Winners at poker never forget one cardinal rule of betting: Yourcards don't love you.

Rational drug designers bemoan the truism: Your peptides don't loveyou.

To be sure, peptides specifically bind disease-related molecules, butthen _ to name only a few peptidic flaws _ they quickly break up inthe bloodstream. Often unkind to the digestive system, and insolublein water, they can't be given by mouth.

So medicinal chemists practice what has been called irrational drugdiscovery. They screen thousands upon thousands of molecules, fromrain forests and other sources, seeking desired functions by tedioustrial and error.

Structural biology offers a more rational approach, but with its owntime-line drawbacks. It was X-ray crystallography that jump-startedbiotechnology by tipping Crick and Watson to the double-helicalstructure of DNA. Now crystallography plus computerized screeningand automated synthesis are claiming a quantum leap toward drugdevelopment in the near future.

That implicit forecast takes the form of a paper in today's Science,titled: "Interaction of a Peptidomimetic Aminimide Inhibitor withElastase." It's the joint work of structural biologists at BrandeisUniversity in Waltham, Mass., and combinatorial chemists at ArQuleInc., in nearby Medford.

Here is how ArQule's scientific founder and senior vice president ofresearch and development, Joseph Hogan Jr. _ a co-author of theScience article _ defines his company's unique approach tocombinatorial chemistry:

"It's making every possible molecule from a given set of buildingblocks." These are real molecules, not virtual ones, he pointed out toBioWorld Today in an interview: "We use the computer as aguidance system. The more information we have, the better our guessat making molecules. So we go back and forth. The computer guidesus in what to make, and once we get bioactivity, what to make next.Then our machines actually synthesize 'em."

In their joint project, company and university teams aimed at creatingan inhibitor of the elastase molecule. As the paper's first author,biophysicist Ezra Peisach, told BioWorld, "Elastase is a class ofprotease molecules linked to emphysema, cardiovascular disease,Alzheimer's and AIDS." He added that the purpose of the experimentreported in Science was not linked to therapy, but rather "a test studyto demonstrate proof-of-concept. The elastase molecule is wellunderstood, and we knew what its inhibitor might look like."

A future project, Peisach suggested, might involve a more expensiveand potentially useful enzyme, namely, "an HIV protease inhibitor."

Such an AIDS therapeutic has long been on ArQule's drawing board,as Hogan reported early this year. (See BioWorld Today, Feb. 8, p.1.). He now said, "We've got this HIV protease inhibitor done, muchthe same as the elastase one, and submitted it for publication in adifferent journal." Hogan observed that "it, too, is a demonstratormodel, but we've got a crystal structure of it too, and it's very muchthe way we predicted. So it conforms to the first one." Whether it hascommercial potential as an AIDS drug, Hogan doesn't know. "Thereare a lot of good inhibitors out there. That's a business question, ifanybody cares to take it to market."

Peptidomimesis is the name of the ArQule/Brandeis game. It consistsof constructing molecules that bind disease ligands the way theoriginal, unloving peptides do, but with biophysical properties thatsidestep their shortcomings.

Hogan explained: "Functionally, what a peptide means to us is that itmakes the same contacts in the same directions, with the same sort ofspatial orientation, the same points of a target, as a peptide would."But these impostor molecules, he added, "are much more water-soluble. In fact, we were able to diffuse this one into the elastasecrystal out of water at reasonably high concentrations, whereas thepeptide was a nightmare to handle."

Moreover, "They are resistant to enzymatic action. The enzymeshave never seen these linkages, so they can't chew them up. It foolsthe enzymes. Being water-soluble and electrically neutral, it's orallyavailable."

The name of their masquerading peptide mimic is aminimide. Thisman-made molecule, Hogan said, "has been around for 20 years." Itsindustrial uses range from tire-cord dips to cosmetic creams.

Peisach noted that it contains an extra nitrogen atom, which canmimic the alpha carbon of an amino acid.

"What's next for ArQule," Hogan concluded, "is to take thiscapability and get to work designing things for real-world targets,with commercial and physiological significance." n

-- David N. Leff Science Editor

(c) 1997 American Health Consultants. All rights reserved.