Viruses

Aravan virus was isolated from the brain of one lesser mouse-eared bat (Myotis blythi) during a survey of 269 bats collected in the Osh region of Kyrghyzstan from 1988 to 1992 (30,32). A direct fluorescent antibody test was conducted. Aravan virus–infected mouse brains were impressed on glass slides, air-dried, and fixed with acetone. To detect the lyssavirus antigen, specimens were stained with fluorescein isothiocyanate (FITC)–labeled anti-rabies globulin (BBL, Cockeysville, MD) or FITC-labeled anti-rabies monoclonal globulin (Centocor Inc., Malvern, PA). FITC-labeled anti-nucleoprotein monoclonal antibodies (NC-MAbs, W502) cross- reactive to lyssaviruses were also used (19).

Amplification of Nucleoprotein cDNA and Direct Sequencing

Total RNA was extracted from virus-infected mouse brain emulsions with a commercial reagent (RNeasy Mini Kit, QIAGEN, Germany). cDNA was obtained with a T-Primed First-Strand kit (Amersharm Biosciences Corporation, Piscataway, NJ). PCR amplification and sequencing of the N gene were performed by using the sense primer AraN-S01 (5´-ATGTACCACCTCTACAATGG-3´, nt 55–74) and an antisense primer AraNC-1400 (5´-TCATGCTCAATTGTAAAAC-3´, nt 1456–1474). The cDNA template (2 μL) was amplified by using primers (AraN-S01 and AraNC-1400), according to the manufacturer’s instruction (Super Taq Premix Kit, Sawady Technology, Tokyo, Japan). PCR reactions were incubated at 94°C for 2 min, subjected to 40 cycles of 94°C for 30 s, 48° for 20 s, and 68°C for 2 min, and a final extension at 68°C for 7 min in a DNA thermal cycler (GeneAmpPCR System 9700 Applied Biosystems, Perkin-Elmer Corporation, Japan) (24,25). PCR products were purified by using a commercial kit (QIAquick PCR Purification Kit, QIAGEN). The sequences of the purified DNA products were determined on an automated sequencer (ABI model 310, Applied Biosystems, Foster City, CA) by using a PRIMS Ready Reaction Dyedeoxy Terminator Cycle Sequencing Kit (Applied Biosystems).

Phylogenetic Analysis

The 1350-nt and the deduced 450 amino acid (aa) sequences of the N gene of the Aravan virus were aligned with 26 lyssaviruses by using Clustal W program (34). A phylogenetic tree was constructed with the computer software MEGA 2 (35). Pairwise evolutionary nucleotide distances, including both transitions and transversions, were estimated according to Kimura’s two-parameter method. Phylogenetic trees were constructed by the neighbor-joining method with 1,000 replicates to generate bootstrap probabilities at each node (36).

Direct Fluorescent Antibody Assay

The three strains used in this study reacted against the Aravan virus infected mouse brain impressions. Fluorescence showed more scattered inclusions than those of the challenge virus standard in the acetone-fixed mouse brain smear (data not shown). The results confirmed that the Aravan virus is a lyssavirus.

Nucleotide and Deduced Amino Acid Sequence Identities among the Aravan Virus and Other Lyssaviruses

The 1350-nt and the deduced 450 aa sequences of the Aravan virus were compared with 26 representative lyssaviruses belonging to seven genotypes (Table 1). We selected 16 representative rabies variants from the eight diverse groups, including rabies variants from geographic areas of Asia near Kyrghyzstan and from bats and raccoons in North and South America (25,37). The nucleotide and amino acid sequence identities among all 27 lyssaviruses, including Aravan virus, were calculated. Then genotype 1 was represented by seven rabies viruses (SRL1032, 86118BRE, 1500AFS, 9218TCH, 8738THA, insectivorous bat/Chile, and PA R89), and genotypes 2, 3, 4, 5, 6, and 7 were represented by Lagos bat virus (8619NGA), Mokola virus (MOK/U22843), Duvenhage virus (86132AS), EBLV-1 (8918FRA), EBLV-2 (9007FIN), and ABLV (Balina/AF006497), respectively (Table 2). The nucleotide sequence identity of Aravan virus with the genotypes 4, 5, 6, and 7 was 77% to 78%; with genotype 1, 75% to 77%; and with genotypes 2 and 3, 72% to 74%. The most extensive nucleotide sequence differences among isolates of genotype 1 were between the raccoon isolate (PA R89) and the African and Asian isolates (82.8% to 82.9% identity). The Aravan virus demonstrated 92% aa sequence identity with genotypes 4, 5, and 7; 89% with genotype 6; and 81% to 85% with genotypes 2 and 3. The maximum variation of amino acid sequences within genotype 1 was exhibited between a vampire bat isolate from Brazil and an African isolate (93.1% to 93.3% identity). Genotype 4 (Duvenhage virus) was most closely related to genotype 5 (EBLV-1) with nucleotide and amino acid sequence identities of 79.8% and 93.3%, respectively. ABLV (genotype 7) was closely related to SRL1032 (genotype 1, Sri Lankan rabies virus) with a 93.1% aa sequence identity. These values were almost same as maximum variation of genotype 1. Based on our present data, we determined that isolates sharing <79.8% nt and 93.1% to 93.3% aa sequence identities belonged to different genotypes. In several studies, thresholds of <80% nt and 92% or 93% aa sequence identities warranted the proposal of a new genotype (1,23,38). Hence, the nucleotide and amino acid percentage identity values demonstrated that Aravan virus should be regarded as a new lyssavirus genotype.

Phylogenetic Analysis

Figure

Figure. Rooted phylogenetic tree showing genetic relationships among Aravan virus and 26 lyssaviruses. Phylogenetic relationships were determined by comparing the 1350-nucleotide sequences of the nucleoprotein (N) gene (a) and the deduced 450-amino-acid...

A phylogenetic tree of 27 lyssaviruses, including the Aravan virus, based on the 1350-nt sequence of the N gene was constructed by using the vesicular stomatitis Indiana virus (VSIV, tsW16B/U13898) as an outgroup (Figure, a). The lyssaviruses divided into two groups: one group consisted of genotypes 2 and 3, and the other consisted of genotypes 1, 4, 5, 6, 7, and the Aravan virus. The latter group was divided into six distinct clusters corresponding to genotypes 1, 7, 6, and 5 (high bootstrap values of 98%, 99%, 100%, and 100%, respectively), then Aravan virus and genotype 4. Moreover, the Aravan virus clustered with genotypes 4, 5, and 6 (low bootstrap value of 59%). Duvenhage virus (genotype 4) and EBLV-1 (genotype 5) formed the same cluster (high bootstrap value of 91%), and are therefore closely related. The Aravan virus occupied the phylogenetic position between genotype 6 and the cluster of genotypes 4 and 5. We also constructed a phylogenetic tree based on the deduced 450-aa sequences of the N gene (Figure, b). Similar to the nucleotide data, the amino acid sequences divided into two large groups and further subdivided into eight groups. One group consisted of genotypes 2 and 3 (bootstrap value of 89%), and the other group consisted of genotypes 1, 7, 6, 4, and 5, and the Aravan virus (high bootstrap value of 100%). The latter group had three distinct clusters corresponding to genotypes 1, 7, and 6 (high bootstrap values of 100%, 99%, and 100%, respectively), genotypes 4 and 5 (same cluster with a high bootstrap value of 98%), and the Aravan virus. The Aravan virus did not group with any other genotypes and is located at a position close to the cluster of genotypes 4 and 5 (bootstrap value of 66%).

These results, along with those in Table 2 and the Figure, suggest that the Aravan virus does not belong to any of the seven Lyssavirus genotypes (rabies, Lagos bat, Mokola, Duvenhage, EBLV-1, EBLV-2, and ABLV). Thus, we propose that the Aravan virus forms an independent cluster and is a new member of the Lyssavirus genus.

In this article, we have reported the first lyssavirus distinct from rabies virus originating on the Asian continent. The Aravan virus was more closely related to genotypes 4, 5, and, to a lesser extent, 6, which circulates among insectivorus bats in Europe and Africa. The lesser mouse-eared bat, from which the Aravan virus was isolated, is widely distributed in northern Africa, the Mediterranean, southern Europe, Crimea, Caucasus, Palestine, southwest Asia, and parts of central and eastern Asia. This information should be considered in the discussion of lyssavirus classification and evolution, as it suggests the possibility of a broader geographic distribution of the Aravan virus. We have no information about human rabies caused by bat exposure from central Asia, and rabies surveillance in this area is not known well. Based on this information and the virus’ misdiagnoses as rabies, we consider that transmission of Aravan virus to humans is possible. Indeed, this finding stimulates interest in new genotypes of lyssaviruses and is important from the viewpoint of public health, necessitating further lyssavirus surveillance of bats on the Asian continent.