Severe acute respiratory syndrome (SARS) is a life-threatening form of pneumonia. In the course of a few months in 2003, an epidemic emerged that has spread from its likely origin in Guangdong Province, China, to 32 countries. By 11 June 2003 more than 8400 cases and 789 deaths had been recorded by the World Health Organization. The rapid transmission by aerosols (and probably also the faecal–oral route) and the high mortality rate make SARS a global threat for which no efficacious therapy is available. There is now clear evidence that SARS is caused by a previously unknown coronavirus, provisionally termed SARS coronavirus (SARS-CoV). Genome sequences of SARS-CoV isolates obtained from a number of index patients have been published recently and provide important information on the organization, phylogeny and variability of the 29·7 kb positive-strand RNA genome of SARS-CoV. By analogy with other coronaviruses, SARS-CoV gene expression is expected to involve complex transcriptional, translational and post-translational regulatory mechanisms, whose molecular details remain to be determined. SARS-CoV genome expression starts with the translation of two large replicative polyproteins, pp1a (486 kDa) and pp1ab (790 kDa), which are encoded by the viral replicase gene (21 221 nt) that comprises ORFs 1a and 1b. Expression of the ORF1b-encoded region of pp1ab is predicted to involve ribosomal frameshifting into the -1 frame just upstream of the ORF1a translation termination codon. The pp1a and pp1ab polyproteins are processed by viral proteases to yield the functional components of the membrane-bound replicase complex. In contrast to most other coronaviruses, which use three proteases activities for replicase polyprotein processing, SARS-CoV is predicted to encode only two proteinases. The replicase complex mediates both genome replication and transcription of a ‘nested’ set of subgenomic mRNAs. These mRNAs encode the structural proteins, S, E, M and N, and a set of accessory proteins whose number and sequence vary among different coronavirus species. The extraordinary size of the coronavirus replicase (poly)proteins, their generally large phylogenetic distance from those of other RNA viruses, and the presence of several predicted RNA processing activities which are not found in other positive-strand RNA viruses, indicate that coronavirus replicases are of an unparalleled complexity. The underlying biological mechanisms and functional constraints that determine the evolution and conservation of these unique activities remain to be elucidated.