Figure 1

Figure 1. Bat collection sites for a study on genetically diverse filoviruses in Rousettus and Eonycteris spp. bats in China. Triangles indicate Jinghong City and Mengla County, Yunnan Province, where 150 apparently healthy...

We captured 150 apparently healthy adult bats from 2 caves in Yunnan Province, China: 1 in Jinghong City in November 2009, and 1 in Mengla County in December 2015 (Table 1; Figure 1). The bat species we collected were Hipposideros armiger, Aselliscus stoliczkanus, Myotis ricketti, Rhinolophus Monoceros, Miniopterus fuscus, Ia io, Eonycteris spelaea, and Rousettus sp. We humanely killed all bats and collected their hearts, intestines, lungs, spleens, kidneys, livers, brains, and blood for testing. We used 2 methods to analyze bat lung tissues for presence of filovirus RNA: first, we used nested PCR with the primers FV F1/R1 and FV F2/R2 (10), and next, we used quantitative PCR (qPCR) with 3 groups of qPCR with primers and probes designed from viral sequences obtained in this study (Technical Appendix Table).

Figure 2

Figure 2. Phylogenetic analysis of filovirus isolates collected in study of genetically diverse filoviruses in Rousettus and Eonycteris spp. bats in China, compared with reference isolates. Analysis was based on a 310-bp segment...

Using degenerate nested PCR, we detected filovirus RNA in 15 fruit bat specimens (E. spelaea and Rousettus sp.); the specimens comprised 10 (23.3%) of 43 E. spelaea and Rousettus sp. collected in 2009 and 5 (11.9%) of 42 collected in 2015. Using qPCR, we detected filovirus RNA in 20 specimens from E. spelaea (n = 4) and Rousettus sp. (n = 16) bats: 10 (23.3%) of the bats were collected in 2009 and 10 (23.8%) in 2015. No filovirus was detected in other bat species studied (Table 1). The 310-bp L gene sequences (GenBank accession nos. KX371873–KX371890) exhibited 65%–99% nt identity among themselves and 61%–99% nt identity with known filoviruses. Phylogenetic analysis showed that the sequences from the bats formed 3 independent groups, groups 1–3. Groups 1 and 2 comprised 6 and 11 sequences, respectively, all of which were obtained in this study (Figure 2). Group 3 comprised 2 sequences, 1 from this study and 1 previously published (10). Pairwise distance analysis indicated that sequences in group 1 share the highest nucleotide identity (75%–78%) with MARV and those in group 2 share the highest identity (69%–74%) with Ravn virus. The 2 sequences in group 3 are highly similar and share 66%–70% nt identity with other filovirus species. Of note, 1 bat specimen (no. 9447) was co-infected with 4 different filovirus strains (BtFiloYN9447–1 to 9447–4) with high divergence (Figure 2; Technical Appendix Figure 1). To further determine the phylogenetic relationship of these viruses with known filoviruses, we amplified more L gene sequence (1,475 bp) for strains BtFiloYN2162 and BtFiloYN9447–1. Similar to the 310-bp sequences, the 1,475-bp sequence of the BtfiloYN2162 shared 99% identity with BtDH04 at the nucleotide level, the 1,475-bp sequence of BtFiloYN9447–1 shared 62%–71% with known filoviruses.

To determine the tissue tropism of these viruses, we performed qPCR with primers and probes designed for each of the 3 different groups (online Technical Appendix Table). Results showed that filoviruses were mainly located in the lung and that genome copy numbers ranged from 29 to 523,582/mg of tissue (Table 2). Only 2 bat blood samples (nos. 2202 and 9447) were positive for filovirus RNA; 5 samples (nos. 2202, 2188, 9434, 9442, and 9447) contained filoviruses with more widespread tissue tropism. We were unable to isolate virus from PCR-positive samples by using Vero-E6 cells.

To detect filovirus IgG and IgM, we expressed His-tagged truncated nucleoproteins from RESTV or ZEBOV in Escherichia coli and used them as antigens (online Technical Appendix). In this experiment, we used 25 bat samples from 2015 that had enough serum volume for testing; 14 samples showed a strong cross-reaction with the ZEBOV nucleoprotein, and among them, 7 showed a weak cross-reaction with RESTV nucleoprotein. We used Western blotting to confirm these results; 11 of the 25 samples were positive for ZEBOV nucleoprotein and 4 for RESTV nucleoprotein (Table 1; Technical Appendix Figure 2). No samples overlapped between those identified as positive by PCR and those identified as positive by serologic testing. Results of a serum neutralization assay with HIV pseudovirus carrying the ZEBOV glycoprotein showed that the ELISA-positive samples had no cross-neutralization activity to the pseudovirus (14).

We detected novel filovirus sequences with high divergence in E. spelaea and Rousettus sp. bats in China. Phylogenetic analysis of partial sequences suggested that at least 3 distinct groups of filovirus are circulating in fruit bats in China. The distances between these sequences indicates that the 3 groups may represent 3 novel species or genera. Of interest, we detected antibodies reacting more strongly to ZEBOV than RESTV nucleoprotein in some filovirus RNA–negative samples, suggesting that the bats were infected with another/other filovirus(es) cross-reactive with ZEBOV nucleoprotein or that nucleoproteins of the novel filoviruses were cross-reactive with ZEBOV and RESTV nucleoproteins.

The bat samples in this study were collected from 2 caves in 2009 and 2015, respectively; the caves are ≈200 km metric apart. Across the 2 different years and locations, we detected closely related viruses and found 1 bat that was acutely co-infected by 4 different filoviruses; this finding suggests that these viruses have been circulating in the 2 bat species and that densely populated bat caves provide opportunity for cross-infection with different viruses. However, considering the migration ability of the fruit bat, we cannot exclude the possibility that there are exchanges of virus between the bats in these two caves. Longitudinal surveillance with tracking tags may help to better understand the spatial–temporal distribution of these viruses in bat populations.

In previous reports, filoviruses were primarily detected in liver and spleen tissues (4,15). In our study, we primarily detected filoviruses in the lung. We suspect that lung tissues are the major target for these bat filoviruses. Thus, these filoviruses may have the potential to be transmitted through the respiratory tract.

These results will be helpful in providing a better understanding of the distribution and diversity of filoviruses, which may have implications for public health. Considering their feeding habitats, fruit bats are often in close contact with domestic animals and human populations. It is therefore necessary to establish long-term and proactive surveillance of these viruses and related diseases.