By SUE VORENBERG
Scripps Howard News Service
November 23, 2005
These cyber-bacteria - models of the interior workings of bacteria cells - could lead the way to a generation of antibiotics that kill the critters from the genetic level up, said Kevin Sanbonmatsu, a scientist at the lab.
During the past two years, Sanbonmatsu has created a computer model of part of a cell called a ribosome on the lab's Q supercomputer - one of the most powerful in the world. The work is partially funded by the Department of Homeland Security.
Ribosomes are in every cell - including those of humans and bacteria - and translate genetic instructions into proteins, which are the base of any cell's functions. The ribosomes in different creatures, however, have different sizes and shapes.
"By understanding the differences between the two types (human and bacteria) of ribosomes, we can tailor antibiotics to attack just the bacteria," Sanbonmatsu said.
The result could be antibiotics with fewer side effects and those that can kill antibiotic-resistant bacteria, which is a growing health problem, he said.
So far, Sanbonmatsu has modeled only the ribosomes of bacteria, which are smaller than those in human cells and easier to simulate. The process has resulted in the largest biological simulation ever done, Sanbonmatsu said.
The model simulates the 2.64 million atoms of a ribosome in motion. The largest simulation before it was of a cell membrane with 420,000 atoms, Sanbonmatsu said.
"We really pushed the limits of Q machine on this one," he said. "It's the only computer in the United States with enough memory to do this sort of thing."
Sanbonmatsu moved the cyber-bacteria ribosome through its protein-making process eight times on Q - using different starting conditions to validate it. The simulations took 2 million computer hours, which translates to about 108 days running Q continuously. The work was done in small chunks over 2 years.
Before the computer models, the only way to study ribosomes was by looking at them in dead cells under powerful microscopes. That doesn't tell scientists how they work and isn't as helpful in designing drugs, said Simpson Joseph, an associate professor of chemistry and biochemistry at the University of California at San Diego.
Joseph is validating Sanbonmatsu's work by studying ribosomes in dead bacteria. Understanding how the ribosome works is important, he said, because more and more bacteria are becoming resistant to the antibiotics we have, and no antibiotics have been discovered in recent years.
Most drug companies aren't working on new antibiotics, because bacteria are so effective at adapting resistance to them.
Bacteria can adapt soon after an antibiotic hits the market, which means developing drugs is not profitable for drug companies, Joseph said.
"But if we can custom-design antibiotics and tailor them for specific bacteria, then it becomes more marketable," he said.
Tailoring antibiotics is already possible, although new ones take from five to 10 years to develop, Sanbonmatsu said.
The key to creating antibiotics that fight bacteria, but not human cells, is a second model Sanbonmatsu wants to create when he has access to even more supercomputing power.
That model will be of a human ribosome, which scientists can then compare to that of a bacterial ribosome.
He expects he will be able to do the simulation within the next two to three years.
Meantime, Sanbonmatsu plans to test his model and adapt it to the shapes and sizes of different types of bacteria.
"One reason Los Alamos is involved in this is for national defense," Sanbonmatsu said. "Antibiotics are very important in fighting biological agents, like anthrax. We could use this model to tailor antibiotics specifically to kill anthrax.
"That's probably what we'll be starting with as the next step."
Publish A Letter on SitNews Read Letters/Opinions
Submit A Letter to the Editor