Volume 4, Issue 30 August 2007
Can't Cut This
Berkeley chemistry professor Jonathan Ellman is also affiliated with the California Institute for Quantitative Biosciences (QB3) and the Department of Cellular and Molecular Pharmacology at UC San Francisco.
When a malaria parasite lands in your blood, one of the first things it does is whip out its scissors. As fast as it can, this protozoan snips the hemoglobin in red blood cells to get the nutrients it needs to survive. Of course, the microbe behind this deadly disease doesn't actually deploy stainless-steel blades. Instead, it uses an array of biochemical scissors known as proteases.
Proteases are enzymes that snip proteins. They recognize certain strings of amino acids on a substrate protein, bind to this area, then break a nearby chemical bond. Proteases can destroy proteins by snipping them in half, as in malaria. They can also activate proteins by lopping off atoms covering a reactive site.
This versatility has made proteases critical to all manner of organisms, from viruses to plants to humans. Over the past 10 years, protease inhibitor drugs have become indispensable in the fight against AIDS, cardiovascular disease and diabetes.
But finding protease inhibitors is no picnic. Humans manufacture tens of thousands of proteins; figuring out which of these a protease targets is extremely challenging and time consuming.
However, "if you can identify the combination of side chains a protease cleaves, that can really help you figure out what its function is, and how to block it," says UC Berkeley Professor of Chemistry Jonathan Ellman.
Ellman has developed several methods to speed the matching of protease to substrate. His techniques to synthesize test molecules and detect good matches are now being used in the development of therapeutics against many diseases.
In collaboration with biochemist Charles Craik of UC San Francisco, Ellman first pioneered a method to create libraries of test molecules quickly and efficiently. Instead of mixing liquid chemicals and painstakingly purifying them again at each step, he attaches his precursor molecules to polystyrene beads resembling sand grains. To add more atoms to this chemical skeleton, he simply adds the beads to another chemical solution. "This allows you to make a lot of different side chains in substrates very rapidly, so you can synthesize many compounds in parallel," Ellman says.
Here, a protease inhibitor (stick figure) identified using Ellman's substrate activity screening method is bound to the protease cathepsin-S (surrounding 3D x-ray structure).
Initially, Ellman and Craik prepared and evaluated mixtures of substrates. This approach has been applied successfully to over 200 proteases. Since then, Ellman has developed a method to evaluate many individual substrates at once. To each substrate, he adds a reporter molecule that fluoresces when broken off. He then binds a dot of each type of substrate to a glass slide. A single slide can accommodate up to 10,000 different substrate droplets. A solution containing the protease gets poured on top. If a substrate is a match, the proteases do their snipping, and the severed pieces will fluoresce. The brightening fluorescent glow provides a clear signal that there is a hit.
"Once you've found a match, you can use a computer to scan the sequences of proteins encoded by the human genome to identify likely candidate targets. This can guide your efforts to better understand what that protease does," Ellman says.
Most new drugs are relatively large and complex, made up of hundreds of atoms. Trying to screen every possible arrangement of atoms in molecules of this size is impossible; the permutations are more numerous than there are molecules in the universe.
Ellman sidesteps this problem by working with molecules a third to two-thirds the size of most drugs. He creates fragments of drug-like substrates, then identifies those cleaved by a target protease. This technique, called substrate activity screening, increases the chances of identifying a match.
Because Ellman builds every substrate, he also knows exactly how the molecules are bound and cleaved by the protease. With this information, he can replace the bond that is normally cut with a more stable structure. The resulting inhibitor molecules act like decoys, keeping the protease occupied and reducing the damage to body proteins.
Ellman has developed a similar version of this assay for another class of enzymes called phosphatases. The bacterium that causes tuberculosis excretes two types of phosphatases into the body of its host. Their presence appears necessary for TB cases to worsen. Working with UC Berkeley Professor of Biochemistry and Molecular Biology Tom Alber, Ellman has identified two molecules capable of inhibiting these enzymes.
"Because these phosphatases are located outside the bacterium, they may be more accessible to drugs than typical bacterial drug targets, making it easier to kill off the bacteria," Ellman says. "By taking these enzymes out, we could provide an assist to the human immune system."
To date, Ellman's methods have been used in a wide variety of biomedical research. One company has employed them to understand why HIV develops resistance to protease inhibitors. Another is using them to develop proteases that could become a whole new class of drugs. And pharmaceutical giant Merck has used his methods to aid in the development of a new drug against type 2 diabetes. The drug was approved by the FDA just last year.