Toxic to Tasty

When a team of HMS geneticists broke out their spades to collect soil samples for a U.S. Department of Energy project called Genomes to Life, they were looking for genetic diversity. The project aimed to speed the conversion of biomass into biofuels and hinged on a hunt for the molecular machinery to do it. The researchers, led by George Church, HMS professor of genetics, found more than they had bargained for.

The first hints that they were on to something unexpected came when they were designing control environments for the biofuels project. To stop biomass conversion, they added antibiotics to the soil, expecting the drugs to kill the microorganisms.

“We were blown away to find that everything we looked at—and with every antibiotic we tried in the experiment—was growing on these antibiotics without a problem,” said Gautam Dantas, HMS research fellow in genetics. “We said, wait a minute. This is not a control experiment. This is the experiment.”

This observation sent co–first authors Dantas and Morten O. A. Sommer, a Harvard biophysics doctoral candidate, on a digging expedition. They suspected that microorganisms from soils with more exposure to antibiotics from human and animal use might show more of an ability to subsist on antibiotics, so they gathered a variety of samples. They collected agricultural samples from Carlisle, Mass., and Pelican Rapids, Minn. Urban soils came from the Boston Fens and Public Garden. Pristine soils were trickier. One sample came from an island in Pennsylvania bounded by lumber companies on opposing riverbanks. In a corporate standoff, both deemed the island off limits, leaving it untouched for a century.

With 11 soil samples in hand, the researchers tested each against 18 clinically relevant antibiotics. They found more than 600 different strains of bacteria across the samples that were able to live on antibiotics as their sole food source. They found no statistically significant differences between the soils (see figure). The results appear in the April 4 Science.

In 2006, Gerald Wright, director of the M.G. DeGroote Institute for Infectious Disease Research at McMaster University in Ontario, Canada, coined the term “antibiotic resistome” to characterize this vast well of antibiotic resistance genes in the soil. His lab focused on strains of actinomycetales, an order of antibiotic-producing bacteria, and found resistance to an average of seven or eight antibiotics.

The Church team’s results extend both the breadth and depth of this antibiotic resistome. Dantas and Sommer randomly selected 75 strains from the 600-plus they had found and tested them for antibiotic resistance. These 75 bacteria on average were resistant to between 17 and 18 of the 18 antibiotics tested. Moreover, a phylogenetic mapping of these strains shows that the antibiotic resistome spans many different orders of bacteria, including several orders containing human pathogens.

“This shows that there is a tremendous amount of genetic diversity out there,” said Wright. In addition, “this tells us that antibiotics are not that special.” Though to humans, antibiotics may be “wonder drugs,”—and in fact, life expectancy rates have climbed by 20 years since the first antibiotics were prescribed in the 1930s—“the reality is they are just regular organic molecules.”

While antibiotics may be just another food source, bacteria also happen to be excellent scavengers. Church, Dantas, and Sommer speculate that this ability to scavenge does not come from genes specialized to metabolize individual drugs, such as penicillin or ciprofloxacin. Rather, said Dantas, these scavengers have an “extremely plastic metabolism” that is able to recognize and “chew off” small bits of organic molecules. This flexible genetic machinery allows them to survive in environments where carbon sources are scarce.

Thinking of antibiotics as a food source forces a change in thinking about resistance. “Anything that can eat antibiotics is by definition resistant to antibiotics,” said Church. This realization led Church’s team to observe what they call “super-resistance,” a bacterium’s ability to resist antibiotics even at doses 50 times higher than the levels typically used to test for resistance.

Though the underlying mechanisms of super-resistance are not yet known, Church said, “organisms that are eating antibiotics are probably special in some way.” For instance, they may be combining multiple resistance strategies to disarm and degrade the antibiotic.

“The really exciting part of what’s next is that we have the capacity to understand these mechanisms,” said Wright.

But there is also an alarming aspect. Human pathogens might pick up these abilities to resist antibiotics through gene transference, said Sommer, because “the mechanisms that are used to transfer resistance genes tend to work better when you’re closely related.” Nearly three quarters of the 75 isolated strains were members of orders that contain human pathogens such as Salmonella, Escherichia coli, and bacteria that cause pneumonia. Moreover, antibiotic pollution in the soil from human and animal use may add selective pressure that favors bacteria that can subsist on these clinically important drugs.

According to Church, however, it is too early to jump to conclusions. “It may or may not have any relevance to agricultural or medical practice. It could just be something that’s always been out there, but we just weren’t tuned in to it. Or it could be something that’s being aggravated by our practices. We just don’t know yet.”

Church is more focused on the possibilities that the antibiotic resistome opens up for using synthetic biology to design organisms that eat toxins. “Bioremediation can be a way to clean up the soil from antibiotics and other chemicals,” said Church. “You just want to do it safely, so whatever organisms you’re using to clean up the toxic waste dump don’t themselves become a pest.”

Church’s lab is working on a safe “chassis” for such synthetic life, though completion is likely several years out.

In the meantime, Church, Dantas, and Sommer intend to investigate some of the more immediate questions that their study has raised. They want to identify the genes behind antibiotic resistance and subsistence and understand their transfer patterns and metabolic biochemistry. “It’s exciting that there is so much to do,” said Dantas, “but it also means there are 600 starting points for every experiment.”