Here’s how it works. Proteins need to fold into very precise shapes to do their jobs, and misfolded proteins are wastes of space. Bacteria dispose of these useless molecules with ClpP—a janitorial protein that digests other proteins. It works with a partner, which recognises misfolded proteins, unfolds them, and threads them through a hole in the middle of ClpP so they can be broken down. But ADEP4 opens ClpP up so it no longer needs its partner. The janitor now becomes an assassin, running amok and chopping up any protein it comes across, misfolded or not.

The Bayer scientists showed that ADEP4 can force fast-growing cells to self-destruct, but Lewis suspected that it would do the same to persisters. Afterall, ClpP’s partner requires energy to do its job, but ClpP itself doesn’t. Once ADEP4 opens it up, it should go about its fatal business even in a dormant cell.

Lewis’ team found that ADEP4 did effectively kills persister populations of Staphylococcus aureus, but the bacteria bounce back. ClpP isn’t essential, so the bacteria just inactivated it to evolve their way around ADEP4. This, says Lewis, is why Bayer stopped working on the drug.

His solution was to pair ADEP4 with another antibiotic called rifampycin. ADEP4 would kill off the majority of the persisters, and if any of the rest started growing again, rifampycin would finish them off. He predicted that the double-whammy would leave very few survivors, maybe just a thousand cells or so.

“That’s not what we saw,” he says. “What we saw was complete sterilisation.”