Finding the right small-molecule compound to counter an enzyme orreceptor associated with a disease can make looking for a needle in ahaystack easy.

Knowing the structure of the protein greatly reduces the trial anderror of drug screening. But pinpointing the complex shapes of thegenetic targets using current methods can take months or years. Themost common approach is X-ray crystallography in which proteincrystals have to be grown and bombarded with X-rays.

Once the protein is defined, computer-aided rational drug designerscan give combinatorial chemists a start in building libraries ofcompounds for screening.

Among the problems, however, are that protein crystals don't alwaysaccurately reflect the structure. Growing them in the weightlessenvironment of space produces the best results. And not all proteinscan be crystallized.

Structural Bioinformatics Inc., a San Diego start-up company formedin March 1996, has a quicker and more accurate way of buildingprotein structures from the gene sequences that define them.

"We have created the operating system that lets the structure-baseddrug design chemist work directly with genomics information," saidEdward Maggio, president and CEO of Structural Bioinformatics.

The company's bioinformatics software converts gene sequence datainto 3-D protein structures, which are stored in the company's databases and can be integrated with drug design computer programs.

The patented high-speed algorithms also calculate the proteinstructures' "dynamic trajectory," Maggio said, referring to theflexibility of proteins, which move and rotate. Knowing what lobes ofthe protein move, he observed, improves the accuracy of drug design.

In addition, Structural Bioinformatics' software identifies points onthe protein that provide the most likely site of interaction for atherapeutic compound, assisting drug designers in making a templatemolecule to create a combinatorial library.

"We give them the protein structure and tell them where to look on itto make their drug," Maggio said.

Structural Bioinformatics, however, has not yet solved the dynamicsof protein folding, Maggio added, and for that reason the technologycan be applied to "certain" types of proteins, such as endothelin,conotoxin, galanin and some growth factors. The company has allthose in its data base and they have been used to create drugcandidates.

For example, Texas Biotechnology Corp., of Houston, identified anendothelin antagonist from Structural Bioinformatics' data base andis developing the drug for pulmonary hypertension and congestiveheart failure.

Maggio would not reveal how many protein structures the companyhas calculated, but he said Structural Bioinformatics eventuallyexpects to have as many as there are gene sequences in the humangenome. Current gene sequence estimates run as high as 150,000.

Maggio said only about 3,500 protein structures are contained inpublic data bases.

His company, he added, intends to sell access to its protein structuredata bases as well as work on specific projects with pharmaceuticalcompanies and genomics firms.

"Our strategy is to be the Incyte [Pharmaceuticals Inc.] of proteinstructures," Maggio said, referring to Palo Alto, Calif.-based Incyte'sbusiness of selling pharmaceutical companies subscriptions to itsgene sequence data bases.

Structural Bioinformatics, he added, is focusing on filling its SBd-Base initially with structures of enzymes, receptors and hormonesknown to be drug targets.

On Tuesday the company received $2.5 million in initial funding andexpects to double that before the close of its first round of financing.

Prior to founding Structural Bioinformatics, Maggio headedImmunoPharmaceutics Inc., of San Diego, and his new companylicensed some technology from his former firm. n

-- Charles Craig

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