Proteins are vital components of every cell. They activate genes, enable motion, catalyze biochemical reactions--the list goes on. Biotechnologists are thus understandably eager to unravel their every secret: only with a thorough comprehension of natural proteins can they engineer novel ones with special properties.

Findings published online today by the Proceedings of the National Academy of Sciences represent intriguing progress on that front. Current efforts to construct new proteins follow two approaches. In the first, scientists painstakingly design strings of the protein building-blocks known as amino acids according to the atomic interactions of the entire molecule. In the second, amino acids are randomly thrown together in all manner of combinations, the logic being that given enough combinations, a promising new protein should result sooner or later.

Michael H. Hecht of Princeton University and his colleagues, the authors of the new report, have developed a tactic that seems to wed the best of both worlds. By imposing a few rules on the so-called combinatorial libraries of amino acid sequences--rules favoring sequences bearing the basic structure of natural proteins--the researchers have been able to eliminate a number of the hopeless molecules from the outset. Specifically, they produced focused libraries of artificial sequences by dictating the pattern of water-loving and water-fearing amino acids--but not the exact identities of the amino acids--so as to encourage proper, protein-like folding while fostering the creation of novel combinations.

To test the viability of the approach, the team analyzed one of the resulting "de novo proteins," S-824. With the help of a spectrometer, the investigators determined that S-824 had among other things just the sort of water-loving surface and water-fearing core characteristic of natural proteins.

Such designer proteins could one day deliver drugs to ailing cells, form the basis of "smart" materials, or serve as superior catalysts. The process outlined by Hecht and his collaborators could streamline the search for man-made molecules up to those lofty tasks.