For decades, scientists have dreamed of designing enzymes from scratch – molecular machines that could catalyze any reaction we need, from breaking down plastics to synthesizing new medicines. But there was always a catch: computationally designed enzymes performed dismally compared to their natural counterparts, requiring extensive laboratory optimization to reach even modest activity levels.

Now, researchers at the Weizmann Institute of Science in Israel have shattered this barrier. Their computationally designed enzymes exhibit catalytic rates matching those found in nature – without any laboratory evolution or optimization whatsoever.

"This is really a game-changer," says Sarel Fleishman, who led the research published in Nature. "We can now program enzymes that work efficiently right from the design stage, which opens up enormous possibilities for biotechnology."

The team focused on designing enzymes for the Kemp elimination reaction, a model chemical transformation that has served as a benchmark for enzyme designers for over 15 years. Previous computational designs required months of laboratory evolution to achieve respectable activity levels. The best designs without optimization showed catalytic efficiencies around 1-400 M⁻¹s⁻¹ – roughly 1,000 times worse than typical natural enzymes.

The new designs blow these numbers out of the water. The team's best enzyme, dubbed Des27.7, achieved a catalytic efficiency of 12,700 M⁻¹s⁻¹ straight from the computer. With a single strategic mutation, this jumped to an astounding 123,000 M⁻¹s⁻¹ – firmly in the range of natural enzymes.

"We're talking about a 100-fold improvement over any previous computational design," notes Fleishman. "And we achieved this without touching a test tube for optimization."

The secret to their success lay in a counterintuitive insight: don't just focus on the active site where the chemistry happens. Previous design efforts concentrated on perfecting the catalytic machinery while neglecting the rest of the protein. This led to floppy, unstable structures where the carefully positioned catalytic groups would shift out of alignment.

The Israeli team took a holistic approach, using their computational tools to ensure the entire protein structure was rock-solid. They assembled protein backbones from fragments of natural proteins, then systematically optimized every amino acid for stability. Only then did they introduce the catalytic machinery.

"It's like building a precision watch," explains Fleishman. "You need a stable frame before you can install the delicate moving parts."

The approach paid off spectacularly. Their designed enzymes are stable at temperatures above 85°C and show remarkable structural accuracy – crystal structures matched the computational models to within fractions of an ångström.
In a surprising twist, the team discovered that a supposedly essential component of Kemp eliminase design – an aromatic amino acid for binding the substrate – wasn't necessary at all. When they replaced this aromatic residue with a simple leucine, the enzyme's efficiency increased tenfold.

"This really shows how much we still have to learn about enzyme catalysis," says Fleishman. "A design principle we've followed for two decades turned out to be holding us back."

The implications extend far beyond academic interest. Enzymes catalyze virtually every chemical reaction in living organisms and are increasingly used in industry for everything from laundry detergents to pharmaceutical synthesis. But finding or engineering enzymes for new reactions has been laborious and expensive. With this computational approach, researchers could potentially design bespoke enzymes for any reaction on demand. Want an enzyme that breaks down a specific environmental pollutant? Or one that synthesizes a complex drug molecule? The new methods could make such designer enzymes a reality. "We're moving toward truly programmable biocatalysis," says Fleishman. "Instead of spending years optimizing enzymes in the lab, we can now design them correctly from the start." The team has made their computational tools freely available, hoping to accelerate enzyme design efforts worldwide. While the current work focused on a model reaction, they're confident the approach will generalize to more complex and useful transformations. As climate change and sustainability concerns drive demand for greener chemical processes, the ability to rapidly design efficient biological catalysts could prove transformative. These molecular machines work under mild conditions, use renewable resources, and produce minimal waste – a stark contrast to traditional chemical manufacturing. The age of designer enzymes may finally have arrived, no evolution required.

Journal reference: Nature DOI: 10.1038/s41586-025-09136-