Log In Join Now
Forums  |  Help  |  Settings  |  Download
News and Media

World Community Grid Completes First Stage of FightAIDS@Home  

By Prof. Arthur J. Olson, Ph.D., Dr. Garrett M. Morris, M.A. (Oxon.), D.Phil. (Oxon),
Dr. William M. Lindstrom, Jr., Ph.D., & Alexandre Gillet, The Scripps Research Institute

World Community Grid has already achieved a significant milestone on FightAIDS@Home, completing Stage 1a of the project, which was launched in November, 2005.

The National Cancer Institute (NCI) maintains a database of chemical compounds, many of which are available as actual samples for testing in test tubes. There are some 230,000 compounds in all. Somewhere in this giant haystack of molecules, there might be one or more "lead" molecules for new HIV protease inhibitors. A "lead" molecule is one that is similar to a final drug, but must first be modified by medicinal chemists to make it less toxic or more soluble in water. The NCI constructed a subset of this database that is supposed to represent the chemical variety of all the molecules in the complete database. This subset consists of nearly 2,000 chemical compounds and is called the "Diversity Set".

All life and viruses have a genetic blueprint, or "genome," that consists of a string of DNA letters. In the DNA alphabet, there are only 4 letters, or "bases" as they are known: A, C, G and T. These letters, when written in a particular order create "genes," genetic instructions on how to build "proteins," the molecular machines of life. The genome of the Human Immunodeficiency Virus (HIV) consists of nearly 10,000 bases, which code for just 9 genes. The most commonly-occurring form of HIV is known as the "wild type" form. When HIV infects its target cells, one of its proteins that it uses to replicate its own genetic instructions makes mistakes, and this gives rise to many "mutant" forms of HIV. Some of these mutants happen to be more resistant to the drugs currently used to treat HIV disease. So it is vital that we discover new drugs that are more robust, that can defeat not just the wild type but also the drug-resistant mutant forms of HIV.

One of the 9 genes in the HIV genome codes for a protein called "HIV protease." A protease is a molecular machine that cuts proteins. When HIV replicates, it converts its genetic instructions into one long chain of proteins, but for the virus to mature properly, it must cut the long "polyprotein" into separate proteins. If we block the active site of the protease with a small molecule, a bit like a key fitting into a lock, we can stop the virus from cutting the polyprotein, and thus prevent the virus from maturing into an infectious virus. This is exactly how clinically approved protease inhibitors work.

Stage 1a involved virtual screening of the 2,000 or so compounds in the NCI Diversity Set against 270 wild type and mutant HIV proteases to discover potential new leads and ultimately new drugs. We have analyzed the results for the wild type and have confirmed that our "positive controls" (molecules that we know bind to HIV protease) do, in fact, bind the most tightly to the wild type form of HIV protease, which verifies that the virtual screening is working. These include clinically-approved HIV protease inhibitors such as Indinavir, Saquinavir and Ritonavir, which are currently prescribed to HIV-infected patients. We have ranked the results from our virtual screening and will be presenting the best anti-HIV protease compounds from the NCI Diversity Set to our synthetic chemists so they can incorporate these "lead" molecules into the design of even better inhibitors than the clinically-approved drugs.

The next step is Stage 1b, where we will be screening 230,000 compounds -- the entire NCI database -- against just wild type HIV protease. This will allow us to validate not only the results of Stage 1a, but also the methodology of screening a diverse subset of a larger database against a target. We may discover new compounds against HIV protease, which we could not have found using just the Diversity Set.

We also are preparing Stage 2, which will involve screening virtual libraries of compounds designed in conjunction with our collaborating synthetic chemists against the broad panel of 270 wild type and mutant HIV proteases, again to find new protease inhibitors that work against more than just the most common form of HIV protease. Additionally, the top hits from Stage 1b will be investigated in more detail against this broad panel of 270 targets.

We are very pleased with the insight that we have gained thus far, which could not have been accomplished in the absence of World Community Grid. Thanks to the more than 170,000 volunteers donating their computing time, we have been able to do more than 2 quadrillion calculations, many of which are far more advanced than we have been able to do previously.

Return to Top