BACTERIAL FACTORIES: Can microorganisms, such as these antibiotic-producing colonies of Streptomcyes, be altered to churn out new drugs? (University of Minnesota)
The battle against disease is not unlike the mythical task of Sisyphus: always uphill and never-ending. Despite the impressive array of modern medicines, new diseases create a constant need for novel cures, even as familiar microbes such as the tuberculosis bacillus are evolving resistance to some widely used compounds.

At the same time, the pace of discovery of new drugs derived from living organisms has slowed since the1970s; most of the natural biologically active compounds flagged in screening programs these days turn out to be molecules that are already known. Pharmaceutical companies are therefore excited about an innovative technique for producing complex molecules not known in nature, compounds that might form the basis of next century's drugs.

One versatile approach for quickly creating novel compounds, known as combinatorial chemistry, assembles molecules from simple entities (see "Cracking the Combination," by Gary Stix). Unfortunately, many drugs have structures so complex that they would be beyond the reach of present combinatorial methods. Hence, the mounting interest in a more subtle technique--engineered biosynthesis--which presses enzymes and microbes into service as assemblers of candidate pharmaceuticals.

Breakthroughs in the past two years have propelled engineered biosynthesis from an esoteric area of basic research to a promising source of new drugs. Substances made this way have been whimsically dubbed "unnatural natural products," but nobody is laughing about their potential value. "Can you use genetic engineering to generate novel, structurally complex molecules? The answer is an unequivocal yes," says Chaitan S. Khosla of Stanford University, one of the lions in the field. Moreover, Khosla states, some of these custom-made molecules have demonstrated biological effects. He has founded a company, Khosan in San Francisco, to exploit the technique.

The goal of engineered biosynthesis is to imitate and modify the chemical construction processes naturally performed by bacteria and fungi. These microorganisms (which have been rich sources of drugs) make antibiotic compounds and some other large molecules through a multistep process. Some of these compounds, such as the antibacterial agent oxytetracycline and the anticancer agent doxorubicin, are fabricated in the microbes by many different enzymes that each do one job in turn.

But some of the most chemically elaborate pharmaceuticals, including a commercially important class of drugs known as macrolide antibiotics, are built inside bacteria by a large enzyme complex that works like an assembly line.

Much of the recent attention has focused on this assembly-line process, specifically the one used by the bacterium Saccharopolyspora erythraea to produce erythromycin, a widely prescribed antibiotic. Trying to figure out how the process works, Peter F. Leadlay of the University of Cambridge and Leonard Katz of Abbott Laboratories first identified the genes encoding the enzyme complex that makes the drug.

The researchers found that the complex--a type 1 polyketide synthase--consists of three pairs of proteins that together include at least 28 sites where synthetic reactions occur. Each of these reactions adds a piece to the growing erythromycin molecule or forges some chemical bond and then passes the molecule on to the next site. The complex gradually assembles a "backbone" of carbon atoms, eventually joining the two ends to make a large ring: the finished erythromycin molecule.

Leadlay and Khosla are racing each other to understand the details of the chemical assembly line. By analyzing the genes encoding the complex, groups led by the two researchers have shown that many of the active sites in S. erythraea consist of structurally similar units, or modules. That observation raised the encouraging prospect that genetic engineering could shuffle the modules to create variant complexes that might make new, medically useful compounds.

Until this idea was tried, nobody was sure whether altered complexes would make anything at all; perhaps the redesigned complexes would simply grind to a halt. But by splicing genes, both the Cambridge group and Khosla's group (along with the latter's collaborators) have succeeded in producing working versions of the erythromycin complex in bacteria that normally lack it. These reengineered complexes produce structurally modified variants of the antibiotic. For instance, the scientists conducted some genetic tinkering that eliminated parts of the complex; the resulting complexes were reduced in size and synthesized ringlike molecules smaller than erythromycin's.

The chemical structures of those rings were just as predicted, confirming earlier indications that the complex is flexible enough to withstand some changes without "jamming" the chemical assembly line. In at least one case, then, a bacterial complex can be reshuffled to create "unnatural" drugs, ones that the bacterium would never synthesize on its own. Other experiments have shown that the products that emerge from the complex can also be altered by supplying different starting materials.

Recently workers in the field have broken even further with nature, getting the erythromycin complex and other drug-producing complexes to function in a test tube--that is, without using bacteria at all. That emancipation makes it easier to do experiments. Modular structures like the one seen in the erythromycin complex seem to occur in many different microbes: the same basic chain-extending units occur in the complex that synthesizes tetracycline in Streptomyces bacteria, for example.

Pharmaceutical companies are salivating at the prospect that engineered bacteria could yield chimeric antibiotics--hybrid drugs that combine the best features of different parent molecules. Because microbes will not have had a chance to evolve resistance to chimeric molecules, such drugs might be more effective than antibiotics now in widespread use. James Staunton, one of Leadlay's colleagues at Cambridge, says his institution has already applied for patents relating to chimeric antibiotics. Pfizer, Glaxo-Wellcome and Abbott Laboratories are also exploring the technology or are preparing to.

It may take years for the dust to settle and for researchers to learn whether unnatural natural compounds will indeed represent an important source of new drugs. Khosla is upbeat, estimating that the technologies needed to construct large arrays of complex molecules in genetically altered systems are already 80 percent developed. He says his group has already created 100 novel molecules using the biosynthesis approach.

Leadlay offers a more cautious assessment. He points out that some experiments have yielded unexpected products, indicating that the rules that govern enzyme complexes may not always be as simple as they at first appear. But even if unnatural natural products do not themselves turn out to be suitable as drugs, Leadlay notes, they might be incorporated into schemes to create variations on a theme by combinatorial chemistry. Either way, the number of possible untapped compounds is huge and sure to include some treasures. "It's a large haystack," Leadlay muses, "but I am confident some things will be possible."