Identification of an Unclassified Paramyxovirus in Coleura afra: A Potential Case of Host Specificity

Bats are known to harbor multiple paramyxoviruses. Despite the creation of two new genera, Aquaparamyxovirus and Ferlavirus, to accommodate this increasing diversity, several recently isolated or characterized viruses remain unclassified beyond the subfamily level. In the present study, among 985 bats belonging to 6 species sampled in the Belinga caves of Gabon, RNA of an unclassified paramyxovirus (Belinga bat virus, BelPV) was discovered in 14 African sheath-tailed bats (Coleura afra), one of which exhibited several hemorrhagic lesions at necropsy, and viral sequence was obtained in two animals. Phylogenetically, BelPV is related to J virus and Beilong virus (BeiPV), two other unclassified paramyxoviruses isolated from rodents. In the diseased BelPV-infected C. afra individual, high viral load was detected in the heart, and the lesions were consistent with those reported in wild rodents and mice experimentally infected by J virus. BelPV was not detected in other tested bat species sharing the same roosting sites and living in very close proximity with C. afra in the two caves sampled, suggesting that this virus may be host-specific for C. afra. The mode of transmission of this paramyxovirus in bat populations remains to be discovered.

Bats have a close evolutionary relationship with several genera of mammalian paramyxoviruses [11]. Otherwise, bat-borne paramyxoviruses are in close relationship to known paramyxoviruses of mammalian. These small mammals are known to harbour a broad diversity of PVs, including emergent henipaviruses (Nipah virus and Hendra virus) and rubulaviruses [Menangle virus, Tioman virus, Mapuera virus, and Tuhoko virus 1, 2 and 3 (ThkPV-1, ThkPV-2 and ThkPV-3)]. A very broad diversity of paramyxoviruses, including Henipa-, Rubula-, Pneumo-and Morbilli-related viruses, have been detected in six of ten tested bat families [11]. Whereas most of the viruses identified in bats do not seem to cause clinical disease in these animals, there have been reports of rabid bats [12,13] and of unusually large numbers of animals succumbing to infection by rabies virus [14].
As part of a large-scale investigation of viral diversity in bats and of associated zoonotic risks, we have previously detected a bat paramyxovirus in one insectivorous African sheath-tailed bat (Coleura afra) [11], exhibiting several hemorrhagic lesions at necropsy. We therefore examined occurrence of this bat paraymxovirus in other bats.

Ethics statements
The study was conducted in the Belinga mountains (northeast Gabon), where Ebola outbreaks occurred in 1994-1996 and 2001-2002. All the capture events, animal handling, euthanasia and transfer of samples across country borders were Gabon. There are no patents, products in development, or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed in the online guide for authors.
performed in accordance with the guidelines of the American Society of Mammalogists (http://www.mammalsociety.org/committees/animal-care-anduse) [15]: bats were captured following recommendations by Kunz and Parsons [16] and identified by trained field biologist. Captured bats were removed carefully from nets as soon as possible to minimize injury, drowning, strangulation, or stress. Safe and humane euthanasia was achieved through the use of inhalant anaesthetic (halothane) prior to autopsy.
All work (capture, euthanasia and autopsy) was carried out with authorization from the Gabonese Ministry of Water and Forestry (Département de la

Bat sampling
Samples of liver, spleen, kidney, lung, heart, gut, brain and salivary glands were collected, stored in liquid nitrogen and transferred to the CIRMF laboratory (Centre International de Recherches Médicales de Franceville, Gabon), where they were stored at 280˚C until analysis. Blood samples were also collected, except from the smallest individuals (body mass ,12 g). A total of 985 bats (Table 1)  Interestingly, at autopsy, only one C. afra out of the 26 individuals, from Batouala cave, exhibited diarrhea and severe hemorrhagic lesions in both thoracic and abdominal organs, along with lung congestion and pleurisy.

Viral PCR screening
Virological screening of the diseased bat was performed to identify the cause of these lesions. The screening process included the whole Paramyxoviridae family and also filoviruses [Marburg virus (MARV) and Zaire Ebola virus (EBOV)], given the nature of the lesions and the known filovirus tropism for bats. Briefly, approximately 100 mg each of this animal's liver and spleen were pooled and crushed in 600 ml of cold PBS in a ball-mill tissue grinder (Genogrinder 2000, Spex Centripep). Total RNA was extracted using a Biorobot EZ1 and the EZ1 RNA tissue mini kit (Qiagen) according to the manufacturer's guidelines. The RNA was then tested for paramyxoviruses, using three heminested reverse transcription-PCR (hnRT-PCR) assays targeting the polymerase gene [17], and also for Marburg virus [18] and Zaire Ebola virus [19]. However, the screening was extended to Arenaviruses, Flaviviruses, Alphaviruses, polyomaviruses, Orthopoxviruses, Parapoxviruses, Influenza viruses, Lyssaviruses and Rhabdoviridae family.

Paramyxovirus detection in bat populations
To further investigate the presence of the virus in bat populations, a strain-specific real-time RT-PCR assay (primers: GB09-478-F, 59-GGCGGCTCTTAAAAGT-GAATG-39; GB09-478-R, 59-GCGGGGTCAAATTGGTCAT-39; probe: GB09-478-P, 59-TCCAGCACAAACATATCCGAGAAGGCTAG-39) was designed within the initial PCR fragment and was used to test total RNA extracted from mixed liver and spleen samples from each of all the other bat species. The amplification was performed in a final volume of 25 ml, containing, 12.5 ml TaqMan R 2X PCR Master Mix (Applied Biosystems), 0.5 ml each primer and probe (10 mM), 1 ml bovine serum albumin (1 mg/ml) (Invitrogen), 5 ml cDNA and RNAse-free water (Invitrogen). Amplification generally involved 2 min at 55˚C, 10 min at 95˚C followed by 45 cycles of 15s at 95˚C and 1 min at 58˚C.

Virus distribution in organs of infected bats
In order to determine the organ distribution of this virus in infected bats, total RNA was extracted from heart, liver, spleen, kidney, lung, intestine and brain samples from all 14 real-time RT-PCR-positive bats, as described previously, and screened, using the same strain-specific real-time RT-PCR assay shown above.

Entomological study
Transmission by hematophagous arthropods was studied. Hematophagous arthropods were collected inside Faucon and Zadié caves. Mosquitoes were sampled using CDC light traps whereas bat-flies were manually collected on bats (Coleura afra, Miniopterus inflatus, Hipposideros cf. ruber and Rousettus aegyptiacus) trapped in the Faucon and Zadié caves between November and December 2009, June 2010, and January, March and April 2011. After the morphologic species determination, insects were crushed by monospecific pools (up to 10 specimens for mosquitoes, and between 1 and 5 for bat-flies) in 300 ml PBS. Total RNA was extracted using 100 ml of the supernatant from each pool by using the RNeasy Mini Kit (Qiagen) and then was tested with the specific realtime RT-PCR assay described above.

Blast and phylogenetic analysis
The viral sequences obtained were first compared to those available in the public database using the algorithm "Blastn" from BLAST program [20] and then aligned with homologous sequences of Paramyxoviridae reference strains from GenBank, using the MEGA program version 5 [21]. Bayesian inference of phylogeny was done using MrBayes V.3.2 software and the GTR+G+I nucleotide substitution model [22] for two million generations with a burn-in of 25%.

Identification and molecular characterization
One of three hnRT-PCR, Respiro-, Morbilli-and Henipaviruses PCR (RMH-PCR) assays yielded an amplicon of 559 base pairs (bp) from diseased bat, in collaborative work with the Bonn Institute of Virology Bonn (Germany). The PCR product was sequenced with dye terminator chemistry (Applied Biosystems).
No other virus was detected in the diseased bat.
Blast and phylogenetic analysis confirmed that the sequence from diseased bat, designated BatPV GB09 478 (Genbank accession number HQ660155), belonged to the Paramyxoviridae family. It showed 65% and 66% nucleotide identity, respectively, with JPV and BeiPV, and had pairwise nucleotide identities of 36-42% with the Henipavirus group and 37-40% with the Morbillivirus group. Phylogenetic analysis indicated that this bat PV, that we named Belinga bat virus (BelPV), clustered with unclassified paramyxoviruses and was located between the genera Morbillivirus and Henipavirus (Fig. 1).
In this study, among all animals tested by the strain-specific real-time RT-PCR assay only 14 C. afra bats, including the diseased animal, were positive. Bat samples positive by specific real-time RT-PCR were then tested with RMH-hemi nested-PCR for molecular characterization. Another 439-bp PCR product was amplified from one liver-spleen pool from a second C. afra specimen, and was sequenced. These two phylogenetically related sequences (Fig. 1), designated BatPV GB09 478 and BatPV GB09 450, displayed 100% of nucleotide identity.

Prevalence and organ distribution of BelPV in Coleura afra
BelPV-specific RNA was detected only in C. afra, with a prevalence of 14.9% (14/ 94 of the C. afra individuals sampled from the Faucon and Batouala caves) ( Table 1). Interestingly, no bats belonging to the other five species tested positive.
BelPV RNA was detected in the heart (8/12) and liver (5/14), at low and variable loads ( Table 2 and Fig. 2). No significant difference in the BelPV RNA detection rate was found between the two organs (x 2 52.476, df51, p.0.20). In the bat exhibiting severe hemorrhagic lesions (GB09 478), both the liver and heart were BelPV-positive, with a higher viral load in the heart (Ct value 528).

Discussion
The BelPV nucleotide sequence obtained showed similarity with the JPV and BeiPV sequences. BelPV has previously been reported to hold a phylogenetic position between the genera Henipavirus and Morbillivirus. The same phylogenetic position had been observed with MosPV and J-V [4].
In this study, organs with high BelPV concentrations are different from those found with high paramyxovirus concentrations in pteropodids and microchiroptera bats. In microchiroptera bats from Brazil, spleen has been found more positive than the others organs with highest viral load, as in Eidolon helvum (megachiroptera bat) in Africa [11]. However, in our study majority of spleen were not available.
The within-host BelPV distribution tended to be organ-specific. BelPV seemed to be restricted to the heart and liver. In contrast, JPV has been isolated from blood, lung, liver, kidney and spleen of experimentally infected laboratory mice [6] but not in heart. The BelPV distribution for the heart and liver, together with the high viral load in heart tissue, could suggest that this virus is likely to be present in the bloodstream and might thus be transmitted during aggressive contacts between bats, or by blood-sucking vectors. Nethertheless, viremia was  not proven. BelPV RNA was not searched from blood because in these small species of bats blood was difficult to collect in the field. We detected BelPV only in Coleura afra and not in other bat species sharing the same roosting sites and living in very close proximity in the two caves sampled. However, it has been shown that bats of different species occupying the same roosting sites can share the same viruses. Marburg virus had been detected in Rousettus aegyptiacus and Hipposideros sp. bats living in Kitaka cave in Uganda [23] and Miniopterus inflatus and Rousettus aegyptiacus bats caught in Goroumbwa mine in the Democratic Republic of the Congo [24]. These bat species are known to live in close proximity. Thus, virus transmission between different bat species is possible [25]. Thus, we can speculate that the failure to detect BelPV in other bat species sharing the same caves would suggest that this virus has strong host specificity for C. afra, as well as restricted intraspecies transmission. Henipaviruses occur naturally in fruit bats belonging to the genus Pteropus [26], and this also appears to be true of severe acute respiratory syndrome-like coronaviruses in Rhinolophus bats [27,11].
In view of our data we can assume that BelPV might have pathogenic potential for its host C. afra. Indeed, high viral load was detected in the heart of the diseased bat, and the lesions were consistent with those reported in wild rodents and mice experimentally infected by JPV [6].
Although BelPV RNA was also detected in asymptomatic bats, pathogenicity may appear in long term under some immunological and/or ecological conditions. Indeed, virus must not induce pathology to persist or adapt within its reservoir host. Many authors suggested that persistence in the absence of pathology or disease appears to be a common characteristic of bat viruses in their natural host population [28,29]. However, a severe immunodepression for instance, may increase the risk of infection with opportunistic pathogens. Under some environmental conditions (cool environments for example), some avirulent pathogens, such as Geomyces destructans, causative agent of white-nose syndrome, may become pathogenic in hibernating bats in North America [30,31]. Nevertheless, infection by BelPV may be mild for bats and thus the pathology observed not directly related. Otherwise, it may also be that this animal had an underlying disease or infection with a different pathogen. Even in this case, we might not draw any conclusions neither establish a link with lesions seen. Therefore, the pathogenicity of the BelPV should be demonstrated by experimental animal infection. Otherwise, viral antigens or RNAs should be detected histologically in the lesions of naturally-infected bats. However, the unavailability of biological tissues from the diseased bat failed to perform these analyzes. Consequently, other captures of Coleura afra species are considered in order to find BelPV again for further studies (pathogenicity to its host, isolation and complete genome characterization). However, Coleura afra is a migratory species living in colonies of several hundred individuals. In Gabon, this species, which has been recently described, is not present all year round in the caves of the north-east of the country, making the studies on this species difficult and thereof partly explaining the lack of virological studies. Some viruses appear to cause clinical disease in wild-living bats; these include lyssaviruses and an ebola-like filovirus named Lloviu virus [32,33,34].
Bats are the natural reservoirs for many viruses, including emerging zoonotic viruses such as SARS-CoV [27], Hendra and Nipah viruses [26,35], Ebola virus [36], Marburg virus [23,37], rabies virus and other Lyssaviruses [11]. In general, humans are infected through an intermediate amplifying host such as palm civets for SARS-CoV, horses for Hendra virus and pigs for Nipah virus [38]. However, in humans Nipah virus outbreaks linked to bats exposure have been reported [39]. It remains to be shown whether the BelPV reported here presents a zoonotic risk. Nonetheless, like most RNA viruses, for example coronaviruses, characterized by high mutation and/or recombination rates [40], PVs may adapt to novel hosts, including humans. A serological test capable of detecting antibodies to this virus in human populations living in the vicinity of these animals is needed to assess zoonotic potential.
All the blood-sucking arthropods collected from bats, as well as mosquitoes collected in the caves where bat sampling took place, were negative for BelPV, in keeping with the lack of known PV vectors [41]. However, BelPV transmission by blood-sucking vectors within the Gabonese population of C. afra cannot be ruled out. Indeed, a haemosporidian parasite (Polychromophilus) was found in a blood parasite vector (Penicillidia fulvida) in Faucon cave in Gabon in 1977 [42] and also in its host M. inflatus (greater long-fingered bat) from the same cave in 2010 and 2011 [43]. In addition, the methodology used to collect flying hematophagous insects (based on light traps) possibly introduced a bias by selecting only those attracted by light. Therefore, we can not exclude that additional sampling techniques could increase the number of mosquitoes species or groups known to colonize caves such as sandflies or biting midges. Hence, the natural mode of transmission of this unclassified paramyxovirus in bat populations, through batbat aggression for example, remains to be determined.
This association between C. afra and BelPV could serve as an interesting model, (i) to evaluate modes of transmission within host populations, (ii) to study hostvirus interactions (pathogenesis and host specificity), and (iii) to evaluate the zoonotic risk of a newly identified virus.
Further studies of C. afra populations and a broader diversity of arthropod vectors, spanning larger areas and time scales, are needed to confirm this apparent host-virus specificity, and to determine the modes of BelPV transmission. Further studies are needed to characterize complete BelPV genome and demonstrate the pathogenicity of this virus for its host Coleura afra.

Author Contributions
Conceived and designed the experiments: GDM MB EML. Performed the experiments: GDM JON NN. Analyzed the data: GDM JFD. Contributed reagents/ materials/analysis tools: GDM MB CD CP JFD EML. Wrote the paper: GDM CP EML SM.