Molecular Survey of Anaplasmataceae Agents and Coxiellaceae in Non-Hematophagous Bats and Associated Ectoparasites from Brazil

The Anaplasmataceae family (order Rickettsiales) encompasses obligately intracellular bacteria of the genera Anaplasma, Ehrlichia, and Neorickettsia. Together with Coxiella burnetii (Coxiellaceae family, order Legionellales), these bacteria represent important causative agents of diseases in humans and animals. The scarcity of studies that investigated the occurrence of these agents in bats and their associated ectoparasites, emphasizes the need to achieve a better understanding of the role of these animals in the maintenance of such bacteria. Herein, 418 samples (133 blood, 135 spleen, and 150 ectoparasites) are collected from 135 non-hematophagous bats belonging to 12 species in a periurban area of Campo Grande city, Mato Grosso do Sul state, midwestern Brazil. In the results, 1.65% (7/418), 12.04% (50/418), and 13.63% (57/418) of samples are positive in PCR assays for Anaplasma spp. (16S rRNA gene), Ehrlichia spp. (dsb gene), and Neorickettsia spp. (16S rRNA gene), respectively. Anaplasma spp. and Neorickettsia spp. are detected in one (5.26%) Ornithodoros hasei tick larva. Ehrlichia spp. is detected in 14% of bat flies (represented by Megistopoda aranea, Trichobius costalimai, and Strebla hertigi), 6% of tick larvae (O. hasei), 12% of Spinturnicidae mites (represented by Periglischrus sp., P. torrealbai, and P. acutisternus), and 38% of Macronyssidae mites (Steatonyssuss sp.). The obtained sequences are observed to be similar to Anaplasma phagocytophilum (97.42–97.6% identified), Ehrlichia minasensis (96.73–100% identified), Neorickettsia risticii (96.7–100% identified), and Neorickettsia findlayensis (95.07–100% identified) by BLASTn analyses, and closely related to Ehrlichia ruminantium by phylogenetic analyses based on the gltA gene. No bat samples (blood/spleen) are positive in the qPCR assay for C. burnetii based on the IS1111 gene. The present work shows, for the first time, the occurrence of Anaplasmataceae in bats and associated ectoparasites (ticks, mites, and bat flies) from Brazil.


Introduction
Approximately 20% of known mammal species belong to the order Chiroptera, which composes the second largest group of mammals in the world [1]. Currently, there are 1400 species described worldwide. From those, 181 bat species are officially registered in a high concentration of the bacteria [48]. This agent can cause abortion and infertility in cattle and goats [49]. In humans, while the acute phase is characterized by hepatitis or, more unusually, pneumonia, the most frequent clinical manifestation in the chronic phase is endocarditis [47,50]. This agent was already described in bats and related ectoparasites from Algeria, South Africa, Chile, and Brazil [51][52][53].
Due to the importance of Anaplasmataceae agents and C. burnetii in animal and human health and the high representativeness of the order Chiroptera in Brazil, the present study aims to molecularly survey these agents in non-hematophagous bats and their ectoparasites sampled in midwestern Brazil.

Results
A total of 135 bats were captured and DNA samples obtained from 135 spleen tissues, 133 blood, and 150 ectoparasites (including 64 bat flies, 46 Spinturnicidae mites, 21 Macronyssidae mites and 19 tick larvae) were positive in the PCR assays for the endogenous genes (Supplementary Table S1).
All DNA samples obtained from spleen fragments, whole blood samples, and ectoparasites, totaling 418 samples, were negative for Ehrlichia spp. and Anaplasma spp. when tested for the groEL gene by a multiplex real-time PCR. Seven (1.67%) out of the 418 samples submitted to the protocol, using conventional nested PCR targeting the 16S rRNA gene for Anaplasma spp., were considered positive, presenting a single band with low intensity (Supplementary Table S2). Out of these, three samples were represented by spleen fragments from Platyrrhinus lineatus individuals, three from blood samples of Phyllostomus discolor specimens, and one from the Ornithodoros hasei tick larvae, which was collected from a negative P. discolor individual. Out of all positive samples, only one amplicon obtained from O. hasei was sequenced, due to the low intensity of bands in agarose gel electrophoresis. When analyzed by BLASTn analysis, the obtained sequence (542 bp) showed 97.6% identity (100% query coverage; E-value = 0.0) with the A. phagocytophilum detected in a deer (Hydropotes inermis) from the Republic of Korea (Accession number: KR611598.1) and 97.42% of identity with A. phagocytophilum and A. bovis detected in Haemaphysalis longicornis (100% of query coverage; E-value = 0.0), sampled in South Korea (Accesion numbers: GU046565.1 / GU064901.1).
Out of the seven samples showing positivity in the nested PCR for Anaplasma spp., none showed positivity neither when submitted to quantitative real time PCR (qPCR) for A. phagocytophilum based on the msp-2 gene nor in PCR assays targeting ITS 23S-5S and groEL. On the other hand, one blood sample obtained from a P. discolor individual was positive in the PCR targeting the gltA gene and sequenced. However, when tested by BLASTn, the sequence obtained (620 bp) showed to be similar (83.15% of identity) to Ehrlichia sp. identified in a tick from China (KJ410275.1) (100% of query coverage; E-value = 4 × 10 −155 ).
Fifty-seven (13.63%) out of 418 samples were positive in the nested PCR assay for N. risticii, based on the 16S rRNA gene: 11 (19.3%) were from animals of the species A. lituratus (five from spleen fragments and six from whole blood), 11 (19.3%) of the species P. lineatus (four from spleen fragments and seven from whole blood), 10 (17.54%) of the species A. planirostris (six from spleen fragments and four from whole blood), nine (15.79%) of the species P. discolor (two from spleen fragments and seven from whole blood), four (7.01%) of the species E. furinalis (one from spleen fragments and three from whole blood), three (5.26%) of the species C. perspicillata (two from spleen fragments and one from whole blood), two (3.5%) of the species Molossus molossus (one from spleen fragments and one from whole blood), two (3.5%) of the species Glossophaga soricina (one from spleen fragments and one from whole blood), and one (1.57%) positive blood sample of the species Myotis nigricans, Molossops temminckii, Eumops perotis, and Chiroderma villosum. The only ectoparasite sample that was positive for this agent was an O. hasei tick larva collected from a negative specimen of P. discolor (Supplementary Table S4). Out of the 34 positive blood samples for the 16S rRNA-based PCR assay, 11 (32.35%) were chosen for sequencing. The samples were chosen for convenience, but in such a way as to represent an individual of each bat species that showed positivity in this protocol. When analyzed by BLASTn, eight sequences (426-524 bp) obtained from blood samples showed similarity (97.9-100%) with N. risticii (Accession numbers MK760560 and CP001431.1) (100% of query coverage; E-value = 0.0); three sequences obtained from a tick larvae (304 bp), one bat spleen (421 bp) and one bat blood sample (385 bp) was observed to be 95.07-100% identical to N. findlayensis (Accession number MH476282) (100% of query coverage; E-value = 7 × 10 −131 to 0.0); finally, two sequences obtained from bat blood samples (212-282 bp) were observed to be similar (96.7-100%) to N. risticii and N. findlayensis (Accession numbers MK580530 and MH476282) (100% of query coverage; E-value = 1 × 10 −143 to 1 × 10 −96 ). All the positive samples were negative in the additional PCR protocols targeting the p51kDa, groEL, and near-full 16S rRNA genes.
All the BLASTn results are summarized and presented in the Supplementary Table S5. A total of 268 DNA samples (133 from whole blood and 135 from spleen fragments) were screened for the presence of C. burnetii DNA and none were positive. Unfortunately, the DNA samples from ectoparasites did not yield sufficient volume for this assay.
One phylogenetic tree was constructed based on a 522 bp alignment of the gltA gene ( Figure 1), using the Maximum Likelihood method and TIM3 + F + I + G4 as the evolutive model. The unique sequence obtained from a blood sample from P. discolor (GenBank Accession Number OK338001) was positioned closely to E. ruminantium. The remaining sequences were not further analyzed as the final alignment length was not suitable for robust inferences.
sequences were not further analyzed as the final alignment length was not suitable for robust inferences.

Discussion
Among all the seven positive samples for Anaplasma spp., only the amplicon obtained from O. hasei tick larva was sequenced, obtaining 97.6% similarity with A. phagocytophilum, which was previously detected in deer of the species Hydropotes inermis from the Republic of Korea. Previously, the occurrence of A. phagocytophilum was described in 22.6% (63/278) of fecal samples of Rhinolophus hipposideros in France [35], 1.7% (1/59) of blood samples of Myotis myotis [54] in Poland, 4.3% (6/138) of I. simplex ticks collected from bats from Hungary and Romania [36], and 14.28% (1/7) of Nyctalus noctula species from Russia [37]. Despite the high occurrence of A. phagocytophilum in bats from France, the remaining abovementioned studies showed low positivity for this agent among bats or associated ectoparasites, corroborating the results found in the present work for Anaplasma spp. (1.67%; 7/418). Noteworthy, even though sequences similar to A. phagocytophilum have been detected in several wild animals from Brazil, negative results have been reported when a qPCR specific for this agent and based on msp-2 gene was used [6,55]. Together, these findings suggest that a putative new Anaplasma species is circulating among wild animals from Brazil. Indeed, new 'Candidatus Anaplasma spp.', namely 'Candidatus Anaplasma brasiliensis' and 'Candidatus Anaplasma amazonensis' have been recently detected in mammals from the Superorder Xenarthra from Brazil [29]. Future studies aiming at obtaining larger fragments of different molecular markers are needed in order to check the phylogenetic relatedness of bat-associated Anaplasma spp. to these newly described 'Candidatus Anaplasma spp.' The unique Anaplasmataceae gltA sequence obtained was observed to be similar to Ehrlichia spp. with low identity obtained when compared to other sequences (<90%). The phylogenetic analysis evidenced that the sequence was closely related to E. ruminantium and Ehrlichia spp., detected in a tick from China. Even though there are reports of the occurrence of E. canis, E. ruminantium, and E. muris-like in humans [8], these agents

Discussion
Among all the seven positive samples for Anaplasma spp., only the amplicon obtained from O. hasei tick larva was sequenced, obtaining 97.6% similarity with A. phagocytophilum, which was previously detected in deer of the species Hydropotes inermis from the Republic of Korea. Previously, the occurrence of A. phagocytophilum was described in 22.6% (63/278) of fecal samples of Rhinolophus hipposideros in France [35], 1.7% (1/59) of blood samples of Myotis myotis [54] in Poland, 4.3% (6/138) of I. simplex ticks collected from bats from Hungary and Romania [36], and 14.28% (1/7) of Nyctalus noctula species from Russia [37]. Despite the high occurrence of A. phagocytophilum in bats from France, the remaining abovementioned studies showed low positivity for this agent among bats or associated ectoparasites, corroborating the results found in the present work for Anaplasma spp. (1.67%; 7/418). Noteworthy, even though sequences similar to A. phagocytophilum have been detected in several wild animals from Brazil, negative results have been reported when a qPCR specific for this agent and based on msp-2 gene was used [6,55]. Together, these findings suggest that a putative new Anaplasma species is circulating among wild animals from Brazil. Indeed, new 'Candidatus Anaplasma spp.', namely 'Candidatus Anaplasma brasiliensis' and 'Candidatus Anaplasma amazonensis' have been recently detected in mammals from the Superorder Xenarthra from Brazil [29]. Future studies aiming at obtaining larger fragments of different molecular markers are needed in order to check the phylogenetic relatedness of bat-associated Anaplasma spp. to these newly described 'Candidatus Anaplasma spp.'.
The unique Anaplasmataceae gltA sequence obtained was observed to be similar to Ehrlichia spp. with low identity obtained when compared to other sequences (<90%). The phylogenetic analysis evidenced that the sequence was closely related to E. ruminantium and Ehrlichia spp., detected in a tick from China. Even though there are reports of the occurrence of E. canis, E. ruminantium, and E. muris-like in humans [8], these agents represent a major concern in animal health. E. ruminantium, which is known to cause heartwater or cowdriosis in wild and domestic ruminants, is mainly transmitted by Amblyomma spp. ticks in Africa and the Caribbean [12,56]. Interestingly, a genotype (Strain Jaguar) related to, albeit distantly, called E. ruminantium, was detected in jaguars and in associated ticks (A. triste, A. sculptum, and Amblyomma spp.) sampled in Pantanal wetland, in midwestern Brazil, and is based on rrs and dsb sequences [57]. Additionally, a similar genotype (fox-ES-1) was detected in crab-eating foxes (Cerdocyon thous) in southeastern Brazil [58].
Herein, 11.96% (50/418) of samples were positive in PCR assays for Ehrlichia spp. based on the dsb gene. Out of these positive samples, 35  Interestingly, the obtained Ehrlichia dsb sequences were observed to be related to E. minasensis. This agent has been described in cattle and R. microplus from Brazil [5,59,60], potentially causing disease in cattle [61]. Recently, Ehrlichia dsb genotypes closely related to E. minasensis were detected in Xenarthra mammals in Brazil [29]. These results suggest that E. minasensis may infect a broader range of hosts than previously recognized.
Even though the multiplex qPCR assay, used herein for detecting a fragment of groEL gene for Ehrlichia and Anaplasma species with high sensitivity and specificicity [62], was originally designed for amplifying DNA from the main Ehrlichia (namely E. canis, E. chaffeensis, E. muris, and Ehrlichia sp., Anan strain/IOE agent), and Anaplasma (namely A. phagocytophilum, A. platys, A. marginale, A. centrale, A. ovis, and A. bovis) species. Considering that bats and associated ectoparasites likely harbor DNA of genetically diverse Ehrlichia and Anaplasma species, the used qPCR assays, based on the groEL gene, which is less conserved than 16S rRNA gene, were not able to catch these putative new genotypes/species of Ehrlichia and Anaplasma. This incongruence between the results of the groEL-based multiplex qPCR and 16S rRNA-based cPCR assays was already reported previously [29,62,63], when working on the molecular detection of Ehrlichia and Anaplasma in biological samples from wild animals from Brazil, which emphasizes the occurrence of putative new genotypes or even species of Anaplasmataceae in the Brazilian wildlife.
It is known that bats not only act as a host for N. risticii-infected trematodes but also as a possible natural reservoir for N. risticii, as this agent has been detected in tissues (e.g., intestine, lung, liver, and spleen) of naturally infected bats [41][42][43][44]. The detection of Neorickettsia spp. in digenean trematodes found parasitizing bats [45] emphasized their contribution to the maintenance of these agents. In addition, vertical transmission of N. risticii from adult trematodes (Acanthatrium oregonense) to eggs was evidenced when these helminthes were collected from E. fuscus bats in the USA [42]. In Europe, Hornok et al. (2018) [46] found 2.96% (4/135) positivity for N. risticii in M. dasycneme individuals sampled in the Netherlands. Future studies should be performed in order to investigate the bat-associated trematodes involved in the epidemiology of N. risticii in Brazil. Considering that it was not possible to confirm which Neorickettsia species is circulating among the sampled bats, epidemiological studies should be performed in an attempt to establish a possible correlation between infected bats and the occurrence or neorickettsiosis in horses in Brazil.
Herein, Ehrlichia DNA was detected in 14% bat flies (M. aranea, T. costalimai, and S. hertigi), 6% tick larvae (O. hasei), 12% Spinturnicidae mites, and 38% Macronyssidae mites. Out of those, Ehrlichia DNA was detected in seven bat flies, three ticks, 18 Macronyssidae and four Spinturnicidae mites, collected from bats that showed to be negative in the PCR assay for this agent. Similarly, the O. hasei tick larva that was positive for Neorickettsia spp. was collected from a specimen of P. discolor that showed to be negative for this agent. Moreover, DNA of Anaplasma spp. was detected in O. hasei larvae collected from a negative P. discolor individual. Considering the impossibility that the larval stage had ever fed on another animal, the transovarial transmission of bat-associated Anaplasma spp. cannot be rule out. Indeed, while the transovarial transmission of non-pathogenic A. phagocytophilum variants has been already reported in Dermacentor albipictus ticks, transovarial transmission of Ehrlichia spp. has not been proved [8,67]. Also, a possible explanation for these results might rely on a false-negative result with the bat samples (blood or spleen). It may occur due to the higher concentration of the host blood meal in the arthropod digestive tract, which may had increased the chances of detection of Ehrlichia DNA in the ectoparasite sample, as previously reported [68]. Lastly, the efficiency of the PCR assays should also be considered in this case. Nevertheless, the real role of ticks, mites, and bat flies in the transmission of Anaplasmataceae agents among bats should be further investigated.
Despite the lack of positivity for C. burnetii among bats sampled in the present work, this agent has already been reported in 3.4% (4/119) of bats in the Atlantic Forest in the states of Santa Catarina and Rio de Janeiro, Brazil, by conventional PCR and nested PCR assays based on the IS1111 gene in which the positive animals belonged to the species A. lituratus (3/4) and A. fimbriatus (1/4) [53]. Furthermore, this agent was detected in 15.8% (3/19) of I. verpertilionis ticks collected from bats in northeastern Algeria using real-time PCR based on the IS1111 marker [51]. More recently, C. burnetii was detected in five individuals (9%) of T. brasiliensis from Chile, also using a qPCR targeting the IS1111 gene [69]. However, it is worth mentioning that Coxiella-like endosymbionts are commonly found in ticks and, recently, it was shown that endosymbionts and C. burnetii are closely related [70]. The authors also suggested that the lineages might have evolved independently. Therefore, the possible participation of bats and specially their ectoparasites in the transmission of C. burnetii should be further investigated, since molecular and serological evidence of infection by C. burnetii has been reported in humans from Brazil [71][72][73].
In conclusion, the present work shows, for the first time, the occurrence of Ehrlichia spp. related to E. minasensis, E. ruminantium, and Anaplasma spp., similar to A. phagocytophilum and Neorickettsia spp., in non-hematophagous bats and associated ectoparasites (ticks, mites, and bat flies) from Brazil.

Material and Methods
A total of 135 spleen, 133 blood and 150 ectoparasites samples were collected from 135 non-hematophagous bats captured and sampled in two different sites ("Centro de Educação Ambiental Polonês" and "Centro de Educação Ambiental Florestinha") situated in the Campo Grande city, Mato Grosso do Sul state, Brazil as described elsewhere [74]. The bats' capture occurred quarterly between June 2017 and March 2018.
The DNA was purified from 50-200 µL blood samples using the Illustra Blood Mini Kit (GE Healthcare ® , Chicago, IL, USA) and from 10 mg of each bat spleen tissue and ectoparasites using the Illustra Tissue Mini Kit (GE Healthcare ® ), according to manufac-turer's instructions. The DNA concentration and absorbance ratio (260/280 nm) were measured using a spectrophotometer (Nanodrop, Thermo Scientific, Waltham, MA, USA).
To check the quality of the extracted DNA and avoid the presence of PCR inhibitors, each spleen and blood DNA extracted was tested by a conventional PCR targeting the mammal gapdh (Glyceraldehyde 3-phosphate dehydrogenase) gene [75]. For the ectoparasites collected from the animals, two different conventional PCR protocols were used. While a PCR protocol based on cytochrome c oxidase subunit I (cox-1) gene [76] was used for DNA samples obtained from bat flies and mites, a protocol based on the 16S rRNA gene [77] was used for DNA samples obtained from tick larvae.
Bat blood and spleen samples positive for the mammals' endogenous gene (gapdh) PCR protocol were submitted to PCR assays for Anaplasmataceae agents and C. burnetii. Ectoparasites positive in PCR assays based on cox-1 and 16S rRNA genes were only submitted to PCR assays for Anaplasmataceae agents, due to the low volume of DNA obtained.
For Anaplasmataceae agents, quantitative real-time (q), nested (n), and conventional (c) PCR screening protocols were performed for Anaplasma spp. and Ehrlichia spp. A multiplex qPCR for Anaplasma spp. and Ehrlichia spp. based on the groEL gene was performed as previously described [62]. Additionally, a nested PCR for Anaplasma spp., based on the 16S rRNA gene [78] and a conventional PCR for Ehrlichia spp., based on the dsb gene [79], were also used as screening protocols. All the positive samples found in the above-mentioned protocols were submitted to qPCR assays for E. canis based on the dsb gene [79], E. chaffeensis based on the vlpt gene [80], and A. phagocytophilum based on msp-2 gene [81]. Conventional and nested PCR assays, based on groEL [17,82], gltA [16], sodB [83], and omp-1 [84] genes as well as based on the intergenic region ITS 23S-5S [85], were performed for molecular characterization.
For molecular detection of N. risticii, DNA samples were submitted to a nested PCR based on the 16S rRNA gene, as previously described [86]. Additionally, the positive samples for the 16S rRNA were submitted to PCR assays targeting the nearly full-length 16S rRNA [87], p51kDa, and groEL [43,63,88] genes.
A quantitative real-time PCR targeting the repetitive element IS1111 was used for C. burnetii detection, as previously described [89].
As positive controls for the PCR assays, DNA samples from A. phagocytophilum and N. risticii kindly provided by Professor John Stephen Dumler (Uniformed Services University of the Health Sciences, Bethesda, MD, USA), and the E. canis Jaboticabal-strain (maintained in DH82 cell culture) were used. The C. burnetii DNA positive control was kindly provided by Professor Dr. Renato Arruda Mortara (Universidade Federal de São Paulo, SP, Brazil).
Firstly, the obtained sequences were submitted to a screening test using the Bioedit version 7.0.5.3 [90] to evaluate the electropherogram quality and to obtain consensus sequences from the alignment of the sense and antisense sequences. The BLAST program [91] was used to analyze the sequences of nucleotides (BLASTn), aiming to compare with sequences from an international database (GenBank) [92]. The consensus sequences obtained in the present study and those retrieved from GenBank were aligned using the ClustalW software [93] via Bioedit version 7.0.5.3 [89]. Phylogenetic inferences were based on Maximum Likelihood (ML). The best evolutionary model was selected by ModelFinder [94] and the trees were constructed using IQTree web server [95] with ultrafast Bootstrap of 1000 replicates [96]. All the trees were examined in Treegraph 2.0 beta software [97].
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/parasitologia1040021/s1, Table S1: Description of the 418 samples (spleen, blood or ectoparasites) collected from bats in Campo Grande, Mato Grosso do Sul, Brazil, and used in the present work for molecular screening of Anaplasmataceae agents and Coxiella burnetii,