PORTLAND, OREGON—The transparent, millimeter-long worm Caenorhabditis elegans may be the world's best understood creature. Thanks to decades of study, scientists know its full genome sequence, exactly how many cells it contains, and even the connections of all of its 302 neurons. But now a team of biophysicists has made a surprising discovery: Nearly all of the nematode's various movements can be reproduced by adding four basic patterns of motion. Meanwhile, to study the neurological basis of the worm's motion, another team has genetically engineered a C. elegans that can be remote controlled with laser light.
Biophysicists Greg Stephens and William Bialek of Princeton University and William Ryu of the University of Toronto in Canada set out to see if they could break the seemingly complex movements of C. elegans into simpler components. Using a computer-controlled rig, they took movies of individual worms wriggling about in a dish. The researchers then tried to reconstruct every pose or contortion that the millimeter-long worm might make by adding up simpler mathematical curves nicknamed "eigenworms," each multiplied by a numerical "weight" that could be positive or negative.
To see how this works mathematically, imagine a horizontal saw-tooth pattern and an upward-sloping line. Adding the two will make an upward-climbing staircase. In principle, any squiggly line can be reproduced by adding up a large enough set of eigenworms.
However, 2 years ago, Stephens and colleagues found that more than 95% of all the configurations that C. elegans ever assumes can be reproduced with combinations of just four eigenworms. In fact, the worm's forward wriggling can be reproduced by cycling between just two eigenworms, whereas its backward motion can be accounted for by reversing the cycle. "What makes the analysis really powerful is that you need only four eigenworms, not 10 or 20," Stephens says.
Please download the latest version of the free Flash plug-in.
Now, the researchers have taken the analysis a step further to account for the curious fact that, although the worm wriggles on timescales of seconds, it randomly reverses direction on timescales of minutes. That might seem to require some sort of decision-making, but it's actually caused by random noise somewhere in the worm's neural circuitry, Stephens reported here yesterday at the March Meeting of the American Physical Society. The worm gets stuck in one cycle until noise nudges it into the opposite one, Stephens explains. Indeed, a mathematical model incorporating this notion neatly explains the distribution of time intervals between reversals, he says.
"That stretched my brain a lot," says Philip Nelson, a biophysicist at the University of Pennsylvania. "I never imagined someone might do that" sort of analysis.
Meanwhile, Andrew Leifer, Christopher Fang-Yen, Aravinthan Samuel, and colleagues at Harvard University have found a way to control these worms themselves. They fiddled with the genes in C. elegans neurons so that they could turn the neurons on or off one by one by gently shining laser light on them. In preliminary studies, the researchers have shown that they can temporarily paralyze the tail of a worm while its head keeps wriggling, or they can force it to start backing up by zapping it at one spot close to its head. That's just a start toward figuring out the functions of all the nerves, Leifer says. "With this tool, we could start to build up a map-which neurons are responsible for speed? Which ones are responsible for turning?"
"What a technological breakthrough!" Nelson says. "You can follow an organism around and tweak it at the single-neuron level." Ultimately, Nelson says, the goal should be to forge a connection between patterns of motion and patterns of neural activity: "You've got to drive the golden spike between these two railroads at some point."