Bringing science and development together through news and analysis

  • Supercomputer models pneumonia drug resistance

Scientists have used large-scale computer models to identify how the organism that causes pneumonia is able to resist drugs.

This knowledge could help researchers design new drugs for the disease that kills 3.5 million people a year, mostly in developing countries. Resistance to existing drugs is on the rise.

Resistance of Streptomyces pneumoniae to drugs is caused by a genetic mutation in one of the bacterium's proteins. But the mutation is situated far from the region where drugs usually interact with the protein, giving scientists no clues as to how the bacterium evades the drug.

A research team led by Peter Coveney at University College London, United Kingdom, created a three-dimensional model of the mutated protein. They did this using information from three other bacteria that have the same protein, combined with what they knew about the mutation in resistant S. pneumoniae.

They then ran computer simulations to see what would happen when the protein from the 'resistant' strain came in contact with drugs used to treat pneumonia.

The team found that the mutation created a tiny difference in the protein's shape, and that this was key to the bacterium's ability to resist drugs.

The simulation showed that in normal strains, the drug bound tightly to the protein. However, the slightly different shape of the mutated protein meant the drug could not bind to it. Instead, it drifted away.

If further experiments confirm the results from these computer models, say the researchers, they could help design new drugs that can target resistant strains of bacteria.

The team are now working on resistance to HIV drugs.

The experiments were done through the UK's 'e-science' initiative, which pools computing power from different locations to create a 'supercomputer'.

Sharing resources this way allows research to be done "that would have taken months to perform and quite probably would never have been done", says Coveney.

The team published their work this week (15 August) in the Philosophical Transactions of the Royal Society.

Reference: Philosophical Transactions of the Royal Society 363, 2055 (2005)