AI uncovers hidden rules of some of nature’s toughest protein bonds

(Press-News.org) (Auburn, AL) Imagine tugging on a Chinese finger trap. The harder you pull, the tighter it grips. This counterintuitive behavior also exists in biology. Certain protein complexes can form catch-bonds, tightening their grip when force is applied. These interactions are essential in processes ranging from how bacteria attach to our cells to how tissues in our body hold together under stress.

But a fundamental mystery has lingered: Do catch-bonds need to be stretched to a certain threshold before they strengthen, or do they activate as soon as force is applied?

In a new study published in Journal of Chemical Theory and Computation, Dr. Marcelo Melo (Colorado State University, formerly Auburn) and Dr. Rafael Bernardi (Auburn University) provide an answer. By combining large-scale molecular simulations with artificial intelligence, they discovered that catch-bonds “switch on” almost immediately after force is applied.

AI as a Molecular Detective

To crack the problem, the team simulated the behavior of a bacterial protein complex called cellulosomes, one of the strongest catch-bond systems known in nature. Using steered molecular dynamics simulations – essentially a computational microscope that stretches molecules atom by atom – they generated hundreds of high-resolution “movies” of the protein under stress.

AI regression models were then trained to predict when the protein complex would rupture. Surprisingly, the AI could make accurate predictions using only short snippets of simulation data, well before the bond actually broke.

“This told us that the proteins already ‘decide’ their level of resilience right after the pulling begins,” said Dr. Bernardi, Associate Professor of Physics at Auburn University. “The catch-bond mechanism is activated almost instantly.”

Why It Matters

Understanding catch-bonds is not just a curiosity. They are central to how bacteria like Staphylococcus aureus resist being washed away, how our immune cells stick to blood vessels, and how tissues such as cartilage endure constant mechanical stress.

“These are systems where life has learned to use force as an advantage,” explained Dr. Bernardi. “By learning from them, we can design new biomaterials, adhesives, and even drug strategies that work with mechanical stress instead of against it.”

A New Blueprint for Bioengineering

The study also highlights the power of AI to make sense of complex biological data. Instead of relying on static structures, the models captured dynamic patterns of motion across protein interfaces, finding the subtle signals that predict stability.

“This is exciting because it shows AI can detect early signs of resilience that humans would miss,” said Dr. Bernardi. “That opens the door to using these tools in drug design, biomaterials, and synthetic biology.”

The research was supported by the National Science Foundation, and it demonstrates the growing role of computational biophysics at the interface of AI and biology.

“This project shows how physics, biology, and artificial intelligence can come together to answer questions that none of us could solve alone,” said Dr. Bernardi.

The paper, “AI Uncovers the Rapid Activation of Catch-Bonds under Force”, was published in Journal of Chemical Theory and Computation.

----

About Auburn’s Computational Biophysics Group

The Computational Biophysics Group at Auburn University uses high-performance computing and advanced AI methods to study how biological molecules behave under force. By revealing the molecular principles of adhesion, resilience, and mechanobiology, the group helps guide the development of new antimicrobial therapies, biomaterials, and computational tools.

END