On a November morning in 2009, a “wild cat” appeared in the garden of a basic health clinic in Moneragala district, Uva Province, Sri Lanka. The animal, which showed aggressive behavior, was suspected to be rabid and was thus killed to prevent transmission of rabies virus to humans. The animal’s head was packed in ice to avoid decomposition and sent to the Medical Research Institute (Colombo, Sri Lanka) for testing. We detected rabies virus in the animal’s brain by using the fluorescent antibody test and extracted viral RNA and DNA by using Trizol (Invitrogen, Carlsbad, CA, USA). The rabies virus from this sample was designated as H-1413-09. The whole genome of the virus was sequenced directly from the sample as described (5).

To confirm the species of the rabid animal, we determined the nucleotide sequence of the mitochondrial cytochrome b (cytb) gene and performed a BLAST search (www.ncbi.nlm.nih.gov/blast/Blast.cgi) for similarity with other sequences. By aligning nucleotide sequences of the cytb gene of mitochondrial DNA of domestic cat, jungle cat, fishing cat, Asiatic golden cat, marbled pole cat, European pole cat, lynx, puma, leopard, African lion, tiger, jaguar, civet, and palm civet with ClustalW2 (www.ebi.ac.uk/clustalw), we designed primer Felis cytb-F, 5′-ATGACCAACATTCGAAAATCACACC-3′ (nt 1–25), and primer Felis cytb-R, 5′-CAATAATGCCTGAGATGGGTATTAG-3′ (nt 1093–1,117). Using these primers, we performed PCR as follows: initial denaturation at 94°C for 2 min, followed by 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min for 35 cycles, followed by a final extension at 72°C for 5 min. PCR generated a 1,117-bp fragment from which a 1,004-nt sequence was determined. Analysis showed that the sequence has 100% identity with the partial (224-nt) sequence of the cytb gene of P. zeylonensis (GenBank accession no. FJ881681); this is the only sequence available for P. zeylonensis. The sequence also has 95% identity with P. jerdoni and 90%–92% identity with P. hermaphroditus.

We performed an evolutionary analysis by using the N gene. We inferred a maximum clade credibility phylogenetic tree by using the Bayesian Markov chain Monte Carlo method available in BEAST version 1.6.1 (6). The analysis used a relaxed (uncorrelated lognormal) molecular clock and GTR + Γ + I model of nucleotide substitution. We selected the model on the basis of Akaike Information Criterion by using jModelTest software (7). All chains were run for 9 × 10⁷ generations and sampled every 3,000 steps. This procedure resulted in an effective sample size of >2,000 for all estimated parameters. The posterior densities were calculated with 10% burn-in and checked for convergence by using Tracer version 1.5 (http://beast.bio.ed.ac.uk/Main_Page). The mean rate of nucleotide substitution estimated for the N gene was 2.2 × 10⁴ substitutions/site/year (95% highest posterior density [HPD] values 1.3–3.2 × 10⁴ substitutions/site/year). This rate is in agreement with previous findings (8). Approximately 155.5 years ago (95% HPD 91.3–249.5 years)—that is, circa 1854 (95% HPD range 1760–1918)—rabies viruses from Sri Lanka and southern India diverged from their most recent common ancestor. Approximately 76.4 years ago (95% HPD 46.4–123.0 years)—that is, in ≈1933 (95% HPD range 1886–1963)—strain H-1413-09 diverged from canine rabies virus in Sri Lanka.

Compared with the genome sequence of rabies virus strain H-08-1320 from Sri Lanka, the genome sequence of strain H-1413-09 had a nucleotide deletion at residue 2,417 and an addition at nt 11,807. As a result, the start and stop signals of the mRNA and the start and stop positions of coding sequences of different genes were altered (Table 1). These altered start and stop codons advanced the coordinates of the intergenic signal of relevant intergenic regions. The P-M region and terminal sequence of strain H-08-1320 were 1 nt shorter (91 nt) and longer (136 nt), respectively. The nucleotide and amino acid identities between the coding regions of strains H-1413-09 and H-08-1320 are shown in Table 1. The substitutions detected in the deduced amino acid sequences of H-1413-09 were compared with the genomic sequence of H-08-1320 (Table 2). By using the MEGA5 Tamura-Nei model (www.megasoftware.net), we determined the genetic distance of the N gene to determine whether strain H-1413-09 is more diverse than other rabies viruses in Sri Lanka (Figure). The rate of variation among sites was modeled with a gamma distribution (shape parameter = 0.5). The genetic distances between strain H-1413-09 and other rabies viruses from Sri Lanka (0.027–0.036) were greater than the genetic distances among other rabies viruses (0.001–0.011). These results support our finding that strain H-1413-09 differs from other rabies viruses circulating in Sri Lanka.

Rabies virus probably survives favorably in the wild because it can infect a large spectrum of animals, thereby maximizing replication and dispersal opportunities (9). Most viruses replicate poorly when transferred to new hosts, but greater genetic variation assists in such species adaptation (10). Increased mutation in an RNA virus like rabies virus can give rise to variants with altered levels of fitness to persist and spread. A large number of substitutions were found in strain H-1413-09 compared with strain H-08-1320; these substitutions might represent changes that resulted from species adaptation. Phylogenetic analysis and comparative sequence data indicated that strain H-1413-09 is a variant rabies virus.

Palm civets are facing extinction in Sri Lanka because the species is losing its habitat, being hunted for its meat, and dying of parasitic diseases (www.sundaytimes.lk/090118/Plus/sundaytimesplus_01.html). Our study indicates that rabies might be another risk factor for extinction of these animals. Identification of a variant rabies virus in wildlife has serious implications for rabies control in Sri Lanka. Identification of such a virus would help provide epidemiologic data about the spread of rabies and its incursion into new geographic regions and would justify allocation of increased resources to help control rabies (11,12).

Several rabies virus variants associated with wildlife are known in the Americas and Africa (1,13–15), and this report identified classical sylvatic rabies in Asia. Whether P. zeylonensis is a reservoir of rabies virus or represents spillover from another animal deserves extensive investigation. The detection of rabies in wildlife indicates that much remains to be discovered in the tropical ecosystem of Sri Lanka. The circulation of a sylvatic variant rabies virus may be another hurdle in the rabies-control effort in Sri Lanka.