Coronaviruses have single-stranded, positive-sense RNA genomes of 26-30 kilobases, by far the largest non-segmented RNA virus genomes currently known. The key functions required for coronavirus RNA synthesis are encoded by the viral replicase gene. The gene comprises more than 20,000 nucleotides and encodes two replicase polyproteins, pp1a and pp1ab, that are proteolytically processed by viral proteases. Over the past years, it has become clear that the unique size of the coronavirus genome and the special mechanism that coronaviruses (and several other nidoviruses) have evolved to produce an extensive set of subgenome-length RNAs is linked to the production of a number of nonstructural proteins (nsps) that is unprecedented among RNA viruses. Many of these replicase cleavage products are in fact multidomain proteins themselves, thus further increasing the complexity of protein functions and interactions. Structural studies suggest that several nsps, following their release from larger precursor molecules, form dimers or even multimers. The various pp1a/pp1ab precursors and processing products are thought to assemble into large, membrane-associated complexes that, in a temporally coordinated manner, catalyze the reactions involved in RNA replication and transcription and, it is presumed, serve yet other functions in the viral life cycle.