Hong Kong SARS sequence.

Abstract

My team of eight graduate students and a postdoc produced one of the first sequences of the SARS coronavirus. In this fast-moving project, we found ourselves striking a balance between time and accuracy. We received the starting viral RNA material during the late afternoon of 7 April. We designed a full set of degenerate primers, as well as using the primers from another Nidovirus PRRSV, so that we could begin sequencing the genome in multiple places. For the completion of the genome sequencing, we also constructed a cDNA library. To characterize the subgenomic transcripts and the leader sequence, we used two separate RACE assays to characterize their 5' ends. The Canadian British Columbia Cancer Agency and the U.S. Centers for Disease Control groups posted their SARS coronavirus genome sequences on the Internet on 12 and 14 April, respectively. On 16 April, at 11:40 p.m., we started to upload our complete HKU-39849 viral genome. Examination of the online SARS coronavirus sequences revealed that the Tor2 sequence from the Canadian group lacked the first 15 nucleotides of the 5' leader sequence but contained the 3' poly-A in the 14 April version. The 16 April version of the Urbani sequence posted by the CDC group lacked the 3' poly-A, but contained the 5' leader sequence, with a T as its starting base pair. The first nucleotide of our 5' leader sequence was an A, confirmed by two independent 5' RACE assays on the genome and 6 subgenomic transcripts. We noticed that the first base pairs of the Urbani sequence released on the CDC Web site on 21 April had been revised to an “A,” matching our original 16 April version. On 1 May, the Canadian group added the 5' leader sequence, also reporting an A as the first nucleotide. I am grateful that Science has allowed us to publish this letter so that we can set the record straight.