Seven brain samples were collected from animals with suspected rabies in 3 districts of Bangladesh (Dhaka, Narayanganj, and Narshingdi) in 2010 (Table 1). A portion of brainstem was removed from each sample and preserved in TRizol (Invitrogen, Carlsbad, CA, USA) at –20°C. Total RNA was extracted from brain homogenate, cDNA was synthesized by using random hexamer primers, reverse transcription PCR was conducted to amplify gene fragments, and nucleotide sequencing of genes was performed (4).

Full-length nucleoprotein (N) and glycoprotein (G) gene sequences from samples were determined. Nucleotide identities of N and G genes were 98%–100%. Amino acid identities of N and G genes were 100% and 98%–100%, respectively. Complete genomic sequencing (11,928 nt) of strain BDR5 was also conducted.

Evolutionary analysis was performed by using full-length N gene. We created a maximum clade credibility phylogenetic tree using the Bayesian Markov chain Monte Carlo method available in BEAST version 1.6.1 (5). Analysis was conducted by using a relaxed (uncorrelated lognormal) molecular clock and a generalized time reversible + Γ + proportion invariant model (6). All chains were run for 90 million generations and sampled every 3,000 steps and an effective sample size >1,383 was obtained for all estimated parameters. Posterior densities were calculated with 10% burn-in and checked for convergence by using Tracer version 1.5 in BEAST.

The mean rate of nucleotide substitution estimated for the N gene was 2.3 × 10⁴ substitutions/site/year (95% highest posterior density [HPD] 1.4–3.1 × 10⁴ substitutions/site/year). This rate is consistent with that of a previous study (7). The phylogenetic tree showed that rabies viruses in Bangladesh belong to Arctic/Arctic-like group 2 (AAL2) (3) also known as Arctic-like-1 (8), in close association with the strain from Bhutan.

Approximately 397.0 years ago (95% HPD 273.5–589.5 years), AAL and cosmopolitan rabies virus segregated from their most recent common ancestor (Figure 1). Approximately 225.6 years ago (95% HPD 157.4–324.2 years), AAL3 segregated. Approximately 187.4 years ago (95% HPD 129.0–271.9 years), AAL1 and AAL2 segregated. The AAL2 clade had a common progenitor that circulated ≈133.1 years ago (95% HPD 91.3–193.4 years), which has evolved into several different lineages. One lineage evolved 91.5 years ago (95% HPD 63.1–132.2 years) and currently circulates in Bangladesh, India, and Bhutan. Separate linages circulate in others countries in this region, including Iran, Nepal, Pakistan, and Afghanistan. AAL2 spread into central Bangladesh 32.3 years ago (95% HPD 18.4–50.6 years) in ≈1978 (95% HPD range 1958–1991).

Compared with the AAL2 strain from India (AY956319), BDR5 had several amino acid substitutions (Table 2). Sizes of their 2 genomes, leader RNA, trailer RNA, and intergenic regions were similar. The N-glycosylation site was predicted by using the NetNGlyc 1.0 server (www.cbs.dtu.dk/server/netnglyc). With the exception of BDR6, the G gene of all strains had potential glycosylation sites at position 37, 146, and 319.

Genetic analysis and phylogenetic studies can contribute to understanding the epidemiology of rabies virus in disease-endemic countries. Molecular analysis of animal rabies viruses showed that AAL2 appeared in central Bangladesh only 32 years ago. A close association between N genes sequences from rabies viruses in Bangladesh and Bhutan indicates that they originated from a common ancestor. If one considers the ease of human movement between countries, AAL2 most likely entered Bangladesh from India rather than from Bhutan.

Circumstantial evidence suggests that rabies virus spread from India to Bhutan (9). AAL2 circulates in many states of India. It has spread into southern India and has replaced older strains (10,11). It is likely that AAL2 is also circulating in states of India that are between Bhutan and Bangladesh. Estimated time of AAL2 spread is based on 7 samples that are representative of central Bangladesh (Figure 2). Therefore, further surveillance might identify the extent to which AAL2 has spread and the diversity of rabies viruses in other parts of Bangladesh that might alter the estimated date of spread. It has been reported that arctic rabies virus and other variants can co-circulate in the same region (12).

The G protein is the major factor responsible for the pathogenesis of rabies virus and contains 2 glycosylation sites (13). The G protein of strains from Bangladesh uniquely evolved to contain 3 potential glycosylation sites, which has been reported in only fixed (laboratory adapted) strains and proposed to be responsible for their reduced pathogenicity (13). However, the site for additional glycosylation differs between Bangladeshi and fixed strains. Detection of an additional glycosylation site and amino acid substitutions deserve further investigations.

AAL viruses could have moved southward from Siberia or other northern regions of the former Soviet Union into Nepal, India, and other countries in Asia by a species jump from fox to dog at some point (3). Another possibility is that AAL viruses first emerged in dogs in southern Asia and subsequently spread to northern climes, where they are now maintained in fox populations (3,8). Extensive surveillance of viruses from Iran, Iraq, Afghanistan, and countries north of them is necessary to determine the origin and spread pattern of AAL rabies virus.

The timeline of divergence of different lineages determined in this study was similar to that previously reported (8). That study and our study used the full-length N gene to determine the time of divergence. Another study reported the timeline of divergence as a more recent event (14). This study used partial sequences of N genes, which might be responsible for different results. Rabies virus from Nepal also belongs to AAL2, and as reported in a previous study (15), seemed to be forming a different lineage. However, the speculation was not supported by a significant a posterior density value (0.6355). Thus, a network of countries is urgently needed to exchange information on molecular typing of circulating strains of rabies virus that might be useful in controlling rabies in this region.