Bioinformatics and protein design

Traditionally, the development of drugs has focused on small molecule therapeutics. However, with recent advances in recombinant protein technology the potential of proteins as therapeutics is starting to be realized. Already there are protein drugs on the market, including naturally occurring proteins, engineered proteins and proteins introduced into the patient by way of gene therapy. The next generation of such drugs are likely to be designed molecules; proteins devised from scratch and specifically tailored to have a required 3-dimensional structure and biochemical function. The key to such de novo design is the extensive bioinformatics knowledge that has been obtained from experimentally determined protein structures during the past 40-50 years. The knowledge is far from complete; for example, the protein folding problem has not yet been completely solved. Despite this, bioinformatics plays a crucial role in protein design, as is outlined in this review, and a number of de novo protein structures have been successfully designed in recent years. Some examples of these successes, which are available in the Protein Data Bank, are presented. They suggest that carefully designed protein therapeutics are a genuine prospect for the future.