In a pair of studies published last year, researchers across Europe used computer simulations to make major advances in our understanding of HIV. Taking advantage of distributed computing networks, they simulated key processes and molecular interactions in the life cycle of the virus, identifying new targets for drug therapy.
In order to invade and infect cells, the HIV virus uses proteins, the molecular machinery of life. Like many human-built machines, proteins accomplish their tasks by being a certain shape; wheels would be much less efficient if they weren’t round, while the specific shape of a key enables it to open a particular lock. A protein’s shape is determined by interactions between the atoms that make it up. Understanding this shape can tell us something about how the protein works or, in the case of an HIV protein, how to stop it from working.
Scientists can measure the shape of a protein using a technique called crystallography, but this only captures a snapshot, freezing a single moment of what is actually a dynamic, variable process. To get an idea of how the shape changes over time, scientists simulate the interactions between the thousands of atoms that make up a protein. The sheer number of interactions makes this an enormous task demanding immense computational power. To run these simulations cost-effectively, researchers use a technique known as grid computing, which combines the processing power from the idle computers of volunteers around the world to create a virtual supercomputer.
One of the questions that’s been tackled using this technology is how newly-formed HIV particles in a cell mature and infect new cells. HIV and other retroviruses produce the proteins they need in a single connected stretch, like a string of pearls, and these proteins have to be cut free before they become active. A special protein called a protease acts like a pair of scissors, cutting the string of pearls, but a problem arises since the protease itself is part of the protein necklace. To make matters worse, just as a scissors needs two blades, the protease only works if two copies come together to act in concert. In other words, for HIV particles to become active and infectious, two strings of proteins have to come together in just the right way for the proteases in them to line up together and snip themselves out, after which they can cut the other proteins free.
Researchers from labs in Germany and Spain used a network called the GPUGRID to understand this complex dance; by combining the power of thousands of volunteers’ computers, they simulated the interactions of all 40,000 atoms in the protease for 400 billionths of a second – a mere fraction of an instant, but long enough to see what happens. In a paper published in the journal PNAS, the team described how the protein contorts to thread one of its ends between the scissor blades and cut itself free. They believe that blocking this crucial step might be a fruitful new strategy in treating HIV.
Once the protease is free, it cuts the other proteins loose. One of these, a “reverse transcriptase”, is the target of many drugs used to treat HIV. In another study, the Spanish lab cooperated with a group in the UK to investigate how these drugs bind to the protein and block its activity. The protein, which copies RNA into DNA, is shaped like an open hand; it works by closing the “thumb” and “fingers” around a RNA molecule to grip it and make a copy. Scientists believe that the drug nestles into the crook of the “thumb”, preventing the protein from working by locking the “hand” into an open position.
The team tested this idea using a network of computers called TeraGRID. In most of their simulations, the drug forced the hand to stay open, but in one case the hand was able to close despite the drug. Even though the thumb folded over the palm in this simulation, the drug forced it to touch the fingers in a different place than it would have otherwise. The researchers suspect that the change in contact between the thumb and fingers affects which bits of RNA the protein recognizes and grips, interfering with its ability to function properly. Based on their new understanding of this process, they suggest that drugs could be designed to force the protein into the non-functional closed state, providing a new target for anti-retroviral therapy.
By harnessing the power of grid computing, scientists have been able to study the development of HIV in greater detail than ever before. Understanding these key molecular processes can help us create new drugs and target existing ones more effectively, offering increased hope to people afflicted with this disease.
Sadiq SK, Noé F, & De Fabritiis G (2012). Kinetic characterization of the critical step in HIV-1 protease maturation. Proceedings of the National Academy of Sciences of the United States of America, 109 (50), 20449-54 PMID: 23184967
Wright, D., Sadiq, S., De Fabritiis, G., & Coveney, P. (2012). Thumbs Down for HIV: Domain Level Rearrangements Do Occur in the NNRTI-Bound HIV-1 Reverse Transcriptase Journal of the American Chemical Society, 134 (31), 12885-12888 DOI: 10.1021/ja301565k