Building chemicals the new-fashioned way.

Building Chemicals the New-Fashioned Way

When organic chemists aim to build a specific chemical structure, they often spend most of their time and money trying to isolate it from a brew of by-products that form during the synthesis. That's why molecule makers admire the chemical mastery of organisms, which make and use complex enzymes to coax reactions into yielding a single product.

Scientists at Harvard University now have devised a versatile strategy of achieving such specificity in the laboratory. they say their efforts will pay off in easier, more efficient and cheaper synthesis of drugs, vitamins and other chemicals. The researchers have made a variety of small, relatively simple molecules--which they call "chemzymes"--that can pull off enzyme-like feats. Like huge and complex biological enzymes, "chemzymes bring together different reactants and force them to react at an accelerated rate," says Chemist Elias J. Corey, head of the research effort.

At this week's National Organic Symposium held at Cornell University in Ithaca, N.Y., Corey reported designing, making and using chemzymes that finetune even today's most widely used molecule-building reactions to churn out only the desired products. Since virtually no by-products form in the first place, many difficult chemical separation and purification steps become unnecessary. "This reaction specificity is the direction toward the future," remarks MIT chemist K. Barry Sharpless, who also is developing reaction-tuning tactics.

Corey's group has designated a chemzyme that catalyzes the first of a series of reactions for making prostaglandins, a family of naturally occurring chemicals that regulate blood flow, blood pressure and other vital signs. Physicians use prostaglandins to induce labor and treat ulcers, among other things.

In the first step of the synthesis, a molecule containing two pairs of double-bonded carbon atoms--a diene--reacts with another molecule, a dienophile, which readily combines with the diene to form a ring of carbon atoms. When Corey first reported a laboratory synthesis of prostaglandins 21 years ago, chemists had no way of controlling the orientation at which these two reactants would approach each other during this step, known as a Diels-Alder reaction. The result was a brew of nearly identical products called isomers, from which chemists had to painstakingly isolate the desired prostaglandin precursor.

With the new approach, says Corey, "the chemzyme brings together the diene and dienophile in a very specific three-dimensional arrangement that gives only one of the 16 possible products. This makes the whole synthesis easier and more cost effective." The importance of controlling reactions in this way mushrooms when the target chemical can have isomers based on many of its carbon atoms. For molecules of medium complexity like prostaglandins, which have about eight such carbon atoms, there can be as many as 256 isomers.

In another example, Corey and graduate student Gregory Reichard designed a chemzyme for a more elegant synthesis of fluoxetine, an antidepressant drug. The drug comes as a mixture of two mirror-image isomers, called enantiomers, only one of which is thought to be therapeutic. Using a chemzyme, the Harvard chemists have designated a reaction sequence that yields either one or the other enantiomer. By eliminating unwanted isomers, drug makers hope to reduce side effects.

At the meeting, Corey described other chemzymes his team has made. "Our approach has been to develop chemzymes for the most powerful synthetic construction reactions [such as the Diels-Alder reaction] because that's where they are needed and will have the greatest impact," he says. Another bonus: Their simplicity should open a new window on basic reaction mechanisms.

Also at the meeting, chemist Philip D. Magnus of the University of Texas at Austin described a different strategy for making new drugs. Many biologically active compounds such as insulin are based on peptide bonds, which connect amino acids into proteins, including enzymes and some hormones. Doctors must inject peptide drugs rather than give them orally, because digestive enzymes snip peptide bonds. Magnus described efforts to make metabolically stable "synthetic proteins" by linking pentagonal pyrrole molecules adorned with the same side groups found on amino acids. So far, he has shown that short pyrrole strings twist into shapes akin to protein helices.