For years, nothing more was seen of the new El Tor strain of cholera. In 1937, in the sultry islands of Indonesia, El Tor reappeared, this time causing a small outbreak of mild diarrheal disease. Once again, it disappeared. Then, in 1961, El Tor resurfaced in Indonesia.

"It had a little bit more of a vengeance to it. It essentially spread much more aggressively. It was causing outbreaks in Asian cities that ordinarily did not have endemic cholera," says John Mekalanos, Harvard Medical School Higgins Professor of Microbiology and Molecular Genetics.

Over the next 30 years, the El Tor strain would kill thousands as it spread from Indonesia to India and other parts of Asia, and from there to Africa and Europe. Most recently, it has flared up in Peru and Rwanda.

What enabled El Tor to turn from a harmless bacterium into a killer? For the past 20 years, Mekalanos has been trying to understand this deadly transformation. Over the past decade, he has been uncovering the genes and proteins responsible for El Tor's infamous rise.

Swimmer's Block

Mekalanos believes that the strategy can be used to design vaccines against all lethal forms of cholera. In fact, he and his colleagues have already crafted a vaccine against the lethal Bengal cholera strain that swept through India in 1993.

"It should be possible to immunize people against all forms of cholera in the world with those two strains [the El Tor and Bengal]," says Mekalanos.

Mekalanos first became interested in cholera in 1972, just as El Tor was beginning its incursion into Africa and Europe. Mekalanos, who was an undergraduate at the time, was swept up not by the raging epidemic but by the attention of a professor.

To colonize the intestine, the cholera bacteria must swim through the mucus coating the intestinal wall. Mekalanos' vaccines are composed of non-motile mutants or elongated strains that can't pass through the intestinal mucus.

"I was noticed by a microbiology professor as someone with promise," he recalled. Mekalanos began working in his professor's cholera lab on a research project that challenged him to learn how cholera bacteria-or vibrios-acquire new genes.

Vibrios, like all bacteria, pick up new genes through a variety of clever mechanisms. For example, when bacteria bump up against one another, often one will put out a tiny appendage across which it sends DNA, a process known as conjugation. Bacteria can also pick up "naked DNA" that has spilled from decomposing bacteria. Or they can pick up new genes from viruses.

Fascinated by the habits of the microbial creature, Mekalanos decided to continue his cholera studies in graduate school. Though cholera was an age-old disease, scientists were only beginning to understand how the classical strain of cholera, Vibrio cholerae 01, caused disease.

Classical vibrios were known to produce a toxin that attacked cells of the small intestine, triggering a diarrheal response so powerful that people can die in a matter of hours if untreated.

"In an epidemic situation, it's a very scary thing," says Mekalanos. Researchers suspected that the El Tor vibrio, which was causing outbreaks in Africa and Europe, was also producing the virulent toxin.

In 1974, when Mekalanos began graduate school, researchers were struggling to understand how the cholera toxin signaled intestinal cells to produce the diarrheal response. However, Mekalanos decided to take another route.

"I wanted to get my hands on the structural gene for the toxin," Mekalanos says.

His initial attempts to isolate the gene were thwarted by a federal prohibition in the mid-1970s against cloning genes for toxins. Resorting to classical genetic techniques, Mekalanos began mutating vibrios-both classical and El Tor-to create a strain that didn't produce the toxin. By comparing the DNA of the mutants with the DNA of normal vibrios, he hoped to identify the DNA segment that was responsible for producing cholera toxin.

But then researchers found that the familiar Escherichia coli bacteria could make a cholera-like toxin, called LT toxin. The gene for LT turned out to be on a plasmid, a small circular piece of DNA separate from the rest of the bacterial DNA.

"Nature had cloned [the LT gene] already," Mekalanos says. Reasoning that the LT gene must be very similar to the cholera toxin gene, researchers-with a federal okay-began looking for a matching region on V. cholerae, where LT could hybridize.

Mekalanos tried a different tack. He took the non-toxin producing mutants, hoping to find one that could not match the LT probe. As it turned out, one of his El Tor mutants rejected the LT probe. Mekalonos then compared the El Tor mutant to normal El Tor vibrios to home in on where the mutation occurred. Finally, he took the normal sequence-the putative cholera toxin gene-and put it into E. coli. His engineered E. coli began producing the cholera toxin.

"So, now we had the cholera toxin gene," Mekalanos says.

Pick-up Genes

"That suggested that maybe there's a factor that's needed for expression of the [cholera toxin] gene in cholera that is not present in E. coli, and that's what we should go after," Mekalanos says.

That hunt led to the isolation of a toxin regulatory gene, called toxR, by Virginia Miller, a student in Mekalanos' lab. The researchers soon discovered that toxR regulated the production of the cholera toxin and 13 other proteins.

"So the bug is getting smarter and smarter. It's picking up genes from other sources but it's establishing them under coordinate regulation to be expressed at the right time, at the right place, which is remarkable in terms of the evolution [of El Tor]," Mekalanos says.

One of these "picked-up" genes is tcpA. It encodes a protein, TcpA pilus, that causes bacteria to clump together in spaghetti-like strands and appears to be key to cholera vibrios' colonization of the small intestine. So important is TcpA pilus that without it vibrios cannot mount an infection in human hosts.

An 1883 wood engraving in Life Magazine, titled "Is this a time for sleep?" shows the guardian "science" sleeping on a New York City dock while the spectre of "cholera" floats across the ocean from Europe.

Having fingered the principal culprits-the toxin gene, toxR and tcpA -Mekalanos and his colleagues began designing a series of vaccines. First, they produced a strain missing the toxin gene. But, to their surprise, the toxin-free strain still caused diarrhea, although not as severe as usual.

Eventually, after a few more frustrating turns, they hit upon a novel strategy. Normally, vibrios enter the intestine and encounter a set of cells known as the M cells. M cells, which are located in a dry patch of the otherwise mucus-lined intestine wall, educate the body's immune cells about the characteristics of the invaders. Once sampled, the mucus-loving vibrios swim down into the lower levels of the intestinal wall where, under the influence of TcpA pilus, they take root on epithelial cells.

If the vibrios could somehow be prevented from making this descent-that is if they could be prevented from swimming-then the vibrios could be made to provoke an immune response without causing infection. The researchers decided to elongate the vibrios.

"Because they're so long they mechanically have a hard time swimming through anything viscous," Mekalanos says. "They're not going zoom, zoom, zoom, zipping around like the [normal] vibrios do."

The strategy worked. When volunteers were fed the elongated, toxin-free vibrios, their symptoms were milder, but their immune systems still launched a response.

Since then, Mekalanos and his colleagues have added a few safety measures to their cholera vaccine. They have tinkered with the vaccine in ways that will prevent the mutant bacteria from picking up new genes and regaining its virulence.

Four months ago, Mekalonos and his colleagues gave the El Tor vaccine strain to 40 volunteers at Johns Hopkins University Medical School. So far, the volunteers have shown no symptoms of cholera. And their immune response looks as good-if not better-than that provided by natural immunization.

"So they have very high titers, indicative of somebody who is convalescing from disease," Mekalanos says.

Ordinarily, the next step in a vaccine trial would be to expose the volunteers to natural vibrios, but that would require testing the vaccine in a cholera-endemic area. "So that's a major hurdle for us," Mekalanos says.

It also could be difficult getting the vaccine from the lab to the countries that most need it, such as India, Rwanda and Peru. "How long before say Peru or Bangladesh can have available to their physicians a cheap, inexpensive vaccine, meaning less than 10 cents a dose? I'd say maybe eight years. It's going to take a while," says Mekalanos.

The first step is producing the vaccine. The vaccine has been patented and licensed by Harvard to a small biotech company, which is currently looking for a partner.

"At that point, if a company can't afford to make the vaccine available for pennies a dose to any country that needs it, my feeling is that the strain should be sent to those countries' individual health departments for them to manufacture the vaccine as cheaply as possible," Mekalanos says.

--Misia Landau