It may sound like science fiction, but the body snatchers are for real. David Hughes has seen them, and trailed them from the jungles of Thailand to the woodlands of South Carolina. He has brought them back to his lab, and cultured them, and begun to unravel their secrets.

Hughes, an assistant professor of entomology and biology at Penn State, is a rainforest ecologist with a special interest in parasites. In particular, he is fascinated by that subset of parasites that accomplishes its ends by mind control: invading the brain of a hapless host and causing that creature to do its bidding. Zombie behavior, biologists call the phenomenon. And the woods, as they say, are full of it.

Hughes has studied crickets compelled by parasitic worms to jump into swimming pools and drown themselves, whereupon the worm emerges wriggling and swims off to find a mate. He has looked at wasps that take orders from small parasitic insects sticking out of their backs, taxiing from flower to flower to spread the parasite's larvae. But the one subject of his research that has made the biggest splash—round-the-world headlines, CBC and BBC documentaries, consultant gigs for Hollywood movies and blockbuster video games—is the case of the zombie ants.

You may have heard the basic outlines. Infected by the fungus Ophiocordyceps unilateralis, a common denizen of the world's tropical forests, individuals of a certain species of tree-dwelling carpenter ant behave in a most peculiar manner. Wandering as if drunk, they leave their nest high in the canopy and stagger or fall to the understory below. There they mill about aimlessly until, at the appointed hour, they bite down hard with their mandibles onto the main vein on the underside of a leaf about 10 inches above the ground. Those jaws remain locked even as the ant dies, its body still clinging to the leaf. A few days later, the victorious fungus pushes a stalk through a hole in the dead ant's head, and the stalk drops spores to infect more unsuspecting ants.

This creepy ritual is not new to science: It was first discovered in 1859 by the great British naturalist Alfred Russel Wallace. But it's only in the last few years that researchers have uncovered its details. During that span, Hughes and colleagues around the world have begun to show just how the fungus brainwashes its victim to accomplish a precise set of behaviors aimed at insuring its own survival.

Evolutionary biologists call it an extended phenotype. In effect, the hijacked host's behavior becomes an expression of the parasite's genes. Or, as Hughes has written: "While the manipulated individual may look like an ant, it represents a fungal genome expressing fungal behavior through the body of an ant."

Parasites Rule

Hughes has been stuck fast on parasites since he was an honors zoology student at the University of Glasgow in the late 1990s. "I was always taken with social insects, the idea of the collective," he remembers. "And then immediately I became interested in how parasites break into that collective, and break it down. It's the intersection between this beautifully orchestrated biology and something that's trying to smash it that interests me."

As a graduate student at Oxford, he worked on "a very beautiful, incredible organism known as Strepsiptera," the wasp-controlling insect, somewhere between a beetle and a fly, described above. This was not, he stresses, a fringe pursuit. "Half of life on earth is parasitic," Hughes says, "and parasites dominate biomass as well. We've only recently realized it, but most of the energy flowing through the environment is flowing through parasites."


Those that manipulate behavior, he notes, are a "tiny, tiny minority, and that only makes sense. It's extremely expensive biologically. The whole point of a parasite is to transmit its genes from one host to another, and so to continue on to the next generation. Most parasites can do a really good job of this without having to control behavior."

In fact, he says, the idea that some parasites control their hosts was long resisted in scientific circles. Its early champions—among them Richard Dawkins, the well-known evolutionary biologist—faced considerable opposition. "People dismissed it as storytelling," Hughes says. "It's only now in the last five years that it's become really accepted."

The difficulty, Hughes explains, has been that "in order to show that a parasite is controlling behavior, you have to show that that behavior is adaptive. That it's actually benefiting the parasite's fitness for survival." This was the task he set himself with the zombie-ant fungus.

He chose Ophiocordyceps mostly for practical reasons. "First, they have a small genome, so you can do a lot of genetics with them," he says. "Second, the beer and yeast industries are based on enzymes from fungi—so we know a lot already about the chemicals they produce. But the most important thing is this." He pulls out a small film canister and lifts off the lid to show two dead ants pinned to a circle of cork, one biting onto a tiny leaf, the other onto a twig. Each ant has a streamer of dried fungus trailing from its head.

"Most parasite-host interactions are ephemeral," he says. "But in this case, as you can see, the behavior is frozen. These ants may have died months before I found them, but I can still see what they were doing in the last minutes of their lives. This allows us to do huge studies all around the world."

Following in Sequence

At first, those studies involved combing the jungle for extended periods, locating ant "graveyards" where hundreds of ant carcasses pile up over time, and then finding and observing live ants. "It isn't rocket science," Hughes says cheerily.

Working in a protected rainforest in Thailand in 2006 and 2007, he and his colleagues showed that fungal infection causes the drunken walking and convulsions suggestive of central nervous system impairment, and ultimately leads the ants to a precise location to die. They showed that that place, outside the nest but above the ground—and even the programmed time of death, solar noon—are optimal for the fungi's growth and reproduction. Examining thin sections of ants with powerful microscopes, they found heads packed with fungal cells, and also atrophy of the jaw muscles, a likely factor in the "death grip" that keeps the dead ant fixed to the leaf.

Theirs was one of the first studies to demonstrate fitness, Hughes says. In the five years since, the study of mind-controlling parasites has boomed. Last year, he co-edited a book for Oxford University Press that offers a broad overview of the field, with a foreword by Richard Dawkins himself. The term "extended phenotype" has come into scientific vogue.

"This has been mainly driven by our ability to understand the chemical evidence," Hughes says. With the powerful gene-sequencing tools now available, he explains, investigators have moved from describing remarkable behaviors observed in the field to explaining their precise chemical mechanisms.

Thus, using the resources of Penn State's Genomics Institute, a part of the Huck Institutes of the Life Sciences, Hughes and his students are currently sequencing the genome and transcriptomes of two species of Ophiocordyceps that manipulate ant behavior, with the aim of comparing them to the genomes of species that don't.A complementary tool is metabolomics: analyzing the bioactive chemicals a given genome produces. Already, Hughes says, this approach has identified in the fungus a compound that likely causes the atrophy of ant jaw muscles. More recently, a post-doc on his team, Charissa de Bekker working with Andrew Patterson and Phil Smith at the Genomics Institute's metabolomics facility, found molecules that play a key role in controlling the ant's brain.

Ultimately, says Hughes, he hopes to move on to reverse genetics, a tactic used successfully by Penn State colleague Kelli Hoover to solve another longstanding insect-behavior mystery. Only last year, Hoover and colleagues, including Hughes, were able to pinpoint the viral gene responsible for tree-top disease, a zombie-like phenomenon observed since the 1890s in gypsy moth caterpillars that are infected with a parasitic virus. She did it by infecting caterpillars in the lab with a version of the virus in which the gene she suspected had been inactivated, and comparing the resulting behavior against that caused by the full-strength version.

Defensive Behavior

In addition to the latest technology, Hughes uses old-fashioned carpentry skills, recreating ant nests in the lab in order to better observe ant behavior. "Because we work in the rainforest, and know the ant habitats well, we can build cages which are realistic," he stresses. "One of the things we want to do is reconstruct the social network inside the ant colony."

This is the flip side of the parasite-host equation: Understanding how the host species defends itself against infection. "In the case of ants," he says, "things are set up to protect the queen, whose life is indispensable to the colony's survival."

More broadly, "We're interested in how societies defend themselves, both prophylactically, by setting themselves up in this way, but also actively," Hughes says. "If we increase the level of infection, does the network change?"

Hughes himself marvels at the ways in which the zombie-ant system has evolved. Apparently, it's had ample time to do so: he has identified the characteristic death-grip bite marks in a fossil leaf over 48 million years old. Through the millennia, the ants have developed a behavioral defense that forces the fungus to leave the colony to transmit its genes to the next generation. In response, the fungus has had to ramp up its own arsenal—the mind-altering chemicals that cause the ant itself to leave the nest. "Bear in mind that this is just a yeast, no different from the one in your beer," Hughes notes. "It can't see the world. It's getting around the world by moving this ant."

Over time, he adds, both opponents have become highly specialized. Hughes and colleagues recently named four new species of Ophiocordyceps fungus, each one associated with a different species of carpenter ant. He speculates that there may be a thousand Ophiocordyceps species in all. Just as remarkably, he says, "of all the hundreds of ant species in the Amazon, only 13 percent are infected by these fungi. So what is it about some ants that makes them potential zombies while others are not?"

Oh, and there's one more twist to these ongoing hostilities. The ants, it turns out, have an ally. Recently Hughes reported on a second type of fungus that lurks in the shadows, moving in to attack Ophiocordyceps as it emerges from the ant cadaver. This so-called hyperparasite—a parasite on another parasite—effectively castrates the zombie-ant fungus, preventing it from spreading its spores and infecting more ants.

In the end, of course, it's all part of a complicated ecological balance. The fact that the hyperparasite keeps Ophiocordyceps in check prevents that fungus from annihilating its host altogether, and thereby short-circuiting its own survival. Instead, Ophiocordyceps kills just enough individual ants to further its cause, while the larger colony on which it relies remains mostly intact.

It's a story of interdependence that, far from science fiction, is something Hughes, who has worked in 11 countries on five continents, has witnessed in the wild time and time again. "There's a million other things you could find out that are as complex and as beautiful as the zombie ant phenomenon," he says. "The problem is so few of us, even biologists, are willing to get down on our hands and knees and spend months in the forest looking at them."