“`html

Researchers from the University of Michigan and Carnegie Mellon University have crafted an innovative tool that enhances the accessibility of eco-friendlier chemistry.

This tool, detailed in a study funded by the U.S. National Science Foundation and released on Oct. 1 in the journal Nature, eliminates a significant obstacle to the broader utilization of biocatalysis.

Biocatalysts, commonly known as enzymes, are a category of proteins that have adapted to execute chemistry that can be intricate and remarkably efficient—usually in aqueous conditions and at ambient temperature—thus negating the necessity for hazardous or costly chemical reagents to perform reactions. However, they are also extremely selective, meaning that they are tailored to interact only with specific initial compounds, or substrates, found in their natural surroundings.

To harness the capabilities of biocatalysts in the lab, chemists must understand which other substrates a protein can interact with and, more specifically, which enzymes will correspond with their desired substrate.

“Biocatalysis presents a more sustainable method to synthesize molecules, and it can also provide access to compounds that we couldn’t create through conventional chemical techniques,” stated Alison Narayan, a chemistry professor in the U-M College of Literature, Sciences, and the Arts, as well as a research professor at the Life Sciences Institute. “However, most known substrates for these biocatalysts originate from nature, which represents only a tiny fraction of the compounds that chemists work with.”

Narayan’s team aimed to bridge the enduring divide between the initial compounds that chemists utilize and the enzymes that could potentially react with those compounds. The project commenced with an initiative to associate proteins with substrates on a large scale. Concentrating on one family of enzymes, Alexandra Paton devised a high-throughput reaction platform that permitted the team to evaluate over 100 substrates against each protein within the entire protein family.

“We uncovered hundreds of new relationships between chemical space and protein space, assembling this varied dataset,” mentioned Paton, a former postdoctoral researcher in Naryan’s lab and the principal author of the study. “That was when we began contemplating more broadly about what we could create with all this information.”

Narayan’s team, together with Gabe Gomes, an assistant professor of chemical engineering and chemistry at Carnegie Mellon University, and Daniil Boiko, who was then a graduate student in Gomes’ lab, utilized this dataset to establish an enzyme recommendation system. The Gomes lab employed its proficiency in machine learning to refine a predictive model capable of navigating between the protein landscape and the chemical landscape.

The resulting open-access CATNIP online platform allows chemists to enter their initial compound and obtain a ranked list of biocatalysts from this protein family that would most effectively facilitate a chemical transformation; conversely, one can start with a chosen enzyme and determine its potential substrates. Boiko likens the platform’s predictive function to a web search, enhancing the results to ensure that the most favorable answers—or the most promising candidates—are presented at the top of the ranked list based on their likelihood of success.

“It serves as an excellent foundational model to enable synthetic projects utilizing biocatalysts,” stated Paton, who is presently an assistant professor of chemistry at the University of Rochester. “Moreover, efforts are already underway to start broadening the database beyond this single enzyme family.”

The research also received backing from the Novartis Global Scholars Program, the Camille Dreyfus Teacher Scholar Award, and the University of Michigan. Other authors of the study include Jonathan Perkins and Nicholas Cemalovic from U-M and Thiago Reschützegger from the Federal University of Santa Maria, Brazil.

“`