Geek Page
Rational Drug Design
You may well live to be 100. The bad news is that you'll have a lot more time to develop degenerative diseases. This opens a potentially vast market for new therapeutics. But traditional methods of drug discovery are in most cases too slow to meet future demands. Rational drug design - using principles of engineering - may provide a solution.
Traditional drug discovery involves blindly testing millions of candidate-molecule mixtures (everything from dirt to sweat samples), selected pretty much at random, in the hope that one might prove to be an effective treatment for some disease. It's a painstaking and laborious process with no guarantee of success. Today, the average cost to discover and develop a new drug is nearly US$400 million, and the average time is 15 years from lab to patient. Improving this even incrementally would represent a huge advance.
A drug is often a small protein molecule that fits receptors on proteins in body tissues. Rational design leverages our scientific knowledge of how drugs work - how two molecules recognize each other's shape and cling together - against the physical and electrostatic characteristics of these proteins. This information helps scientists predict the shape the drug might take and thus how it might work. With rational design, drugs are built to fit particular molecular locks - the drug's specific target - and the tighter and more specific the fit, the better.
Molecular biology 101
The genetic code in everything from molds to mammals is carried by nucleic acids - DNA and RNA. DNA is transcribed into messenger RNA, which is translated into specific amino acids, which are the building blocks of proteins. All proteins in all species are formed from the same set of 20 amino acids. In the last 20 years, we have begun reading and cataloging these blueprints.
This data is now pouring into public and private databases at exponential rates. It is being used to build databases of DNA, RNA, amino acid, and protein sequences, and three-dimensional structures of protein molecules. (See www.ncbi.nlm.nih.gov).
There are about 3 billion base pairs of DNA in the human genome, including perhaps 100,000 to 120,000 distinct genes. At the moment, about 3 percent of this information has been captured, and that information is scattered and spotty at best. Given the rate at which the data is being gathered, that won't be the case for long - even 3 percent is useful in the hunt for new drugs.
Arranging atoms
Researchers comb the databases for potential molecular candidates and then use computer models to further refine the search. For example, scientists created a structural model of HIV-1 protease and found a similar protease (a protein-cutting enzyme) in an unrelated virus. They took a known inhibitor of that protease and engineered it to block HIV-1 protease. Stopping its action prevents the virus from replicating and infecting new cells.
The Assisted Model Building with Energy Refinement model, developed by the Kollman group at UC San Francisco (www.amber.ucsf.edu/), is a computer-based force-field method that can be used to approximate the final shape of a short amino acid chain. AMBER uses several different types of atomic models (varieties of carbon, oxygen, nitrogen, phosphorous, sulfur, and hydrogen) and builds the amino acid atom by atom. It takes into account the way the atoms are arranged, the bonds among them, and their interacting electrons to approximate the shape of a receptor molecule or the drug that might fit that receptor exactly. For a drug based on a known class of enzymes (like the protease inhibitors), this can show the precise mechanism by which the drug works.
Once the researcher has come up with a strong candidate, the process moves from the computer into the lab. This might involve cloning the gene for the protein and obtaining enough of it to start testing its effect on the receptors in living organisms. At this point, the work of fine-tuning and characterizing the drug's reaction begins.
Site-directed mutagenesis is a process by which molecular biologists tweak the shape of these drug candidates in the lab. By modifying the DNA sequence very precisely, you can change the protein in a predictable way.
Proteins are made up of one or more chains of amino acids; these linear molecules then bind to each other, often folding to form a compact, globular structure. This arrangement means that changing one amino acid can change the shape of the entire protein. Each amino acid has a particular shape and electrostatic field; a large amino acid can't fit where a small one might, so the rest of the protein refolds to accommodate the addition.
By substituting one amino acid for another, the drug might bend outward and away from its receptor, allowing it to be degraded more quickly and reducing its strength. The reverse substitution might make the drug last longer.
Getting results
Even though rational design eliminates a large amount of preliminary lab work, the modeling process is far from perfect. One reason is that protein structures are not rigid; they squish and change shape, so the calculations to predict their shape can become arbitrarily complex.
Currently, a short amino acid sequence can be modeled to within 4 or 5 percent of the measured shape of the molecule. This is good enough to select potential drug candidates but not good enough to get a final product without traditional lab testing.
If current trends continue, rational drug design will become the dominant method of discovering new therapeutic agents. Billions of dollars are being invested by large companies such as SmithKline Beecham and Merck to further develop this field, assisted by smaller outfits such as Incyte Pharmaceuticals, Molecular Applications Group, and Human Genome Sciences.
The first few drugs designed with the help of rational methods, like the HIV-1 protease inhibitors, are just entering the market, and more are coming. Which means that seeing - and enjoying - your 100th birthday is more likely than not.
_________________
Andrew MacBride(amacbride@yahoo.com) divides his time between database research and molecular biology.
