Researchers at Penn created these fly insomniacs by altering a gene they have dubbed "Sleepless." They are investigating whether humans carry the same genes.

The goal, said Penn neurobiologist Amita Sehgal, was not necessarily to circumvent our persistent need to sleep away a third of our lives, but to figure out what happens in the brain during sleep and understand the series of mysterious molecular switches that extinguish the inner lights.

The researchers announced their discovery in today's issue of the journal Science.

Of course there was a downside for these flies who grew sleepless in West Philly: They lived only half as long as their slumbering counterparts, Sehgal said.

Sehgal said that a few years ago, researchers weren't sure whether fruit flies really slept or were merely in repose. "Now we know it is sleep," she said, in part by measuring whether it takes a much louder noise or a harder jostle to make them stir.

The finding follows a similar discovery in 2005 by researchers at the University of Wisconsin. They created flies that slept one to two hours a night by altering a gene called "Shaker."

Chiara Cirelli, of the University of Wisconsin team, said as recently as 2000, sleep researchers were skeptical they could learn anything useful from fruit flies. But that year her colleague Guilio Tononi dispelled those doubts by implanting tiny electrodes in fruit flies' heads, proving that their brain-wave patterns resembled those of sleeping humans and other mammals.

The study of sleep in fruit flies and other animals could help doctors better understand how much sleep people really need, and how to treat sleep disorders.

"The central issue in sleep research is what sleep is for," said UCLA neurobiologist Jerome Siegel. "It seems ridiculous that we can't answer this question. But we can't," at least not yet.

The Penn team created its night-owl fruit flies by making a whole variety of mutants, exposing them to what are called transposable elements - stretches of DNA that can move into the flies' genomelike viruses and alter the existing genes.

The researchers then look through these mutant flies for ones with unusual sleep patterns.

What they found were a few that slept just a fraction of their usual 12 hours - many getting by with just one to two hours.

That pattern was similar to what the Wisconsin researchers saw with their gene, but there were some differences.

Some Penn mutants appeared never to sleep at all, Sehgal said, though she conceded that their tests of fly sleep aren't perfect, so the flies could be sneaking in some snooze time.

By examining the DNA of these flies, Sehgal found all had damage to one particular gene - the one the researchers dubbed Sleepless.

To bolster the evidence that they had isolated a crucial sleep gene, they gave some mutant flies the equivalent of gene therapy, adding back copies of the sleep gene. Those flies stopped pulling all-nighters and started napping normally.

Sehgal said they did not know yet whether humans shared this gene with the fruit fly. But it wouldn't be surprising since both species share a number of other genes, including two called "Per" and "Clock," which regulate our circadian rhythms.

More genes have been found in mice that affect sleep - reducing it by 20 or even 50 percent, said Wisconsin's Cirelli, but only Shaker and Sleepless have shown a greater effect on sleep.

She said the Shaker gene was associated with something called a potassium channel - a conduit for charged particles that allows signals to travel through the brain. Sleepless appears to determine whether these channels are open or closed.

The genes seem to work in concert, contributing to a mechanism for periodically shutting down connections between brain cells.

"Sleep is not just this passive stage, at least at the molecular level," said Allan Pack, a sleep researcher at Penn who was not associated with this latest study. "It's a very active stage. But why do we need to do so much of it?"

One theory, he said, is that sleep is a time when your brain creates new proteins and repairs damage. Others have proposed that it helps replenish energy.

Some scientists also argue that sleep is required for the building and rerouting of connections among brain cells that form and consolidate memories.

Whatever sleep does, it's not unique to human beings, said UCLA sleep researcher Siegel. It was long thought sleep was related to human intelligence and creativity - especially the phase known as REM sleep, when we do most of our dreaming.

In Freudian psychology, he said, dreams carry important symbolism. "Because of the connection to dreaming, it's natural to think REM sleep has some higher cognitive function," he said.

But as many dog owners know, canines indulge in plenty of REM sleep, in which they can seem to be chasing squirrels or other creatures in their dreams.

Siegel believes scientists may be reading too much into sleep. Natural selection will favor any trait that helps animals survive and reproduce, and most animals are simply safer sleeping some of the time - especially if they are only able to see well in the daytime or at night.

Most people with kids would consider them safer at night if they're asleep in their beds rather than roaming around the streets, he said. Our relentless daily need for sleep could have evolved because it kept many of our animal and human ancestors out of trouble.

In which case, sometimes a dream is just a dream.