SARS/avian coronaviruses.
In the hunt for the aetiology of the SARS outbreak in 2003, a newly developed virus DNA micro-array was successfully used to hybridise PCR products obtained by random amplification of nucleic acids extracted from a cell culture infected with material from a SARS patient. The SARS agent was found to hybridise with micro-array probes from both coronaviruses and astroviruses, but one of the coronavirus probes and the four astrovirus probes contained redundant sequences, spanning a highly conserved motif, named s2m, found at the 3' end of the genomes of almost all astroviruses, one picornavirus, and the poultry coronaviruses. The three other coronavirus probes, that hybridised with the SARS agent, were located in the replicase gene, and it could be concluded that the SARS agent was a novel coronavirus, harbouring s2m. The presence of this motif in different virus families is probably the result of recombinations between unrelated viruses, but its presence in both poultry and SARS coronaviruses could suggest a bird involvement in the history of the SARS coronavirus. A recent screening of wild birds for the presence of coronaviruses, using a pan-coronavirus RT-PCR, led to the identification of novel coronaviruses in the three species studied. Phylogenetic analyses performed on both replicase gene and nucleocapsid protein could not add support to a close relationship between avian and SARS coronaviruses, but all the novel avian coronaviruses were found to harbour s2m. The motif is inserted at a homologous place in avian and SARS coronavirus genomes, but in a somewhat different context for the SARS coronavirus. If the presence of s2m in these viruses is a result of two separate recombination events, this suggests that its particular position in these genomes is the only one that would not be deleterious for coronaviral replication, or that it is the result of a copy-choice recombination between coronaviruses, following an ancestral introduction in the coronavirus family by an unrelated virus. In conclusion, the relative high frequencies of recombination observed both experimentally and in the natural evolution of RNA viruses, indicate that horizontal gene transfer does occur, even between unrelated viruses. This might represent a challenge in the rapid identification of novel pathogens with DNA micro-array techniques.