Earlier this year, Seth Herzon, Ph.D., professor of chemistry and member of Yale Cancer Center, found—quite inadvertently—an arresting opening slide for his lectures: an image of himself in the hospital with an infection he had picked up from a cat bite. “All right—here it is,” he tells his audiences. “Here’s why we need to do antibiotics.”

There’s no denying the medical need for new and better antibiotics—in 2013, the Centers for Disease Control and Prevention (CDC) reported that 23,000 Americans die of drug-resistant bacterial infections each year. Herzon worried he might be one of them when oral antibiotics failed against his infection.

“My biggest fear was not death itself, but what would be on my tombstone if I died,” he said. “It would be ‘Herzon, ’79–’17 (cat bite).’ ” Fortunately, his infection was cleared up with an IV antibiotic. Not everyone’s is.

Our species has been in a continuous evolutionary arms race with bacterial pathogens for decades now, and they seem to be gaining the upper hand. Bacteria evolve at an alarming pace—some species can produce a new generation in half an hour—and can even exchange genetic information that confers defenses against our antibiotics. Today, some infectious bacteria have evolved such a degree of resistance that they’ve outpaced the production of fresh antibiotics.

Herzon’s lab has recently developed a way to chemically produce pleuromutilin, a fungal antibiotic agent. It was isolated in the 1950s, but the means of creating it “from scratch” in the lab—an alluring possibility that would enable the production of many variations of the drug—had always posed a challenge to researchers.

“Bacteria can’t find a way to evolve resistance to pleuromutilins without killing themselves,” Herzon explained. The site that pleuromutilin binds to in pathogenic bacteria is essential for making proteins—and without proteins, the cells cannot survive. To this point, bacteria have not devised a way to evolve resistance without also injuring their own protein centers.

“We have a way now to access sites in pleuromutilin that were entirely inaccessible before,” Herzon said. “And the structural data tell us if we modify these sites, we’re going to get better antibiotics.”

There are currently one pleuromutilin derivative approved to treat infections in humans, and two used for animals in the United States. Herzon’s technique opens the possibility of developing many more to treat a broader range of bacterial species. Bacteria are classified as either Gram-positive or Gram-negative (more simply, as having one cell wall or two). Gram-negative bacteria are more difficult to treat, and the existing pleuromutilins fight only Gram-positive ones.

“There’s some evidence that if we change the structure of pleuromutilin, we can get activity against Gram-negatives,” Herzon said. “So that’s what we’re really going for.”

The quest for novel antibiotics has fallen almost entirely onto universities’ and philanthropies’ shoulders. Big pharmaceutical companies are reluctant to invest in antibiotic treatments (despite the urgency), because even successful antibiotics offer a meager return on investment.

“We’ve got a unique opportunity here to make a dent in this problem using the chemistry that we’ve developed,” Herzon said. “It’s important work to do. Resistance to antibiotics has been in the literature for 30 years. People have been aware of it for 30 years—and it just keeps getting worse.”