Discovery and Characterization of Bukakata orbivirus (Reoviridae:Orbivirus), a Novel Virus from a Ugandan Bat

While serological and virological evidence documents the exposure of bats to medically-important arboviruses, their role as reservoirs or amplifying hosts is less well-characterized. We describe a novel orbivirus (Reoviridae:Orbivirus) isolated from an Egyptian fruit bat (Rousettus aegyptiacus leachii) trapped in 2013 in Uganda and named Bukakata orbivirus. This is the fifth orbivirus isolated from a bat, however genetic information had previously only been available for one bat-associated orbivirus. We performed whole-genome sequencing on Bukakata orbivirus and three other bat-associated orbiviruses (Fomede, Ife, and Japanaut) to assess their phylogenetic relationship within the genus Orbivirus and develop hypotheses regarding potential arthropod vectors. Replication kinetics were assessed for Bukakata orbivirus in three different vertebrate cell lines. Lastly, qRT-PCR and nested PCR were used to determine the prevalence of Bukakata orbivirus RNA in archived samples from three populations of Egyptian fruit bats and one population of cave-associated soft ticks in Uganda. Complete coding sequences were obtained for all ten segments of Fomede, Ife, and Japanaut orbiviruses and for nine of the ten segments for Bukakata orbivirus. Phylogenetic analysis placed Bukakata and Fomede in the tick-borne orbivirus clade and Ife and Japanaut within the Culicoides/phlebotomine sandfly orbivirus clade. Further, Bukakata and Fomede appear to be serotypes of the Chobar Gorge virus species. Bukakata orbivirus replicated to high titers (106–107 PFU/mL) in Vero, BHK-21 [C-13], and R06E (Egyptian fruit bat) cells. Preliminary screening of archived bat and tick samples do not support Bukakata orbivirus presence in these collections, however additional testing is warranted given the phylogenetic associations observed. This study provided complete coding sequence for several bat-associated orbiviruses and in vitro characterization of a bat-associated orbivirus. Our results indicate that bats may play an important role in the epidemiology of viruses in the genus Orbivirus and further investigation is warranted into vector-host associations and ongoing surveillance efforts.


Introduction
Serological and virological evidence documents the exposure of various East African bat species to several arboviruses including Rift Valley fever, dengue, and yellow fever virus [1,2], however, little is known about the potential role of bats as arbovirus reservoirs or potential amplifying hosts. Orbiviruses (Reoviridae:Orbivirus) are 10-segmented, dsRNA, vector-borne viruses that cluster phylogenetically by arthropod vector group [3]. Recent research has also reported a novel orbivirus that may represent the first recognized insect-specific virus in the genus Orbivirus [4]. While mostly recognized as veterinary pathogens (i.e., bluetongue virus, African horse sickness), several orbiviruses have been associated with neurologic disease in humans [5][6][7][8].
Prior to this study, four orbiviruses had been isolated from wild bats. Japanaut virus (JAPV) was isolated from the blood of a southern blossom bat (Syconycteris crassa) and a pool of mixed culicine mosquitoes in the Sepik District, New Guinea in 1965 [9,10]. Heramatsu orbivirus was isolated from the blood of an eastern long-fingered bat (Myotis macrodactylus) trapped in a mine in Heramatsu, Kagoshima, Japan in 1965 [11]. The genome of Heramatsu orbivirus (isolate KY-663) has been partially sequenced [12]. Eight isolates of Ife virus (IFEV) were isolated from the blood and organs of straw-colored fruit bats (Eidolon helvum) in Nigeria, Cameroon, and the Central African Republic in 1971 and 1974 [13]. Gambian pouched rats (Cricetomys gambianus), African grass rats (Arvicanthis niloticus), and domestic ruminants in Nigeria were also found to be seropositive for IFEV [14][15][16]. Fomede virus (FOMV) was isolated from the brain, liver, and spleen of a dwarf slit-faced bat (Nycteris nana) in Kindia, Guinea in 1978, and has been repeatedly isolated from Nycteridae bats in Guinea [17][18][19]. Additionally, serologic evidence exists for exposure of Bolivian bats (genera Myotis and Noctilio) to Matucare virus, an orbivirus isolated from Ornithodoros ticks in 1963 [20]. Australian fruit bats were found to be seropositive to Elsey virus, a serotype of the mosquito-borne Peruvian horse sickness virus [21].
In 2013, a novel orbivirus, tentatively named Bukakata orbivirus (BUKV), was isolated from an Egyptian fruit bat (Rousettus aegyptiacus leachii (A. Smith, 1829)) (ERB) captured in Kasokero Cave, Uganda ( Figure 1). This is the fifth orbivirus isolated from bats; however, aside from partial sequencing of Heramatsu, no bat-associated orbivirus has been genetically characterized. The specific aims of this project were to 1) determine the genome sequence of the four bat-associated orbiviruses without published existing genetic information (JAPV, IFEV, FOMV, & BUKV) and conduct phylogenetic analyses to ascertain their potential arthropod associations, as orbiviruses cluster phylogenetically based on their arthropod vector, 2) determine the replication kinetics of BUKV in multiple vertebrate cell types, and 3) determine the prevalence of BUKV RNA in additional archived field samples from Uganda.  [2].  RNA was extracted using the MagMax 96 total RNA isolation kit (Applied Biosystems/Ambion, Austin, TX, USA). Eluted RNA from splenic samples was tested in a quantitative reversetranscriptase PCR (qRT-PCR) assay using TaqMan Fast Virus 1-Step Master Mix (ThermoFisher Scientific, Foster City, CA, USA). Each sample was run in duplicate using primers and probe (5′-3′) (F′: GCAGACTGTATCGCGGAAAG, R′: TAAGTTTCGCTTTCCTCCCGA, probe: CTGAAACTCGATCTCCGCAACGTTCTT) targeting the VP1 gene (RdRp) of BUKV and as single reactions using primers and probe targeting the GAPDH gene (F′: GTCGCCATCAATGACCCCTTC, R′: TTCAAGTGAGCCCCAGCC, probe: CCACCCATGGCAAGTTCAAAGGCACA) to ensure RNA integrity. Each reaction contained 5 µL splenic RNA, 5 µL TaqMan Fast Virus 1-Step Master Mix, 500 nM each primer, 250 nM probe, and 7.5 µL H2O. All samples were run on a QuantStudio 3 thermocycler using the following cycling parameters: 50 °C for 5 min; 95 °C for 20 s; 95 °C for 3 s, 60 °C for 30 s (40×).

Viruses and Cells
JAPV (MK 6357), IFEV (IbAn 57245), and FOMV (DakAnK 654) viruses were all sourced from the Arbovirus Reference Collection at the Centers for Disease Control Arbovirus Diseases Branch in Fort Collins, CO. BUKV (UGA432), isolated during this study, was deposited in the Arbovirus Reference Collection following isolation on Vero cells (African green monkey kidney epithelial cells) (ATCC CCL-81).

Bat Capture and Sampling
Bats were captured from multiple locations throughout Uganda during 2011-2013 [2]. Seventy-one additional combined liver and spleen tissue RNA samples from ERBs captured previously at Maramagambo Forest in 2009 were provided for inclusion in this study (Table 1, Figure 1). All bat captures were conducted under the approval of IACUC protocols 1731AMMULX (samples from Maramagambo Forest) and 010-015 (all other samples). Bats were captured using harp traps or mist nets. Sampling locations, numbers and species captured, and blood collection and serological results from these samples are described elsewhere [2,25]. All bats were euthanized according to approved IACUC protocols in accordance with AVMA guidelines to harvest tissues for virus isolations. Tissues collected included lung, intestine, liver, spleen, and oral and fecal swabs. Tissue sections were immediately placed in cryotubes and liquid nitrogen dry shippers. Duplicate aliquots of serum and liver/spleen were analyzed first for filovirus RNA at the CDC Viral Special Pathogens laboratory in Atlanta, GA USA. Filovirus-negative liver and spleen specimens were homogenized for virus isolation. Approximately 0.5 cm 3 sections of tissue were mechanically homogenized in a 2.0 mL snap cap tube containing 1 mL BA1 medium (Hanks M-199 salts, 0.05 M Tris pH 7.6, 1% bovine serum albumin, 0.35 g/L sodium bicarbonate, 100 U/mL streptomycin, 1 µg/mL Fungizone) and one or two stainless steel 5 mm beads in a Qiagen mixer mill (Qiagen, Valencia, CA, USA) at 25 cycles/sec for four minutes. Homogenates were clarified by centrifugation at approximately 12,800× g for 8 min at 4 • C and stored at −80 • C. A 100 µL aliquot of supernatant was inoculated directly onto Vero cell monolayers, with one sample per well on a 6-well plate for virus isolation by double-overlay plaque assay [26]. A second overlay containing neutral red was added four days post infection, and plates were observed for plaques up to 10 days post infection. Cells from plaque-positive wells were harvested into 1 mL DMEM + 10% fetal bovine serum and clarified by centrifugation and the infected supernatant was stored at −80 • C. The viral RNA from plaque-positive samples was extracted from 200 µL of the supernatant and eluted into a final volume of 140 µL AE buffer using the Qiagen BioRobot EZ1 Workstation using the EZ1 Virus Mini Kit v2.0. From the infected bat, additional sections of liver, lung, intestine, and oral and fecal swabs were also processed for virus isolation as above.

Sequencing and Bioinformatics Analysis
Initial sequencing of the novel virus isolate and bioinformatics analysis was performed following published methods [27]. RNA was extracted from FOMV, IFEV, and JAPV virus isolates using the MagMAX Viral RNA Isolation Kit (Life Technologies, Carlsbad, CA, USA) and DNAse treated using the TURBO DNA-free™ kit (Ambion, Austin, TX, USA). Library preparation was carried out using ScriptSeq™ v2 RNA-Seq Library Preparation Kit (Epicentre Biotechnologies, Madison, WI, USA) and ScriptSeq™ Index PCR Primers (Epicentre Biotechnologies, Madison, WI, USA). Libraries were pooled and sequenced on the Illumina MiSeq using the MiSeq Reagent Kit v2 (300 cycles) (Illumina, San Diego, CA, USA). RNA was Trizol extracted from Vero cell supernatant for the novel orbivirus isolate, BUKV (UGA432), and the library was prepared using the Kapa RNA HyperPrep kit (Kapa Biosystems, Wilmington, MA, USA). The library was sequenced on the Illumina MiSeq using the MiSeq Reagent Micro Kit v2 (300 cycles) (Illumina, San Diego, CA, USA).
Low-quality bases and barcoded sequencing adapters were removed using Cutadapt, duplicate reads were collapsed using cd-hit, and host reads were filtered using Bowtie2 [28][29][30]. De novo assembly was performed using SPADES assembler and contigs were searched for homology against the NCBI Viral RefSeq database at the nucleotide (blastn) and amino acid (blastx) levels [31]. Contigs representing orbivirus genomic segments were validated by remapping the quality-and host-filtered reads using Bowtie2 [28]. The resulting alignments were visually inspected to confirm that mapping depth and base-calls were sufficient for accurately determining the sequence of each open reading frame for each genomic segment. Putative open reading frames were characterized using open reading frame (ORF) prediction (Geneious v11.1.15, Biomatters, Auckland, New Zealand) and comparing the output (ORF start/stop/length etc.) to related orbiviruses. This workflow resulted in a coding-complete consensus sequences for all ten segments of JAPV, FOMV, and IFEV, and for nine of ten segments for BUKV.

Phylogenetic Analysis
Genome regions of interest included the highly conserved RNA-dependent RNA polymerase (VP1), sub-core shell protein (T2 (VP2/VP3)), and the major-core surface protein (T13 (VP7)) [32,33]. Multiple alignment and phylogenetic analysis were analyzed for these three genes using the newlyderived consensus sequences obtained for the four bat-borne orbiviruses in addition to previously genetically characterized orbivirus sequences obtained from GenBank ( (Table S1). Putative open reading frames were translated and aligned in SeaView version 4 [34] using Clustal Omega [35], and then back-translated to nucleotide sequence. Nucleotide alignments were trimmed using TrimAl in order to remove poorly aligned regions [36].
The best-fit substitution models for nucleotide alignment and protein alignment were determined using jModelTest 2.1.10 [37,38] and ProtTest 3.4.2 [38,39], respectively. Best-fit substitution models were selected based on lowest BIC and AIC scores. The best-fit substitution models selected for amino acid multiple alignment as determined by ProtTest was LG with gamma distribution including estimation of invariant sites (LG + G + I) for VP1, T2, and T13. The best-fit substitution models selected for nucleotide multiple alignment as determined by jModelTest were generalized-time reversible with gamma distribution including estimation of variant sites (GTR + G + I) for VP1 and T2, and generalized-time reversible with gamma distribution (GTR + G) for T13. Nucleotide trees were prepared using the Bayesian Markov Chain Monte Carlo method, as implemented in MrBayes 3.2.5 [40]. The analysis was performed for five million steps, with sampling every 1000 steps and discarding the first 10% as burn-in. Amino acid maximum-likelihood trees were prepared in MEGA7 with 1000 bootstrap replicates [41,42].

Multi-Step Growth Curves
Confluent T-12.5 flasks of Vero cells, BHK-21 [C-13] cells, and R06E cells were inoculated with BUKV in triplicate at a MOI of 0.01. Triplicate mock-infected flasks served as negative controls. An additional flask for each cell line was prepared to count cells and ensure an accurate MOI calculation. Cells were allowed to incubate with virus for 60 min at 37 • C, and then washed with 1× phosphate buffered saline (PBS) three times prior to replacing maintenance media (DMEM with 2% FBS and 1% P/S for Vero cells and BHK-21 [C-13] cells, DMEM-F12 with 2% FBS and 1% P/S for R06E cells). Immediately, 100 µL was taken for the 0 h post-infection timepoint and proportion FBS was increased to 20% prior to freezing at −80 • C. Additional time points were collected at 12, 24, 48, 72, and 96 h post-infection. Back-titrations were performed in duplicate to confirm each inoculum was within two-fold of the desired titer. Viral quantification was assessed by plaque titration as previously described using Vero cell plaque assay [43].
As a positive control, supernatant from BUKV-infected Vero cells was 10-fold serially diluted and from each dilution, RNA was extracted using the Mag-Bind Viral DNA/RNA 96 kit (Omega Bio-Tek, Norcross, GA, USA) on the Kingfisher ® extraction system (Thermo Scientific, Rockford, IL, USA). Additionally, plaque assays of the dilution series were performed in duplicate on Vero cells to quantify viral titer [43]. A linear regression was performed to correlate viral titer to cutoff threshold value determined by QuantStudio™ Design & Analysis Software run on the QuantStudio™ 3 polymerase chain reaction system (Thermo Scientific, Rockford, IL, USA). The same parameters were used to screen the tick samples. Archived RNA extracted from pooled Ornithodoros faini soft ticks collected in Python Cave, Maramagambo forest, Uganda, were screened for BUKV RNA following the same protocol as for archived bat sample testing [44].

Virus Isolation
Four days post-inoculation on Vero cells, plaques were visible for the spleen of bat #432 (UGA432, JCK8050, FMNH 223820) ( Figure S1), a young male ERB captured in Kasokero cave in January 2013 (Table 1, Figure 1). Of the additional tissues harvested from the infected bat and tested for BUKV, only liver and spleen were positive for infectious virus by plaque assay; lung, oral and fecal swabs, and intestine were negative.

Genome Sequencing
Complete coding sequences were obtained for all ten segments of FOMV, IFEV, and JAPV, and for all segments besides segment 3 of BUKV. Each segment possessed one gene encoding one protein, with the exception of segments 9 and 10, which each contained a second shorter ORF similar to other orbiviruses [45,46]. The size and GC% content for each ORF, in addition to its closest match to available data on GenBank using BLASTX (predicted translation product from six reading frames) and its GenBank accession number are listed in Tables 2-5.
Pairwise distances generated in Geneious using MAFFT for both nucleotide and amino acids are listed for each of the four newly-sequenced orbiviruses in addition to the other previously-sequenced orbiviruses obtained from Genbank in Figure 5 (T2), Figure S5 Table S1.
Pairwise distances generated in Geneious using MAFFT for both nucleotide and amino acids are listed for each of the four newly-sequenced orbiviruses in addition to the other previously-sequenced orbiviruses obtained from Genbank in Figure 5 (T2), Figure S5

Growth Curves
Propagation of BUKV was successful in all three mammalian cell lines tested (Vero, BHK-21 [C-13], and R06E cells) with viral titers peaking at 24 hpi in all three cell lines, and the highest titer achieved in BHK-21 [C-13] cells (1.6 × 10 7 PFU/mL at 24 hpi) ( Figure 6). Cytopathic effect was detected in all three cell lines between 12 and 24 hpi.

Growth Curves
Propagation of BUKV was successful in all three mammalian cell lines tested (Vero, BHK-21 [C-13], and R06E cells) with viral titers peaking at 24 hpi in all three cell lines, and the highest titer achieved in BHK-21 [C-13] cells (1.6 × 10 7 PFU/mL at 24 hpi) ( Figure 6). Cytopathic effect was detected in all three cell lines between 12 and 24 hpi.

Testing of Additional Bat and Tick Samples
None of the 171 bat samples screened via qRT-PCR resulted in amplification of BUKV VP1 RNA, though six were suspected or weak positive. None of the six suspect-positive bat samples were confirmed positive for BUKV RNA by nested PCR. Of the 171 samples, GAPDH RNA was successfully amplified from 86% (147/171) confirming RNA integrity. Of the 513 tick pools tested, 16s rRNA internal extraction control was amplified from 485, and none were positive for BUKV VP1 RNA.

Discussion
Bats are known to host a number of emerging zoonotic viruses highly pathogenic to humans in the absence of overt pathology within the bat host. Limited information exists, however, about the ability of bat species to harbor and transmit medically important arboviruses. Despite extensive serologic evidence suggestive of exposure of multiple bat species to arboviruses from different families (Flaviviridae, Togaviridae, and Bunyavirales), few studies have resulted in the isolation of arboviruses from wild-caught bats. Molecular and in vivo characterization of novel viruses isolated from wild-caught bats, in addition to enhanced surveillance efforts targeting suspected hosts helps clarify the role of bats as reservoirs for emerging arboviruses.
This study provides complete coding sequences of three previously uncharacterized orbiviruses isolated from bats (JAPV, IFEV, and FOMV) and one novel bat-associated orbivirus, BUKV, isolated

Testing of Additional Bat and Tick Samples
None of the 171 bat samples screened via qRT-PCR resulted in amplification of BUKV VP1 RNA, though six were suspected or weak positive. None of the six suspect-positive bat samples were confirmed positive for BUKV RNA by nested PCR. Of the 171 samples, GAPDH RNA was successfully amplified from 86% (147/171) confirming RNA integrity. Of the 513 tick pools tested, 16s rRNA internal extraction control was amplified from 485, and none were positive for BUKV VP1 RNA.

Discussion
Bats are known to host a number of emerging zoonotic viruses highly pathogenic to humans in the absence of overt pathology within the bat host. Limited information exists, however, about the ability of bat species to harbor and transmit medically important arboviruses. Despite extensive serologic evidence suggestive of exposure of multiple bat species to arboviruses from different families (Flaviviridae, Togaviridae, and Bunyavirales), few studies have resulted in the isolation of arboviruses from wild-caught bats. Molecular and in vivo characterization of novel viruses isolated from wild-caught bats, in addition to enhanced surveillance efforts targeting suspected hosts helps clarify the role of bats as reservoirs for emerging arboviruses.
This study provides complete coding sequences of three previously uncharacterized orbiviruses isolated from bats (JAPV, IFEV, and FOMV) and one novel bat-associated orbivirus, BUKV, isolated from an ERB. Phylogenetic analyses place BUKV and FOMV in the tick-borne orbivirus clade, and IFEV and JAPV cluster with the Culicoides/sandfly-borne orbiviruses. This study also provides documentation for in vitro propagation of a bat-associated orbivirus in a number of different vertebrate cell lines, one of which was derived from the ERB. The screening of additional archived bat and tick samples resulted in negative findings, but reflects a critical step in the investigation process into the host-vector relationships supported by phylogenetic analyses.
Of the three segments analyzed, the topology of the phylogenetic analysis consistently placed BUKV and FOMV within the tick-borne orbivirus subclade along with Chobar Gorge virus (CGV). Past studies indicated that FOMV is a serotype of the Chobar Gorge virus species based on results of complement fixation, and it is considered as such by the International Committee on the Taxonomy of Viruses [3,47,48]. This is consistent with past isolations of FOMV in field-caught Ixodid ticks [17]. CGV has been isolated from Ornithidoros spp. ticks in Nepal, and antibodies have been detected in humans and domestic ruminant species in the same region [49]. Additionally, the clustering of BUKV, FOMV, and CGV within the same subclade of tick-borne orbiviruses and high degree of nucleotide and amino acid similarity regardless of protein analyzed suggests they are three different serotypes of the same species Attoui et al. suggested that the amino acid identity for T2 of <91% should be the criteria for designating a species within the genus Orbivirus [50]. According to that criterion BUKV and FOMV viruses are on the border for consideration as new species. BUKV possesses 95.27% amino acid similarity to FOMV and 91.56% similarity to CGV, and Fomede possesses 90.87% amino acid similarity to CGV. Bukakata and Fomede viruses may be ecologically unique from Chobar Gorge in having been isolated from bats, however it is not known whether or not Chobar Gorge virus is also found in bats. Further characterization into the evolutionary relationship of these bat-associated and potentially tick-borne orbiviruses should involve exploration into in vitro growth kinetics in invertebrate cell lines in addition to the potential for serologic cross-reactivity and in vitro reassortment potential.
JAPV and IFEV cluster with the Culicoides/sandfly-borne orbivirus clade and have not yet been approved as species of orbiviruses, but novel genetic sequence obtained during this study indicates that their listing should be revised. Neither JAPV nor IFEV possess the requisite >76% nucleotide identity to any other orbivirus in their conserved T2 gene, indicating they are each their own individual species [3] (Figure 5). In the analysis of all three segments, IFEV is very distantly related to all other orbiviruses and may represent its own species due to the low level of nucleotide (maximum 55.7%) and amino acid similarity (maximum 59.3%) to any other orbivirus when analyzing the gene encoding sub-core shell T2 protein ( Figure 5). Interestingly, the BLASTX results for IFEV virus segments reveal that it is most closely related to Heramatsu orbivirus. Heramatsu virus was obtained from a Japanese eastern long-fingered bat in 1965 and was partially sequenced in 2013 [11,12]. However, due to lack of complete genome information and access to an archive isolate, this virus was not included in our phylogenetic analyses.
Due to their segmented genome, orbiviruses are known to undergo reassortment during co-infection [51,52]. Comparing the placement of JAPV within the VP1 and T2 phylogenies to its placement in the T13 phylogeny suggests that it may have undergone reassortment; however, definitive conclusions surrounding its potential as a reassortant virus are difficult to make due to low bootstrap values and posterior probabilities (Figures 2-4, Figures S2-S4). Interestingly, our phylogenetic analyses indicate JAPV clusters with the Culicoides/sandfly-borne orbiviruses, though it was isolated from a pool of Culicine mosquitoes in New Guinea [10]. Further investigation is required to better characterize potential vector-host associations for JAPV and its potential as a reassortant orbivirus.
BUKV grew to high titers in all three vertebrate cell lines in which multi-step growth curves were conducted. Two of these cell lines, Vero cells and BHK-21 [C-13] cells, are deficient for the interferon pathway, while the R06E pathway has an intact interferon response [53]. The Type I interferon response is the first line of antiviral defense in the mammalian immune system [54]. Interestingly, viral titers in interferon-competent R06E cells were comparable to interferon-deficient BHK-21 [C13] and Vero cells ( Figure 6). The immune system of some bat species is highly unique in its constitutive expression of IFN-α [55]. A recent study by Pavlovich and colleagues indicate that unlike Pteropus alecto, transcriptomic analysis of the ERB does not provide evidence of constitutive interferon expression [56]. Analysis of interferon expression over the course of infection in bat cells and other interferon-competent vertebrate lines would be an informative way to analyze the presence of this constitutive expression in existing bat cell lines.
None of the 171 bat samples or 513 tick pools tested positive for BUKV viral nucleic acid. However, GAPDH mRNA was tested for each bat sample and samples for which amplification of the GAPDH was not obtained were not included in the denominator of the total tested samples. While all samples were negative for BUKV RNA, some of the bat samples had very high CT values or were nearing the cycle threshold and as such, were considered to be suspect and subjected to a nested PCR protocol. The six suspect samples tested using this nested PCR were also confirmed negative. Samples types screened (spleen and/or liver) are consistent with the organs from which the virus was originally isolated. Testing additional bat species for viral RNA could yield additional information on the circulation of this virus.
While this study provides valuable information regarding potential vector-host associations among the orbiviruses, limitations restrict certain conclusions. Each orbivirus segment contains 1-2 genes, with untranslated regions on either end of the ORF. Due to decreased coverage at the ends of the reads, variable coverage was achieved throughout the length of each segment and only the complete coding genome of IFEV, JAPV, and FOMV were obtained. The coding complete sequence of segments 1-2 and 4-10 were obtained for BUKV but due to low coverage at the 5 end of segment 3, the start codon was not obtained ( Table 2). Individual orbivirus species possess conserved 5 UTR and 3 UTR terminal sequences and as such, higher coverage in the untranslated regions would have provided additional information surrounding level of relatedness between these and previously sequenced orbiviruses [3]. Field surveillance efforts were opportunistic and retrospective, and only ERB RNA was tested. The testing of additional bat species from nearby geographic areas sharing similar ecological habitats would provide additional information surrounding vertebrate host range. The tick pools tested were also opportunistic and retrospective, and originated in Python Cave, a cave with analogous ecological characteristics to Kasokero Cave, where BUKV was isolated, yet 213 km away (Table 1, Figure 1).

Conclusions
This study provided coding of the complete genome sequence on three bat-associated orbiviruses that have been discovered to date, and coding the complete sequence for nine of ten segments for a newly isolated bat-associated orbivirus. Further, in vitro replication kinetics were described in three vertebrate cell lines. From the phylogenetic analyses, inference regarding potential arthropod vector associations can be drawn regarding transmission of bat-associated orbiviruses and supports ticks as a potential vector of BUKV. While the archived bat and tick samples tested for BUKV RNA were negative, this study provides a strong framework for comprehensive viral characterization, including initial discovery, in vitro characterization, and the screening of samples collected from the potential vertebrate host and the purported invertebrate vector. The isolation of five orbiviruses from distinct bat species located across space and time and phylogenetically-associated with different arthropod vectors indicates a strong association between orbiviruses and bats, and further investigation into the public health impact of these orbiviruses is warranted.