After 10 years of hard work punctuated by serendipity, leaps of faith, and intuition, an unsinkable team of scientists devise an ingenious system to clean water contaminated with carbon tetrachloride, a toxic chemical.

 By Marina Chicurel
 Illustrated by Thomas Tarpley

An ominous toxic threat lurks beneath the ground at a remote site in south-central Washington. Born in 1955, it is a liquid plume of toxin spreading slowly underground. Already it reaches down 18 stories into the soil and sprawls across an area larger than 1,000 football fields.

The liquid is the waste product of the extraction process used to recover plutonium for making nuclear weapons. From 1955 to 1973, government workers dumped more than 100,000 gallons of carbon tetrachloride, a toxic chemical, into trenches and directly onto the soil at the U.S. Department of Energy site in Hanford, Washington. The chemical seeped into the ground and formed a huge plume that is slowly creeping through the underground water.

A deceptively clear and sweet-smelling liquid, carbon tetrachloride causes headaches, weakness, nausea and vomiting. Even worse, it is a probable human carcinogen that can seriously damage the liver and kidneys.

In an attempt to contain the Hanford disaster, in 1996, the government set up a system of pumps that removes the carbon tetrachloride from the water. But the system can handle only a small fraction of the plume. As of now, there is no way to fully restrain it, much less clean it up.

The carbon tetrachloride at Hanford is spreading through an unpopulated area slowly enough that, for the moment, there is no real danger of it affecting nearby communities. The chemical can linger for thousands of years, however. Sooner or later, if left undisturbed, it is bound to cause serious injury to humans and damage to the environment.

The Hanford site is not the only underground aquifer contaminated with carbon tetrachloride. Although the exact extent of the problem remains unknown, aquifers in virtually every community in the U.S. could be polluted. Carbon tetrachloride is used only rarely these days because of its known toxicity and because, like the better-known chlorofluorocarbons used until recently in refrigerants, it destroys the ozone layer. But in the past, it was produced in large amounts to make insecticides, refrigerants, dry-cleaning agents, and solvents for oils, varnishes and resins.

The first steps toward a solution of this widespread problem, however, are well under way. Craig Criddle, an environmental engineer now at Stanford University, and his collaborators at Michigan State University have brought together two novel approaches: a bacterium that degrades carbon tetrachloride and an ingenious system of wells that exposes the contaminated water to the microbe. The system is deceptively simple in appearance—a row of perforated plastic pipes sunk 80 feet into the ground, connected to an aboveground set of reservoirs and mixers. What’s impossible to glean from viewing the apparatus, however, are the 10 years of hard work punctuated by serendipity, leaps of faith, and intuition that went into its creation.

An abandoned bucket

In 1988, Craig Criddle, then a graduate student at Stanford, was hired by professor of environmental engineering Dunja Grbic-Galic to search for microorganisms that could break down carbon tetrachloride. “I took the job because the work was consistent with my research project, because I needed the money, and because Dunja was a very nice person to work with,” says Criddle.

Using pollution-eating microbes to clean up the environment, also known as bioremediation, was just beginning to gain momentum. Bioremediation promised to overcome some of the problems faced by more traditional cleaning methods. Bacteria are free, inexhaustible, and run themselves, whereas traditional remedies can be expensive and labor-intensive. For example, many traditional methods use activated charcoal to trap contaminants. This process cleans the water, but leaves contaminated charcoal behind. Cleaning the charcoal is possible, but it is a costly process that requires heating it above 15000 F. Bioremediation promised to completely eliminate contaminants in a single, cheap step.

Today, one form of bioremediation — in situ bioremediation — is common. In this approach, the addition of nutrients spurs bacteria, already present in the aquifer, to degrade contaminants. In situ bioremediation has efficiently cleaned up aquifers polluted with diesel fuel and gasoline, for example.

A less common alternative, called bioaugmentation, requires adding non-native microbes to aquifers. Criddle was skeptical of bioaugmentation. He doubted that an added microbe would be able to compete with other local bacteria for space and food.

But before deciding which kind of bioremediation the team would aim for, Criddle first had to determine whether a suitable microbe even existed. He incubated bacteria from several different aquifers into separate bottles with a food solution and carbon tetrachloride. Just as he was finishing setting up his rows of bug-filled bottles, Criddle paused. “As an afterthought, I remembered some old aquifer material sitting in a bucket at the back of a cooler,” he says. “I figured I’d just give it a try for the fun of it.”

A couple of days later, Criddle went to measure the levels of carbon tetrachloride in his bottles and found that the contaminant had disappeared from the bottle containing bacteria from the old bucket. Even more importantly, he found no chloroform in the bottle.

Other scientists before Criddle had discovered bacteria that break down carbon tetrachloride. For every molecule these bacteria broke down, however, they produced a molecule of chloroform — a probable carcinogen that can cause breathing and heart problems. Criddle’s abandoned bucket seemed to contain a bacterium that degraded carbon tetrachloride without producing chloroform. In addition, it seemed to be getting rid of nitrate, a harmful substance that can cause methemoglobinemia — a disease, which infants are particularly susceptible to, in which the blood fails to carry its normal load of oxygen. Because nitrate is used as a fertilizer and carbon tetrachloride was used in grain silos as an insecticide, these chemicals are often found together in aquifers near agricultural areas.

Criddle isolated the bacterium from the bottle and eventually identified it. In honor of his wife, Karrie Criddle, he dubbed it KC. Although he became more excited about the project, Criddle remained cautious.

“At the time I thought that it was unlikely that we had isolated anything unique,” he says. So he ordered previously isolated strains of the same type of bacteria from the largest repository of known microorganisms, the American Type Culture Collection in Virginia. None of them could perform KC’s trick of degrading carbon tetrachloride and getting rid of nitrate without producing chloroform.

KC was one of a kind — an exceptional find. It provided the first hope that an insidious and widespread pollutant might be eventually conquered through bioremediation. Criddle was delighted, but unaware of his discovery’s full implications — as yet he could not imagine the role KC would play in his future.

Catering to KC’s finicky tastes

Like a trainer who discovers a gifted athlete, Criddle was eager to hone KC’s talents and test it in the field. The next step appeared simple: KC’s performance had to be tested in a real aquifer water sample. So Criddle added KC directly to some carbon tetrachloride-contaminated water from a site at Moffett Field near Mountain View, California.

KC disappointed Criddle by eating very little carbon tetrachloride. For some reason, the bug could devour the pollutant when growing in the original food mixture, but not when placed in the contaminated water sample.

Divining KC’s finicky tastes proved a challenge. Criddle spent months testing each component of his food solution. He discovered that it wasn’t enough to have a solution that merely allowed KC to grow. For example, iron stimulated KC’s growth but inhibited its carbon tetrachloride-degrading abilities. So for each component Criddle tested, he had to measure both KC’s growth and its performance. He also found out that he couldn’t just test each component individually, because the combination of certain components had effects that couldn’t be predicted from the individual tests alone. He had to mix and match different components to tease apart KC’s tastes. Finally, Criddle concluded that two conditions had to be balanced carefully for KC to grow and transform carbon tetrachloride: acidity and amounts of iron.

Criddle realized how lucky he had been when he performed that first experiment. By chance, he had chosen the abandoned bucket and used the right degree of acidity and amounts of metals in preparing his feeding solution. What was not chance or luck, however, were the years of hard work that followed to understand just what had happened that first time.

Soon after this discovery in 1990, Criddle graduated and took a faculty position at Michigan State University in East Lansing. Although enthusiastic about his own future, Criddle did not regard KC’s as golden. As promising as KC had once appeared, it now seemed too demanding to ever qualify for practical use. “I doubted that it would ever be more than a laboratory curiosity,” says Criddle.

A winding road to success

As the young head of a lab, Criddle had many affairs to attend to, and although he maintained an academic interest in KC, it was not a priority. For a year, the project languished. Then Greg Tatara, a graduate student at the time, and Michael Dybas, a postdoctoral fellow, joined Criddle’s lab and enthusiastically took over the exploration of KC’s predilections.

Tatara was attracted to the project because it involved a recently discovered organism — KC’s lifestyle, genes, and physiology were a mystery, uncharted territory that he could adventurously explore.

Dybas’s original plan was to isolate the genes that underlie KC’s carbon tetrachloride-eating abilities. The genes, he thought, would provide them with a powerful handle to begin solving the problem of carbon tetrachloride pollution. Dybas planned to search for KC mutants that were unable to degrade carbon tetrachloride. By comparing the genes of those mutants to normal KC genes, he hoped to identify the carbon tetrachloride-degrading genes.

But when Dybas and Tatara arrived in the lab, they couldn’t get normal KC to degrade carbon tetrachloride consistently. “Out of 50 batches, only five would work," says Tatara. KC’s fastidious nature was rearing its ugly head again.

Criddle began to worry once more about KC’s ability to successfully compete with other local bacteria for food and space. But this didn’t slow down Dybas. “Mike was just a crazy man,” says Criddle. “One of the great things about having new people in the lab is their ignorance of the ‘impossible.’”

Joining forces, Dybas and Tatara set out to more carefully determine what KC needed to degrade carbon tetrachloride efficiently. They soon discovered KC secreted molecules that could degrade carbon tetrachloride on their own, even when there were no bacteria present. This suggested they might be able to use these simple molecules and avoid dealing with the complicated KC.

Unfortunately, the molecules were devilishly sensitive to acidity. Dybas spent no less than six months testing the ways in which tiny differences in acidity affected the molecules’ activity. With unflinching persistence he found out exactly how to coax those molecules into doing their very best job.

And then something terrible happened. Without any obvious explanation, the molecules stopped working. Even when Dybas adjusted the conditions precisely, he could not wring any activity out of them. “The last time I had activity was the last day of George Bush’s presidency,” says Dybas. “Since Bill Clinton’s been in office I’ve never been able to get it. I can’t directly correlate that with him, but I have my suspicions.”

Dybas was disappointed, but not defeated. He decided to set the molecules aside and go back to working with the whole bacterium. He repeated the experiment Criddle had tried three years earlier, testing KC’s ability to degrade carbon tetrachloride in an aquifer sample of contaminated water. Dybas now knew more than Criddle had known then, however.

Dybas understood, better than anyone else, that KC needed very particular conditions to do its job. So instead of just dumping KC into the water as Criddle had initially done, he first adjusted the acidity of the sample and added a food supplement for KC to munch on. Finally, KC came through. It passed its carbon tetrachloride-eating test with flying colors.

This was the first experiment to show that KC didn’t need a carefully prepared lab solution to survive and destroy carbon tetrachloride; with a little help, it could work its magic in real aquifer water. Since there were other, naturally occurring bacteria in the water, it was also the first indication that KC could effectively compete with other bacteria for space and food. The researchers stood on the threshold of developing a new approach to environmental restoration. All they needed now was a site where they could challenge their wonder-microbe in a real-world setting.

KC moves out of the lab and into the real world

At about the same time that Dybas’s experiments were meeting with success, a student in one of Criddle’s classes provided just what the team needed. Tim Mayotte had worked for six years as the head hydrogeologist on a carbon tetrachloride-contaminated aquifer in the village of Schoolcraft in southwest Michigan. When he learned of Criddle’s carbon tetrachloride-decontaminating bugs, he told Criddle there might be an opportunity for testing KC.

But before dumping KC into the Schoolcraft aquifer, the team had to do more work. What if KC itself was toxic? They decided to have KC’s metabolic waste products tested for cancer-causing compounds. The results came back negative. The team also tested the effects of KC on the germination of corn seedlings, since corn is grown in a field close to Schoolcraft. KC seemed to allow germination just fine. Finally, since the groundwater flows into a lake, they tested KC’s effects on fish. Dybas even set up an aquarium in his office where fish swam in a soup of KC. The fish seemed fine. Although none of these tests proved KC is harmless, they at least suggested the bacterium would not cause an environmental disaster at Schoolcraft.

Besides testing KC’s environmental friendliness, the team had to resolve another issue before injecting KC into the Schoolcraft aquifer. Criddle and his co-workers had done all their tests in bottles with watery solutions. The Schoolcraft aquifer is primarily made up of sand saturated with water that moves slowly through it — more like slush than a water solution. Would the notoriously discriminating KC cooperate under these conditions?

Much to its delight, the team rapidly found out that KC was capable of attaching to the sand, growing on its surface, and breaking down carbon tetrachloride. Criddle and his group set up a small field experiment. They drilled a well and fitted it with a four-inch diameter PVC pipe ending in a strainer-like screen that allows water and bacteria through, but keeps the sand out. They then injected KC and its feeding solution into the pipe and let it seep through the screen and into the surrounding ground. The group also drilled several other holes around this injection well to draw out water samples for monitoring KC’s whereabouts and accomplishments.

KC did not disappoint the team on its field debut. It quickly started feasting on the carbon tetrachloride. But after a few months, chloroform started to show up in the monitoring wells. Like a schoolyard bully, the natural population of bacteria in the ground was eating KC’s food and converting carbon tetrachloride into chloroform. Criddle’s initial concern about KC’s inability to compete well with other bacteria was playing out.

To solve the problem and increase the size of the operation at the same time, the team would have to figure out how to distribute enough KC across a larger area, and how to keep it well-fed so it wouldn’t be squeezed out by neighboring bugs.

“We really had our back against the wall when we wanted to scale up,” said Criddle. “We realized that it was a huge challenge.”

A new idea

Just in time, geologist David Hyndman and civil engineer David Wiggert, both of Michigan State University, breathed new life into the project. Brainstorming with Criddle’s group, they envisioned a curtain of bacteria intercepting the plume and destroying it at its leading edge. Wiggert suggested drilling a row of closely spaced wells to inject KC and its feeding solution into the aquifer. The idea was unprecedented.

For months, Hyndman ran computer simulations to determine if and how this idea would work. Based on his calculations, the scientists sketched out a working model. They would drill a row of 15 one-inch-diameter wells, spaced about 3 feet apart, with strainer-like screens along most of their lengths. Into each well, they would inject KC, letting the bacteria seep out into the aquifer and attach to nearby grains of sand. The researchers would use the slotted screen pipes to periodically supplement KC’s diet with feeding solution. To make sure the entire biocurtain was well-fed, they would recirculate the feeding solution by using alternate wells to inject and extract the solution in the aquifer.

Trap-and-Treat

The key to this system is that it’s passive. Within the aquifer, groundwater moves like a slow-flowing river at six inches a day. Because it sticks to the sand, the carbon tetrachloride contaminating the water moves even more slowly — about two inches a day. Most bioaugmentation engineers seeking to flush the contaminant out view its stickiness as a problem that interferes with their goal. But the approach of Criddle’s team actually takes advantage of the stickiness.

The group’s system uses the aquifer’s natural water flow and the carbon tetrachloride-trapping properties of the sand to accumulate carbon tetrachloride for the bacteria to munch on leisurely. Spacing the wells 3 feet apart creates a uniformly dense curtain of KC, enabling it, through sheer numbers, to can outcompete the natural bacterial inhabitants. Within a day or two, KC can break down most of the carbon tetrachloride immediately surrounding it. During the next six days, water is left to naturally drift through the KC curtain. This stage requires no pumping. As the contaminated water flows by, the cleaned sand acts as a brake on the sticky carbon tetrachloride. At the end of the six days, a pump delivers fresh food and the cycle repeats.

Plumbing diagram of the carbon tetrachloride-decontaminating system. KC is grown in the inoculum vessel and pumped into the aquifer through an array of 15 wells. The bacteria seep into the ground through the perforated wells. Every week, a feeding solution is pumped and recirculated between neighboring wells to ensure uniform distribution of food.

This trap-and-treat technology is not speedy, but it is very cheap and potentially very effective. Conventional techniques for treating contaminated water are often limited by the cost of running pumps continuously as well as the expense of the labor to maintain them. In addition, conventional techniques are not very effective at cleaning sticky contaminants like carbon tetrachoride, which can’t be easily flushed out with water. Hyndman and Wiggert’s calculations suggested that by properly spacing the wells, they could get away with only six hours of pumping a week, at the paltry rate of 40 gallons per minute —only about eight times faster than the rate at which a garden hose fills a bucket.

Criddle found it very hard to imagine that such a low level of pumping could clean an entire aquifer. “Even now when I say it to people, I’m amazed,” says Criddle. “It shocked me.” Criddle admits he was nervous about investing money and effort into this next step. But the Michigan Department of Environmental Quality, looking for ways to reduce their costs for cleaning up aquifers, decided to fund the team. “It was a big risk,” says Criddle. “I guess our funding agency had faith in us.”

During the summer and fall of 1997, Criddle and his team worked with three private companies to build a full-scale system. Workers drilled 15 bore holes reaching eight stories below the ground, and fitted each with a one-inch diameter PVC pipe. Except for the uppermost 30 feet, the rest of the pipes’ lengths were riddled with slots to let KC and its food out into the aquifer. They also installed over 100 other wells to monitor KC and its decontaminating progress.

Viewed from across the field that overlies the aquifer, the installation resembles a 50-foot-long diamondback snake. Each square of the diamond is a wooden trapdoor that, when opened, reveals the upper portion of a pair of wells. Each well is connected by underground tubing to the feeding vats, KC reservoirs, and mixing system which are housed in a small building nearby.

 

A bird’s-eye view of the system for cleaning the carbon tetrachloride-contaminated aquifer in Schoolcraft, Michigan. Polluted water is transformed into clean water as it flows through the KC biocurtain created by an array of wells.

On January 7, 1998 the team injected KC into the system along with “Mike’s recipe for KC à la carte,” as Criddle calls the feeding solution developed by Dybas. At first, samples from the monitoring wells didn’t show much of a drop in carbon tetrachloride concentrations. But by then, Criddle and his group had learned patience. After two months, carbon tetrachloride levels began to drop significantly and they have continued to drop steadily ever since. Criddle says the system now degrades more than 95 percent of the carbon tetrachloride contained in the 2,500 gallons that drift through the biocurtain every day.

“There were a lot of challenges. There were a lot of quick decisions,” says Dybas. “And moments of panic and moments of victory.”

Today, Dybas is the project manager of the KC decontaminating team at Michigan State, and Tatara works as a consultant and a key member of the team. “Most people don’t get to see their Ph.D. work go into the field,” says Tatara. “It was a once-in-a-lifetime opportunity.” Criddle returned to Stanford where he is now an associate professor in the department of civil and environmental engineering.

What lies ahead

An important test of the biocurtain system hinges on its evaluation by other experts in the field. But the preliminary results are encouraging. The time it will take his system to clean the entire aquifer depends on how many KC biocurtains are set up. Criddle calculates that with the single biocurtain they have now, it would take approximately 25 years and cost $1.5 million. Working at an equivalent rate, pump-and-treat — an approach that requires continuous pumping and multiple steps to destroy carbon tetrachloride — would cost about $6 million. Criddle says the group is now working on ways of automating the system to further reduce costs. “Although I started this journey as a bioaugmentation skeptic,” says Criddle. “I now find myself a believer.”

But more significant than its low cost is its potential for improvement. Lycely Sepulveda-Torres, a graduate student of Criddle, and Ron Crawford and his coworkers at the University of Idaho, have recently cloned the genes that endow KC with its carbon tetrachloride-eating capability, and are now attempting to put them into other organisms. If successful, they may be able to engineer less-demanding and faster-growing bacteria that produce more carbon tetrachloride-degrading molecules.

Encouraged by all this, Criddle is preparing to use the trap-and-treat method to tackle other contaminants such as chromium, tetrachloroethylene, and trichloroethylene. Other researchers have already found bacteria that degrade some of these contaminants, so setting up these new decontaminating systems may be straightforward. In other cases, he will have to search for entirely new strains of bacteria, but his experience with KC will no doubt serve as a valuable guide.

The problem of carbon tetrachloride contamination, however, is far from solved. The monstrous plume at the Hanford site remains out of control. It is so vast that, right now, neither the new KC biocurtain nor conventional pump-and-treat systems stand a chance of making a significant dent in it.

Nevertheless, in looking to the future, it is worth remembering the unpredictable path that research often follows. “The story of KC shows that the microbial world is full of unexpected surprises,” says Criddle. “And that serendipity can play a big role, maybe bigger than we care to admit.”

 

BIOs

 

WRITER Marina Chicurel

B. Sc., basic medical research, Universidad Nacional Autonoma de Mexico; Ph.D., neurobiology, Harvard University.
Internship: New Scientist (Latin America correspondent).

ILLUSTRATOR Thomas A. Tarpley

B.S., biology, Freed-Hardeman University, 1995.
Internship: American Museum of Natural History, New York.