Jie Dong, Shuyi Zhang, Qi Jin, Junpeng Zhang, Guimei He, Xianwen Ren, Zhiqiang Wu, Fan Yang, Yafang Zhu, Lilian Sun, and Li Yang
To the Editor: Middle East respiratory syndrome coronavirus (MERS-CoV), a novel lineage C betacoronavirus, was first described in September 2012, and by April 16, 2014, the virus had caused 238 infections and 92 deaths in humans worldwide (1). Antibodies against MERS-CoV in dromedary camels were recently reported (2), as was the full genome of MERS-CoV from dromedary camels (3). Finding the natural reservoir of MERS-CoV is fundamental to our ability to control transmission of this virus to humans (4). We report a novel lineage C betacoronavirus identified from Vespertilio superans bats in China. The full-length genome of this betacoronavirus showed close genetic relationship with MERS-CoV. Together with other evidence of MERS-CoV–related viruses in bats (5–8), our findings suggest that bats might be the natural reservoirs of MERS-related CoVs. In June 2013, we collected anal swab samples from 32 V. superans bats from southwestern China. A small proportion of each sample was pooled (without barcoding) and processed by using virus particle–protected nucleic acid purification and sequence-independent PCR for next-generation sequencing analysis with the Illumina (Solexa) Genome Analyzer II (Illumina, San Diego, CA, USA). Redundant reads were filtered, as described (9), from the raw sequencing reads generated by the genome analyzer and then aligned with the nonredundant protein database of the National Center for Biotechnology Information (ftp://ftp.ncbi.nlm.nih.gov/blast/db/) by using BLAST (http://blast.ncbi.nlm.nih.gov). The taxonomy of these aligned reads was parsed by using MEGAN 4 (http://ab.inf.uni-tuebingen.de/software/megan/). On the basis of the BLAST results, 8,751,354 sequence reads 81 nt in length were aligned with the protein sequences of the nonredundant protein database: 72,084 of the reads were uniquely matched with virus proteins. Of these 72,084 reads, 32,365 were assigned to the family Coronaviridae, primarily to lineage C of the genus Betacoronavirus, and found to share 60%–97% aa identity with MERS-CoV. The MERS-CoV–related reads were extracted and assembled by using SeqMan software from the Lasergene 7.1.0 program (DNASTAR, Madison, WI, USA), resulting in a draft CoV genome. Reverse transcription PCR selective for the partial RNA-dependent RNA polymerase (RdRp) gene of this novel lineage C betacoronavirus suggested that 5 of the 32 samples (≈16%) were positive for the novel betacoronavirus, and the PCR amplicons shared >98% nt identity with each other. Using a set of overlapped nested PCRs and the rapid amplification of cDNA ends method, we determined the full-length genome of 1 strain of this V. superans bat–derived betacoronavirus (referred to as BtVs-BetaCoV/SC2013, GenBank accession no. {"type":"entrez-nucleotide","attrs":{"text":"KJ473821","term_id":"627792518","term_text":"KJ473821"}}KJ473821). The betacoronavirus strain had a genome length of 30,413 nt, excluding the 3′ poly (A) tails, and a G+C content of 43.1%. Pairwise genome sequence alignment, conducted by the EMBOSS Needle software (http://www.ebi.ac.uk/Tools/psa/emboss_needle/) with default parameters, suggested that the genome sequence of BtVs-BetaCoV/SC2013 showed 75.7% nt identity with that of human MERS-CoV (hCoV-MERS); this shared identity is higher than that for other lineage C betacoronaviruses (from bats and hedgehogs) with full genomes available. hCoV-MERS showed 69.9% nt identity with bat CoV (BtCoV) HKU4-1, 70.1% nt identity with BtCoV-HKU5-1, and 69.6% nt identity with hedgehog CoV EriCoV-2012–174. Compared with those lineage C betacoronaviruses, which had an 816-bp partial RdRp sequence fragment available, BtVs-BetaCoV/SC2013 shared 96.7 % aa identity with hCoV-MERS. Pipistrellus BtCoVs found in Europe (BtCoV-8-724, BtCoV-8-691, BtCoV-UKR-G17) shared 98.2 % aa identity with hCoV-MERS, and Eptesicus BtCoV found in Italy (BtCoV-ITA26/384/2012) and other lineage C betacoronaviruses shared 96.3 % aa and