Enzymes as viscoelastic catalytic machines

Enzymes are not only chemical catalysts, but finely tuned mechanical machines. Enzymes must bind substrates, reshape themselves, reach transition states, and release products; these steps often depend on coordinated motions that extend far beyond the active site. In our study, we asked whether the mechanical properties of an enzyme are encoded in its amino-acid sequence, and whether mutations far from the active site can disrupt function by changing the enzyme’s internal flexibility rather than by directly altering chemistry.

We focused on guanylate kinase, an enzyme that undergoes a large conformational change during its catalytic cycle. The Tlusty group uses structural comparisons between open and substrate-bound forms, and they identified regions that experience high mechanical strain during function. These regions act like hinges or flexible joints, enabling the protein to move smoothly between conformations. They also used AlphaFold-based structure predictions to estimate how single amino-acid mutations deform the protein, providing an independent way to identify mechanically important sites. Together, these approaches allowed us to map “mechanical sweet spots” in the enzyme: positions where small sequence changes can have large effects on motion and activity.

To test these predictions experimentally, we generated 34 single-mutant variants of guanylate kinase. Some mutations were placed in substrate-binding regions, some in high-strain regions distant from the active site, and others in low-strain control regions. We then measured enzyme activity and, for a subset of variants, directly probed mechanical response using nano-rheology. This method applies tiny oscillating forces to proteins and measures how they stretch, relax, store mechanical energy, and dissipate energy. The results showed that mutations in high-strain regions make the enzyme stiffer and more viscous, reducing the conformational flexibility needed for catalysis. By contrast, mutations in low-strain, non-binding regions were often nearly neutral.

A key result is that mechanical disruption can impair catalysis even when the mutation is far from the active site. Mutations in binding residues predictably reduced activity, but mutations in high-strain regions also strongly reduced enzymatic function, whereas most control mutations did not. This demonstrates that enzyme activity depends on long-range mechanical networks within the protein. In this view, an enzyme resembles a precisely engineered viscoelastic material: it must be elastic enough to return to the right shape, yet viscous enough to dissipate energy and avoid uncontrolled motion. The characteristic mechanical timescales measured in the study were comparable to the enzyme’s catalytic rate, suggesting that the enzyme’s material properties are tuned to its biological task.

These findings open new opportunities for protein design. Traditional protein engineering often emphasizes active-site chemistry or sequence conservation, but our results show that mechanical strain and viscoelasticity are also central design variables. By identifying residues that control flexibility, stiffness, and long-range force transmission, researchers may be able to engineer enzymes with altered activity, specificity, or mechanical responsiveness. More broadly, the study supports the idea that proteins are evolutionary materials: their sequences encode not only folded structures, but also the collective motions and mechanical properties required for life’s molecular machines to work. Combining mechanics, AI-based structure prediction, and experiments could help guide the creation of programmable molecular machines with functions inspired by, and eventually extending beyond, natural proteins.

Enzymes as viscoelastic catalytic machines
Weinreb E., McBride J. M., Siek M., Rougemont J., Renault R., Peleg Y., Unger T., Albeck S., Fridmann-Sirkis Y., Lushchekina S., Sussman J. L., Grzybowski B. A., Zocchi G., Eckmann J., Moses E., & Tlusty T. (2025) Nature Physics. 21(5), 787-798.
Article, Supplementary