The NS3 protein of Hepatitis C virus contains, within its N-terminal 180 amino acids, a serine proteinase domain that shows only limited homology to cellular serine proteinases (~15% sequence identify). Despite this low sequence conservation, we built a homology model of this domain that has helped us in directing experimental work, even in the absence of detailed structural information. The three-dimensional structure of the HCV NS3 proteinase domain has now been solved by X-ray crystallography, both alone and complexed with the NS4A cofactor. A comparison between the homology model and the newly determined three-dimensional structure reveals that the general topology, as well as several unique structural features of the enzyme, were correctly predicted by our model. Among the features of the NS3 proteinase domain that the model has accurately predicted is the positions of the residues that determine the shape of the S₁ substrate binding pocket. In addition, we predicted the presence of a tetrahedral metal binding site formed by three cysteines and one histidine that lie on the opposite sides of the molecule from the active site. The recent discovery of HCV related viruses sharing about 30% sequence identity with HCV in the serine proteinase domain allows the possibility of modelling these proteins and deducing their biochemical properties.