The four major classes of protease enzymes1-4 (aspartic, serine, cysteine, and metallo) selectively catalyze the hydrolysis of polypeptide bonds. Their control over protein synthesis, turnover, and function enables them to regulate physiological processes such as digestion, fertilization, growth, differentiation, cell signaling/ migration, immunological defense, wound healing, and apoptosis. Proteases of these classes are also crucial for disease propagation, and inhibitors of such proteases are emerging with promising therapeutic uses3,5 in the treatment of diseases (Table 1) such as cancers,6-8 parasitic, fungal, and viral infections (e.g. schistosomiasis,9,10 malaria,11,12 C. albicans, 13,14 HIV,15-17 hepatitis,18,19 herpes20,21), and inflammatory, immunological, respiratory,22-25 cardiovascular,26 and neurodegenerative disorders including Alzheimer's disease.27 There are now many designed potent and selective protease inhibitors that slow or halt disease progression, inhibitors of the human immunodeficiency virus protease (HIV-1 protease) being notable for the speed with which they became available to humans.17,28-31 To be effective as biological tools, protease inhibitors must be not only very potent but also highly selective in binding to a particular protease. As potential drugs, protease inhibitors must in addition have appropriate pharmacokinetic and pharmacodynamic properties. Clues to how specific proteases selectively recognize small molecules most often come from peptide substrates for proteases. Although peptides display a diverse range of biological properties, their use as drugs is however usually compromised by their instability, low bioavailability, and poor pharmacological profiles. To be effective drugs, protease inhibitors need to have minimal peptide character, high stability to nonselective proteolytic degradation, good membrane permeability, long lifetimes in the bloodstream and in cells, low susceptibility to elimination, high selectivity for a protease, and good bioavailability (preferably by oral delivery). These properties usually require the compounds to have a low molecular weight (e1000 Da).32 Protease inhibitors have been traditionally developed by natural product screening for lead compounds with subsequent optimization or by empirical substrate-based methods,17 involving truncating polypeptide substrates to short peptides (<10 amino acids), replacing the cleavable amide bond by a noncleavable isostere, and optimizing inhibitor potency through trial and error structural modifications that progressively reduce the peptide nature of the molecule. This substrate-based drug design has been substantially improved in recent years with the availability of three-dimensional structural information for proteases, permitting receptor-based design. This involves using structural information about the active site of the receptor (or protease) and fitting into it selections of designed molecules with the aid of computers. Combinatorial chemistry also presents opportunities both to discover new molecular entities for assaying and to optimize lead structures for development of protease inhibitors. Most proteases are sequence-specific, the size and hydrophobicity/hydrophilicity of enzyme sites defining possible binding amino acid side chains of polypeptide substrates. The standard nomenclature33 used to designate substrate/inhibitor residues (e.g. P3, P2, P1, P1′, P2′, P3′) that bind to corresponding enzyme subsites (S3, S2, S1, S1′, S2′, S3′) is shown in Figure 1. Recently it has been convincingly demonstrated for a wide range of proteases that aspartic, serine, cysteine, and metallo proteases universally bind their inhibitors/ substrates in extended or â-strand conformations; that is, the peptide backbone or equivalent is drawn out in a linear arrangement.34,35 This common conformational requirement for recognition by proteases suggests new efforts to develop conformationally restricted inhibitors that adopt receptor-binding conformations and thus are entropically advantaged for binding to a protease. On the other hand, most of the many thousands of protease inhibitors that have been developed to date are relatively flexible molecules that have to use energy to rearrange into a protease-binding conformation. A possible trend in the development of more selective and potent protease inhibitors may be the use of more conformationally restricted molecules that are fixed in the protease-binding conformation. We now describe some of the better-studied small molecule inhibitors of aspartic, serine, cysteine, and metallo proteases, illustrate briefly how some of them bind to proteases, report their inhibitor potencies, and comment upon their pharmacological properties where available and their clinical prospects (Table 1).