AMHERST — A research team led by the University of Massachusetts Amherst developed two techniques to accelerate the search for tuberculosis drugs. The research findings were published in the journal Nature Microbiology.

One technique measures which chemical compounds can cross the outer membrane of Mycobacterium tuberculosis, the bacterium that causes tuberculosis. The second technique uses membrane permeability measurements to predict which other compounds can enter the Mycobacterium tuberculosis cell. Tuberculosis was responsible for 1.23 million deaths in 2024. The bacterium has a unique outer cell membrane called the mycomembrane, which contributes to its resilience against the human immune system and antibiotics.

Sloan Siegrist, an associate professor of microbiology at the University of Massachusetts Amherst and a senior author of the paper, stated that Mtb is unique. "Mtb is unique. Not only does it have two membranes that protect the cell from antimicrobial chemical compounds that we might use to kill it, its outer membrane is unlike any other biological barrier out there," Siegrist said. In 2023, Siegrist coauthored a paper with Marcos Pires announcing a technique called Peptidoglycan Accessibility Click-Mediated AssessmeNt (PAC-MAN). Pires is a professor of chemistry at the University of Virginia. The PAC-MAN technique allows researchers to test multiple chemical compounds in parallel for entry into Mycobacterium tuberculosis cells.

Siegrist collaborated with computational biologists and chemists, including Anna Green, to predict compound uptake for unknown chemicals. Green is an assistant professor in the Manning College of Information and Computer Sciences at the University of Massachusetts Amherst and a senior author of the paper. "Small molecules can be particularly difficult to analyze computationally. Because they come in all different sizes with a wide range of molecular connections, you can't describe them with a single measurement-by weight, say, or size," Green said. Green and her lab designed a machine learning model called the Mycobacterial Permeability neural Network (MycoPermeNet).

The MycoPermeNet model was trained on PAC-MAN screening data. It predicts compound permeability through the mycomembrane based on chemical structure, and identifies physical properties and molecular substructures that help compounds penetrate the mycomembrane. Using PAC-MAN and MycoPermeNet, the research team identified attributes that predict a compound's ability to penetrate the mycomembrane. The research team found that these predictive features correlate with a compound's ability to kill Mycobacterium tuberculosis in large datasets.

"The mycomembrane lets some molecules through and keeps others out. There must be something about this membrane, and about the chemistry of each molecule, that decides which ones get in-and our combined tools help us figure out which ones can get through, and why," Green said. Irene Lepori and Nelson Evbarunegbe of the University of Massachusetts Amherst, along with Zichen Liu of the University of Virginia and Shasha Feng of Lehigh University, were co-lead authors of the study. Joel Freundlich of Rutgers University–New Jersey Medical School, Wonpil Im of Lehigh University, and Pires were also senior authors of the study. The work received support from the National Institutes of Health, the University of Massachusetts Amherst's Institute for Applied Life Sciences, and the Gates Foundation.