Identification of a Novel Coronavirus in Bats

Sample collection.The animal surveillance program was performed between the summer of 2003 and the summer of 2004. The study was approved and supported by the Department of Agriculture, Fisheries and Conservation, Hong Kong, Special Administrative Region, People's Republic of China. Small mammalian, avian, and reptile species living in natural reservoirs or country parks in Hong Kong were studied. Animals were trapped, and respiratory and fecal swab samples were collected. Before samples were taken, all animals were examined by a veterinary surgeon and confirmed to be free of overt disease. All captured animals were released after samples were taken. Samples were kept in viral transport medium (Earle's balanced salt solution, 0.2% sodium bicarbonate, 0.5% bovine serum albumin, 200 μg of vancomycin per liter, 18 μg of amikacin per liter, 160 U of nystatin per liter) at 4°C. In addition, blood samples were drawn from captured Himalayan palm civets for neutralization assays for SARS-CoV (8). The neutralization assays were performed as described previously (8).

RNA extraction and reverse transcription.RNA from 140 μl of the sample was extracted by QIAamp virus RNA mini kit (QIAGEN) following the manufacturer's instructions. Extracted RNA was eluted in 50 μl of RNase-free water and stored at −20°C. cDNA was generated as described previously (19). Briefly, 10 μl of eluted RNA samples was reverse transcribed by 200 U of Superscript II reverse transcriptase (Invitrogen) in a 20-μl reaction mixture containing 0.15 μg of random hexamers, 10 mmol of dithiothreitol per liter, and 0.5 mmol of deoxynucleoside triphosphate per liter. Reaction mixtures were incubated at 42°C for 50 min, followed by a heat inactivation step (72°C for 15 min). Reverse-transcribed products were stored at −20°C.

PCR and sequencing.A pair of consensus primers targeted to the conserved region of coronavirus RNA polymerase (RNA-dependent RNA polymerase [RdRp]) sequences was used to screen the RNA samples (PCR 2 in Table 2). In a typical PCR, 2 μl of cDNA was amplified in a 50-μl reaction mixture containing 0.2 mmol of deoxynucleoside triphosphates per liter, 3 mmol of MgCl₂ per liter, 0.5 μmol of forward primer per liter, 0.5 μmol of reverse primer per liter, and 0.25 U of AmpliTaq Gold (Applied Biosystems). PCR was performed as follows: (i) 10 min at 95°C; (ii) 40 cycles, with 1 cycle consisting of 1 min at 95°C, 1 min at 58°C, and 1 min at 72°C. Amplified DNA products were analyzed by agarose gel electrophoresis and DNA sequencing. To avoid possible contamination, RNA extraction, reverse transcription-PCR (RT-PCR), and gel electrophoresis were preformed in separate laboratories. In addition, water controls were included in each run of the RT-PCR assay, and no false-positive result was observed in the negative-control reactions.

In subsequent experiments, primers targeted to the RdRp, helicase-ExoN, and S-encoding sequences were used to determine the Bat-CoV sequence (Table 2, PCRs 1 and 3 to 6). PCR conditions of these assays were identical to the PCR assay described above. DNA products with expected sizes were purified by QIAquick PCR purification kit (QIAGEN) and were cloned into DNA vectors (pCR-TOPO; Invitrogen). DNA inserts in purified plasmid were sequenced by BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems). Sequencing products were analyzed by ABI PRISM 3700 DNA analyzer (Applied Biosystems). Both sense and antisense sequences of these PCR products were sequenced at least once.

Data analysis.Deduced viral sequences were analyzed and aligned by BioEdit, version 5.0.9 (http://www.mbio.ncsu.edu/BioEdit/bioedit.html ). Phylogenetic trees were constructed by the neighbor-joining method, and bootstrap values were determined by 1,000 replicates in MEGA 2.1 (http://www.megasoftware.net ). Potential glycosylation sites of the S protein were predicted by NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/ ). Stablecoil 1.0 (http://www.pence.ca/stablecoil/ ) was used to detect the heptad repeat (HR) regions of the S protein. Reference sequences used in the study are FIPV (GenBank accession number AB088222 ), TGEV (NC002306 ), HCoV-229E (AF304460 ), HCoV-NL63 (AY567487 ), HCoV-OC43 (NC005147 ), CCoV (AY342160 ), PEDV (AF353511 ), IBV (NC005147 ), MHV (NC001846 ), and SARS-CoV (NC004718 ).

Nucleotide sequence accession numbers.The deduced sequences from this study were deposited in GenBank under accession numbers AY864196 to AY864198 .

Surveillance.A total of 162 swab samples from 12 bat species and 176 swab samples from 32 other animal species (Table 1) were collected. To identify coronaviruses from these specimens, a pair of consensus primers that can cross-react with a number of coronavirus RdRp sequences were used to screen the field samples (Table 2, PCR 2). Positive PCR amplicons were detected in three different bat species (Miniopterus spp. [Table 1]). In particular, 12 of 19 (63%) fecal swab samples from Miniopterus pusillus were positive in the screening test. These specimens were collected from three different geographical locations (Fig. 1A).

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Protein sequence alignment of coronavirus RdRps. The conserved motifs for RdRps are indicated above the sequences. Specimens collected from M. pusillus at geographical site 1 (*), M. pusillus at geographical site 2 (¶), and M. schreibersii (specimen 86) and M. magnater (specimens 88 and 96) at site 3 (+) are indicated. 229e, HCoV-229E. (B) Phylogenetic analysis of RNA sequences encoding RdRp (partial sequence).

Identification of a Bat-CoV.To characterize the RdRp sequence deduced from PCR 2 (Table 2), amplicons from each animal were subjected to DNA sequencing. All the sequences generated from M. pusillus were highly similar (Fig. 1A). Although there are sequence polymorphisms among these sequences, these sequences form a distinct branch within the clade of group 1 coronavirus RdRp sequences (data not shown). These results suggested that the coronavirus circulating in M. pusillus is a novel virus. More interestingly, viral sequences deduced from Miniopterus magnater (Fig. 1A, samples 88 and 96) and Miniopterus schreibersii (Fig. 1A, sample 86) were also highly similar to those sequences from M. pusillus (Fig. 1A), indicating that viruses isolated from these three bat species are the same virus or are of the same lineage.

Attempts to isolate this virus in cell cultures were made. Fecal and respiratory samples were used to infect Madin-Darby canine kidney (MDCK), fetal rhesus kidney (FRhK4) and African green monkey kidney (Vero E6) cells. However, no evidence of virus replication was detected in these cells by cytopathic effect or by RT-PCR.

Characterization of Bat-CoV RdRp, helicase-ExoN, and S sequences.A representative sample from M. pusillus (specimen 61 [Fig. 1A]) was selected for further sequence analysis. As mentioned above, our preliminary data indicated that this novel virus is a group 1 coronavirus. On the basis of the conserved regions of group 1 viruses (HCoV-229E, HCoV-NL63, PEDV, TGEV, FIPV, and CCoV), 30 sets of primers were generated for the determination of the viral sequences. Of these PCRs, five additional viral RNA sequences were deduced (Table 2, PCRs 1 and 3 to 6). Alignments of these deduced sequences generated three RNA fragments containing partial sequences for the (i) RdRp gene, (ii) genes C terminal of the helicase gene and N terminal of the hypothetical exonuclease (ExoN) gene, and (iii) S genes (Table 3). The percent sequence identity of each RNA fragment to group 1 to 3 viruses was determined (Table 4). All these RNA sequences shared the highest identity with group 1 coronaviruses (54 to 75%). Phylogenetic analyses of these viral sequences resulted in consensus trees with similar topologies (Fig. 1B, Fig. 3A, and Fig. 4A). In all cases, Bat-CoV clustered with group 1 coronaviruses. In particular, the RdRp and S sequences were most closely related to those of PEDV.

Sequence analysis of the partial sequence of the S gene reveal that this RNA sequence encodes the S2 protein subunit (2, 23). The S protein of coronavirus is known to be heavily glycosylated, and 11 potential N-glycosylation sites were identified (Fig. 2). The deduced sequence contains the HR1 region and part of the sequence of the HR2 region (Fig. 3B and C). These HR regions were separated by an interhelical domain of ∼130 amino acid residues (Fig. 2). As with other group 1 coronaviruses, Bat-CoV also has 14 “additional” amino acid residues in both HR regions (Fig. 3C) (2).

• Open in new tab
• Download powerpoint

FIG. 2.

Protein sequence alignment of coronavirus S proteins (partial sequence). The HR1 and HR2 regions are indicated. The locations of potential N-glycosylation sites in the Bat-CoV sequence are marked by asterisks. TGV, TGEV; NL63, HCoV-N63; 229E, HCoV-229E; OC43, HCoV-OC43.

• Open in new tab
• Download powerpoint

FIG. 3.

(A) Phylogenetic analysis of RNA sequences for the S gene (partial sequence). (B) Predicted coiled-coil regions in the deduced S-protein sequence. The coiled-coil regions were predicted by Stablecoil 1.0 with a 35-residue window width. The HR1 and HR2 regions are indicated. (C) HR1 and HR2 in the S protein of Bat-CoV. The a and d positions of the strongest predicted coiled-coil heptad repeats are indicated. The 14-amino-acid residue insertions that are unique in group 1 viruses are underlined.

The partial RdRp protein sequence of Bat-CoV contains several conserved motifs of RdRps (motifs A to C, G, and F in Fig. 1A) (18, 25). Motif A has two conserved Asp residues separated by four residues and is known for metal ion binding and for recognition of the ribonucleoside triphosphate sugar ring. Motif B contains the highly conserved Ser, Gly, Thr, and Asn residues and is known to be involved in selection of the correct ribonucleoside triphosphate substrate. Motif C contains the highly conserved SDD sequence and is associated with metal ion and 3′-primer terminus binding. Motif F contains several conserved positively charged basic residues and is associated with nucleoside triphosphate binding. This motif could be divided into three submotifs (F1 to F3) (25). Like other coronaviruses, the F motif of Bat-CoV lacks the F2 submotif. The biological functions of the F2 submotif in other viruses are yet to be determined. The G motif had a conserved SXGXP sequence and is known to be involved in positioning of the 5′ template strand in other RdRps.

The RNA fragment for the helicase-ExoN junction encodes the last 54 amino acid residues of the RNA helicase and the first 143 amino acid residues of the putative exonuclease (Fig. 4B) (21). The partial sequence for the helicase protein contains the conserved helicase motifs V and VI (12). The partial sequence for the putative ExoN protein contains motif 1 of the DEDD exonuclease superfamily (21). At the junction of these two proteins, a conserved cleavage signal for 3CL proteinase was identified (LQS at positions P2 to P1′), which suggests that the junction may be cleaved by 3CL proteinase expressed by Bat-CoV.

• Open in new tab
• Download powerpoint

FIG. 4.

(A) Phylogenetic analysis of RNA sequences coding for helicase-ExoN (partial sequence). (B) Protein sequence alignment of coronavirus helicase-ExoN. The conserved motifs for helicases and the first motif for ExoN (DEDD motif I) are indicated. The invariant acidic residues in DEDD motif 1 are labeled with white stars below the sequences. The inverted open triangle above the sequences marks the predicted 3CL proteinase cleavage site.