Enzymes are biological catalysts that accelerate reactions and carry them out with a high degree of specificity. Enzymes are usually proteins. Over time, evolution has produced a wide variety of protein folds for enzymes: when a protein is formed, a long chain of amino acids is initially produced, which then folds into a defined three-dimensional structure. It is precisely this structure that determines how the enzyme functions. The so-called TIM barrel fold plays a prominent role. It is found in around 10 per cent of all known enzymes and is capable of facilitating almost all types of reaction.
Although artificial TIM barrels have already been designed on a computer and confirmed experimentally, these so-called de novo proteins, unlike their natural counterparts, possessed no enzymatic activity whatsoever; they merely exhibited the same structure. This rendered the proteins unusable for application in biological reactions. A research team led by Prof. Dr. Birte Höcker, Chair of Biochemistry III at the University of Bayreuth, has tackled this problem in collaboration with the research group of Prof. Dr Roberto Chica at the University of Ottawa. Using a new workflow called CANVAS, the researchers have transformed the non-functional scaffolds into active enzymes.
“In our study, we combined various computer-based methods. This enabled us to specifically extend the artificial TIM barrels to include a tailor-made active site,” says Dr. Julian Beck, lead author of the study and research assistant at the University of Bayreuth’s Biochemistry III research group. This newly inserted active site confers an exceptionally high measurable activity on the previously non-functional enzymes.
“We chose the Kemp elimination as a test reaction for the new enzymes – a classic non-natural reaction in protein design that is easy to measure,” says Höcker. In a single design round, the researchers achieved exceptionally high activity with the engineered enzyme KempTIM1: The catalytic efficiency – which describes how ‘good’ an enzyme is – was already seven times better for KempTIM1 than for comparable enzymes in other recent publications, without any further experimental optimization. Subsequent optimization steps have yielded a new variant called KempTIM4b, which even exceeds the activity of KempTIM1.
“Overall, our research shows that it is possible to use de novo proteins as a starting point for new enzymes, thereby expanding the existing repertoire of available enzymes. This opens up new possibilities in biotechnology and green chemistry by enabling the design of new, tailor-made proteins for specific reactions,” said Beck.
Original publication: Customizing the structure of minimal TIM barrels to create efficient de novo enzymes. Julian Beck, Benjamin J. Smith, Mark Kriegel, Niayesh Zarifi, Emily Frend, Ahana G. Harsha, Jan Hartmann, Roberto A. Chica & Birte Höcker. Nature Chemical Biology (2026)
DOI: https://doi.org/10.1038/s41589-026-02250-w
