Isolation of Kaeng Khoi virus from dead Chaerephon plicata bats in Cambodia -- Osborne et al. 84 (10): 2685 -- Journal of General Virology

A virus isolated from dead Chaerephon plicata bats collected near Kampot, Cambodia, was identified as a member of the family Bunyaviridae by electron microscopy. The only bunyavirus previously isolated from Chaerephon species bats in South-East Asia is Kaeng Khoi (KK) virus (genus Orthobunyavirus), detected in Thailand over 30 years earlier and implicated as a public health problem. Using RT-PCR, nucleotide sequences from the M RNA segment of several virus isolates from the Cambodian C. plicata bats were found to be almost identical and to differ from those of the prototype KK virus by only 2·63·2 %, despite the temporal and geographic separation of the viruses. These results identify the Cambodian bat viruses as KK virus, extend the known virus geographic range and document the first KK virus isolation in 30 years. These genetic data, together with earlier serologic data, show that KK viruses represent a distinct group within the genus Orthobunyavirus.

Bats are the only flying mammals and have been identified as the source of various viruses of public health importance, particularly viruses of the genus Lyssavirus, family Rhabdoviridae (e.g. rabies virus), and genus Henipavirus, family Paramyxoviridae (e.g. Nipah and Hendra viruses). As such they represent an interesting target for pathogen discovery efforts aimed at identifying viruses with the potential to cause human disease or novel viruses related to known human pathogens.

This investigation was part of a larger project focusing on the potential role of bats in human and veterinary health in Cambodia (Olson et al., 2002). Field locations were chosen on the basis of known or suspected bat roosts. The focus of this investigation was Kampot Cave, approximately 15 km north of the town of Kampot in southern Cambodia. The cave is used by local villagers for harvesting bat guano for fertilizer. Large numbers of bats use the cave as a roost and the villagers had reported an apparent decline in the colony over time. On 30 November, 2000, 46 bats were collected, 12 of which were dead or moribund while the rest appeared healthy. Live bats were restrained, sedated by the administration of ketamine hydrochloride (5 mg) via the intramuscular (i.m.) route and euthanized under sedation by cardiac bleeding. At necropsy, brain tissue was removed, frozen at -70 °C and the carcass placed in 10 % buffered formalin for archival storage.

No lyssavirus antigens were detected by direct fluorescent antibody test in any of the 46 bat brains examined (Whitfield et al., 2001). Virus isolation was attempted from all 12 dead bats and one brain collected from an apparently healthy bat by intracerebral (i.c.) inoculation into groups of mice. Briefly, the bat brains were homogenized with sterile grinders to a 10 % (w/v) suspension in sterile PBS and 2 % equine serum, clarified by low-speed centrifugation and 0·03 ml of the supernatant inoculated by the i.c. route into groups of three to five weanling (4-week-old) female ICR mice. Mice were observed daily for 30 days for any adverse clinical signs and were euthanized by CO₂ intoxication when illness appeared. Mouse brains were removed by sterile technique and stored at -70 °C for virus isolation attempts and further processing.

Neither illness nor death was observed in the mouse group inoculated with the healthy bat brain material. Deaths occurred in 11 of the 12 mouse groups inoculated with material from the dead bats. In each of these groups, 80100 % of the mice showed signs of encephalitis within 511 days of inoculation. Similar results were obtained after passing brain supernatant (from three of the dead bats) through 0·45 and 0·22 µm filters prior to inoculation. Three brain supernatants from dead bats were also administered to groups of mice by the i.m. (50 µl, gastrocnemius) and oral (50 µl) routes. In the i.m. group inoculated with two different brain suspensions, 2040 % became sick and died within 2 weeks. In the mice inoculated orally, 67 % died. None of the mouse brains had evidence of lyssavirus antigens when tested by direct fluorescent antibody test, regardless of group. In the course of preparing reference ascitic fluids, a representative virus isolate (6472) was inoculated by intraperitoneal (i.p.) or subcutaneous route in seven adult mice. Two of the seven i.p.-inoculated mice developed paralysis by day 2 post-inoculation and died by day 3 (29 % mortality). Similarly, two of the seven mice inoculated subcutaneously died by day 4.

Virus isolation was attempted using the mouse brain homogenates from the initial i.c.-inoculated mice to infect Vero E6 cell monolayers. Thin-section electron microscopy of infected cells revealed numerous extracellular virus particles consistent with features of the family Bunyaviridae. Virions were mostly round, with an average diameter of 67 nm and consisted of a moderately dense, finely granular core enclosed within the viral envelope and surrounded by surface projections (Fig. 1). While the virus appearance was consistent with that of known bunyaviruses, the particle size was unusually small for viruses of this family (Nichol, 2001).

(56K): Fig. 1. Electron micrograph of virus isolate 6472 from bats in Kampot, Cambodia. The isolate obtained from field material Ph P. 6472 (SPB VSP 809082) was selected as representative of the virus isolates. Two days after infection, Vero E6 cells were washed with 0·2 M phosphate buffer, fixed with 2·5 % glutaraldehyde and pelleted by centrifugation. Cell pellets were post-fixed in buffered osmium tetroxide, stained with uranyl acetate, dehydrated and embedded in Epon-substitute and Araldite resins (Mullenhauer, 1964). Thin sections were stained with uranyl acetate and lead citrate and viewed using a Philips EM410-LS transmission electron microscope. Extracellular enveloped virions were filled with a finely granular matrix and surrounded by fine surface projections. Occasionally, cross-sections of viral ribonucleoprotein strands could be seen (arrowheads). Bar, 50 nm.

A representative mouse brain sample was selected for RNA extraction and genetic analysis from each group in which death occurred after the primary i.c. inoculation. Total RNA was extracted from 0·2 ml 10 % mouse brain homogenate with 1 ml Tripure (Roche), and RNA was isolated using an RNaid extraction kit following the manufacturer's protocol (Q-Biogene). RT-PCR was performed using the primer pairs that have been used to characterize Garissa virus (Bowen et al., 2001): M14C and M619R (M segment), BUNYA1 and BUNYA2 (S segment), BUNCAL1 and BUNCAL2 (S segment) and M13CBUNL1C and BUNL605R (L segment). PCR products were analysed essentially as described previously (Bowen et al., 2001). A PCR product was obtained from all of the 11 mouse-lethal agents using primer pair M14C and M619R, which was designed to amplify the M segment of viruses of the genus Orthobunyavirus. No RT-PCR products were obtained using the other primers tested. RT-PCR products were sequenced using the primers with which they had been amplified, using standard protocols on an ABI 377 automated DNA sequencer (Applied Biosystems), and chromatograms and sequences were analysed using Sequencher 3.1.1 sequence analysis software (Gene Codes). A maximum of 1·0 % difference in nucleotide sequence identity was seen among the virus isolates, with five of the isolates having an identical sequence and three containing the same single nucleotide difference. Isolate 6472 was chosen for further analysis. Comparison of the M segment fragment nucleotide sequence with the equivalent region of other characterized members of the family Bunyaviridae demonstrated that these viruses were indeed members of this family. At both the nucleotide and deduced amino acid levels, the closest match was with members of the California encephalitis serogroup of the genus Orthobunyavirus. However, these Cambodian bat viruses were quite distinct, with a maximum of only 57 % nucleotide and 48 % deduced amino acid identity within the coding sequence observed. Such genetic differences would be consistent with these viruses being members of the genus Orthobunyavirus but representing a group other than the currently characterized serogroups.

According to the International Catalog of Arboviruses (), the only virus belonging to the genus Orthobunyavirus reported to be present in Chaerephon species bats in South-East Asia is Kaeng Khoi (KK) virus. The virus was isolated from Chaerephon plicata and Taphozous theobaldi bats in a cave in Thailand in 1969 (Neill, 1985) and again from C. plicata bats in the same region in 19701971 (Williams et al., 1976), and was identified as an unassigned member of the Bunyamwera supergroup within the Orthobunyavirus genus on the basis of electron microscopy and serology (Neill, 1985). Thus, despite the fact that there were no sequence data available for the KK virus isolated 33 years earlier in Thailand, there were several factors that together suggested that the Cambodian bat virus isolate might be related to KK virus: (i) the relatively close (550 km) geographic proximity of the isolation sites; (ii) isolation of both viruses from the same bat species (C. plicata); and (iii) characterization of both viruses as unassigned members of the Orthobunyavirus genus. KK virus prototype strain S-19-B was obtained from Robert Shope (University of Texas Medical Branch, Galveston, Texas, USA), grown in Vero E6 cells, and the cells and medium harvested at 4 days post-infection, frozen and thawed and total RNA extracted as before. RT-PCR analysis of the prototype KK virus RNA genome was performed using the same M segment-specific primers and protocols that had been successful with the Cambodian bat viruses. Sequence analysis showed that the nucleotide sequence of the 1969 KK virus isolate was 96·897·4 % identical to those of the Cambodian bat virus isolates and demonstrated that the Cambodian bat viruses appeared to be minor variants of KK virus.

Detailed phylogenetic analysis of the nucleotide sequence differences among members of the genus Orthobunyavirus has shown that the genetically characterized members of the group can be divided into five distinct genetic lineages containing members of the Bunyamwera, California encephalitis, Simbu and Group C serogroups and KK viruses (Fig. 2). The genetic divergence between these lineages would suggest that KK virus represents a distinct group within the genus Orthobunyavirus and is consistent with earlier antigenic analysis indicating that it does not fit within the known serogroups of the genus (Neill, 1985).

(27K): Fig. 2. Phylogenetic relationship of the M RNA segments of Orthobunyavirus genus viruses. Maximum-likelihood analysis of the aligned sequences of a 574 nucleotide region of virus M segments was performed using PILEUP (Wisconsin Package Version 10.2; Genetics Computer Group) followed by PAUP 4.0b10 (Sinauer Associates) software programs. Viruses included Garissa virus isolates 9800521 and 9800535 (GenBank accession nos AF398348 and AF398347, respectively), Bunyamwera virus (BLCMA), Germiston (BLCMPOLY), Cache Valley virus (AF082576, AF186241-3), Trivittatus virus (AF123491), California encephalitis virus (AF123483), Keystone virus (AF123489), Serra do Navio (AF123490), Melao virus (U88057), South River virus (AF123488), Inkoo virus (U88059-60), Jamestown Canyon virus (U88058, AF123487), Oropouche virus (AF312381) and Jatobal virus (AF312380). Maximum-likelihood analysis was carried out using a substitution model approximately corresponding to Hasegawa et al. (1985), including empirical nucleotide frequencies (A=0·34334, C=0·17758, G=0·20311, T=0·27596) and estimations of transition : transversion ratios, proportion of invariable sites and distribution shape parameter. The analysis resulted in a -ln likelihood of 9545·03363 with an estimated transition to transversion ratio of 2·09548 and shape of 0·4202. Kaeng Khoi virus forms a genetic lineage distinct from members of other serogroups.

The apparent genetic stability of KK viruses is similar to that seen for other viruses of the Orthobunyavirus genus. For KK viruses, we have observed up to 3·2 % variation in nucleotide identity (for the M segment fragment analysed) over more than 30 years and over a broad geographical area (Fig. 2). For the equivalent genome fragment, only 0·9 % nucleotide variation in Cache Valley viruses has been observed over 39 years (19561995) from Utah, Texas, Missouri and North Carolina (Brockus & Grimstad, 2001). A similar lack of diversity exists among La Crosse viruses from over 33 years (19601993) and over the midwestern and north-eastern USA (Huang et al., 1997). These data suggest that these viruses exist in a very stable ecological niche in nature with a lack of strong positive selection driving their evolution. As has been suggested previously for other arboviruses, it may be that the need to grow and compete well in both invertebrate and higher vertebrate hosts places a high level of constraint on the evolution of these viruses.

The Cambodian virus isolates were found approximately 550 km away from the small region in Thailand where the earlier KK virus isolates were collected, indicating that the virus is present over a larger geographical area than previously suspected. This is consistent with the broad geographical range of C. plicata bats, which is known to include India, much of South-East Asia, the Philippines and Indonesia. In addition, KK virus or antibodies to the virus have been found in another bat species. Together these data suggest that the distribution of KK virus may be even greater than that documented here.

The data presented here suggest that KK virus may display a higher level of pathogenicity than that normally associated with viruses of the Orthobunyavirus genus (Karabatsos, 1985). In our study, KK virus was only isolated from dead bats, and all of the dead bats that were assayed for the presence of KK virus tested positive. In addition, considerable levels of mortality were seen in the groups of adult mice inoculated with KK virus-infected bat brain material by i.p. (29 %) and subcutaneous (29 %) routes. High mortality was also seen in weanling mice inoculated by i.c. (100 %), i.m. (2040 %) or oral (67 %) routes. The i.c.-inoculated mice exhibited encephalitis 511 days post-inoculation. It is interesting to note that Williams et al. (1976) reported isolating KK virus from live C. plicata bats and detected antibodies in adult bats, suggesting the virus is not always lethal to bats. What effect KK virus may have on local bat populations is unknown. Despite the suggestions of villagers of increasing bat deaths, no objective data were obtained and multiple causes may be related to true population declines.

The public health importance of KK virus is unclear, but it is noteworthy that KK virus-neutralizing antibodies were detected in 29 % of bat guano collectors who worked in a specific bat cave in Thailand (Neill, 1985). KK virus was subsequently isolated from bedbugs in the region, including inside the cave in which the guano collectors worked, leading to the implication of bedbugs as biological vectors of the virus (Williams et al., 1976). The workers believed that bedbug bites were the cause of an influenza-like illness, typical of infection by a number of members of the genus Orthobunyavirus, including members of the California encephalitis and Bunyamwera serogroups (Gonzalez-Scarano et al., 1996). Interestingly, the caves in which infected bats were found in the previous studies were not usually frequented by people, but the caves investigated in this study are a tourist destination and are visited by both locals and travellers. This raises the possibility of frequent human infection and spread of the virus and any associated human disease.