Computer simulations are a vital tool in the fight against HIV, Judy Redfearn reports
Large-scale computer simulations performed using the National Grid Service are revealing how the human immunodeficiency virus evades the action of drugs. The simulations involve modelling how 30,000 atoms and surrounding water molecules move relative to each other. The atoms belong to an enzyme that the virus uses for replication (HIV protease) and the drug saquinavir, which blocks the enzyme's action.
The virus resists the drug by mutating to produce a variant of the enzyme that the drug no longer blocks. It is a subtle problem as the drug still attaches to the enzyme, but the nature of the interaction between the two has changed to reduce the drug's effectiveness.
Kashif Sadiq at University College London has thrown light on the matter by comparing computer simulations using enzymes from normal and mutant viruses. His simulation tool was a molecular dynamics code used by RealityGrid, an e-Science project funded by the Engineering and Physical Sciences Research Council, to pilot the development of key e-Science tools and techniques. Each simulation takes 24 hours to run, using 32 processors, and Sadiq has performed 60 of them. He has reduced the time they take by using NGS computing resources and the national high-performance computing facility (HPCx) at the Daresbury Laboratory.
"The NGS has enabled me to submit a whole load of jobs and then just look at them from time to time. I get an e-mail back when they are done," he says.
The simulations have revealed subtle differences in the interaction between the drug molecule and the normal and mutant enzymes. The drug molecule is far more likely to bind tightly to atoms at the centre of the normal enzyme and loosely to atoms at the edge of the mutant enzyme.
"The frequency of conformations changes - it is a subtle difference. The mutations seem to be leading in the direction of pushing the drug out of the active site, although we have not observed that yet," Sadiq says. "With a better understanding of how the enzyme adapts to the drug, there is a good chance we can design a better drug."