Detection and identification of putative bacterial endosymbionts and endogenous viruses in tick cell lines

As well as being vectors of many viral, bacterial, and protozoan pathogens of medical and veterinary importance, ticks harbour a variety of microorganisms which are not known to be pathogenic for vertebrate hosts. Continuous cell lines established from ixodid and argasid ticks could be infected with such endosymbiotic bacteria and endogenous viruses, but to date very few cell lines have been examined for their presence. DNA and RNA extracted from over 50 tick cell lines deposited in the Roslin Wellcome Trust Tick Cell Biobank (http://tickcells.roslin.ac.uk) were screened for presence of bacteria and RNA viruses, respectively. Sequencing of PCR products amplified using pan-16S rRNA primers revealed the presence of DNA sequences from bacterial endosymbionts in several cell lines derived from Amblyomma and Dermacentor spp. ticks. Identification to species level was attempted using Rickettsia- and Francisella-specific primers. Pan-Nairovirus primers amplified PCR products of uncertain specificity in cell lines derived from Rhipicephalus, Hyalomma, Ixodes, Carios, and Ornithodoros spp. ticks. Further characterisation attempted with primers specific for Crimean-Congo haemorrhagic fever virus segments confirmed the absence of this arbovirus in the cells. A set of pan-Flavivirus primers did not detect endogenous viruses in any of the cell lines. Transmission electron microscopy revealed the presence of endogenous reovirus-like viruses in many of the cell lines; only 4 of these lines gave positive results with primers specific for the tick Orbivirus St Croix River virus, indicating that there may be additional, as yet undescribed ‘tick-only’ viruses inhabiting tick cell lines.


Introduction
Ixodid and argasid ticks are vectors of many viral, bacterial, and protozoan pathogens of worldwide medical and veterinary importance (Jongejan and Uilenberg, 2004). Some of these microorganisms are entirely dependent on ticks for transmission between vertebrate hosts in nature, such as tick-borne encephalitis virus (TBEV), Borrelia spp. spirochaetes, the obligatory intracellular genera Rickettsia and Ehrlichia, and the protozoan parasite genera Babesia and Theileria. Others, such as the obligatory intracellular bacterial genera Coxiella, Francisella, and some species of Anaplasma, can be naturally transmitted between vertebrates both by ticks and by other routes (direct contact, transplacental, biting flies) (Parola and Raoult, 2001;Aubry and Geale, 2011). A third group, including arboviruses such as Crimean-Congo haemorrhagic fever virus (CCHFV) and African swine fever virus (ASFV), are transmitted between wild vertebrates by ticks in a 'sylvatic cycle' without causing overt disease, but once they infect hosts such as humans or domestic pigs, respectively, direct vertebrate-vertebrate transmission can occur (Whitehouse, 2004;Costard et al., 2009). A fourth group of microorganisms cause persistent infections in ticks, presumably through transovarial transmission, but have no known vertebrate host. Many of these are only known through molecular detection by PCR and have not been visualised or cultivated in vitro, for example the Francisellalike endosymbionts detected in Dermacentor, Amblyomma, and Ornithodoros spp. ticks (Scoles, 2004).
Continuous cell lines derived from ixodid and argasid ticks are playing an increasingly important role in research on ticks and pathogenic and symbiotic tick-borne microorganisms (Bell-Sakyi et al., 2007. To date, very few of the around 60 currently available tick cell lines (Bell-Sakyi et al., 2011) have been screened for the presence of endosymbiotic bacteria and endogenous viruses. The Dermacentor andersoni embryo-derived cell line DAE100 was found to harbour the non-pathogenic endosymbiont Rickettsia peacockii (Simser et al., 2001), and a new species, Rickettsia hoogstraalii, was discovered in several cell lines derived from embryonic Carios capensis (Mattila et al., 2007). Najm et al. (2012) detected intermittent presence of DNA of the intramitochondrial symbiont Candidatus Midichloria mitochondrii in Ixodes ricinus and Rhipicephalus (Boophilus) decoloratus cell lines. A previously undescribed Orbivirus was found to infect the IDE2 cell line derived from embryonic Ixodes scapularis (Attoui et al., 2001); St Croix River virus (SCRV) is possibly the first 'tick only' virus to be discovered (Nuttall, 2009). Unidentified 'reovirus-like particles' were seen in electron micrographs of the Rhipicephalus appendiculatus nymph-derived cell line RA243 (Munz et al., 1987), but at the time there were no molecular tools available to assist in identification of this unknown, possibly endogenous virus. The effect of the presence of these endogenous microorganisms on growth of tick-borne pathogens in the infected cell lines is largely unknown, and similarly there is no information on their incidence in the remaining majority of tick cell lines.
Therefore, using PCR-based methods, we screened a representative panel of tick cell lines derived from embryonic, developing nymphal, or developing adult ixodid or argasid ticks for the presence of viruses and bacteria. DNA was extracted from the cells and screened with pan-bacterial 16S primers. RNA was used to generate cDNA which was screened using primers known to amplify tick-borne virus genera. PCR products were sequenced to aid identification. Selected cell lines were further examined by transmission electron microscopy (TEM) for presence of intracellular bacteria and viruses.

Tick cell lines
Fifty continuous tick cell lines derived from species of the ixodid genera Amblyomma, Dermacentor, Hyalomma, Ixodes, and Rhipicephalus, and the argasid genera Carios and Ornithodoros (Table 1), were cultured in either 2.2 ml medium in flat-sided culture tubes (Nunc) or 5 ml medium in 25-cm 2 flasks (Nunc), with weekly medium changes. The incubation temperatures used for each cell line are given in Table 1. For DNA extraction, cells were harvested either from actively growing cultures by pipetting followed by centrifugation at 200 × g for 5 min or from cryopreserved stabilates which were thawed rapidly, diluted 1 in 10 in appropriate complete medium and centrifuged as above. Cell pellets were resuspended in PBS. For RNA extraction, cells were harvested from actively growing cultures as described above; RNA extraction was not attempted from cryopreserved stabilates.

DNA and RNA isolation
Total genomic DNA was prepared from harvested tick cells using the DNeasy Blood and Tissue Kit (Qiagen Ltd, Crawley, UK) following the manufacturer's protocol. Purified DNA was eluted from the spin column with 400 l TE buffer (two successive 200-l elutions) and stored at −20 • C until use. RNA was extracted using an RNeasy Midi Kit (Qiagen Ltd, Crawley, UK) following the manufacturer's instructions. Purified RNA was eluted using 300 l (2 × 150 l) RNase-free water stored at −80 • C until use.

cDNA synthesis
First-strand cDNA synthesis was carried out using the Super-Script III System (Life Technologies, Inc.). Five hundred nanograms of total RNA and 1 l of a 50-M solution of random hexamers in a total volume of 13 l were incubated for 10 min at 65 • C and chilled on ice. After adding 4 l of 1st Strand Buffer, 1 l of DTT (0.1 M), 2 l of deoxynucleotide triphosphate mix (10 mM), and 1 l of SuperScript reverse transcriptase III (200 units/l), the reaction was incubated for 5 min at 25 • C, 1 h at 50 • C, and, finally, for 15 min at 70 • C. The resulting cDNA was used directly as a template for PCR amplification.

PCR amplification
DNA and cDNA were amplified by PCR using primers listed in Table 2. Each 50 l polymerase chain reaction (PCR) contained 36.85 l molecular biology grade water (Sigma), 1 l dNTP mix (10 mM of each dNTP), 10 l 5 × PCR buffer, 0.2 l of each primer (100 M), 0.4 U of Promega GoTaq TM DNA polymerase, and 2 l of template. Each PCR was carried out in an Applied Biosystems thermal cycler. Amplification was carried out with an initial 3-min denaturation at 95 • C followed by 40 cycles (55 cycles for the amplification of the Nairovirus S segment N ORF) of denaturation at 95 • C for 30 s, annealing at 50-60 • C for 30 s, and extension at 70 • C for 1 min 30 s. The amplification was completed by holding the reaction mixture for 7 min at 70 • C to allow complete extension. The PCR products were visualised by UV illumination on a 1% agarose gel stained with ethidium bromide.

DNA sequencing and analysis
Positive PCR products of, or close to, the expected size were purified using a QIAquick PCR purification kit (Qiagen Ltd, Crawley, UK) following the manufacturer's recommendations. DNA sequencing in the forward and reverse directions was performed by DNA Sequencing & Services, MRCPPU, College of Life Sciences, University of Dundee, Scotland (www.dnaseq.co.uk) using Applied Biosystems Big-Dye Ver 3.1 chemistry on an Applied Biosystems model 3730 automated capillary DNA sequencer. Homology searches were performed in the NCBI database using the BLAST search programme (Altschul et al., 1990). Sequences were aligned using ClustalW software (Larkin et al., 2007) for multiple sequence alignment. Nucleotide sequence identities were calculated using Lasergene software (DNASTAR Inc., Madison, USA).

Transmission electron microscopy
For TEM, samples of resuspended cells from selected cell lines were washed once in PBS, fixed in 3% glutaraldehyde in cacodylate buffer for 2-3 h, post-fixed in 1% osmium tetroxide in cacodylate buffer, dehydrated in acetone, and embedded in Araldite resin. Sections were cut on a Reichert OMU4 ultramicrotome (Leica), stained in uranyl acetate and lead citrate, and viewed in a Phillips CM120 transmission electron microscope. Images were taken on a Gatan Orius CCD camera.

Molecular identification of putative bacterial endosymbionts in tick cell lines
When screened with pan-bacterial 16S rRNA primers, 5 of the tick cell lines from the genera Amblyomma and Dermacentor yielded PCR products in the expected size range of 528 bp (Table 2). BLAST analysis of the sequenced PCR products placed the putative bacteria in the genera Rickettsia (from the 2 A. variegatum cell lines) and Francisella (from 3 Dermacentor spp. cell lines). Further analysis with genus-specific primers (Table 2) yielded PCR products whose partial gene sequences were 98.2-99.8% identical to published sequences for, respectively, Rickettsia africae in AVL/CTVM13 and AVL/CTVM17, Francisella-like endosymbiont of D. albipictus in Table 1 Continuous tick cell lines examined in this study. All the lines are deposited in the Tick Cell Biobank (http://tickcells.roslin.ac.uk).   Deyde et al. (2006) DALBE3 and Francisella-like endosymbiont of D. variabilis in DVE1 (Table 3). The sequence amplified from the D. nitens cell line ANE58 using pan-bacterial 16S rRNA primers showed nearly 97% similarity to sequences from endosymbionts of O. moubata and D. variabilis (Table 3). The sequences obtained from AVL/CTVM17 (16S rRNA, ompB and sca4 genes), ANE58 (16S rRNA), and DALBE3 (16S rRNA and lipoprotein genes) have been deposited in GenBank with accession numbers JX101606, JX101599, JX101598, JX101604, JX101605, and JX101603, respectively.

Molecular detection of endogenous viruses in tick cell lines
As expected (Attoui et al., 2001;Bell-Sakyi et al., 2007), the I. scapularis cell lines IDE2 and IDE8 were PCR-positive for SCRV. The remaining 3 lines from this species, IDE12, ISE6, and ISE18 were negative for SCRV. Two R. appendiculatus cell lines, RA243 and RA257, were also positive in the SCRV PCR. All the SCRV PCR products were 98-99% identical to a 358-bp section of the SCRV segment 2 VP2 gene (Fig. 1). The sequences obtained from IDE8, RA243, and RA257 have been deposited in GenBank with accession numbers JX101600, JX101601, and JX101602, respectively.
Thirteen of the cell lines derived from the tick species R. (B.) decoloratus (BDE/CTVM12, 14, and 16), R. appendiculatus (RAE/CTVM1, RAN/CTVM3, and RA243), H. anatolicum (HAE/CTVM8 and 9), I. scapularis (IDE12), C. capensis (CCE1) and O. moubata (OME/CTVM21, 22, and 27) yielded PCR products of, or near the expected 400 bp size with the pan-Nairovirus primers from both of 2 cDNA samples generated on separate occasions. When sequenced, most were found to contain a short segment which was identical to nucleotides 431-463 of the nucleocapsid gene of the S genome segment of CCHFV. However, the remainder of each of the tick cell line-derived sequences did not match well with any other part of the genomes of CCHFV or other published nairoviruses, and it was not clear whether the amplification was specific or non-specific. In some cases, the sequence obtained with the forward primer differed from that obtained with the reverse primer. Considering the sequences obtained from both the forward and reverse primers, all the sequences were different from each other (Table 4) apart from the 3 R. (B.) decoloratus and the 2 H. anatolicum cell lines which yielded consensus sequences from both primers which were, respectively, 98% and 100% identical within tick species, but only 40% identical between species and with less than 38% identity to the published Nairovirus sequences (Fig. 2). PCR using CCHFV-specific primers did not amplify any specific products from cDNA generated from 10 of the cell lines, representative of all the tick species positive with the pan-Nairovirus primers (data not shown). The pan-Flavivirus primers did not reveal the presence of any endogenous viruses in the tick cell lines (data not shown).

Detection of endogenous viruses in tick cell lines by electron microscopy
Samples of 35 of the ixodid tick cell lines were examined by TEM (Table 1), and 25 lines were found to harbour endogenous viruses which morphologically resembled reoviruses (Table 1, Fig. 3). No convincing bunyavirus-like particles were seen in any of the samples. The cell lines which were PCR-positive for Rickettsia (AVL/CTVM13 and AVL/CTVM17) or Francisella-like endosymbionts (DALBE3, DVE1, and ANE58) were included in this study, but no convincing evidence of intracellular bacteria was seen in any of the 5 lines.
Although we detected the presence of Rickettsia and Francisella spp. DNA in several tick cell lines, we were unable to detect bacteria in these cells by light (data not shown) or electron microscopy. We have attempted to infect Amblyomma americanum (AAE2) and R. (B.) microplus (BME/CTVM2) cells with the putative Rickettsia from the A. variegatum cell lines, also without success (data not shown). Therefore, we cannot say with certainty that these endosymbiotic bacteria are growing in the host cell lines. A similar problem arose in the study of Najm et al. (2012), who intermittently detected a small fragment of a Candidatus M. mitochondrii gene in 2 tick cell lines (BDE/CTVM14 and IRE/CTVM19), but failed to amplify a larger section of a different Candidatus M. mitochondrii gene from the same samples. Incorporation of segments of Wolbachia genes in invertebrate host genomes has been reported by McNulty et al. (2010) in nematodes and Martin et al. (2010) in woodlice; it is possible that a similar phenomenon occurs in ticks. On the other hand, endosymbiotic bacteria may be present in tick cell lines at such low levels as to be only intermittently detectable even by PCR, which is much more sensitive than microscopy. Further studies, involving PCR targeting additional genes and more extensive subinoculation experiments, are required to definitively characterise and, if possible propagate in vitro, the putative endosymbiotic bacteria identified in the Amblyomma and Dermacentor spp. cell lines. The availability of systems for propagation of such endosymbionts in tick cell lines will aid in their characterisation, providing a useful model system for study of the interaction between the bacteria and their arthropod vectors. Understanding the mechanisms by which ticks and endosymbionts interact might reveal ways in which the non-pathogenic endosymbionts can be manipulated for control of both the ticks and the pathogens that they also harbour, as demonstrated for mosquito cell-adapted Wolbachia pipientis which, when transfected into Aedes aegypti mosquitoes, interfered with replication of the human pathogens Dengue virus, Chikungunya virus, and Plasmodium (Iturbe-Ormaetxe et al., 2011). Nuttall (2009) lists 61 viruses transmitted by ticks -apart from the Orbivirus SCRV, these are viruses that have been detected because they infect vertebrate cells. To date, SCRV is the only characterised tick virus (Attoui et al., 2001); but it is likely that there are others (Nuttall, 2009). SCRV is known to chronically infect 2 I. scapularis cell lines, IDE2 (Attoui et al., 2001) and IDE8, both derived from progeny of field ticks collected from deer in Wisconsin, USA (Munderloh et al., 1994). Interestingly, we found that a third cell line, IDE12, which originated from the same area, was negative for SCRV, while 2 cell lines derived from the African species R. appendiculatus were positive in the PCR targeting segment 2 of this virus. The latter lines were both established in a laboratory in the UK by Varma et al. (1975) many years before the SCRV-positive I. scapularis lines were established by Munderloh et al. (1994) in a laboratory in the US; however, both RA243 and RA257 were subsequently cultivated in laboratories where IDE2 and/or IDE8 are also maintained, before deposition in the Tick Cell Biobank. On the other hand, reovirus-like particles were reported to be present in RA243 maintained in a laboratory in Germany by Munz et al. (1987), and we have successfully amplified a PCR product with 99% sequence identity to a 358-kb fragment of segment 2 of SCRV from a sample of RA243 cryopreserved since 1982 at Texas A&M University, where I. scapularis cell lines have never been cultivated (authors' unpublished data; Patricia Holman, pers. communication). Further study is required to establish the origin and identity of the virus(es) in RA243 and RA257, and their relationship to the SCRV in IDE2 and IDE8.
We found reovirus-like particles in all of the I. scapularis cell lines examined except IDE12, but did not observe any in the RA243 cells, suggesting that if SCRV is replicating in this cell line, it is present at a low level compared to the viruses in the I. scapularis cells, and/or it only infects a small proportion of the total cell population. We cannot definitely say that the reovirus-like particles observed in IDE2 and IDE8 cells are in fact SCRV, and the viruses in ISE6, ISE18, and the 21 other cell lines in which we saw reoviruslike particles remain at present unidentified. We could not find any reports of TEM detection of reoviruses or reovirus-like particles in tick cell lines, apart that from Munz et al. (1987). However, structures described as 'deposits of beta particles of glycogen' in salivary glands of partially-fed R. appendiculatus ticks infected with Theileria parva by Fawcett et al. (1982) closely resemble some of our putative endogenous viruses. Other TEM studies report presence of glycogen in perineurial cells of engorged female R. (B.) microplus (Binnington and Lane, 1980) and midgut cells of unfed adult A. americanum (Jaworski et al., 1983) but, in these reports, the magnification of the electron micrographs is insufficient to detect any structural detail within the glycogen particles. Although beta particles of glycogen as visualised by TEM appear to be of a similar texture to some of our putative viruses (particularly those shown in Fig. 2b, d, and f), they resemble random agglomerations of glycogen alpha particles, occur in a wide range of sizes, and are highly pleomorphic whilst retaining similar electron density throughout, as illustrated previously (Rhodin, 1974). In contrast, our putative viruses are generally roughly spherical, radially or at least bilaterally symmetrical with regular arrangements of external protrusions, and are of a similar size and electron density when in the same plane of section within each group of particles. A possible approach to resolving the identity of these virus-like particles would be to apply differential fixation protocols to render glycogen particles invisible in TEM preparations as described by Binnington and Lane (1980).
Many of the known tick-borne viruses belong to the families Bunyaviridae (genus Nairovirus), Flaviviridae (genus Flavivirus), and Reoviridae (genus Orbivirus). Therefore we carried out a preliminary screening of the tick cell lines using one set each of pan-Nairovirus and pan-Flavivirus primers. Due to time restrictions, we were unable to include additional primer sets such as those of Palacios et al. (2011) targeting orbiviruses. The high incidence of reovirus-like viruses detected by TEM suggests that these pan-Orbivirus primers are likely to aid us in detection and identification of endogenous viruses in many of the tick cell lines. Nuttall (2009) lists 19 tick-borne bunyaviruses, of which most either belong in the Nairovirus and Phlebovirus genera, or are unassigned. Of the 7 primer pairs designed by Lambert and Lanciotti (2009) for detection of different Bunyavirus groups, we only tested the Nairovirus primers. In future, their primer pairs targeting phleboviruses and orthobunyaviruses, of which at least one member is tick-borne, should also be used to screen the tick cell lines. Using the Nairovirus primers, we amplified PCR products from 13 cell lines derived from 3 ixodid and 2 argasid tick genera, but due to the high degree of diversity and low level of sequence identity with the few published Nairovirus sequences, it was difficult to conclude anything from the sequences of many of them, except that they indicated the possible presence of one or more novel nairoviruses in each of the cell lines. On 2 occasions, an almost identical sequence was amplified from the 3 R. (B.) decoloratus cell lines, which were all derived from the same pool of 4 egg batches, suggesting that the same virus may be present in all 3 lines. Similarly, almost identical sequences were amplified from the 2 H. anatolicum lines suggesting that they may also harbour the same virus, although they were derived from different egg batches (Bell-Sakyi, 1991). In contrast, sequences amplified using the forward and reverse primers differed for the 3 O. moubata lines and 2 of the R. appendiculatus lines (RAE/CTVM1 and RAN/CTVM3), suggesting that these lines could each harbour more than one virus. The negative results obtained using the CCHFV-specific primer sets confirmed the absence of this particular Nairovirus in the tick cell lines. Further analysis is needed to identify and characterise these new putative nairoviruses and to determine whether they are arboviruses with the potential to infect vertebrate hosts, or tick-only viruses. The primers targeting the L polymerase-encoding region of the Nairovirus genome, used by Honig et al. (2004) to amplify and sequence PCR products from 14 different tick-borne nairoviruses, could assist us in this future analysis.
The pan-Flavivirus primers which we used are reported to consistently amplify several flaviviruses (Johnson et al., 2010); however these authors only included 2 closely-related tick-borne viruses, TBEV and louping ill virus, in their test panel, so it is possible that this assay might not detect endogenous flaviviruses harboured by the tick cell lines. Screening with additional broad-spectrum primers amplifying different portions of the Flavivirus genome and proven to detect a wide range of tick-borne flaviviruses (Moureau et al., 2008;Maher-Sturgess et al., 2008), should be carried out in future. Ngoye virus, a novel Flavivirus, was detected in Rhipicephalus spp. ticks using such a PCR-based method (Grard et al., 2006). This virus failed to replicate in cell lines from vertebrate and tick species, including a non-host Rhipicephalus sp., suggesting the possibility that it might be another 'tick virus'. Moreover, the recent detection of ASFV-like gene sequences in human serum (Loh et al., 2009) suggests that the tick-borne DNA virus ASFV may not be the only member of the family Asfaviridae and that screening DNA from the tick cell lines for ASFV-related sequences might also be productive.
Over half of the known tick-borne viruses have been propagated in tick cells, along with many insect-borne arboviruses (Bell-Sakyi et al., 2011). Infection of tick cell lines with arboviruses generally results in low-level persistent infection, in the absence of detectable cytopathic effects. The cellular and molecular mechanisms responsible for this tolerance of arbovirus infection by tick cells are unknown; it is possible that ticks have evolved to coexist with a community of endogenous viruses. The continuous presence of these tick viruses may moderate or suppress the tick cell innate immune response to arbovirus infection, allowing persistent infection to occur. Ongoing studies in our laboratories to dissect the tick cell innate immune response to arbovirus infection (Bell-Sakyi et al., 2011) will help to elucidate the role of endogenous viruses in these processes.
to Matthew Dalby. We would like to thank Timothy Kurtti and Ulrike Munderloh of the University of Minnesota, Patricia Holman of Texas A&M University, and Libor Grubhoffer of the University of South Bohemia for provision of tick cell lines, Peter Simmonds and Colin Sharp for advice on sequencing data analysis and presentation, Lukas Mühlbauer for technical assistance and Claudia Rückert, Sabine Weisheit, and Steven Mitchell for help with electron microscopy.