Map of essential-gene interactions in tuberculosis bacterium points to new drug pairings

Researchers at Rockefeller University have built what they describe as the most comprehensive map yet of how essential genes interact in Mycobacterium tuberculosis (Mtb), the bacterium that causes tuberculosis. Led by Jeremy Rock's Laboratory of Host-Pathogen Biology and published in Nature, the work identified nearly 6,500 genetic interactions among 762 genes, shedding light on genes of unknown function and suggesting new antibiotic combinations.

Why essential genes are hard to study

Tuberculosis claims more than a million lives each year, and treatment is demanding. Patients typically take several antibiotics for months, and drug-resistant infections take longer still. Mtb has a robust cell wall that limits drug entry, and it grows and dies slowly. As co-first author Isaac Kresse noted, the longer an infection lasts, the more opportunity bacteria have to acquire resistance.

Researchers have long known which Mtb genes are essential for survival, but explaining why they are essential has proved harder. Conventional CRISPR editing cuts DNA and disables a gene, which can kill the cell and leave nothing to study. Many essential genes therefore remain uncharacterised, and how they cooperate is poorly understood.

Silencing two genes at once

The team adapted CRISPR interference (CRISPRi), which turns down a gene's expression without cutting or permanently changing the DNA. Building on a single-gene CRISPRi system previously developed in the same laboratory, they created a version able to suppress two genes simultaneously.

This enabled genetic interaction mapping. Each gene is weakened on its own, then both are weakened together, and the observed effect on growth is compared with the effect expected from the individual results. Growth that is better or worse than predicted indicates a functional link between the genes. The researchers refined the method in Mycobacterium smegmatis, a related model species, before applying it to Mtb.

New clues to gene function

By comparing the interaction patterns of poorly understood genes with those of well-characterised ones, the team could place unknown genes within known biological pathways. Several previously uncharacterised genes appeared to take part in the electron transport chain, the machinery involved in cellular energy generation. Follow-up experiments confirmed that these genes contribute to the function of a major respiratory protein complex.

The researchers also combined the interaction data with AlphaFold 3, an AI system that predicts molecular structures, to explore how some of these gene products might work together at the molecular level.

Implications for drug combinations

TB drug regimens have largely been developed by trial and error, and the number of possible combinations is far too large to test exhaustively. The interaction maps offer a way to narrow the search. If suppressing two genes reduces growth more than expected, drugs aimed at those genes may work especially well together, according to co-first author Natalie Sayegh.

The team identified four synergistic drug combinations, including an experimental compound that boosted the killing activity of ethambutol, an established TB antibiotic. They also found one antagonistic pair, in which the two drugs weakened each other's effect. Co-senior author Michael DeJesus, who leads the computational biology work, said systematic data of this kind could usefully inform the design of new regimens.

The maps may also point to genes and processes not previously regarded as promising drug targets, opening a part of the genome that Sayegh described as largely off limits to drug discovery until now.

Limitations and next steps

The source reports the combinations as findings from the mapping and follow-up experiments; it does not describe testing in animals or patients, so clinical relevance remains to be established. The map covers essential genes only, and gene interactions may differ in the conditions bacteria face during human infection.

The team plans to extend the approach to nonessential genes and to study how interactions change under conditions closer to those in the body. The maps have been made publicly available, allowing other groups to test predictions and mine the data for further targets.

Source: Genetic interaction map could lead to better antibiotic pairings for tuberculosis (Phys.org, 5 November 2026).

Image: AI-generated illustration.