Only a few viral diseases are presently treatable because of our limited knowledge of specific viral target molecules. An attractive class of viral molecules toward which chemotherapeutic agents could be aimed are proteases coded by some virus groups such as retro- or picornaviruses (poliomyelitis, common cold virus). The picornavirus enzymes were discovered first, and they have now been characterized by a combination of molecular-genetic and biochemical approaches. Several laboratories have expressed the picornaviral enzymes in heterologous systems and have reported proteolytic activity, as well as the high cleavage fidelity diagnostic of the viral proteases. After dealing with several technical difficulties often encountered in standard genetic engineering approaches, one viral protease is now available to us in quantity and is amendable to mutagenic procedures. The initial outcome of the mutagenesis studies has been the confirmation of our earlier work with inhibitors, which suggested a cysteine active-site class. There is a clustering of active-site residues which may be unique to these viruses. The requirement for an active-site cysteine-histidine pair in combination with detailed information on the viral cleavage sites has permitted design of selective inhibitors with attractive antiviral properties. Future goals include investigation of the structural basis for selective processing and application of the cleavage specificity to general problems in genetic engineering.