A Visit to IBM: Rhodospin - The Light Receptor

Most of you know my partner Zha works for IBM in RL. Last night I was over on her work island to build a little quiet spot for meditation for Zha and her guests and colleagues. The IBM sims are pretty interesting places but are not really open to the public for obvious reasons. I took a few photos of this massive structure that floats over the lake in the center of the island and wanted to share them here.

This is Zha sitting on the structure. Note the red tea house on the ground in the background. It looks familiar? You mean it looks exactly like the tea house I built for our sim? Imagine that. ;)

Zha at ThorneBridgeTown

Here is the structure under a fancypants windlight sky after I messed around with the location of the sun, the colors and the clouds. Cool, right?

Molecule in the Sky at ThorneBridgeTown


The structure is a huge molecule Rhodopsin, which is the protein that allows us to see. From the notecard written by the talented creator of this Rhodospin model, Rez Tone.

Rhodopsin - The Light Receptor

The large molecule floating over the lake in front of you is Rhodopsin. Rhodopsin is the protein that allows us to see. It is densely packed in the rod cells of the eye, and is primarily responsible for night vision. Color vision utilizes proteins closely related to rhodopsin. Common to all vision is a small molecule attached to the protein core called retinal, which actually captures a photon of light. Retinal changes its shape upon photon absorption, which drives further more profound changes in protein structure leading to activation. A cascade of signaling follows rhodopsin activation, which ultimately results in the perception that you saw something.

As we zoom in close, we see a structure of tubes in varying colors. The colors indicate which particular atom or element lies at the tube junctions:

Green is Carbon
White is Hydrogen
Red is Oxygen
Blue is Nitrogen
Yellow is Sulphur

Studying rhodopsin has importance beyond the process of vision. Rhodopsin is a member of the class of signaling proteins known as G-protein coupled receptors (GPCRs). Over half (> 50%) of the drugs on the market today target GPCRs. GPCRs represent the most important class of therapeutic targets for the treatment of disease. Rhodopsin is, however, the only GPCR where the structure is experimentally known in atomic detail. This allows detailed studies of rhodopsin to uncover secrets of how it works, and perhaps gain functional insight into the broader class of GPCRs.

One way to gain insight into how rhodopsin and perhaps other GPCRs function is to study their motion in a native-like environment. This is very difficult to do experimentally, but advances in supercomputing have made it practical to study the early stages of rhodopsin activation with a simulation method called molecular dynamics. Though the method itself is not new, breakthroughs in massively parallel hardware and applications, such as Blue Gene and Blue Matter, can be applied to the motion of rhodopsin on the microsecond scale, in full atomic detail. The ability to simulate atomic motion of membrane proteins on the microsecond scale is unprecedented, and represents an increase in a factor of 100 for the practical length of a membrane protein simulation.

The structure you see floating over the lake represents the results of one of the longest all-atom molecular dynamics simulations in history. Starting from the experimentally determined positions of the atoms in rhodopsin, the protein motion was simulated on the Blue Gene supercomputer at IBM's T. J. Watson Research Center, with the retinal in the activated state. This immense calculation revealed details of how the light-activated retinal transduces activation of the protein.

In this myriad of detail, a discovery was made. An important prediction from the simulation has been confirmed by experiment. The simulation showed a key step in the activation of rhodopsin is for water to flood into the core of the protein, and drive further change in the protein structure. The experimental confirmation of the predicted involvement of water in the activation of rhodopsin, and perhaps other GPCRs, is a powerful demonstration that large scale molecular dynamics simulations can offer deeper insight into biological processes. Simulation can yield deeper insight into biology by providing atomic-level detail, beyond the capabilities of
contemporary experiment.

0 comments: