New view of a biological antibiotic assembly line reveals how machinery could be exploited to build better antibiotics

University of Michigan researchers have determined how the final machine in a biological assembly line completes the production of an important group of antibiotics. The discovery, published Aug. 5 in ACS Catalysis, opens the door to tinkering with the machinery to produce new antibiotics. 

Macrolides are a class of antibiotics commonly used to treat a range of bacterial infections, including pneumonia, tonsilitis, and sexually transmitted infections such as chlamydia and gonorrhea. These drugs are derived from soil-dwelling, single-celled microorganisms that have evolved specialized enzymes to efficiently build the compounds. 

When a specific starting material, or substrate, is introduced to the assembly line, the enzymes systematically add new chemical groups to churn out one final product. But researchers have struggled to coax the enzymes to build new, modified antibiotics needed to combat bacteria’s growing resistance to existing antibiotics. 

A major sticking point in the assembly line has been the final machine: an enzyme called a thioesterase. Thioesterases serve as gatekeepers, completing a final quality check and folding the linear compound into the ring-shaped structure that gives it much of its bacteria-fighting strength. 

“To build better antibiotics, we need to tinker with the structure of the substrate that the enzymes work on,” explains Janet Smith, Ph.D., a structural biologist at the University of Michigan and senior author of the new study. “But when we make changes to the substrate, the thioesterase doesn’t do its job: it won’t create the ring.” 

To learn more about this final step, and thus potentially manipulate it, scientists need to observe the details of how it takes place. But the bond between the substrate in the enzyme is very unstable, making it challenging to visualize or even to predict using AI tools.

Researchers in Smith’s lab at the U-M Life Sciences Institute have now overcome that challenge by swapping out one of the building blocks of a thioesterase. With just one block changed, the enzyme trapped its substrate in place long enough for the team to reveal the molecular structure of the two bound together.

“Once we see how the natural substrate fits in there, then we can really understand how this final step takes place and why it has not worked with other substrates,” says Vishakha Choudhary, a graduate student in Smith’s lab who worked on the project. “We're already thinking about how to engineer the protein to accept other non-natural substrates.”

“The findings enable a path to modified thioesterase enzymes optimized to generate new macrocycles that can be efficiently converted to next-generation macrolides,” adds LSI faculty member David Sherman, Ph.D., whose lab collaborated with Smith on this research.