(FORTUNE Magazine) – Charles Arntzen, white-haired and engaging, has dropped in on the vaccine research project he brought to Cornell's Boyce Thompson Institute for Plant Research back in the mid-'90s. He runs from meeting to meeting, room to room. The tomato plants in the greenhouse are growing nicely, although a few have been infested with mites and placed in quarantine. Baby plants are germinating in the growth rooms--some withering, others thriving. The visit is only his first in eight weeks. It's not that he doesn't care about the mites in the greenhouse or the tomato plants in the growth rooms. It's just that, with as much chatty energy as he has, the way he can help most is to be away, pounding the pavement.

For ten years now Arntzen has been evangelizing a new, inexpensive vaccine (it costs less than 1 cent per dose) that he believes could prevent the occurrence of hepatitis B, one of the top two causes of cancer in the developing world. You'd think Arntzen's PR job would be easy (who's against helping sick people?), but there's a twist: The vaccine he's championing doesn't come in a pill or in an injection or on a sugar cube. It's grown in genetically modified fruits--and that scares the heck out of people.

Some of those people are concerned that fiddling with the genes could make food dangerous to eat. Others worry that genetically modified crops could disrupt ecosystems if they crowd out or cross-pollinate with nonmodified crops. So Arntzen--who is convinced the potential benefits outweigh the risks--has to spend about a quarter of his time trying to sway skeptics and win the approval and the money he needs to complete his work. "I'm a cheerleader," he says.

Arntzen, 59, came up with the idea for his vaccine project in 1990, around the time he took a trip to Thailand and watched a young mother feed a banana to her infant. A plant biologist, he had just begun to turn his attention to vaccine research. He knew that injectable hepatitis B vaccine, which is beyond the means of 40% of the world's population, was cultured in yeast cells. "I thought, If yeast can do it, why can't a green plant?" he says. "Why couldn't that mother vaccinate her baby with that banana?"

Rejiggering plant DNA to generate vaccine was a far-fetched notion, associates told him, but not beyond the realm of possibility. The idea has obvious advantages. Traditional vaccines, cultured in yeast or animal cells, require expensive purification and refrigeration; plant-based vaccines do not. Standard vaccines must often be injected; plant vaccines are simply gobbled down. And even if it sounded crazy at first, within five years it seemed doable. By 1995, Boyce Thompson researchers, led by longtime Arntzen collaborator Hugh Mason, had gotten plants to produce an E. coli vaccine that worked in mice. By 2000 they had performed their first human trial using potatoes that had been engineered to stimulate immune responses against the Norwalk virus, which causes intestinal disease. And today they are analyzing data from human trials of the hepatitis B vaccine.

But the successes haven't translated into bigtime backing. The team has drummed up scant interest from deep-pocketed pharmaceuticals companies, which generally don't get Viagra-sized profits from lowly vaccines, especially when they're destined for sale in the Third World. Vaccines make up just 1% of pharma's $145 billion in annual U.S. sales, according to IMS Health in Westport, Conn. A 1998 joint venture between Arntzen's institute and British biotech firm Axis Genetics ended in bankruptcy, partly because of political resistance in Europe to modified crops. Some believe Arntzen's focus on developing nations (he's working with public health groups in Mexico, China, and India) is little more than a scheme to test out a risky technology on disenfranchised people.

The technology is not yet ready for prime time, either. Dosing is difficult; it's hard to know how to deliver the right amount of fruit vaccine to people--not to mention how to get the right amount of the right protein to grow in the fruit in the first place. The dream was a mother feeding her child a sweet banana; the reality in trials thus far has been nurses feeding steel-stomached volunteers baseball-sized portions of uncooked potato. "We had a naive notion in the beginning that the village shaman would hand these [bananas] out from his backyard garden," says Mason. Most likely, both Mason and Arntzen now say, the vaccine will need to be processed in some way--it might be whipped into a banana pudding or freeze-dried into a banana chip--and will have to be distributed by medical professionals in a controlled fashion.

The roadblocks don't seem to bother Arntzen. If his critics accuse him of pushing unsafe medicines on poor people, he'll silence them by winning approval in the U.S. first. "What I want to do in the next five years is get at least one of these products licensed by the FDA," he says. "Not produced. Just licensed. All I need is $10 million to $20 million to do this. That's cheap."

Arntzen is also making headway with his bananas. He retired from Boyce Thompson in the fall of 2000 to head Arizona State's Arizona Biotechnology Institute, in Tempe. The new job has proved a boon to his work in an unexpected but important way. "It's easier to get large plots of land there," he says. Arntzen already has plans to build 10,000 square feet of greenhouses for his next generation of bananas.

--Eryn Brown