Antibiotics Don’t Kill With Reactive Oxygen Species

Just when it seemed that scientists had a handle on how antibiotics kill bacteria and how their power could be amplified, new evidence has emerged that casts doubt on that approach.

In 2007, James J. Collins and coworkers at Boston University presented evidence that all bacteria-killing antibiotics do their jobs by inducing the formation of reactive oxygen species (ROS) (C&EN, Sept. 10, 2007, page 8; Cell, DOI: 10.1016/j.cell.2007.06.049). The reactive molecules damage bacterial cells.

The three antibiotics the Collins group used—norfloxacin (a quinolone), ampicillin (a β-lactam), and kanamycin (an aminoglycoside)—target different bacterial functions. The scientists hoped to show that a common mechanism could provide a way to make bacteria more sensitive to antibiotics.

But now it seems the research results may have been too good to be true. Two studies from other research groups report that the same antibiotics used in Collins’ study don’t kill bacteria with ROS after all.

Kim Lewis and coworkers at Northeastern University’s Antimicrobial Discovery Center measured ROS with hydroxyphenyl fluorescein (HPF), which was also used in Collins’ study to show that antibiotics kill bacteria with ROS. Biologists thought HPF fluoresces selectively in the presence of ROS, but Lewis and coworkers found that not to be the case. They observed that HPF lights up under anaerobic conditions, in which the formation of ROS would be impossible, and they found no correlation between HPF fluorescence and bacterial death (Science, DOI: 10.1126/science.1232688).

In the other study, James A. Imlay and Yuanyuan Liu of the University of Illinois, Urbana-Champaign, used several different experiments to determine whether antibiotics kill bacteria with ROS (Science, DOI: 10.1126/science.1232751). When treating bacteria with antibiotics, they observed no formation of hydrogen peroxide (a hallmark of ROS), no activation of the bacteria to fight oxidative damage, and no oxidation of bacterial DNA.

Both the Lewis and Imlay studies show that antibiotics kill bacteria under ROS-hostile anaerobic conditions as well as aerobic circumstances.

Lewis hopes both studies will bring clarity to the field of antibiotic research. “A lot of people I know are trying to figure out the details of the mechanism of antibiotic killing through ROS and how that can be exploited for potentiating antibiotic action,” he says. “I think our papers will redirect the effort to something more productive.”

The two studies are “thorough and together very convincing,” says Floyd E. Romesberg, who studies antibiotics at Scripps Research Institute. “Not only do they show that in general ROS are not responsible for antibiotic activity, but they also provide data and clear explanations for previously observed artifacts.”

Collins, however, says the new papers are flawed. “We have serious methodological concerns about these studies, which we are addressing in a new paper that we plan to submit in the coming weeks, and we have extensive new data to include that further support our case,” he says. “We are now working to utilize this mechanism to develop effective means to enhance our existing antibiotic arsenal and treat resistant bacterial infections.”

He even suggests the skeptics could be test subjects. “I am confident these therapies will work, even on the microbiologists who don’t believe in the underlying mechanism,” Collins says.