A novel Coltivirus-related virus isolated from free-tailed bats from Côte d’Ivoire is able to infect human cells in vitro

Bats were captured with mist nets in November and December 2006 in the Taï National Park in Côte d’Ivoire. Blood was taken from the wing vein and mixed with EDTA to prevent coagulation. Animals were released immediately after the procedure. All samples were stored in liquid nitrogen, transported on dry ice and transferred to −80 °C at the Robert Koch-Institut in Berlin for long-time storage. Primary virus isolation was achieved on VeroE6 cells. For inoculation 50 μl of blood from each of three individuals Ch. aloysiisabaudiae was pooled. Cells were grown to 70–80% confluence in 25 cm² cell culture flasks. Prior to inoculation cells were washed with phosphate buffered saline (PBS) followed by medium without supplements. Cells were inoculated with blood diluted in PBS to a final volume of 2 ml. After 1–2 h 3 ml of medium including 2% Fetal Calf Serum, 1% L-glutamine, 1% penicillin/streptomycin was added and cells were incubated at 37 °C with 5% CO₂. After 24 h cells were washed with PBS and medium was replaced. Cells were then incubated for 7 days and monitored daily microscopically for the presence of a cytopathic effect (CPE). After 7 days 300 μl supernatant was used to infect fresh cells. A CPE was observed at day five in the second passage. To prepare virus stocks virus was grown in a 175 cm² flask and virus particles isolated by ultracentrifugation through a 36% sucrose solution for 4 h at 28,000 rpm at 4 °C in a Beckmann ultracentrifuge (rotor SW 32 Ti, Beckman Coulter).

For visualization of viral particles cells grown in a 25 cm² flask were fixed with glutaraldehyde (final concentration 2.5%) over night at 4 °C as soon as a CPE became visible. Photographic documentation of samples was performed on a FEI Tecnai G2 transmission electron microscope (TEM).

Viral RNA was isolated using the QIAamp Viral RNA Mini Kit (Qiagen) without carrier RNA and initial sequence information was generated by Particle-associated Nucleic Acid PCR [24] and subsequent cloning and Sanger sequencing of the amplicons. Of the resulting sequences nine contigs and three individual reads showed homologies with two viruses, CTFV and EYAV, when compared to GenBank entries. The CTFV genome was used as reference genome to map contigs of the novel virus so as to get a possible arrangement, based on which sequence gaps were closed with primers directed outwards from known fragments (described in detail in [25]). Segment ends of three identified segments (S1, S2, S9) were obtained by rapid amplification of cDNA ends (Thermo Fisher). In parallel, RNA was applied to 454 next generation sequencing (Roche), resulting in a total of 222,953 reads. Reads were trimmed using Trimmomatic version 0.33 [26] and all trimmed reads shorter than 250 bases were discarded. The resulting 135,911 reads were separated into pathogen and host reads using RAMBO-K version 1.2 [27] (background reference: Chlorocebus sabaeus, GCF_000409795.2; foreground references: Eyach virus, NC_003696-NC_003707, and Colorado tick fever virus, NC_004180-NC_004191) using a k-mer length of 4 and a cutoff score of −2.5, yielding 71,962 potential foreground reads. These reads were assembled together with the available Sanger sequences using SPAdes version 3.8.1 [28]. Of the 250 contigs reconstructed by SPAdes, only 38 were longer than 300 bases. BLASTX of these 38 contigs was performed against a custom database of the Eyach virus and Colorado tick fever virus genomes using Geneious Pro version 10 [29], with 10 contigs showing significant similarity to 9 known segments. In an iterative process, all reads were mapped back to the contigs, and the mapping reads were re-assembled, until no new unambiguously mapping reads could be identified and contigs could not be extended anymore. Finally, a last round of mapping all reads to all contigs was performed for error correction.

To identify putative proteins encoded by the different segments possible open reading frames (ORFs) were identified using the ORF finder in Geneious Pro. Putative gene functions were predicted based on homology to CTFV and the according UniProtKB entries [30].

For phylogenetic inference sequences of RNA dependent RNA polymerase (RdRp) genes from representative viruses across different genera within the Spinareoviridae were downloaded and aligned on protein level using the muscle algorithm [31] as implemented in SeaView version 4.6.1 [32]. To increase alignment quality conserved blocks were selected using the Gblocks server [33], which resulted in an alignment of 263 amino acids (aa). Maximum likelihood (ML) tree was achieved using the PhyML 3.0 server [34] using the RtREV +G + F model as determined by Smart Model Selection as implemented on the website. Sequences of available putative RNA-methyltransferase proteins from reoviruses were used for phylogenetic inference using the same pathway as described above. The resulting alignment consisted of 667 aa and the model used for phylogenetic inference was LG + G + F.