Molecular characterization of two hantavirus strains from different rattus species in Singapore

All animals were handled in strict accordance with good animal practice, as defined by The Animal Research Ethics Committee of DSO National Laboratories, 20 Science Park Drive, Singapore 118230.

Indirect Enzyme-linked immunosorbent assay (ELISA) was performed using a recombinant truncated nucleocapsid protein of the Seoul virus (M34881) and serum samples of collected rodents. The analysis was done as described earlier [23]. The calculated cut-off value (0.135) was set as 3 times the value of sera derived from negative lab. rats (R. norvegicus).

Primers with the potential to detect all known rodent-borne hantaviruses were designed towards conserved regions identified from multiple alignments of hantavirus L segment sequences by the BioEdit program package [24]. Using the primer pair HantaL_2_fwd_3312-40 5'- TYTTTGARTTTGCHCAYCAYTCWGATGATGC and HantaL_2_rev_3481-53 5' - TCATGNARRTTRAACATRCTYTTCCACA for PCR, tissue samples (lungs and kidneys) of all rodents were analyzed, as described below. The PCR was done in Mg⁺⁺ free PCR buffer supplemented with 2 mM of MgCl₂, 0.2 mM of dNTP, 1 μM of the primers and 1.25 U of AmpliTaq^® (Applied Biosystems Inc., USA), in a total volume of 25 μl. After a preincubation for 5 min at 95°C, 35 cycles at 53°C for 30 sec, 72°C for 45 sec and 45 sec at 95°C were done. After a final incubation for 10 min at 72°C the result was examined by agarose gel electrophoresis. A resulting PCR product of 170 bp was considered PCR positive.

RNA was extracted from tissue samples of kidneys and/or lungs of individual rodents using Qiagen RNeasy Mini kit (Qiagen, Inc., Germany) or the TRIzol reagent (Invitrogen Inc., USA), according to manufacturer's instructions. Reverse transcription of RNA was performed with SuperScript III (Invitrogen Inc., USA) and random hexamers. For direct sequencing of overlapping amplimers, numerous PCR primers were constructed using the PriFi software [25]. Specific cDNA fragments obtained by PCR were purified using QIAquick Gel Extraction kit or QIAquick PCR Purification Kit (Qiagen, Inc., Germany) before sequencing was conducted. Thereafter, the sequences of the segments were confirmed using a new set of primers targeting the determined cDNA sequences. However, the 15-20 nucleotides at the 5' and 3' termini of the S and M segments were set by the primers used for PCR. Finally, the resulting sequence data was assembled as chromatograms, using SeqMan (LaserGene Inc., USA) and edited with the BioEdit software package [24]. With the exception of the conserved 3' and 5' pan-handle sequence the S and M and partial L sequences of these hantavirus strains of Singapore have been made available (Genbank accession no's: GQ274936 - GQ274945).

The resulting sequences were aligned by ClustalW [26], as implemented in the BioEdit software package before bootstrapped maximum parsimony and trees were calculated using the SeqBoot, DNApars (setting: search for best tree, more thorough search, use input order, multiple data sets 1000 and 10 jumbles) and Consence softwares. Branch lengths of the trees were calculated using DNAml of the Phylip 3.67 software package as distributed by the author [27]. The tree was visualized by the TreeView software [28]. The similarity plots were performed by Stuart Ray's SimPlot 3.5.1 software using a window of 200 bp and steps of 20 bp, GapStrip: on, Kimura (2-parameter), T/t: 2.0 [29].

The phylogeny-based analysis of the cytochrome b gene identifies the rodent hosts of the Seoul virus strain Singapore (RN41 and 46) as R. norvegicus. Furthermore, the rodent hosts carrying the virus strain Jurong TJK/06 (RT49 and 50), (in this paper denoted R. tanezumi) clearly clusters with R. rattus diardi, R. tanezumi and R. kandianus of the diardii clade of the R. rattus complex (Additional file 1, Figure S1) as described by Robins et.al [18].

RT-PCR screening of kidney samples from rodents with Pan-hantavirus primers for the L segment, together with a succeeding PCR targeting the S segment revealed that 2.1% (21/996) of R. norvegicus and 4.5% (7/156) of R. tanezumi were PCR positive for hantavirus RNA. An observation was made when comparing the optical density OD values of the ELISA with the PCR results. It was noticed that rodents with the highest OD values were more likely to be PCR positive too. Among the rodents with OD values more than 1.0, 70% (7/10) were also found PCR positive (Fig 1). In contrast, none of the 17 R. tanezumi and only 2 of the 335 seropositive R. norvegicus having OD values less than 1.0 were also PCR positive. The exceptions included one rodent with a slightly lower ELISA OD value (0.96) and one ELISA negative but PCR positive R. norvegicus. It is likely that the latter rodent was infected close to the time of trapping.

For hantavirus sequence analysis, two PCR positive lung samples of each of the two rodent species (R. norvegicus and R. tanezumi) were selected: rodent # 41, 46 and 49, 50, respectively. When comparing the complete open reading frames of the S segments from the two rodent species of Singapore with other complete or partial hantavirus S segment sequences (Additional file 2, Table S1), the hantavirus of the Singaporean R. tanezumi (strain Jurong) showed the highest nucleotide sequence identity (95.0%) to the Serang virus, obtained of R. tanezumi from Indonesia [13]. In addition, the Jurong sequence also demonstrated a high nucleotide sequence identity to partial sequences of the S segment of rodent samples of Cambodia (85.8%) [11] and full length S sequences of the Thailand virus (83.6%), isolated from B. indica [30].

Sequence comparisons of M segments also revealed the Serang virus as most similar (94.5%) to the Jurong strain. However, M and L segment of the Cambodian hantavirus strains are not available in public databases and only 343 bp and 412 bp of the Serang virus are currently accessible for sequence comparisons. As shown in Additional file 2, Table S1, analyses of nucleotide and amino acid sequences of the S and M segments strongly suggest that the Serang and Cambodian isolates, together with the Thailand virus share the closest identity to the Jurong virus strains of Singapore. When comparing the L segment sequences of the Jurong strains with the Serang virus, a nucleotide sequence identity of 91.2% was found, slightly lower than between the corresponding S or M segments. Though, the deduced amino acid sequences of the S segment encoding N, the M segment encoding Gn and Gc and the L segments encoding RdRp are nearly identical over the overlapping regions.

SimPlot comparisons of the complete open reading frames of the Jurong virus with the Seoul/Hantaan/Dobrava types of viruses demonstrated a significant non-conserved region of approximately 300 nt near the middle of the S segment [29]. However, such drop in nucleotide sequence identity was not noticed towards the Thailand virus sequences or against partial nucleotide sequences of the corresponding Cambodian strains or towards the Indonesian Serang virus (Fig 2). Moreover, comparisons of the amino acid sequences revealed a similar trend, although the decrease in similarity in the non-conserved region was less prominent (data not shown).

Due to the lack of available sequences in public databases, a comprehensive SimPlot comparison of the M segment could only be performed towards the Thailand virus along with a subset of less related hantaviruses. As expected, the M segment of the Jurong strain was most similar to the Thailand virus, especially when comparing the amino acid sequence of the glycoprotein Gn. In addition, two highly conserved regions consisting of approximately 300 amino acids each were found near the middle of the glycoproteins of the Jurong and Thailand viruses (data not shown).

Phylogenetic analysis of S segment sequences showed that the Jurong virus strains and the Thailand virus isolates cluster together (Fig 3). Other hantaviruses carried by rodents of the Cricetidae family, Arvicolinae, Neotominae or Sigmodontinae subfamilies (e.g. Puumala virus, Tula virus, Sin Nombre and Andes virus) or hantaviruses carried by family Muridae subfamily Murinae, such as Seoul virus, Hantaan virus and Dobrava-Belgrad virus are genetically more distant (Fig 3). When partial sequences were included to the phylogenetic analysis, a novel clade consisting of the Thailand, Jurong, Serang and Cambodian hantaviruses was found. According to the phylogenetic trees and distance matrix, these sequences are clearly strains of the same hantavirus species. A common origin seems evident in such case.

As indicated by the phylogenetic trees for the viral S and M segments (Fig 3 and 4, respectively) the Jurong virus strains discovered in Singapore rats is genetically similar and form a clade with the Serang virus of Indonesia, Cambodian strains (e.g. Camb174, Camb117, Camb96) and hantavirus sequences of B. indica from Thailand (Thailand virus). Interestingly, the four closely related viruses; Jurong, Cambodia, Serang, and Thailand are carried by two or more different rodent species. Hantavirus strains of R. norvegicus, such as the Seoul Singapore and corresponding Cambodian isolates (e.g. Camb32, 41, 58 and 180) cluster closely together with other Seoul virus sequences, see Fig 3. The closest identity of the Seoul Singapore strain was observed towards the Cambodian strain 41 (Camb 41) with 99.1% identity for the overlapping S segment sequences.