Scientists publish 'atlas' of how proteins flex
change shape and interact with other molecules, which could prove
an invaluable tool to drug developers.
Scientists from the Institute for Research in Biomedicine (IRB), Barcelona, Spain, used molecular dynamics simulations on Europe's most powerful supercomputer to discover structural changes possible in a range of proteins. The research is part of an ambitious wider project called MoDel (Molecular Dynamics Extended Library).
"MoDel aims to establish a 'fourth dimension' for protein structures, thereby providing a complete landscape of possible conformations for the entire proteome (the complete network of protein interactions in a cell), over time," said project director Prof Modesto Orozco.
Proteins carry out their function through binding to other molecules and so determining the structure of a protein is a key step to establishing not only their function but what happens when they function incorrectly.
However, protein structures determined using experimental techniques are merely rigid snapshots that don't reflect the true dynamic nature of the molecules. For example, many proteins are unfolded before another molecule binds to them and others change their shape as a result of binding.
The protein 'atlas' allows scientists to predict the different shapes a protein can adopt, facilitating the understanding of how they interact with other molecules and opening up vast possibilities for the design of new drugs.
Molecular Dynamics (MD) has been used for this purpose since the 1970s and remains one of the most powerful methods of predicting changes in a protein's shape. Although not new, the cost of simulations and the diversity of the computer algorithms (so called force fields) that underlie them means that this research is the first time a consensus view of protein dynamics has been published.
"In the near future, a biochemist will be able to understand the behaviour of a protein, or design a drug that can interact with that protein, drawing on not only the knowledge of a single structure, but of an entire repertory spontaneously occurring in physiological conditions," said Prof Orozoco.
The research, published in the US journal Proceedings of the National Academy of Sciences, was conducted in collaboration with scientists from the National Institute for Bioinformatics (INB) and the Barcelona Supercomputing Center (BSC), Spain.
A protein is built up from sets of relatively rigid sub-structures (such as alpha-helices and beta-sheets) linked by more flexible strings of amino acids. Certain types of structural fold are common to many proteins and the team set out to map the dynamics of the most popular of these so called metafolds, theoretically allowing them to predict structural changes for a vast number of proteins. The calculations, using four different force fields, are equivalent to 57 years of research on a personal computer.
"To attempt this project with the MareNostrum supercomputer would have simply been impossible," said Prof Orozco.
