Search

Navigation and service

Unravelling the Influence of Protein Dynamics on Molecular Recognition

“Everything the living things do can be understood in terms of the jiggling and wiggling of atoms”

Richard Feynman

Principal Investigator: Bert de Groot, Computational Biomolecular Dynamics Group, Max Planck Institute for Biophysical Chemistry, Göttingen (Germany)
HPC Platform: SuperMUC of LRZ

From Feynman’s quote above it is clear that in order to understand biochemical processes we need to observe the behavior of biomolecules at an atomistic level. We require an instrument to observe e.g. proteins at an atomistic level. Wet-lab experimental techniques have come a long way to achieve this goal. There are several spectroscopic techniques such as NMR (Nuclear Magnetic Resonance), FRET (Fluorescence Resonance Energy Transfer), super-resolution microscopy etc. that reveal a glimpse of the jiggling and wiggling of atoms. Unfortunately, all of those methods only provide us with indirect information and do not enable us to observe directly at an atomistic detail the behavior of complex biomolecules. This is where computer experiments come into play.

The laws that determine the behavior of molecules at an atomistic level are known and can be integrated numerically, enabling us to simulate the time evolution of biomolecules starting from a static experimental picture. The “movie” generated can be used to understand the molecular processes. Moreover, we can easily simulate the impact of a perturbation, for example the impact of mutations, which enables us to design and validate new molecules with altered or improved properties. However, the numerical integration of a complex molecular system at atomistic detail requires a tremendous computational effort. One should take into account that a simple protein easily contains several thousand atoms. Moreover, to generate a realistic simulation the natural environment needs to be taken into account as well. Overall, this adds up to several ten-thousands of atoms all interacting (in a complex manner) which each other. And that is not all: due to rugged nature of the energy landscape a small integration time step is required. This is why supercomputing is mandatory to allow addressing biologically relevant questions.

In the current project, run by a team of scientists of the Computational biomolecular dynamics group at the Max Planck Institute for Biophysical Chemistry in Göttingen, the aim was to deliver a proof of principle that modulation of protein dynamics can serve to modulate molecular recognition. The idea is that some states of a biomolecule are compatible to interact with other biomolecules while others are not. Thus, modulating (by introducing mutations) the population of the compatible and non-compatible states is expected to result in a related change in binding affinity. Simulations of mutated proteins were conducted to quantify and understand the mechanism of the change in population of binding compatible versus non-compatible states. This resulted in a predicted change in binding affinity which is a property that can be validated experimentally. The experimental validation demonstrated the power of the approach and confirmed the computational results.

Overall, the story demonstrates that computer simulations are indispensable in the understanding and rational design of (bio)molecules. The successful simulation of such complex and large molecular systems was only possible by continuously pushing the limits of computing, both in terms of hardware and software.

Petascale system SuperMUC of Leibniz Supercomputing Centre in Garching near Munich served as computing platform for this project.

Unravelling the Influence of Protein Dynamics on Molecular RecognitionCopyright: Max Planck Institute for Biophysical Chemistry, Göttingen

Scientific Contact:

apl. Prof. Bert de Groot, Ph.D.
Computational biomolecular dynamics group at the
Max Planck Institute for Biophysical Chemistry
Am Fassberg 11
D-37077 Göttingen/Germany
e-mail: bgroot@gwdg.de

January 2015