National Center forBiotechnology Information (NCBI) GenBank database withdefault parameters, while excluding uncultured or environ-mental sample sequences.

Phylogenetic reconstruction

Phylogenetic reconstruction was conducted using MEGAsoftware version 5.1 (Tamura et al. 2011). Alignments wereperformed with the MUSCLE algorithm (Edgar 2004) withdefault parameters. Phylogenetic relations were inferred us-ing the neighbor-joining method (Saitou and Nei 1987). Theevolutionary distances were computed using the Jukes–Cantor model ( Jukes and Cantor 1969). The partial gltA andrpoB sequences of the novel dromedary isolates were in-cluded in this analysis, along with the respective nucleotidesequences of valid Bartonella species and Candidatus B.melophagi. Bootstrap analysis (Felsenstein 1985) was carriedout on 1000 replications of the dataset.

Biochemical phenotypes

The enzymatic profiles of Bartonella isolates were ob-tained using Vitek�2 (bioMerieux, Durham, NC), an auto-mated bacterial identification system using GN ID and NH IDcards (product numbers 21341 and 21346, respectively).Oxidase and catalase tests were performed using standardbacteriological procedures.

Electron microscopy

Bacteria were submitted to electron microscopy, as de-scribed elsewhere (Bermond et al. 2002), with minor modifi-cations. Briefly, bacteria were scraped from chocolate agarplates, suspended in sterile 1 · PBS (Bio-Lab, Jerusalem, Is-rael), and adsorbed to Formvar-coated copper grids. Gridswere then stained with 1% (wt/vol) uranyl acetate and airdried. Negative-stained micrographs of cell morphology wereanalyzed using a transmission electron microscope (TEM)Tecnai 12 (Phillips, Eindhoven, The Netherlands) at 100 kVequipped with MegaView II CCD camera and Analysis� v. 3.0software (Soft Imaging System GmbH, Munstar, Germany).

776 RASIS ET AL.

Multilocus sequence typing

Sequence data of dromedary Bartonella isolates was ob-tained for nine genetic loci listed in Table 1, similarly to themultilocus sequence typing (MLST) scheme previously de-fined for B. bovis (Bai et al. 2013). For determination ofallelic variations, nucleotide sequences were aligned usingClustal Omega multiple sequence alignment program withdefault parameters (www.ebi.ac.uk/Tools/msa/clustalo).

Nucleotide sequence accession numbers

Nucleotide sequences were deposited in the NCBI Gen-Bank database with the accession numbers listed in Table 2.

Results

Prevalence of Bartonella sp. bacteremia

Fifty-one animals were included in this study; 41% were< 10 years old, 53% 10–20 years old, and 6% > 20 years old.Thirty-three (65%) of the 51 camels were females. Nine of 51(18%) camels sampled were found to be bacteremic withBartonella sp., with concentrations ranging from five to> 1000 colony-forming units (CFU)/mL (median = 63).Round, small (3–5 mm in diameter), homogeneous, opaque,grey-white bacterial colonies appeared after 14.5 – 1.6 daysof incubation. Subcultures growth time was approximately 7days under the same conditions.

Molecular identification and phylogenetic analysis

Sequence comparison analysis of the partial gltA and rpoBsequences of the dromedary Bartonella isolates demonstrated*96% identity to B. bovis for each of the amplified se-quences, contrary to 100% identity demonstrated for a re-cently characterized bovine B. bovis isolate from Israel(Rudoler et al. 2014), using the same targets. Given theconsiderable difference (4%) between the gltA and rpoBsequences of the dromedary isolates and B. bovis and theresulting homologies, which are at the low range of theidentity values proposed for species differentiation (La Scolaet al. 2003), we decided to expand our analysis. Sequencecomparison analysis of the housekeeping genes ftsZ, ribC,and groEL showed the highest homology to B. chomelii,

Ta

ble

1.

Descriptio

no

fa

Mu

ltilo

cu

sS

eq

uen

ce

Ty

pin

g(M

LS

T)

Sch

em

eD

esig

na

ted

fo

rD

ro

med

ary

Ba

rto

nella

Iso

la

tes

Gen

etic

locu

sF

orw

ard

pri

mer

sequen

ce(5

¢to

3¢)

Rev

erse

pri

mer

sequen

ce(5

¢to

3¢)

Len

gth

of

PC

Rpro

duct

(bp)

Len

gth

of

analy

zed

sequen

ce(n

t)

No.

of

vari

able

site

sN

o.

of

all

eles

rpoB

CG

CA

TT

GG

CT

TA

CT

TC

GT

AT

GG

TA

GA

CT

GA

TT

AG

AA

CG

CT

G893

852

23

nuoG

GG

CG

TG

AT

TG

TT

CT

CG

TT

AC

AC

GA

CC

AC

GG

CT

AT

CA

AT

366

328

22

glt

AG

CT

AT

GT

CT

GC

AT

TC

TA

TC

AG

AT

CY

TC

AA

TC

AT

TT

CT

TT

CC

A790

753

01

ftsZ

AT

TA

AT

CT

GC

AY

CG

GC

CA

GA

AC

VG

AD

AC

AC

GA

AT

AA

CA

CC

935

859

43

ribC

TA

AC

CG

AT

AT

TG

GT

TG

TG

TT

GA

AG

TA

AA

GC

TA

GA

AA

GT

CT

GG

CA

AC

AT

AA

CG

588

535

63

Ssr

AG

CT

AT

GG

TA

AT

AA

AT

GG

AC

AA

TG

AA

AT

AA

GC

TT

CT

GT

TG

CC

AG

GT

G301

267

32

gro

EL

GA

AC

TN

GA

AG

AT

AA

GT

TN

GA

AA

AT

CC

AT

TC

CG

CC

CA

TT

C1459

939

12

ITS

CT

TC

AG

AT

GA

TG

AT

CC

CA

AG

CC

TT

CT

GG

CG

GA

AC

CG

AC

GA

CC

CC

CT

GC

TT

GC

AA

AG

332–348

317

01

16S

rRN

AA

GA

GT

TT

GA

TC

MT

GG

CT

CA

GC

GG

YT

AC

CT

TG

TT

AC

GA

CT

T1352

1337

84

Tota

l6187

26

21

ITS

,in

tern

altr

ansc

ribed

spac

er.

Table 2. List of GenBank Accession Numbers

of Sequences Originating from This Study

Geneticlocus

DromedaryBartonella isolates(n = 9; this study)

B. bovis cattleisolate #234

(Rudoleret al. 2014)

rpoB KJ909798-KJ909806 KJ909808nuoG KM371025-KM371033 KM371088gltA KM371034-KM371042 KM371089ftsZ KM371043-KM371051 KM371090ribC KM371052-KM371060 KM371091SsrA KM371061-KM371069 KM371092groEL KM371070-KM371078 KM371093ITS KM371096-KM371104 KM37109416S rRNA KM371079-KM371087 KM371095

ITS, internal transcribed spacer.

FIRST REPORT OF Bartonella INFECTION IN CAMELS 777

B. capreoli, and B. birtlesii, respectively (Table 3). Thesedata demonstrate that the dromedary isolates are not B. bovis.

A phylogram based on the concatenated sequences of gltAand rpoB demonstrated that all dromedary Bartonella iso-lates comprise a distinct monophyletic group with B. bovis(bootstrap test = 81%) (Fig. 1). As expected, additionalruminant-associated Bartonella spp. (B. capreoli, B. scho-enbuchensis, B. chomelii, and Candidatus B. melophagi) wereclustered together in the adjacent sister clade, which suggests acommon evolutionary ancestor to Bartonella spp. that infectruminants (bootstrap test = 100%). This phylogenetic analysisclearly indicates that the dromedary isolates found in this studyare closely related to ruminant-derived Bartonella spp., andB. bovis is the nearest evolutionary relative.

Dromedary Bartonella isolates are flagellated

Electron micrographs of three representative dromedaryisolates (one from each herd) (Fig. 2a–c), as well as a B. bovisisolate from Israeli beef cattle (Rudoler et al. 2014) (Fig. 2d)were obtained. All isolates submitted to TEM examinationwere grown and harvested under the same conditions. Thebacterial cells were found to be rod-shaped and approxima-tely 850 nM in length and 450 nM in width. As expected, adouble-layer membrane characteristic of Gram-negativebacteria was also observed. All three dromedary isolates werefound to carry unipolar flagella in contrast to the bovineB. bovis isolate, which was nonflagellated. A B. bovis strainpreviously isolated from a French cow was also shown to benonflagellated (Bermond et al. 2002), in agreement with ourobservation. Finally, wet-mount slide tests confirmed that the

dromedary Bartonella isolates are motile, in contrast to thenonflagellated B. bovis.

Biochemical phenotypes

All dromedary isolates were initially characterized as ox-idase ( - ), catalase ( - ), Gram-negative bacilli. Using theVitek� 2 system, representative isolates were found to haveamino acid arylamidase activity with the following aminoacids: Arginine, phenylalanine, proline, glycine, tyrosine,and leucine. One of the isolates was also shown to have lysinearylamidase activity; however, they did not have alanine-phenylalanine-proline arylamidase activity. The bacterialacked the ability to hydrolyze mono- or disaccharides (i.e.,glucose, galactose, mannose, trehalose, maltose, sucrose,tagatose, ribose, and xylose), had no urease activity, and wereunable to hydrolyze N-acetyl-b-d-glucosaminide. The iso-lates were also able to hydrolyze phenylphosphonate, andtherefore can use it as a single source of phosphorus under theconditions examined. Taken together, these results demon-strate that the dromedary isolates have peptidase activity andconsequently might utilize amino acids as a carbon or ni-trogen source. Moreover, these fastidious bacteria do not useeven simple carbohydrates as carbon or energy sources underthe conditions examined, as was previously shown forB. bovis isolates (Bermond et al. 2002, Erol et al. 2013).

Multilocus sequence typing

To further characterize the novel dromedary Bartonellaisolates and determine the prevalent sequence types (STs), a

Table 3. Sequence Homologies of Nine Genetic Loci Originating from the Novel Dromedary Bartonella

Isolates Compared to an Israeli Bovine B. bovis Isolate

Geneticlocus

Description orgene product

Query length(nucleotides)

Dromedary Bartonellaisolates (n = 9;

this study)

Bartonella bovis cattleisolate #234 (Rudoler

et al. 2014)

Best score BLASTmatch (GenBank

accession number) % identity

Best scoreBLAST match

(GenBankaccession number) % identity

rpoB RNA polymerase b-subunit 852 B. bovis(KF218224)

95.9–96.0 B. bovis(KF218222)

100.0

nuoG NADH dehydrogenasec-subunit

328 B. bovis(EF659938)

95.4–96.0 B. bovis(EF659938)

99.4

gltA Citrate synthase 327 B. bovis(KF199898)

96.0 B. bovis(KF199897)

100.0

ftsZ Cell division protein 859 B. chomelii( JN646675)

96.5–96.7 B. bovis(KF193410)

100.0

ribC Riboflavin synthase 535 B. capreoli(AB703134)

94.9–95.3 B. bovis(AY116637)

100.0

SsrA SsrA RNA (tmRNA) 267 B. bovis(KF218228)

97.7–98.0 B. bovis(KF218226)

100.0

groEL Heat-shock chaperoninprotein

939 B. birtlesii(AF355773)

97.1–97.2 B. bovis(KF212449)

100.0

ITS 16S–23S rRNA intergenicspacer

317/347 B. bovis(KF218233)

93.0 B. bovis(KF218232)

99.7

16SrRNA

16S ribosomal RNA 1337 B. schoenbuchensis(KJ639882)

B. bovis (NR_025121)

99.6–99.9 B. bovis(NR_025121)

100.0

NADH, nicotinamide adenine dinucleotide; ITS, internal transcribed spacer.

778 RASIS ET AL.

MLST analysis was performed. Twenty-six of 6187 (0.42%)nucleotide positions analyzed were variable, suggesting closegenetic relatedness among the examined isolates. Eachdromedary isolate was characterized by a series of nine in-tegers that correspond to the alleles at the internal fragmentsof nine housekeeping loci, and a total of seven STs wereassigned. Of note, the partial gltA sequences of the drome-dary isolates were identical. ST1 was the prevalent ST, andthe only ST found in more than one camel farm (Table 4).

Tick infestation

High tick infestation rate (80%) was found in herd B, asopposed to the other two herds not infested with ectopara-sites, probably due to prior anti-ectoparasite treatment.

Eighteen ticks, representing one or two ticks from each of the11 camels out of 12 infested animals, were classified by anentomologist as Hyalomma spp. Three ticks were males andthe rest were females at different stages of feeding.

Discussion

In this study, we isolated and characterized a novel Bar-tonella sp. from dromedary camels inhabiting camel farmslocated in the Negev desert and central Israel. To the best ofour knowledge, this is the first report of a Bartonella sp. fromcamels. Bartonella sp. bacteremia was found in all of thecamel herds examined with an overall prevalence of 18%.Bartonella dromedary isolates displayed phenotypic traitscharacteristic of the genus Bartonella: They were isolated

FIG. 1. A phylogenetic tree highlighting the position of Bartonella isolates form Israeli dromedary camels relative toother type strains within the genus Bartonella. Concatenated gltA and rpoB sequences were aligned using MUSCLE (Edgar2004), and phylogenetic interferences were obtained using the neighbor-joining method (Saitou and Nei 1987). The optimaltree with the sum of branch length = 1.99 is shown. GenBank accession numbers are indicated in parentheses as (rpoB,gltA). The tree was rooted using Brucella melitensis as an outgroup. Numbers at the nodes are the percentage of replicatetrees in which the associated taxa clustered together in the bootstrap test (1000 replicates) (Felsenstein 1985). The tree isdrawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetictree. Branches drawn with a dashed line represent the dromedary isolates described in this study. The evolutionary distanceswere computed using the Jukes–Cantor method ( Jukes and Cantor 1969), and the scale bar indicates 10% nucleotidesequence divergence. The analysis involved 42 nucleotide sequences. There were a total of 1093 positions in the finaldataset. All evolutionary analyses were conducted in MEGA5 (Tamura et al. 2011).

FIRST REPORT OF Bartonella INFECTION IN CAMELS 779

from blood, grew slowly (14 days for primary isolation and 7days for subcultures), and produced small, round, homoge-neous, grey-white colonies on chocolate agar plates, withstraight and slender Gram-negative rods. These isolates canbe briefly described as small, aerobic, fastidious, oxidase ( - )and catalase ( - ), Gram-negative bacteria, which are non-fermenters and biochemically inert except for the productionof peptidases. Such enzymatic profiles are characteristic ofthe Bartonella genus (Bermond et al. 2002, Maillard et al.2004); consequently, biochemical profiles cannot be usedroutinely for the differentiation of Bartonella spp.

We initially applied a straightforward and well-knownmolecular approach for the identification of newly encoun-tered Bartonella isolates, as was suggested by La Scola andcolleagues (La Scola et al. 2003). DNA sequence compari-sons of the partial gene fragments of two Bartonella house-keeping genes (gltA and rpoB) of the dromedary isolatesrevealed 96% identity to B. bovis sequences for each one ofthe targets (Table 3); the latter were chosen for this analysis

due to their good discriminatory power. According to LaScola’s criteria, a newly encountered Bartonella isolateshould be considered a new species if a 327-bp gltA fragmentshares less than 96.0% sequence similarity and a 825-bp rpoBfragment shares less than 95.4% sequence similarity withthose of the validated species (La Scola et al. 2003). Basedexclusively on these criteria, the dromedary isolates de-scribed here should be considered as B. bovis. However, amore comprehensive sequence comparisons analysis basedon nine genetic loci suggested that the dromedary isolatesrepresent a novel Bartonella sp., which is evidently differentfrom, although evolutionarily related to, B. bovis (Table 3 andFig. 1).

We suggest that a sequence-based approach for the differ-entiation of Bartonella spp. that utilizes the partial sequencesof gltA and rpoB does not have sufficient discriminatorypower in relation to ruminant-related Bartonella spp. If that isthe case, then the La Scola paradigm should be revised.Accordingly, to fully differentiate members of this group,more robust methods should be taken, such as sequence de-termination of multiple targets with high discriminatorypower or whole-genome sequence comparisons. A MLSTscheme, on the basis of nine genetic loci, originally devel-oped to compare genetic variants of B. bovis strains (Bai et al.2013), was modified in our study and identified seven STsof the new dromedary isolates (Table 4). This MLSTscheme may serve as a future reference for comparing camel-associated Bartonella isolates.

Dromedary Bartonella isolates were also found to beflagellated (Fig. 2). Motility is not a known trait of B. bovis, asa previously reported bovine B. bovis isolate from France(Bermond et al. 2002) as well as a B. bovis isolate from cattlein Israel (Rudoler et al. 2014) were shown to be non-flagellated. Thus far, ruminant-derived Bartonella spp. B.chomelii, B. schoenbuchensis, B. capreoli, and Candidatus B.melophagi were shown to carry flagella (Dehio et al. 2001,Bermond et al. 2002, Maillard et al. 2004, Maggi et al. 2009).B. bovis is an exception, because it is the only nonflagellatedmember among the species comprising lineage II of thephylogenetic tree of the genus Bartonella (Fig. 1 and Harmsand Dehio 2012). The dromedary Bartonella isolates werefound to be closely related to all of these species, becausethey share the same evolutionary lineage.

Camels have adapted to desert life, which presents tem-perature extremes and a scarce supply of food and water.Seasonal variations and the hydration/dehydration status of

FIG. 2. Dromedary Bartonella isolates are flagellated.Transmission electron micrographs of negatively stainedcells of Bartonella isolates 24C (a), 31C (b), and 47C (c)from Israeli dromedary camels, and nonflagellated B. bovisisolate 234 (d) isolated from beef cattle in Israel (Rudoleret al. 2014). Scale bars indicate 200 nm. Magnification,46,000 · (a, c) and 59,000 · (b, d).

Table 4. Multilocus Sequence Typing of Dromedary Bartonella Isolates from Israel

Based on Nine Genetic Loci

Isolate Host Camel farm rpoB nuoG gltA ftsZ ribC SsrA groEL ITS 16S rRNA ST

24C Camel (Camelus dromedarius) A 1 1 1 1 1 1 1 1 1 ST138C Camel (Camelus dromedarius) B 1 1 1 1 1 1 1 1 1 ST131C Camel (Camelus dromedarius) B 2 2 1 1 1 1 1 1 1 ST247C Camel (Camelus dromedarius) C 1 1 1 2 2 2 2 1 1 ST354C Camel (Camelus dromedarius) C 3 2 1 3 3 3 1 1 2 ST465C Camel (Camelus dromedarius) C 1 1 1 1 1 1 2 1 1 ST567C Camel (Camelus dromedarius) C 1 1 1 1 1 1 2 1 1 ST568C Camel (Camelus dromedarius) C 1 1 1 1 1 1 2 1 3 ST669C Camel (Camelus dromedarius) C 3 2 1 3 3 3 1 1 4 ST7

ITS, internal transcribed spacer; ST, sequence type.

780 RASIS ET AL.

the camel were shown to have significant effects on numeroushematological indices, endocrine function, as well as bloodminerals and metabolites concentrations of dromedaries. Inaddition, the camel’s body temperature fluctuates dramati-cally (34�C–42�C) during the day to minimize water loss(Amin et al. 2007, Ouajd and Barhoumi 2009). It is reason-able to assume that the unique physiology of these mammalsrequired host-specific adaptations of a Bartonella sp. infect-ing camels. One may also speculate that expression of flagellain the novel dromedary Bartonella isolates is host dependentand tightly regulated by a yet unknown environmental sig-nal(s). In a previous report, a B. melophagi isolate from asheep ked was shown to be flagellated, whereas no flagellawere observed by TEM of a Candidatus B. melophagi humanisolate, identified on the basis of sequence homology of therpoB, gltA, 16S rRNA, and internal transcribed spacer (ITS)fragments (99.2%, 97.9%, 99.7%, and 99.2% similarity, re-spectively) (Maggi et al. 2009). This is an example of strainvariation within a Bartonella sp. that is manifested by dif-ferential phenotypic expression of flagella. This phenomenonmight be strain or host dependent and warrants furtherinvestigation.

Conclusions

To conclude, we show for the first time that dromedarycamels may serve as natural hosts for a Bartonella sp. Onthe basis of distinct reservoir host and ecological niche,sequence-based analyses, and expression of flagella, wedesignate dromedary Bartonella isolates described herein asa novel Bartonella sp. named Bartonella dromedarii sp. nov.after its host, and assign isolate 24C as the type strain of thisspecies.

The notable prevalence of B. dromedarii bacteremia foundin Israeli dromedary camels combined with the close human-to-camel contact and observed abundance of ticks constitutea potential risk for zoonotic transmission. Continuous sur-veillance of Bartonella infection in camels and in personswho come in close and frequent contact with camels (e.g.,camel handlers) are needed to understand the potential role ofB. dromedarii as a zoonotic agent.

Acknowledgments

Electron microscopy work was done at the Bio-ImagingUnit, the Alexander Silberman Institute of Life Science, theHebrew University, Jerusalem, Israel. The authors thank Dr.Yuval Gottlieb-Dror from Koret School of Veterinary Med-icine, the Hebrew University of Jerusalem for morphologicalclassification of ticks.

Author Disclosure Statement

No competing financial interests exist.

References

Alagaili AN, Briese T, Mishra N, Kapoor V, et al. Middle Eastrespiratory syndrome coronavirus infection in dromedarycamels in Saudi Arabia. mBio 2014; 5:e00884–e00814.

Amin A, Abdoun K, Abdelatif A. Seasonal variation in bloodconstituents of one-humped camel (Camelus dromedarius).Pak J Biol Sci 2007; 10:1250–1256.

Bai Y, Cross PC, Malania L, Kosoy M. Isolation of Bartonellacapreoli from elk. Vet Microbiol 2011; 148:329–332.

Bai Y, Malania L, Alvarez Castillo D, Moran D, et al. Globaldistribution of Bartonella infections in domestic bovine andcharacterization of Bartonella bovis strains using multi-locussequence typing. PLoS One 2013; 8:e80894.

Bemis DA, Kania SA. Isolation of Bartonella sp. from sheepblood. Emerg Infect Dis 2007; 13:1565–1567.

Bermond D, Boulouis HJ, Heller R,Van Laere G, et al. Barto-nella bovis Bermond et al. sp. nov. and Bartonella capreolisp. nov., isolated from European ruminants. Int J Syst EvolMicrobiol 2002; 52:383–390.

Boutellis A,Veracx A, Angelakis E, Diatta G, et al. Bartonellaquintana in head lice from Senegal. Vector Borne ZoonoticDis 2012; 12:564–567.

Breitschwerdt EB, Kordick DL. Bartonella infection in animals:Carriership, reservoir potential, pathogenicity, and zoonoticpotential for human infection. Clin Microbiol Rev 2000; 13:428–438.

Chang CC, Chomel BB, Kasten RW, Heller RM, et al. Barto-nella spp. isolated from wild and domestic ruminants in NorthAmerica. Emerg Infect Dis 2000; 6:306–311.

Dehio C, Lanz C, Pohl R, Behrens P, et al. Bartonellaschoenbuchii sp. nov., isolated from the blood of wild roedeer. Int J Syst Evol Microbiol 2001; 51:1557–1565.

Edgar RC. Muscle: Multiple sequence alignment with highaccuracy and high throughput. Nucleic Acids Res 2004;32:1792–1797.

Erol E, Jackson C, Bai Y, Sells S, et al. Bartonella bovis iso-lated from a cow with endocarditis. J Vet Diagn Invest 2013;25:288–290.

Felsenstein J. Confidence limits on phylogenies: An approachusing the bootstrap. Evolution 1985; 39:783–791.

Giladi M, Ephros M. Chapter 160: Bartonella infections, in-cluding cat-scratch disease. In: Longo DL, Fauci AS, KasperDL, Hauser SL, eds. Harrison’s Principles of Internal Med-icine. New York: McGraw Hill Professional; 2011:1314–1320.

Greub G, Raoult D. Bartonella: New explanations for old dis-eases. J Med Microbiol 2002; 51:915–923.

Haagmans BL, Al Dhahiry SH, Reusken CB, Raj VS, et al.Middle East respiratory syndrome coronavirus in dromedarycamels: An outbreak investigation. Lancet Infect Dis 2014;14:140–145.

Harms A, Dehio C. Intruders below the radar: molecularpathogenesis of Bartonella spp. Clin Microbiol Rev. 2012;25:42–78.

Higgins JA, Radulovic S, Jaworski DC, Azad AF. Acquisitionof the cat scratch disease agent Bartonella henselae by catfleas (Siphonaptera:Pulicidae). J Med Entomol 1996; 33:490–495.

Jacomo V, Kelly PJ, Raoult D. Natural history of Bartonellainfections (an exception to Koch’s postulate). Clin Diagn LabImmunol 2002; 9:8–18.

Jukes TH, Cantor CR. Evolution of protein molecules. In:Munro HN, ed. Mammalian Protein Metabolism. New York:Academy Press; 1969:21–132.

Kim CM, Kim JY, Yi YH, Lee MJ, et al. Detection of Barto-nella species from ticks, mites and small mammals in Korea.J Vet Sci 2005; 6:327–334.

Kumsa B, Parola P, Raoult D, Socolovschi C. Bartonella me-lophagi in Melophagus ovinus (sheep ked) collected fromsheep in northern Oromia, Ethiopia. Comp Immunol Micro-biol Infect Dis 2014; 37:69–76.

FIRST REPORT OF Bartonella INFECTION IN CAMELS 781

La Scola B, Zeaiter Z, Khamis A, Raoult D. Gene-sequence-based criteria for species definition in bacteriology: TheBartonella paradigm. Trends Microbiol 2003; 11:318–321.

Maggi RG, Kosoy M, Mintzer M, Breitschwerdt EB. Isolationof candidatus Bartonella melophagi from human blood.Emerg Infect Dis 2009; 15:66–68.

Maillard R, Riegel P, Barrat F, Bouillin C, et al. Bartonellachomelii sp. nov., isolated from french domestic cattle (Bostaurus). Int J Syst Evol Microbiol 2004; 54:215–220.

Maillard R, Petit E, Chomel B, Lacroux C, et al. Endocarditis incattle caused by Bartonella bovis. Emerg Infect Dis 2007;13:1383–1385.

Ouajd S, Barhoumi K. Physiological particularities of drome-dary (Camelus dromedarius) and experimental implications.Scand J Lab Anim Sci 2009; 36:19–29.

Rudoler N, Rasis M, Sharir B, Novikov A, et al. First de-scription of Bartonella bovis in cattle herds in Israel. VetMicrobiol 2014; 173:110–117.

Saisongkorh W, Rolain JM, Suputtamongkol Y, Raoult D.Emerging Bartonella in humans and animals in Asia andAustralia. J Med Assoc Thai 2009; 92:707–731.

Saitou N, Nei M. The neighbor-joining method: A new methodfor reconstructing phylogenetic trees. Mol Biol Evol 1987;4:406–425.

Tamura K, Peterson D, Peterson N, Stecher G, et al. Mega5:Molecular evolutionary genetics analysis using maximumlikelihood, evolutionary distance, and maximum parsimonymethods. Mol Biol Evol 2011; 28:2731–2739.

Vayssier-Taussat M, Le Rhun D, Bonnet S, Cotte V. Insights inBartonella host specificity. Ann NY Acad Sci 2009; 1166:127–132.

Address correspondence to:Michael Giladi

The Bernard Pridan Laboratory for MolecularBiology of Infectious Diseases

Tel Aviv Sourasky Medical Center6 Weizmann Street

Tel Aviv 64239Israel

E-mail: giladi@tlvmc.gov.il

782 RASIS ET AL.