Prevalence and phylogeny of Chlamydiae and hemotropic mycoplasma species in captive and free-living bats

Background Bats are hosts for a variety of microorganisms, however, little is known about the presence of Chlamydiales and hemotropic mycoplasmas. This study investigated 475 captive and free-living bats from Switzerland, Germany, and Costa Rica for Chlamydiales and hemotropic mycoplasmas by PCR to determine the prevalence and phylogeny of these organisms. Results Screening for Chlamydiales resulted in a total prevalence of 31.4%. Positive samples originated from captive and free-living bats from all three countries. Sequencing of 15 samples allowed the detection of two phylogenetically distinct groups. These groups share sequence identities to Chlamydiaceae, and to Chlamydia-like organisms including Rhabdochlamydiaceae and unclassified Chlamydiales from environmental samples, respectively. PCR analysis for the presence of hemotropic mycoplasmas resulted in a total prevalence of 0.7%, comprising free-living bats from Germany and Costa Rica. Phylogenetic analysis revealed three sequences related to other unidentified mycoplasmas found in vampire bats and Chilean bats. Conclusions Bats can harbor Chlamydiales and hemotropic mycoplasmas and the newly described sequences in this study indicate that the diversity of these bacteria in bats is much larger than previously thought. Both, Chlamydiales and hemotropic mycoplasmas are not restricted to certain bat species or countries and captive and free-living bats can be colonized. In conclusion, bats represent another potential host or vector for novel, previously unidentified, Chlamydiales and hemotropic mycoplasmas.


Background
Bats of the order Chiroptera are of increasing interest as potential reservoirs and vectors of pathogens. They possess unique characteristics among mammals, such as the ability to fly. Their extensive mobility, combined with their roost plasticity, nesting behavior and broad food range allows transport of pathogens to many different animal species in various locations [1]. It has been shown that bats are hosts for a multitude of different microorganisms that include viruses, bacteria, parasites and fungi. Several of these infectious agents are common to humans and domestic animals. The pathogenic potential of some bacterial species has been confirmed for bats, but the knowledge regarding the impact of such microorganisms on bat hosts is limited. In addition, knowledge on the natural microbiota of bats is scarce [2].
Chlamydiae are highly successful animal and human pathogens. Their taxonomic structure is composed of the order Chlamydiales, which consists of the nine families Parachlamydiaceae, Waddliaceae, Simkaniaceae, Rhabdochlamydiaceae, Criblamydiaceae, Piscichlamydiaceae, Clavichlamydiaceae and Parilichlamydiaceae, collectively referred to as Chlamydia-like organisms, and the Chlamydiaceae [3][4][5]. Chlamydiae are obligate intracellular bacteria with a biphasic developmental cycle. This cycle is characterized by the infectious but metabolically less active elementary body, which infects susceptible host cells, and the intracellular reticulate body, which undergoes binary fission [6]. As an inclusion fills with progeny, the reticulate bodies condense back into elementary bodies and are released by host cell rupture or by fusion of the inclusion membrane with the host cell plasma membrane. Chlamydiae are disseminated by aerosol or contact, requiring no alternative vector [4].
In 2005 and 2015, two novel Chlamydia-like organisms, Waddlia malaysiensis and Waddlia cocoyoc were detected in fruit bats in Malaysia and Mexico, respectively [7,8]. In 2016, members of the order Chlamydiales were detected in the fecal bacterial microbiota of Daubenton's bats in Finland [9].
Hemotropic mycoplasmas (hemoplasmas) are considered to be emerging or re-emerging zoonotic pathogens. They are classified in the order Mycoplasmatales, family Mycoplasmataceae and genus Mycoplasma. They are small, pleomorphic bacteria without a cell wall and parasitize red blood cells [10,11]. The transmission routes of hemoplasmas are not yet fully understood, but they are thought to be transmitted through blood, saliva and possibly also arthropods [10,[12][13][14]. Hemoplasmas are able to cause acute infectious anemia or various chronic diseases in farm animals [15,16], wild animals [17,18], pets [14] and humans [19]. The extent of clinical manifestations ranges from asymptomatic to lifethreatening [10].
The objective of this study was to investigate the prevalence and phylogenetic positioning of Chlamydiae and hemotropic mycoplasmas species in 475 captive and free-living bats from six families and 28 species from Switzerland, Germany, and Costa Rica (Table 1).

Results
Polymerase chain reaction (PCR) and sequencing results for Chlamydiales A total of 166/1021 DNA samples (16.3%) originating from 149/475 bats (31.4%) were positive for Chlamydiales DNA by real-time PCR. Of these 166 samples, 15 samples had a Ct value below 35.0 and were therefore sequenced, resulting in 17 consensus sequences comprising two to four sequences (Table 2). A BLASTn analysis and phylogenetic trees ( Fig. 1) of these sequences revealed two groups and suggest that these sequences may represent two novel species-level lineages.
Group two contained 9 sequences from bacteria within four captive and five free-living bats (   Chlamydial sequences assigned to group 2 Fig. 1 Phylogenetic trees based on partial sequences of the 16S and the 23S rRNA genes that show the relationships of the sequences obtained in this study and publicly available sequences of Chlamydia species and selected published sequences found in bats reflecting the phylogenetic relationship between known Chlamydia species based on nine genes [24]. a Phylogeny of chlamydial 16S rRNA gene, 284 bp covering V1 -V2, including all novel sequences and illustrating that they fall together within a novel clade. b Phylogeny of chlamydial 16S rRNA gene, 200 bp covering V3, relating available novel sequences to those previously found in bat samples and illustrating that these samples are closely related to previous bat samples over this region. c. Phylogeny of chlamydial 23S rRNA gene, 530 bp, relating available novel sequences to those previously found in bat samples and illustrating differences between the novel samples and previous bat samples across this region. Bat samples were selected from those published in Hokynar et al. [9] to reflect closely related samples and outgroup "Rhabdochlamydiaceae-like" samples. Bootstraps of 100 replicates are shown on key branches. Scale bar shows number of substitutions per site. Samples from this study are shown in bold closely related to sequences from uncultured Chlamydiales and from the Rhabdochlamydiaceae bacterium isolate P gt1 (GenBank accession number MF620051.1) ( Table 2). Not all sequences gave clearly interpretable traces: one 16S rRNA sequence gave sufficient quality for phylogenetic analysis and shows this relationship to previously characterized sequences from bats [9] (Fig. 1b).  Table 2). The phylogenetic tree constructed from these 16S rRNA gene sequences (Fig. 2) shows that they are closely related to previously published sequences from vampire bat samples [23] and sequences from Chilean bat samples [25]. The remaining nine sequences were closely related to Proteobacteria by BLAST analysis (Table 2 "Others").

Discussion
Bats are known to be important vectors and reservoirs for a multitude of different microorganisms such as lyssaviruses, henipaviruses [2], Leptospira sp. [26] and Bartonella spp. [27], but the knowledge regarding their potential role as carriers of other clinically significant pathogens or less common bacteria is limited. To date, the only Chlamydiales identified in bats are two Chlamydialike organisms: Waddlia malaysiensis [7] and Waddlia cocoyoc [8], isolated from fruit bats in Malaysia and Mexico, respectively, and members of the order Chlamydiales, isolated from a common bat species (Myotis daubentonii) in Finland [9].

PCR and sequencing results for Chlamydiales
Phylogenetic analysis of the sequences obtained in the present study resulted in the formation of two groups: Chlamydiaceae-like and Chlamydiales-like. Separation in these groups had been previously described in the study from Hokynar et al. [9]. However, a precise classification according to the scheme from Pillonel et al. [24] was not possible based on short amplicons retrieved in this study and also because only 16S and 23S rRNA gene products were sequenced. The phylogenetic trees showed that five samples from group one (Sample IDs F18-0155.54, F18-0155.98, E 148/07, E 161/07 and E 18/07) are closely related to the samples from Hokynar et al. and four of them (Sample IDs F18-0155.54, F18-0155.98, E 148/07 and E 161/07) form a novel clade within the Chlamydiaceae family. These five samples originated from free-living bats from Switzerland and Germany belonging to the bat species Myotis myotis, Pipistrellus pipistrellus, Nyctalus noctula and Eptesicus serotinus. Therefore, it can be said that this clade first discovered in Finland occurs also in other parts of Europe and colonizes various species of free-living insectivorous bats. In the present study, these Chlamydiae were detected directly in the inner organs; therefore, these bats might actually act as a host, meaning that the Chlamydiae are not merely prey-borne, as hypothesized by Hokynar et al. The latter study only investigated bat feces and insect material, inner organs were not available. Although the present study indicates the presence of chlamydial DNA in inner organs, such as the intestine, detection of replicating chlamydial organisms could not be assessed; Detection of replication would have required the attempt of isolation, which was hampered by the unavailability of fresh frozen tissue material. Chlamydial DNA might originate from the diet or represent colonization and/or real infection [9]. If the latter holds true, this would suggest that the different chlamydial species documented in bats might have a pathogenic role for the bats or that these bats act as vectors.
In a study by Hornok et al. [28], 196 individual and 25 pooled fecal samples collected from 19 bat species from Hungary and the Netherlands were investigated for the presence of chlamydial DNA; they all tested negative indicating that the prevalence of Chlamydiae in bats is highly variable in studies from neighboring countries. This observation suggests that, although this new Chlamydiae clade may occur in some parts of Europe, colonization or infection of bats is influenced by factors that remain to be defined. Methodological differences may play a part of these factors.
Also a Rhabdochlamydiaceae-like sequence was identified in a sample (E 18/07) taken from a free-living bat from Costa Rica, which formed another novel clade with bat samples from Hokynar et al. [9] in the phylogenetic tree (E18_07). The remaining bat samples, similar to chlamydial sequences from environmental samples retrieved from the NCBI database, originated from captive and freeliving bats from Switzerland and Germany. Chlamydialike organisms are commonly present in water and inside amoebae [29][30][31], and have also been observed in ticks [32,33] and fleas [34]. Thus, it is difficult to guess, which of these chlamydial putative vectors (amoebae or ticks) is the more likely at play as reservoir and vector.
In previous studies from Chua et al. [7] and Piérle et al. [8], two novel Chlamydia-like organisms belonging to the Waddliaceae family and named Waddlia malaysiensis and Waddlia cocoyoc were isolated from fruit bats from Malaysia and the municipality Cocoyoc in Mexico. These regions have a tropical climate, whereas a temperate climate predominates in Switzerland and Germany. In this study, bats from a tropical climate (Costa Rica) were analysed as well, but compared to the two other studies (n = 206 and n = 38) our sample size from Costa Rica (n = 17) was small. Another reason for the absence of Waddliaceae in bats of this study might the different sample type. Waddlia malaysiensis was first detected in urine samples from bats [7] but we did not investigate urine in this study. Waddlia cocoyoc was detected in DNA from the skin and in infected Vero and BHK 21 cells, but also caused severe lesions in lungs and spleen [8]. Therefore, it can be assumed that bacterial DNA would have been detected in these two organs if bacteria were present. The two isolates from Mexico and Malaysia might be bound to the regions and/or the climate.
In this study as well as in the studies from Hokynar et al. and Chua et al., only short fragments of the 16S rRNA and 23S rRNA genes were amplified, as many of the samples in the present study were only available as FFPE samples because the bats had died in the wild and were subsequently collected and sampled, therefore preventing detailed classification.

PCR and sequencing results for Mycoplasmatales
In the current study, three sequences were closely related to uncultured Mycoplasma sp. and nine sequences were closely related to Proteobacteria. The amplification of Proteobacteria-like sequences might be the result of unspecific primer binding yielding false-positive PCR amplification.
The three sequences related to uncultured Mycoplasma sp. originated from the samples of free-living bats from Germany and Costa Rica belonging to the bat species Nyctalus noctula, Vespertilio murinus and Glossophaga commissarisi. Consequently, different species from the bat families Vespertilionidae and Phyllostomidae can harbor hemotropic mycoplasmas. The presence of hemoplasmal DNA could be only confirmed in freeliving bats, which is in agreement with the presumed transmission pathways of hemotropic mycoplasmas through blood, saliva and arthropods [10,[12][13][14]. The captive bats were isolated from the environment and thus arthropod contact was at least partially excluded or reduced. Moreover, none of them were sanguivory bats and therefore do not bite other animals living in the enclosure. Consequently, transmission by blood, saliva and arthropods is very unlikely in captive bats.
The phylogenetic tree shows a close relationship between the sequences of the German bats among each other (E17307 and E19007) and a sequence from a Chilean bat [25]. The sequence from Costa Rican bat (E7006), on the other hand, was more closely related to sequences of vampire bats from Peru and Belize [23]. Vampire bats and the Commissaris's long-tongued bat (Glossophaga commissarisi) belong to the family of leafnosed bats and Peru, Belize and Costa Rica are geographically close to each other strengthening the high degree of relationship in the phylogenetic tree.
In other studies by Ikeda et al. [20], Mascarelli et al. [21], Millan et al. [22] and Volokhov et al. [23], a much higher mycoplasmal prevalence (18.5 to 97%) was found than in the present study (0.7%), despite partially similar numbers of samples. A similar prevalence (1.8%) was only reported in the study by Hornok et al. [28]. Millan et al. suggested that subclinical infections with mycoplasmas in bats are common. This may be related to distinct climatic conditions and thus different arthropod activity, which are suspected to act as vectors of hemotropic mycoplasmas. In all studies, only fragments of the 16S rRNA encoding gene were sequenced and then combined into contigs to optimize the chances of getting a larger 16S rRNA encoding gene sequence for further analysis. This study gives insight into the diversity of hemotropic mycoplasmas in different bat species from three geographical regions, although a more detailed description of the sequences and a more precise taxonomical classification were not possible.
Bats belong to the superorder of the Laurasiatheria and therefore are more closely related to carnivores, even-toed and odd-toed ungulates, than to rodents and primates, which belong to the superorder of the Euarchontoglires. While the relationship of other hemotropic mycoplasmas represented in the phylogenetic tree (Fig.  2) roughly corresponds to that of the mammalian classification, Mycoplasma of bats are the only ones to classify quite differently. This is likely due to the high resistance of bats to pathogenic microorganisms, somehow similar to some rodents (mice, rats). Moreover, this suggests that bats do not get infected by exposure to meat, but rather by exposure to water (and free-living amoebae) as well as to ectoparasites colonizing bats (such as Spinturnix). The latter hypothesis is supported for Chlamydialike organisms by the recent work of Thiévent et al. that demonstrated the occurrence of Chlamydia-related bacteria in several Spinturnix species [35].

Conclusions
Chlamydiaceae, Chlamydia-like organisms and hemotropic mycoplasmas have been identified in humans and various animal species but only a few studies investigated their presence in bats. This study assessed the prevalence and phylogeny of chlamydial and hemoplasmal DNA in captive and free-living bats from Switzerland, Germany and Costa Rica. Chlamydial and hemoplasmal sequences, similar to sequences obtained from bats and their prey investigated in other studies and from environmental samples, were identified. Newly described sequences indicate that the diversity of these bacteria in bats is broader than previously thought. Neither Chlamydiales nor hemotropic mycoplasmas are restricted to certain bat species or countries, and captive and free-living bats can be colonized.

Bat sampling and DNA extraction
A total number of 475 bats belonging to six bat families and 28 bat species were sampled (Table 1). Lungs, liver, intestines and spleen from 89 captive bats from Switzerland, 28 captive and 55 free-living bats from Germany, and 17 free-living bats from Costa Rica were investigated. Thereof, bat samples from Germany and Costa Rica were obtained from the Department of Wildlife Diseases, Leibniz Institute for Zoo and Wildlife Research, Berlin, Germany and samples from Switzerland were obtained from the Institute of Animal Pathology, Vetsuisse-Faculty, University of Bern, Switzerland, from the Stiftung Papiliorama, Kerzers, Switzerland and from Dr. Danja Wiederkehr, collection curator from the Swiss bat preservation organization. The intestines were selected based on a previous study [9], where Chlamydiae were detected in bat feces. For this study, jejunum was available from the captive bats of Switzerland. Small and large intestines from Swiss free-living bats, captive and free-living bats from Germany and free-living bats from Costa Rica were pooled for DNA extraction. The spleen was selected as target organ for hemotropic mycoplasma detection. Lung and liver were selected because these are common target organs for chlamydial infections in mammals and birds. Moreover, we were interested to investigate the systemic (hematogenous) spread of Chlamydiae in these organs.
DNA extraction from bat samples from Switzerland was performed using the DNeasy Blood and Tissue Kit #69506 (Qiagen, Hilden, Germany) following manufacturer's instructions, whereas DNA extraction of the samples from Germany and Costa Rica was performed using either the NucleoSpin® DNA RapidLyse (lung, liver, spleen) or the NucleoSpin® Tissue Kit (intestine) (MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany). Additionally, inner organs including lungs, liver, intestines and spleen, from 285 free-living bats from Switzerland were obtained either as FFPE blocks or fixed in 4% formalin. The DNA of the FFPE blocks was extracted directly, while the organs fixed in 4% formalin were first embedded in paraffin according to routine procedures. The DNA extraction was then performed using the QIAamp DNA FFPE Tissue Kit #56404 (Qiagen) following manufacturer's instructions. DNA quantity and quality of all samples was evaluated by spectrophotometry with the Nanodrop-1000 (Witec AG, Luzern, Switzerland).

PCR analysis for Chlamydiales DNA
A total of 1021 DNA samples was screened for the presence of Chlamydiales DNA using two different real-time PCRs targeting sequentially the 23S rRNA gene (Chlamydiaceae family-specific) and the 16S rRNA gene (pan-Chlamydiales order-specific).
The Chlamydiaceae-specific real-time PCR targeting the 23S rRNA gene (Chlam23S quantitative polymerase chain reaction, qPCR) [36] used primers Ch23S-F, Ch23S-R and probe Ch23S-p (Microsynth, Balgach, Switzerland) described by Ehricht et al. [37]. The internal amplification control eGFP amplified with primers eGFP-1-F, eGFP-10-R and probe eGFP-Hex (Microsynth) was added to each reaction [38]. The PCR was conducted on a Thermocycler 7500 Fast ABI (Thermo Fisher Scientific). All samples were tested in duplicate and samples with a cycle threshold of < 38 in duplicate PCR reactions were considered positive. Quantitation was performed using 10-fold dilutions (10 7 copies to 10 copies/μL) of the Chlamydia abortus genomic DNA positive control (standard curve).
For the pan-Chlamydiales real-time PCR targeting a partial sequence of the 16S rRNA-encoding gene (16S-pan-qPCR) [41], primers 16S-panCh-F, 16S-panCh-R and probe 16S-panCh (Eurogentec, Seraing, Belgium) were applied in a StepOne Plus real-time PCR system (Thermo Fisher Scientific). All samples were tested in duplicate, and if a single replicate was positive (Ct ≤ 37), the corresponding sample was considered positive. Quantification was performed using a 10-fold-dilution of a plasmid control tested in duplicate, constructed with the sequence of interest based on the Parachlamydia acanthamoebae 16S rRNA encoding gene, cloned with the TOPO TA Cloning Kit for Subcloning with One Shot TOP10 chemically competent E. coli cells (Thermo Fisher Scientific). Molecular-biology-grade water was used as a negative control in all PCR reactions.
All PCR primers and probes, the targeted genes and amplicon sizes used in this study are summarized in Table 3. All reaction mix compositions and cycling conditions are shown in Table S1.

PCR analysis for hemotropic mycoplasma DNA
All spleen and FFPE block samples (n = 475) from 462 bats were screened for the presence of hemoplasma DNA using a universal hemotropic mycoplasma-specific SYBR Green real-time PCR (Hemoplasma SYBR Green qPCR) targeting the 16S rRNA encoding gene, including primers Mhae_sybr.359f, Mcocc_sybrF, Mhae_sybr.432r and Cmhae_Sybr.493r (Microsynth) [42]. Samples within the same melting temperature range as the positive controls were considered positive.
Samples positive by SYBR Green real-time PCR were then further analysed by using real-time PCRs specific for M. haemofelis-like and 'Candidatus M. haemominutum'-like organisms and two conventional PCRs targeting an 871-bp and 1030-bp region of the 16S rRNA encoding gene.
'Candidatus Mycoplasma haematoparvum' was used as a positive PCR control.
The conventional PCR targeting an 871-bp fragment of hemoplasmal 16S rRNA encoding gene (HemMycop41/ 938-PCR) was performed using primers HemMycop16S-41 s and HemMycop16S-938as (Microsynth). Primers HemMycop16S-322 s and HemMycop16S-1420as (Microsynth) were applied for the conventional PCR targeting a 1030-bp fragment of the hemoplasmal 16S rRNA encoding gene (HemMycop322/1420-PCR) [21]. Molecularbiology-grade water was used as a negative PCR control and DNA from both Mycoplasma wenyonii and Mycoplasma haemocanis were used as positive PCR controls. Cycling was performed on a Biometra T-personal Thermal Cycler (Biolabo Scientific Instruments, Châtel-Saint-Denis, Switzerland) and PCR products were analysed by gel-electrophoresis on a 1.5% agarose gel.
All PCR primers and probes, the targeted genes and amplicon sizes used in this study are summarized in Table 3. All reaction mix compositions and cycling conditions are shown in Table S1.
Sequencing and analysing of Chlamydialesand Hemoplasma-positive PCR products For sequencing, amplicons of samples positive by conventional PCRs 16S-IGF/IGR-PCR, 16S-pan-PCR, 23SIG-PCR, HemMycop41/938-PCR or HemMycop322/ 1420-PCR were purified using the GeneJET PCR Purification Kit (ThermoFisher Scientific) or the GeneJET Gel Extraction Kit (ThermoFisher Scientific) according to manufacturer's instructions. The forward and reverse strands of the PCR products were sequenced using the respective primers of the positive PCR reaction. Microsynth performed all sequencing reactions using Sanger sequencing.
Amplicons of samples positive by 16S-pan-qPCR with a Ct value ≤35.0 were a) either purified using an MSB Spin PCRapace kit (Invitek, Berlin, Germany) with a subsequent PCR reaction using a BigDye Terminator v1.1 cycle sequencing kit (Applied Biosystems) [41] or b) purified and sequenced by Microsynth using specifically designed inner primers panFseq and panRseq.
Sequence traces were visualized and assembled in CLC Genomics Workbench v10.1.1; assemblies were compared to known sequences in the NCBI database by BLAST analysis. Phylogenetic analyses on the assemblies and related database sequences were performed using muscle with default parameters in Seaview (44) with manual correction where necessary to create alignments, and using PhyML in Seaview with default parameters including GTR model, and 100 bootstrap replicates, to create phylogenetic trees.