Fluorescent proteins, which are compounds that can absorb and then emit light, have become a powerful instrument in the cell biologist's toolkit—so powerful, in fact, that the discovery and development of green fluorescent proteins from jellyfish earned the 2008 Nobel Prize in Chemistry. (Here's a Q&A with one of the winners, Columbia University's Martin Chalfie, about his work.) These proteins have limitations, however: They need to be excited with the blue to orange part of the visible spectrum, at wavelengths of 495 to 570 nanometers. These wavelengths of light are too short to penetrate tissue very well, and so green fluorescent proteins are mainly used in test tube studies to watch cell division or to label certain cell types.

But one of the 2008 Nobelists, Roger Y. Tsien of the University of California, San Diego, and his U.C.S.D. colleagues report in today's issue of Science that they have developed a new fluorescent protein that could enable scientists to tag and visualize cellular activity as it happens inside a live animal. The protein, after absorbing light from the far-red part of the spectrum, shines in the near-infrared, at wavelengths of around 700 nanometers.

These longer wavelengths can penetrate mammalian tissue and even pass through bone. "Say you label a tumor with a green fluorescent protein, and if this labeled tumor is buried inside the animal, then you barely can get green fluorescence out," says lead researcher Xiaokun Shu. "But if you label this deeply buried tumor by infrared fluorescent proteins, you will get a stronger signal because infrared penetrates tissue more efficiently."

Tsien's group derived the infrared fluorescent protein from a hardy bacterium called Deinococcus radiodurans, famous for its ability to survive extreme environments. Bacteria do not actually use this class of proteins, called bacteriophytochromes, to emit light. "They use these bacteriophytochromes to control gene expression,” Shu says—the proteins convert absorbed light into energy to signal certain genes to turn on or off.

The initial challenge for researchers was to re-engineer the protein so that absorbed light would be re-emitted instead of being used as a source of power. They accomplished the feat by deleting the part of the protein that converts the absorbed light into chemical energy; as a result, this truncated and mutant form gave up its absorbed energy as an infrared glow. The scientists incorporated the engineered bacterial protein into mammalian cells—specifically, into the liver of a live mouse, which lit up with infrared light.

The finding paves the way for in vivo visualization of a wide range of biochemical processes and internal organs in animals. (Its use in humans is unlikely, as it would require gene therapy and the ethically dubious transplantation of bacterial genes into humans.)

“This is so important," comments David James, a cell biologist from Australia's Garvan Institute in Sydney, "because a lot of knowledge at the moment is confined to individual cells grown on a glass coverslip," leaving open the question of whether that knowledge "is transferable to an animal." The infrared version could also solve the problem of naturally occuring fluorescence from other biological molecules, which tend to glow at wavelengths similar to conventional fluorescent protein markers and thereby create a lot of “background noise," James says.

But even greater potential lies in harnessing the bacteriophytochrome's original function, namely, powering gene expression. It should be feasible, Tsien thinks, to put back in the signal-controlling properties of the phytochrome. Then it could be possible with animals to “switch on genes and control biochemistry" with light, he says.

For example, you want to explore the effects on mouse behavior of switching on a particular gene that controls some aspect of brain function, but, thankfully for the mouse, you do not want to open up its skull or stick a needle in its brain. "If the infrared fluorescent protein can be made to turn back into an infrared phytochrome, you could have the switch all ready and just waiting for enough infrared light," Tsien speculates. Because infrared light can penetrate the skull, it can reach the phytochrome and remotely switch the gene on, resulting in observable changes in the mouse's behavior.

It's the next evolutionary step for fluorescent proteins, remarks Tsien, who believes that phytochromes represent a class of proteins with enormous potential. If he's correct, then in the coming years, expect more scientists to see the (infrared) light.