In November 2002, a deadly respiratory infection first appeared in the Guandong Province of China. In the ensuing months, unprecedented international health efforts moved toward isolating and identifying the source of SARS. From late March to early April 2003, research groups from various parts of the world closed in on the cause, a member of the coronavirus family.[1] Almost immediately following, the 30 kb genome sequence of the SARS-associated coronavirus was published by researchers at the Centers for Disease Control and Prevention (CDC) in Atlanta,[2] and a consortium of Canadian agencies including the British Columbia Cancer Agency (BCCA) Genome Sciences Center in Vancouver.[3]

The speed of the accomplishment is credited to massive efforts at mobilization and collaboration spearheaded by the World Health Organization. "Laboratories that could be considered under normal circumstances as competitors were brought together and worked together in an exemplary way ... and that made it so that the whole sequence of this coronavirus was cracked within two weeks," says virologist Albert Osterhaus of Erasmus University, the Netherlands, a coauthor on the CDC paper.

The sequence papers, covered in this issue's Hot Papers, are just two of a slew of highly cited papers on the cause of SARS. Although SARS has largely left the health-crisis crosshairs, the sequences enlivened research for a virus that had previously been considered harmless, and spurred research to stave off similar outbreaks. "We had over a million hits on our Web site during the weekend that the sequence was derived," notes coauthor Steven Jones, head of bioinformatics at the BCCA Genome Sciences Center. He credits the speed to publication as well as free access to the articles provided by Science for bolstering the paper's impact. With the sequences published online on May 1, 2003, SARS was still making headlines, Jones says.

The sequencing speed was a result of sheer hard work rather than the product of new tools and methods, says Columbia University neurobiologist and virologist W. Ian Lipkin. CDC virologist and coauthor Paul Rota concurs, explaining that his team did their sequencing the "old-fashioned way," by primer walking, in which information from a known part of a sequence is used to design primers for a neighboring sequence. The Canadian group developed and sequenced a cDNA library, and used this to fuel a shotgun-sequencing approach. They assembled the rest into a final sequence.

Though the two groups sequenced different strains of the virus, the resulting sequences differed by only eight nucleotides. This has been credited to the fact that each strain was derived from a late-phase outbreak. "Retrospective sequencing by the groups in China have found quite a bit more genetic diversity in some of the strains that were circulating earlier in the outbreak," Rota says.

Lipkin notes that the most fundamental observations were those that made the initial link between the newly discovered coronavirus and SARS itself. Before the disease's cause was known, scientists had speculated that Chlamydia, human metapneumovirus (hMPV), or a paramyxovirus was the etiological agent behind it. A number of papers, first published online in April 2003, began to solidify the coronavirus connection, however. Hot Papers in their own right, some of these articles have been cited more than 450 times.[4-6]

That a coronavirus could cause lethal infections in humans was surprising at the time. "The previously known human respiratory coronaviruses are generally considered to be very benign [and] cause very limited respiratory illness," notes molecular virologist David Wang of the Washington University School of Medicine, St. Louis, a coauthor on the CDC sequence paper.

The papers offered slightly different insights into the molecular workings of the virus. "The CDC group had some information on the actual transcripts that were being formed by the virus," says Jones, who adds that his group predicted the presence of a stem loop II motif (s2m) RNA element that was also conserved in other viruses. William Scott and colleagues at the University of California, Santa Cruz, recently confirmed and solved the three-dimensional crystal structure of the SARS s2m element. By comparing this to known RNA tertiary structures, the authors proposed a mechanism by which the virus may hijack host protein synthesis.[7]

Others are using the sequence to develop vaccines, drugs, and diagnostic tools. Epidemiologist Ralph Baric and colleagues at the University of North Carolina assembled a full-length infectious cDNA clone of the Urbani strain of the virus; SARS has an RNA-based genome that cannot be manipulated by genetic engineering.[8] "By having an infectious cDNA clone, it is [now] possible to perform reverse genetics," says Luis Enjuanes of the National Center for Biotechnology, Madrid, whose laboratory has also developed an infectious cDNA clone of SARS.

The sequence can tell stories of the virus' progression, too. Edison Liu and colleagues at the Chinese National Human Genome Center in Singapore sequenced early-stage isolates of the virus from Singapore and compared these sequences to those of later-stage isolates to help determine the origin and evolution of the disease.[9] Recently, a team led by Guoping Zhao of the Chinese National Human Genome Center in Shanghai compared sequences isolated from humans and civet cats in Guangzhou, China, during the early stages of the outbreak and identified mutations that may have been responsible for making the virus capable of human-to-human transmission.[10] "The significance of our work was the idea of employing the comparative genomics knowledge to study the epidemiology of the disease and, conversely, employing the epidemiology information to study the evolution of the virus," explains Zhao.

Although SARS research hardly ended with the outbreaks themselves, many scientists note that funding and activity has slowed. "The emergency nature of SARS research has declined, so it's still going on but at a more relaxed pace," says Liu, who is currently developing diagnostic tests for SARS. Baric says he thinks SARS research has continued at a steady pace. Since the identification of SARS, two new human coronaviruses that cause severe pneumonia have been identified. Thus it's likely that human coronaviruses are associated with more disease than previously believed.

Much still needs to be learned about the virus, including its source. Zhao's research demonstrated that even though SARS infected civet cats, these creatures were not the natural reservoir for the disease. "Even now, we don't have a fix on the natural reservoir, ... and without this information we have no way to adequately prevent transfer. We don't know what to monitor," says Lipkin.

Improved animal monitoring is crucial. Lipkin says, "Unless we can break down some of the barriers between comparative medicine and human medicine and think more about diseases of animals, we're going to be caught unawares with other agents like SARS, which will be coming both from the wild as well as from domestic populations."

Indeed, avian flu, which currently cannot be passed from human to human, may ultimately emerge in a virulent form. But the lessons learned from SARS may not be applicable to an avian flu pandemic. Malik Peiris of the University of Hong Kong explains: "There were features of SARS that allowed it to be contained. Most of the time it was not airborne and much of the transmission occurred at the late stage, so we had a window of four to five days to detect and isolate cases and so could interrupt transmission. With a disease like flu a lot of the transmission works in the first few days of the disease, so public health measures are not so effective."