The paper in Nature Microbiology is here: http://go.nature.com/2xaJxNY
Involvement of reactive oxygen species (ROS) in thymineless death evolved from Collins’ discovery that ROS commonly contribute to killing by a variety of diverse antibacterials. For many years we have been interested in how inhibitors of bacterial replication kill cells, and a contribution of ROS provided a new direction. We supported and extended Collins’ work with follow-up studies, but controversy arose when the Lewis and Imlay groups failed to repeat Collins’ initial observations. We were unconvinced by the challenges because the known concentration dependence of the major stressor (norfloxacin) could easily account for laboratory-to-laboratory differences. Moreover, follow-up work by Collins addressed the criticisms.
We next asked whether non-antimicrobial lethal stress also leads to ROS-mediated self-destruction. An obvious study choice was thymineless death, since it also involves inhibition of DNA replication. Thymineless death was also interesting because the molecular events were not fully understood, even after 60 years of study. None of the existing explanations of thymineless death fully accounted for the rapid and severe lethality, which led Kuzminov to suggest that a key component was missing from the death pathway arising from thymine starvation. We identified ROS as the missing and probably the crucial component, since thymineless death was completely abolished when ROS was kept at background levels. An interesting outcome was that DNA repair creates a more vulnerable substrate for ROS attack (persistent single-stranded DNA regions). In essence, two pathways join to give the lethal lesion (double-stranded DNA breaks). Since an ultrasensitive substrate for lethal ROS attack is created, very high levels of ROS are not required to kill cells; thus, Imlay’s speculation that endogenous levels of ROS are insufficient to kill bacteria can be readily explained.
We have also used ROS involvement in cell death to address the quinolone paradox, a phenomenon in which very high concentrations fail to kill cells even though moderate concentrations are highly lethal. At high concentrations of nalidixic acid, ROS fail to accumulate, even though lesions created by the drug persist and are readily attacked by exogenous peroxide.
One expectation of a death pathway involving ROS is that it can be blocked by anti-oxidant treatment after removal of the stressor -- bacteria thought to have been killed by the primary stressor can be resuscitated by addition of an ROS scavenger to agar used for post-stress viability determination. We found this to be the case with several lethal stressors. A key point is that ROS accumulation is the cause of death, not the result of it (had ROS been the consequence of cell death, addition of ROS scavenger to agar would have no effect on bacterial survival). Moreover, these data suggest that ROS-mediated post-stress self-destruction accounts for ROS contributing to the lethality of diverse stressors. The existence of an active death pathway in single-cell organisms raises evolutionary questions regarding the selective advantage of suicide following severe stress. Self-destruction also suggests new ways to control bacterial populations if small-molecule adjuvants are found that stimulate ROS accumulation during antimicrobial treatment.