Francisella-Like Endosymbionts and Rickettsia Species in Local and Imported Hyalomma Ticks

ABSTRACT

Hyalomma ticks (Acari: Ixodidae) are hosts for Francisella-like endosymbionts (FLE) and may serve as vectors of zoonotic disease agents. This study aimed to provide an initial characterization of the interaction between Hyalomma and FLE and to determine the prevalence of pathogenic Rickettsia in these ticks. Hyalomma marginatum, Hyalomma rufipes, Hyalomma dromedarii, Hyalomma aegyptium, and Hyalomma excavatum ticks, identified morphologically and molecularly, were collected from different hosts and locations representing the distribution of the genus Hyalomma in Israel, as well as from migratory birds. A high prevalence of FLE was found in all Hyalomma species (90.6%), as well as efficient maternal transmission of FLE (91.8%), and the localization of FLE in Malpighian tubules, ovaries, and salivary glands in H. marginatum. Furthermore, we demonstrated strong cophylogeny between FLE and their host species. Contrary to FLE, the prevalence of Rickettsia ranged from 2.4% to 81.3% and was significantly different between Hyalomma species, with a higher prevalence in ticks collected from migratory birds. Using ompA gene sequences, most of the Rickettsia spp. were similar to Rickettsia aeschlimannii, while a few were similar to Rickettsia africae of the spotted fever group (SFG). Given their zoonotic importance, 249 ticks were tested for Crimean Congo hemorrhagic fever virus infection, and all were negative. The results imply that Hyalomma and FLE have obligatory symbiotic interactions, indicating a potential SFG Rickettsia zoonosis risk. A further understanding of the possible influence of FLE on Hyalomma development, as well as on its infection with Rickettsia pathogens, may lead to novel ways to control tick-borne zoonoses.

IMPORTANCE This study shows that Francisella-like endosymbionts were ubiquitous in Hyalomma, were maternally transmitted, and cospeciated with their hosts. These findings imply that the interaction between FLE and Hyalomma is of an obligatory nature. It provides an example of an integrative taxonomy approach to simply differentiate among species infesting the same host and to identify nymphal and larval stages to be used in further studies. In addition, it shows the potential of imported Hyalomma ticks to serve as a vector for spotted fever group rickettsiae. The information gathered in this study can be further implemented in the development of symbiont-based disease control strategies for the benefit of human health.

INTRODUCTION

Hyalomma (Acari: Ixodidae) hard ticks include approximately 30 species and several subspecies (1) among which taxonomic relationships are constantly being reevaluated (2). Ticks of this genus infest a wide range of vertebrate hosts and are prevalent on migratory birds, which can be important in the dissemination of emerging Hyalomma-borne zoonoses, such as spotted fever group rickettsiae (SFGR) and Crimean Congo hemorrhagic fever virus (CCHFV) (3–7).

Other than pathogenic agent infections, Hyalomma spp. carry Francisella-like endosymbionts (FLE) (8–10) that may be obligatory primary symbionts supporting the restricted blood diet, as with other ticks and hematophagous insects (11). Several endosymbiotic bacteria in arthropod vectors were shown to interact with pathogens and affect their host's susceptibility to infection, including in tick hosts (12–14). These symbiotic interactions can lead to the development of novel vector-borne disease control (15), as in the case of dengue virus (www.eliminatedengue.com). The potential symbiosis between FLE bacteria and Hyalomma ticks can serve as a target for the novel development of Hyalomma-borne disease control; however, there is no current study testing the nature of FLE in these vectors or its interaction with potential pathogens, such as Rickettsia species.

The genus Rickettsia is composed of several groups, including the spotted fever group rickettsiae (SFGR) found in ticks (16). SFGR include at least 15 species described as the causative agents of rickettsioses mostly transmitted to humans by ticks of the genera Dermacentor, Amblyomma, Rhipicephalus, and Hyalomma (17). Israel is a region endemic for Israeli spotted fever caused by Rickettsia conorii israelensis, where fatal cases have been reported (18, 19). SFGR, such as R. felis, R. sibirica mongolitimonae, R. africae, R. massiliae, and R. aeschlimannii, were also found in various tick species in Israel, including in Hyalomma species (20–22). However, the distribution of the SFGR species among the various Hyalomma species and those introduced to Israel was never investigated.

Hyalomma ticks are the main vector of CCHFV, the most widespread tick-borne zoonosis worldwide that can cause fatal hemorrhagic fever in humans (23). Despite the existence of Hyalomma ticks in Israel, human and animal infections were never demonstrated in the country. CCHFV has been isolated from H. marginatum and other Hyalomma ticks collected from migratory birds in Morocco and Turkey (3, 4), and the ability of H. rufipes ticks to become infected from birds inoculated with the virus has been experimentally proven (24). Since Israel is situated at one of the largest migration routes in the world, where about 4% of the estimated 5 billion birds migrating from Africa to the Western Palearctic fly through every spring (25), it is suitable to identify which Hyalomma species are introduced into the country and determine their infection status, as was performed in other countries where yearly migratory events occur (26–28).

This study aimed to characterize the Hyalomma species distribution on various hosts in Israel and on birds migrating through Israel and to determine the role of FLE as a potential primary symbiont in these ticks and the prevalence of SFG Rickettsia and CCHFV in Hyalomma species endemic and imported to Israel. The knowledge gained in this study is expected to enable an assessment of the zoonotic risk imposed by these ticks and to open novel possibilities for their control.

RESULTS

Ticks sampled locally.

Out of 130 adult ticks collected during 2012 from 29 camels, 99.3% were Hyalomma dromedarii and 0.7% were H. excavatum. Out of 151 adult ticks collected from 56 horses during 2011 and 2014, 47.6% were H. marginatum and 52.4% were H. excavatum. All 28 ticks collected from seven tortoises were identified as H. aegyptium.

Ticks sampled from migratory birds.

One hundred fifty-six ticks were collected from 89 migratory birds over the spring migration of 2014 and 2015, of which 150 (96.1%) ticks were of the genus Hyalomma (H. rufipes in 2014 and H. marginatum complex in 2015), and the remaining samples (3.9%) were five Amblyomma species ticks and one Rhipicephalus guilhoni (collected in 2014). Out of 81 ticks collected in 2014, one was identified as an adult tick, 12 were larvae, and 68 were nymphs. Four of the nymphs were found with their molted skins, meaning molting occurred while attached to the bird host. In 2015, 75 ticks were collected, with 14 being larvae and 61 being nymphs, of which 5 ticks were found with their molted skins. Most ticks were at least partially engorged.

Of all the birds examined for ticks in 2014 and 2015, 19 (0.29%) and 4 (0.1%) ticks, respectively, were found infested in the Jerusalem Bird Observatory (JBO), while 27 (0.47%) and 39 (0.59%) ticks were found infested in Eilat. The average infestation prevalences (0.22% [23 out of 10,232 in JBO] and 0.54% [66 out of 12,190 in Eilat]) differ significantly between stations (P < 0.001). Ticks at larval and nymphal stages were found on 17 bird species, mostly of the order Passeriformes but also from the order Accipitriformes (Table S1 in the supplemental material). The species Sylvia atricapilla (Eurasian blackcap) was the most commonly infested bird (33.7%) in both years and was prevalent in both locations, followed by Iduna pallida (eastern olivaceous warbler) and Sylvia curruca (lesser whitethroat), which amount to 16.8% and 13.4% of all infested birds, respectively.

Hyalomma molecular identification by restriction fragment length polymorphism (RFLP).

In order to ease the identification of ticks found on the same host or those in early developmental stages, we used a restriction enzyme analysis. The enzymatic reactions correlated to the morphological taxonomic identification: the HpaI restriction analysis was successful for 93% of H. marginatum samples, the enzyme DraI successfully cut all samples morphologically identified as H. excavatum, and the HindIII restriction analysis was effective for 96.6% of the H. rufipes samples (Fig. S1).

Ticks of immature stages, collected from birds in 2015, were morphologically identified with uncertainty regarding their species (H. marginatum or Hyalomma turanicum). Using the restriction analysis, 88% of samples were cleaved and resulted in two fragment patterns: 11 samples showed a 650-bp fragment (HpaI in H. marginatum), and 55 samples showed a 600-bp fragment (HindIII in H. rufipes).

Hyalomma phylogeny.

The cytochrome c oxidase I (COI) consensus sequences were compared to the GenBank and BOLD databases (http://www.boldsystems.org/). The sequences of H. dromedarii, H. marginatum, H. excavatum, and H. rufipes matched sequences corresponding to the same species with high similarities (>97%). The sequence H. aegyptium 1 was similar to a sequence from Romania of the same species (GenBank accession no. JX394192.1, BOLD accession no. ACH4580), with a 99.5% identity match. The sequence H. aegyptium 2 was 99.04% similar to an H. aegyptium sequence from Belgium (GenBank accession no. AF132821.1, BOLD accession no. AAX2350). According to a pairwise analysis of evolutionary divergence, the most closely related sequences were H. aegyptium 1 and H. aegyptium 2, with 3.88% divergence, followed by H. marginatum and H. rufipes, with 4.07% divergence; the average distance between species was 10.29%, with a standard error of 0.01% (Table 1). A phylogenetic tree was constructed and used for further analysis, as described below.

TABLE 1.

Matrix of mt cytochrome c oxidase I gene sequence divergence between species based on the percentage of unshared nucleotides on pairwise comparisons^a

^amt, mitochondrial; standard deviations are in gray.

Prevalence and taxonomy of FLE.

Of the 257 Hyalomma ticks screened for the presence of FLE (Table S2), 90.66% of the samples were found to be positive. The prevalence varied between tick species, with 84.6% positive H. marginatum, 90.5% H. excavatum, 89.8% H. dromedarii, 100% H. aegyptium, and 90.4% H. rufipes samples. The difference in prevalence between species and life stages was not statistically significant (P = 0.996 and 0.931, respectively). A statistically significant difference in prevalence between males and females was found only for H. dromedarii (P < 0.001), with 64.7% prevalence in females and 96.77% prevalence in males.

The consensus sequences obtained from H. marginatum, H. excavatum, H. dromedarii, H. aegyptium, and H. rufipes ranged from 688 to 702 bp. The FLE sequences of H. marginatum, H. excavatum, and H. aegyptium were closely related to the FLE sequence of H. truncatum (accession no. JF290387.1), with 99% identity. The FLE sequence of H. dromedarii was also similar to the H. truncatum sequence and to that of an Ornithodoros moubata symbiont (accession no. AB001522.1), with 99% identity. There were no FLE sequences that were homologues to the H. rufipes sequence in GenBank.

Cophylogeny of FLE with Hyalomma hosts.

The phylogenetic tree inferred from FLE and other Francisella sequences (Fig. 1A) and the tree inferred from Hyalomma COI sequences (Fig. 1B) are shown with links between corresponding host and FLE branches. The FLE sequences are clustered together, resembling a monophyletic group separated from pathogenic Francisella sequences (bootstrap value, 97%). However, the inner splits in this monophyletic group are poorly supported (53% and 54%), with little divergence between FLE sequences (0 to 1.6%). The FLE sequences of H. marginatum and H. rufipes appear to be the closest, since their sequences were identical.

FIG 1.

Phylogenetic trees based on the maximum likelihood method with 1,000 bootstraps. (A) 16S rRNA of FLE from Hyalomma species (this study), Amblyomma maculatum (accession no. LNCT01000002), Francisella tularensis (accession no. CP017155.1), and Francisella philomiragia (accession no. NR_114925.1). (B) Corresponding Hyalomma host species COI gene.

The FLE and Hyalomma phylogenies show obvious similarities (Fig. 1), and the global goodness-of-fit statistic obtained from the Procrustes Approach to Co-phylogeny (PACo) analysis was significant (m²_XY = 0.007355749, P = 0.00094), which shows that the apparent dependence of the symbiont phylogeny on the host phylogeny is unlikely to be incidental. The agreement between them can be visualized with a Procrustes superimposition plot (Fig. S2). The bar plot obtained through the PACo analysis (Fig. 2) provides a representation of the contribution of each host-symbiont association to the global cospeciation fit, measured by means of jackknife estimation of their respective squared residuals (e²_i) at a 95% confidence interval. Links that represent a smaller fraction of the sum of squares are more likely to be the result of cospeciation, meaning that the lower bars, showing the link between H. rufipes and its FLE, the link between H. aegyptium 1 and 2 and their FLE, and the link between H. dromedarii and its FLE or even H. marginatum and its FLE, which are very close to the median, represent links where coevolution most likely occurred. The link between H. excavatum and its FLE appears to contribute less to the coevolutionary congruence between the phylogenetic trees.

FIG 2.

Bar plot showing the contribution of each host-symbiont link to the Procrustean fit; the bars represent jackknifed squared residuals, the error bars represent the upper 95% confidence intervals, and the dashed line represents the median squared residual value.

Maternal transmission of FLE.

All five egg-laying H. marginatum females were positive for FLE. The samples obtained from the DNA extraction of 110 single eggs (22 laid by each female, Fe1 to Fe5) were screened for FLE. From Fe1, 21/22 (95.4%) eggs were positive for FLE, as were 21/22 (95.4%) eggs from Fe2, 19/22 (83.3%) eggs from Fe3, and 20/22 (90.9%) eggs from Fe4 and Fe5, with an overall vertical transmission efficiency of 91.81%.

Localization of FLE in tick organs.

High densities of FLE can be seen in Malpighian tubules (Fig. 3A), with clusters apparently surrounding the cell nuclei (Fig. 3B). FLE were also observed in the poles of the oocytes (Fig. 3D and E), and strong signals of the FLE probe were recorded scattered on salivary gland acini (Fig. 3G), where the bacteria appear to surround cell nuclei (Fig. 3H). Assays with no probes were performed as controls in order to rule out autofluorescence as the source of recorded signals (Fig. 3C, F, and I). Images of antisense-EUB control, of FLE, and of EUB338 probes can be viewed separately in Fig. S3 to S6.

FIG 3.

FLE within Hyalomma marginatum females' Malpighian tubules (A to C), ovaries (D to F), and salivary glands (G to I), using whole-mount organ fluorescence in situ hybridization (FISH) viewed under a confocal microscope. Red, FLE-specific probe; blue, direct DNA staining using DAPI; yellow/orange, EUB338 probe for general bacteria simultaneously labeled with FLE probes. All scale bars represent 50 μm.

Prevalence of SFGR and CCHFV.

Out of 280 ticks screened (Table S3), 114 (40.7%) ticks were found to be positive for SFGR. The prevalence of SFGR infection was distributed among the various Hyalomma species as follows: 54.5% H. marginatum, 10% H. excavatum, 2.4% H. dromedarii, 19.2% H. aegyptium, 34.6% H. rufipes, and 81.3% H. marginatum complex were positive (Fig. 4). The difference in the prevalences of Rickettsia was statistically significant among tick species (chi-square test df = 5, P < 0.001) and tick hosts (chi-square test df = 3, P < 0.001) but not between locations. Sixty-eight of the positive samples were randomly chosen and Sanger sequenced; these were identified as spotted fever group rickettsiae (SFGR). Of the 68 samples, 64 samples were identified as R. aeschlimannii (accession no. HQ335157.1), with a 100% identity match, and 4 samples were identified as R. africae (accession no. HQ335137.1), with a 99% identity match (Fig. 4).

FIG 4.

Prevalence of Rickettsia in Hyalomma using diagnostic ompA PCR. Black, R. aeschlimannii; dark gray, R. africae; light gray, nonsequenced positive samples. H. m., H. marginatum.

None of the 249 tick samples collected from migratory birds, camels, and tortoises in various regions in Israel (Table S4) were found to be positive for CCHFV.

Maternal transmission of SFGR.

Three of the five H. marginatum females screened for the Rickettsia ompA gene were found to be positive, and maternal transmission of SFGR was demonstrated by screening 22 individual eggs from each positive female. The overall efficiency of maternal transmission was 95.3%: one of the females transmitted Rickettsia to all of the offspring analyzed, while the other two females transmitted the bacterium to 21 (95.45%) and 19 (90.47%) of the eggs. In order to ascertain the veracity of the positive results from egg samples, 4 of them were sequenced and identified as SFGR of the species R. aeschlimannii, with a 99% identity match using the BLAST algorithm in GenBank.

DISCUSSION

In this study, we determined the phylogeny, prevalence, and potential symbiont-host relationships of FLE in five different Hyalomma species either found on local hosts or imported via migratory birds. Our results support the assumption that FLE are obligatory symbionts of their Hyalomma hosts.

First, FLE was found to be highly prevalent across all Hyalomma species screened, including H. marginatum, whereas other studies reported lower FLE prevalence (9, 29). This difference, however, might be explained by method sensitivity, or the possibility that FLE have a different importance in Hyalomma ticks from different areas. In H. dromedarii, we found higher FLE prevalence in males, which is in contrast to other studies that reported the prevalence of symbionts to be typically higher in females (8, 30–32). The sampling effort (17 females and 62 males) could explain this deviation, and a quantitative approach may ascertain the accuracy of this observation.

Second, the 16S rRNA-based phylogenetic analysis demonstrated that the FLE from Hyalomma ticks are clustered together as a monophyletic group separated from pathogenic Francisella spp., in agreement with other FLE phylogeny analyses (9, 29, 33, 34), and the phylogenetic congruence between FLE and their Hyalomma hosts is significant. The sequences of FLE from H. marginatum and H. rufipes were identical, as was shown previously (10, 29). These two tick species, which are closely related, may also harbor the same endosymbionts. Also, the distance between FLE from H. aegyptium and H. marginatum corresponds to that shown in studies from Hungary, Ethiopia, and Yemen (9, 10, 29), indicating that FLE strains are comparable in different locations. Congruence between host and symbiont phylogenies has been shown for many obligatory symbiosis systems where the association is required to support the dietary needs of the host (35), and it was also shown for Rhipicephalus and Coxiella-like endosymbiont lineages (36). The FLE sequences of Hyalomma are close to an FLE sequence from Amblyomma maculatum which was hypothesized to have recently evolved from pathogenic Francisella species (37). These findings suggest that FLE in some or all Hyalomma species evolved in the same manner; however, the congruency with their hosts' phylogeny suggests that the transition to endosymbiosis occurred earlier in Hyalomma than in A. maculatum. In contrast, other studies on FLE phylogeny concluded that the symbiont-host association is relatively recent, although no specific analyses were applied to test the hypothesis in these studies (10, 34), excluding Dermacentor reticulatus and its FLE that formed a separate phylogenetic subgroup (38). As in the current study, the phylogenies of FLE and their Hyalomma hosts were analyzed on a relatively large scale, and the results of other studies are not necessarily contradictory.

Third, we demonstrated efficient maternal transmission of FLE to offspring (91.8%). In Dermacentor and Amblyomma ticks, transovarial transmission was tested in two pools of 10 larvae from 16 ticks and revealed a high transmission rate of 95 to 100% (39). Transmission may be even higher in Hyalomma, since individual eggs were tested, and the detection of low bacterial concentrations by standard PCR might not be optimal. In addition, we demonstrated FLE in the tick ovaries, suggesting a specific mechanism to ensure transmission. High densities of FLE were found in the Malpighian tubules, organs which have been shown to harbor main endosymbionts in ticks (8, 40). Interestingly, ubiquitous sporadic clusters of FLE were found in the salivary glands, which raises questions regarding the pathogenic potential of FLE (8). Nevertheless, FLE and other bacterial endosymbionts have been known to colonize salivary glands, and their presence in this organ does not ensure transmission to vertebrate hosts (41). Similar observations have been made for other hard ticks and their obligatory symbionts (36, 40), suggesting a potential nutritional role of the symbionts in supplying B vitamins missing in the tick blood meal (11, 42).

In this study, we also developed an RFLP identification method based on the mitochondrial COI gene in order to discriminate 3 Hyalomma species (H. marginatum, H. excavatum, and H. rufipes) that can be found on the same host (horse) and screened the ticks for the major zoonotic agents they may transmit, Rickettsia and CCHFV. For each morphologically identified Hyalomma species, a single consensus sequence was generated, except for H. aegyptium, which is divided into two branches with a relatively high intraspecific divergence (3.88%) (39). This difference could represent geographic distribution, as was shown for the African H. rufipes (43); however, the sample size in our study is too low to conclude the same. While the average interspecific sequence divergence between species in this study was 10.29%, only 4.07% divergence was found between H. marginatum and H. rufipes. These two species were considered part of the Hyalomma (Euhyalomma) marginatum Koch complex until recent years (2), and perhaps the low divergence reflects recent species separation. The RFLP analysis was also useful for the identification of immature stages of Hyalomma ticks, which is a limiting factor for the identification of ticks collected from migratory birds (3, 4, 27, 44). The main species found on migratory birds were H. rufipes and H. marginatum. The first is common through sub-Saharan Africa, while the second is common in North Africa, as well as southern Europe and Asia Minor to western Iran (2). The bird species with higher tick infestations in this study were S. atricapilla, I. pallida, and S. curruca (33.7%, 16.8%, and 13.4%, respectively), which are known to feed on the ground and forage in low scrub (45) and thus are exposed to ectoparasites (46, 47).

The role of migratory birds in the dissemination of disease by importing pathogen-infected ticks over a broad geographic range has become evident (48). Thus, we tested the ticks for CCHFV and SFGR. While no virus was detected, as was the case in similar surveys (27, 44, 46), Rickettsia was prevalent in H. rufipes and H. marginatum complex ticks from birds (34.6% and 81.3%, respectively). The sequenced SFGR samples were identified as R. aeschlimannii, a prevalent tick-borne SFGR in the African continent, previously detected in ticks from migratory birds arriving to Europe (49, 50). Rickettsia aeschlimannii is more common in northern Africa, and its high prevalence might point to the wintering location of the tick-infested birds (49). The most commonly tick-infested bird species in this study, Sylvia atricapilla, may occasionally winter in the area (45), suggesting that the majority of the ticks inspected originated from northern Africa. In local hosts, the prevalence of R. aeschlimannii varied. It was low in H. dromedarii ticks from camels (2.4%) and higher in H. excavatum and H. marginatum ticks from horses (10% and 54.5%, respectively). This is in agreement with the low prevalence found previously in H. dromedarii ticks from camels and in contrast to the prevalence in H. excavatum ticks from a camel and a horse (0.38% and 3.92%, respectively) (20). Rickettsia aeschlimannii was also detected in H. aegyptium ticks from tortoises, as has been shown previously (51). The ability of R. aeschlimannii to be transovarially transmitted from H. marginatum ticks to their offspring (52) with high efficacy, as shown here in individual eggs, strengthens the notion that H. marginatum may act as a reservoir for the disease in Israel. Rickettsia africae was detected in a few local ticks from northern Israel. Previous studies showed R. africae in Hyalomma ticks collected in southern Israel, and the authors hypothesized that it was imported to Israel from tick-infested camels arriving from Egypt (20, 22). Thus, our findings may indicate that R. africae propagated since its original introduction to the country via Hyalomma ticks. The clinical symptoms for most spotted fever rickettsioses are nonspecific and may be misdiagnosed (19). Reported cases of spotted fever in Israel are currently minor, with 6 and 7 reported cases in 2016 and 2017, respectively, with no specific rickettsial agent diagnosis (http://www.health.gov.il). Assuming this study presents the actual infection prevalences of both H. marginatum and H. excavatum ticks from horses, with which humans have frequent and prolonged interactions, Hyalomma and SFGR control and surveillance strategies should be considered.

Understanding the relationship between ticks and their bacterial endosymbionts is a first step toward symbiont-based control plans (15). Nonpathogenic bacteria in ticks may interact with pathogens and affect a host's susceptibility to infection (8); thus, the obligatory nature of the symbiosis between Hyalomma and FLE established in this study and finding of high-prevalence SFGR can be further implemented in zoonotic disease risk assessment, as well as in the development of control strategies for the benefit of human health.

MATERIALS AND METHODS

Tick collection.

Hyalomma ticks actively search for their hosts; therefore, rather than flagging vegetation, we removed the ticks directly from hosts in representative locations in Israel. Ticks were collected from horses, camels, tortoises (Testudo graeca), and migratory birds during 2011 to 2015 (Fig. 5). Ticks were morphologically identified using taxonomic keys (1, 53–55), preserved in vials containing 100% ethanol, and kept at −20°C until further processing, except for ticks from tortoises which were preserved in dry vials at −80°C until further use.

FIG 5.

Collection sites and hosts of all Hyalomma ticks in this study. Map generated using ArcMap version 10.0 (ESRI, Redlands, CA).

Nucleic acid extraction.

Prior to nucleic acid extraction, each tick was washed in 3% sodium hypochlorite, 70% ethanol, and finally twice in sterile Dulbecco's phosphate buffer saline solution (Biological Industries, Israel) in order to reduce external contaminants. DNA and RNA were then extracted using either the RTP DNA/RNA virus minikit or the RTP pathogen kit (Stratec, Germany), according to the manufacturer protocols, with a few modifications. Following sample disinfection, ticks were cut to four pieces on a sterile petri dish using a sterile scalpel blade and placed in the provided extraction tube with lysis buffer; fully engorged ticks were extracted in two separate tubes. After lysis in the prefilled extraction tube (with lyophilized lysis components), the tubes were centrifuged, and the supernatant was used for the remainder of the protocol. DNA and RNA were quantified in a NanoDrop ND1000 spectrophotometer (NanoDrop Technologies, Denmark) at A₂₆₀/A₂₈₀, and the samples were stored at −80°C until further use.

For DNA extraction of individual eggs, these were placed separately in a 0.2-ml microtube and crushed against the bottom of the tube with a sterile needle. Each microtube was then filled with 12 μl of extraction buffer containing 1 mg/ml proteinase K (Sigma-Aldrich, USA), 0.01 M NaCl, 0.1 M EDTA, 0.01 M Tris-HCl (pH 8.0), and 0.5% Nonidet P-40 substitute (Sigma-Aldrich, USA) (56); after a spin down of the tubes, they were incubated for 15 min at 65°C, followed by 10 min of inactivation at 95°C.

Standard PCR and real-time RT-PCR.

All standard PCR amplifications were performed on a T-gradient 180 basic thermocycler (Biometra, Germany), according to published protocols, with suitable primers (Table 2) in a final volume of 25 μl, containing 2× GoTaq Green master mix (Promega, USA), 0.2 μM forward and reverse primers, genomic DNA (gDNA) template, and nuclease-free water. Negative controls with no template were included in all reactions. Real-time reverse transcription-PCR (RT-PCR) for CCHFV detection was performed according to Wölfel et al. (57).

TABLE 2.

Primers and probes used in this study^a

Primer name	Platform	Specificity	Sequence (5′ to 3′)	Product size (bp)	Reference or source	Notes
F- CCHF S1	RT-PCR	CCHFV S segment	TCTCAAAGAAACACGTGCC	122	71
R- CCHF S122			CCTTTTTGAACTCTTCAAACC
Probe			(FAM)ACTCAAGGKAACACTGTGGGCGTAAG(BHQ1)
NC-Fran16S-F	PCR	FLE 16S rRNA gene	CAACATTCTGGACCGAT	373	33	Used for screening, 40 cycles for egg samples and 35 cycles for others
NC-Fran16S-R			TGCGGGACTTAACCCAACAT
NC-Fran16S-F		FLE 16S rRNA gene and universal primer 16S rRNA	CAACATTCTGGACCGAT	728	33	Used for phylogeny with R1494 under same conditions as screening the PCR
R1494			CTACGGCTACCTTGTTACGA		30
Rr190.70F		Rickettsial OmpA-encoding gene	ATGGCGAATATTTCTCCAAAA	630	72	Extension for 1 min at 72°C
Rr190.701R			GTTCCGTTAATGGCAGCATCT
TY-J-1449		5′ region of COI	AATTTACAGTTTATCGCCT	860	73	Used to obtain consensus sequences
C1-N-2312			CATACAATAAAGCCTAATA
HCO2064		5′ region of COI	GGTGGGCTCATACAATAAATCC	860	73	Used for restriction analyses
HCO1215			GCCATTTTACCGCGATGA
/Cy5/Eub338	FISH	Bacterial 16S rRNA	GCTGCCTCCCGTAGGAGT		74
/Cy3/anti-sense Eub338		No bacterial detection	ACTCCTACGGGAGGCAGC		74
/Cy3/FLE		FLE 16S rRNA	ACTCCAACAGCTAGTACTCA		This study

^aFAM, 6-carboxyfluorescein; BHQ1, black hole quencher 1; FISH, fluorescence in situ hybridization.

Sequencing and analysis.

PCR products were purified with ExoSAP (New England BioLabs, USA). Each reaction mixture contained 0.1 μl of exonuclease I, 0.2 μl of Antarctic phosphatase, 1 μl of exonuclease I buffer, 1 μl of Antarctic phosphatase buffer, 7.7 μl of nuclease-free water, and 10 μl of PCR product; the reaction mixture was incubated for 30 min at 37°C, followed by 5 min of inactivation at 70°C. The purified samples were Sanger sequenced (Macrogen, Holland) from both directions. All sequences were manually edited and aligned using BioEdit version 7.2.5 (http://jwbrown.mbio.ncsu.edu/BioEdit/bioedit.html) (58). The sequences were then compared to the GenBank database using the BLAST algorithm to confirm their identity.

RFLP analyses.

The species H. marginatum, H. excavatum, and H. rufipes, which can be found on the same host, were chosen for RFLP discrimination. After morphological identification of all samples, and based on the cytochrome c oxidase I (COI) consensus sequences obtained from representative samples, as described above, restriction enzymes were chosen in silico using the Webcutter 2.0 software (59) based on a distinct cleavage site (HpaI for H. marginatum, resulting in a 650-bp fragment; DraI for H. excavatum, resulting in a 550-bp fragment; and HindIII for H. rufipes, resulting in a 600-bp fragment). The restriction reactions were performed on 25 μl of PCR template added directly to a mixture containing 1 μl of restriction enzyme, 3 μl of NEBuffer 2.1 (New England BioLabs, USA) for HindIII or CutSmart buffer (New England BioLabs, USA) for HpaI and DraI, and 1 μl of diethyl pyrocarbonate-treated double-distilled water (DEPC-DDW). The mixture was incubated for 60 min at 37°C, and restricted versus unrestricted amplicons were visualized in a 1.5% agarose gel. All analyses were performed on previously morphologically identified and sequenced samples. Since the restricted fragments of each species differ in size, it allowed for the simultaneous identification of samples with all enzymes in the same reaction: 25 μl of PCR template, 0.66 μl of enzyme each, 4 μl of CutSmart buffer, and 9 μl of DEPC-DDW, followed by 2 h incubation at 37°C and 20 min of heat inactivation at 80°C.

Maternal transmission assay.

Seven fully engorged H. marginatum females were collected live from horses (location 6, Fig. 5), brought to the laboratory, and kept separately in a chamber under suitable conditions for laying eggs (60). The relative humidity of 75 to 85% was maintained using monopotassium phosphate (KH₂PO₄) diluted in DEPC-DDW in a sterile petri dish at 23 to 25°C. During the following 21 days, five of the 7 ticks laid eggs, and all were transferred to −80°C until further processing.

Phylogenetic analyses.

For Hyalomma COI, PCR products from 10 H. marginatum, 10 H. rufipes, 10 H. dromedarii, 9 H. aegyptium, 4 H. excavatum, and 75 unidentified Hyalomma ticks from migratory birds were sequenced as described above. Consensus sequences were manually determined for each species, except for H. aegyptium, which segregated into two different consensus sequences.

For FLE, 4 sequences from each tick species, obtained as described above, and partial 16S rRNA Francisella sequences (FLE of Amblyomma maculatum [accession no. LNCT01000002], Francisella tularensis [accession no. CP017155.1], and F. philomiragia [accession no. NR_114925.1]) were used for analyses with Molecular Evolutionary Genetics Analysis version 7.0 (MEGA7) for bigger data sets (61). Sequence alignment was performed with MUSCLE (multiple sequence comparison by log-expectation) (62) using default parameters, and phylogenetic trees were inferred based on the maximum likelihood method with 1,000 bootstraps. The distance between sequences was calculated based on the number of base differences per site between sequences using pairwise comparisons with 1,000 bootstraps; this simple method was applied, since most distance methods are comparable and give similar estimates (63–65).

Cophylogeny was tested with the tick COI tree and the FLE tree using the Procrustes Approach to Co-phylogeny (PACo) (66) in R version 3.3.1 (https://www.R-project.org/), with the vegan and ape packages (67, 68).

Fluorescence in situ hybridization.

To target FLE in tick organs, the probe/Cy3/FLE was designed based on the FLE 16S rRNA partial gene sequence (Table 2). The specificity for FLE of the probe was tested in silico using the ARB probe design tool against the SILVA database (69). Individual female ticks were disinfected as described above and dissected inside a droplet of sterile Dulbecco's phosphate-buffered saline (PBS), as described in reference 40. Salivary glands, ovaries, and Malpighian tubules were removed and placed in a droplet of sterile PBS. Pools of organs were transferred into cell strainers (Corning, USA) and placed in 6-well plates filled with FAA (5% acetic acid, 5% formaldehyde, 90% ethanol) in a vacuum chamber for 1 h. After 24 h at room temperature, the organs were dehydrated in increasing ethanol concentrations of 50% for 10 min, and then 50%, 80%, and 100% for 15 min each step, and finally the organs were left to dry at room temperature for 10 min. Staining was performed on separate pools of salivary glands, ovaries, and Malpighian tubules, as described before (40), with a few modifications, in that cell strainers were used to transfer the organs from buffer to buffer. The probes /Cy3/FLE and /Cy5/Eub338 were diluted in the hybridization buffer (20 mM Tris-HCl [pH 8.0], 0.9 M NaCl, 35% [vol/vol] formamide, 0.01% [wt/vol] sodium dodecyl sulfate [SDS]) to a concentration of 6.25 ng/μl. Later, 4′,6′-diamidino-2-phenylindole (DAPI; Invitrogen, USA) staining was applied to label DNA. After the procedure, samples were mounted on a Superfrost plus slide (Baor-Naor Ltd., Israel) with a drop of CitiFluor (EMS, USA). The whole procedure was also performed on organs using only the probe /Cy5/Eub338, as well as no-probe and antisense-EUB probe (Table 2) for fluorescence control. Samples were viewed under Leica's CTR65000 confocal laser scanning microscope (Leica Microsystems, Germany).

Statistical analyses.

Pearson chi-square tests were performed to ascertain differences between groups, and the analyses were performed using WinPEPI version 11.65 (70).

Accession number(s).

All COI consensus sequences were deposited in GenBank: H. dromedarii, accession no. KY548842; H. excavatum, accession no. KY548843; H. marginatum, accession no. KY548844; H. rufipes, accession no. KY548845; H. aegyptium 1, accession no. KY548846; and H. aegyptium 2, accession no. KY548847. All FLE sequences were deposited in GenBank under accession numbers KY469285 to KY469289.