An evolved enzyme makes only the TB-active form of an oxazolidinone building block

Directed evolution is a screening problem before it is a chemistry problem. Mutations go into the gene that encodes an enzyme, thousands of variants are made and tested for the trait you want, the best performers are kept, and the cycle runs again ^[13]. The method dates to the early 1990s and won Arnold the 2018 Nobel Prize in Chemistry ^[14].

The loop started from something weak. Reading "a few hundred" as about 300, the single hit came out of roughly 6 percent of the freezer collection ^[23]. "It wasn't necessarily a very good starting point, but we iteratively improved this enzyme to ultimately get to a final point where it's high yield and high selectivity," Li said ^[12].

Both mirror images hold the same atoms, arranged differently at one stereogenic center, and that arrangement sets the molecule's shape, which decides how it meets proteins and other biomolecules in the body ^[7]. Building the center correctly from the start, enzymatically, instead of making a mixture and losing one half ^[10], is worth at most a factor of two in usable intermediate at that step ^[22]. The paper's listed title describes the chemistry as a biocatalytic aziridination ^[24].

The drug does the same thing to a bacterium either way. The compound is the same compound, and the route to it wastes less. The cost claim attached to it is qualitative: better cost efficiency and less waste, across clinically relevant and discovery-stage versions of the compound ^[17]. "But with these tools, we can make enzymes that do so much more efficiently and with little waste, reducing the cost of producing what we need for our daily lives," said Arnold, who directs the Donna and Benjamin M. Rosen Bioengineering Center at Caltech ^[15][16].

More than 80% of TB cases occur in low- and middle-income nations, where drug-resistant strains are on the rise and the cost of medicines is a burden ^[2]. An enzyme that stops throwing away half of one step's output works on that cost at the manufacturing step, and the intermediate here has uses beyond TB drugs ^[6]. Li is direct about the distance between this result and a supply chain. "We solved the first step to prove that the enzymes can make these molecules," Li said ^[19].