12 APRIL 2007
More than 300 years after Anton Van Leeuwenhoek first glimpsed single-celled life, some scientists are abandoning the microscope and booting up their computers to literally enter the world of the cell.
Harnessing the nation's beefiest supercomputers, scientists are building digital simulations of cellular activity, millions of atoms at a time. In the process, they are getting a better understanding of the inner workings of a cell's micro-machinery. Eventually, their work may bring a fully realized simulation of a one-celled organism to "life."
Image courtesy Theoretical and Computational Biophysics Group
Peering into the cell never looked so good -- or revealed so much. If Leeuwenhoek could see it, he would flip his powdered wig!
Klaus Schulten is a master of these digital recreations, known as molecular dynamics simulations. As the director of the Theoretical Computational Biophysics group at the University of Illinois in Urbana-Champaign, Schulten's team helped write the book on simulating the tiny components of life.
According to Schulten, these simulations provide "a microscope into the molecular world of the living cell." He says they resolve minute details that conventional microscopes can't. Best of all, they describe the moving and shaking of cellular phenomena, unlike electron scanning microscopy and other high-powered magnification tools that only yield static images.
"The cell is, in a way, a society of molecules," says Schulten. "I would like to understand that society. I would like to understand aspects of these molecular societies that go beyond the individual. How does the cell assemble its molecules into functional units? How do these units work?"
By building these atomic-scale models, Schulten hopes to answer these questions. This is no easy task-his team has a penchant for tackling some seriously complex molecular societies.
All images courtesy Theoretical and Computational Biophysics Group
The bouncer of the nucleus
The nuclear pore complex (NPC) is a microscopic monster. Built of hundreds of proteins, these may be the largest molecular machines in the cell.
In every cell, the nucleus contains the genetic material that dictates biological destiny. Like an exclusive downtown club, the nucleus is snootily selective about who gets in and who stays on the sidewalk. Each NPC acts as a bouncer, waving in the appropriate proteins -- but not the riff-raff.
Schulten is fascinated by the pore's selectivity. To get a closer look at how it sorts the good from the bad, his team built a digital facsimile of parts of the nuclear pore, together with the transport receptors that shuttle molecules through the pore. Then, using special software to visualize the interactions between the proteins and transport receptors, they were able to see the machinery in action.
The simulations helped the team find many new binding spots on the transport receptors -- areas where the pore complex can bind and recognize the desired proteins. After Schulten published his work, an experiment with real cells by other scientists immediately confirmed the existence of one of the new binding spots.
In other words, silicon matched reality. "It was a nice verification that what we were doing on the computer was very relevant in vivo [in a real, live critter]," says Tim Isgro, a doctoral student at Schulten's lab and lead investigator of the NPC simulations.
Isgro says that if the NPC investigators learn enough about the interactions between the proteins in the complex and taxi molecules that drive through it, they will get a deeper understanding of molecular transport, an essential function in all forms of life. He speculates that this knowledge could eventually lead to some clever reverse engineering.
"There may be some eventual application in nanotechnology. The NPC filters large molecules in a very controlled way. One might be able to design a 'customized nuclear pore' . . . that filters out toxins, for instance, or separates solutions of other large chemical species," says Isgro.
Simulating life on even the tiniest scale isn't easy. Schulten's team requires mammoth amounts of data-crunching power. To simulate tens of billionths of a second of nuclear pore complex activity requires many days or even months of processing at facilities like the National Center for Supercomputing Applications or the Pittsburgh Supercomputing Center, which house some of the most powerful supercomputers in the world. Soon, Schulten will need even more processing power. For his next simulation, he will attempt to simulate one of the most enigmatic structures in the cell, the ribosome.
The molecular protein chef
As genetic messages get shipped from a cell's nucleus, they go to thousands of ribosomes, which use the instructions as recipes to cook up essential proteins. "In many ways, the ribosome is the most famous and important machine you find in all living cells," says Schulten. It's also complicated -- Schulten's digital ribosome model is almost three million atoms strong and took a painstaking year and a half to build.
Using the model, Schulten hopes to answer some critical questions about the protein-synthesizing workhorse. Specifically, how does the ribosome stitch proteins together with such accuracy?
Unfortunately, Schulten's team won't be able to observe virtual protein synthesis in action. Even the fastest supercomputers can only depict such atomic complexity for a few dozen nanoseconds. But as supercomputing centers ramp up their processing power, this will change.
Instead, Schulten's unique approach allows the team to observe much longer time frames. First, the team looks at portions of the ribosome at the atomic scale, a fine-grained approach, as he calls it. As time goes on, they will shift to a "coarse-grained" approach to observe the ribosome on the protein scale. In this way, the scientists can use the molecular dynamics simulations to connect the dots between different experiments and create a clearer picture of the vital ribosome at work.
As other cellular machines get the sim treatment, researchers will be able to paint a picture of the cell that's deeper and more accurate than ever.
"We want to develop the physical basis for what really makes a cell the wonderful entity it is, something that is capable of duplicating itself, creating a daughter cell, repairing itself and maintaining itself for a long time," says Schulten. "That is what we want to understand in the end."
— Adam Dylewski