Enzymes are nature’s tiny machines, responsible for everything from digesting food to breaking down environmental pollutants. But what if we could design custom-made enzymes to tackle real-world challenges—like neutralizing harmful chemicals or detecting diseases—more efficiently than ever before? Jing Pan Ph. D., a faculty member in the Department of Mechanical and Aerospace Engineering at the University of Florida, is pioneering an innovative approach to enzyme engineering. By combining artificial intelligence (AI) with molecular simulations, his team is developing nanoscale enzyme machines with unprecedented precision and functionality.
Through a collaboration with Wenjun Xie Ph. D., from UF’s College of Pharmacy and funding from the Defense Advanced Research Projects Agency (DARPA), Dr. Pan’s research is pushing the boundaries of what’s possible in molecular design. By leveraging AI to predict protein structures and refine their performance through advanced simulations, this work has the potential to transform fields ranging from medicine to environmental protection to defense.
Engineering Next-Generation Biosensors
Pan’s lab specializes in biosensors capable of continuously monitoring human physiology. At the heart of these biosensors are specialized molecules, such as enzymes, that function as nanoscale machines, detecting and responding to biochemical changes within the body. However, designing artificial enzymes that rival natural ones in efficiency remains a major hurdle. Traditional protein design relies heavily on trial-and-error experimentation, but the vastness of the protein design space far exceeds what can be physically tested in the lab.
To address this challenge, Pan and Xie’s teams are developing an AI-driven engineering pipeline that automates and optimizes enzyme design. By leveraging AI to predict protein structures and using molecular simulations to refine functionality, they aim to significantly improve the success rate of enzyme development while reducing the need for extensive laboratory experimentation.
A Proof-of-Concept: Improving Dehalogenase for Industrial and Defense Applications
As an initial test case, the research team is focusing on haloalkane dehalogenase, an enzyme capable of breaking down harmful halogenated compounds. This enzyme has broad applications in industrial and defense settings, such as:
- Chemical Weapon Neutralization: Mustard gas, a toxic halogenated agent used in chemical warfare, could potentially be neutralized through enzymatic hydrolysis using optimized dehalogenase.
- Environmental Bioremediation: Many halogenated pollutants, including pesticides and industrial waste, persist in soil and water. Engineered dehalogenase could be deployed to detoxify groundwater and clean up industrial contamination sites.
While this project primarily serves as a proof-of-concept, the overarching goal is to establish a generalized AI-driven approach for enzyme engineering. By demonstrating the ability to create artificial enzymes with tailored properties, this research could open doors to a wide range of applications, including therapeutic protein design and infectious disease control.
Interdisciplinary Collaboration: AI Meets Experimental Biochemistry
The success of this project is rooted in interdisciplinary collaboration. Dr. Xie, an expert in AI-driven enzyme design, generates candidate protein structures using generative AI. Dr. Pan’s lab then synthesizes and experimentally validates these proteins, to characterize their performance. This synergistic approach allows for iterative refinement of enzyme designs, ensuring that the final proteins meet the desired functional criteria.
“AI gives us an unprecedented ability to explore the vast landscape of protein possibilities, but experimental validation is crucial to ensuring these designs function in real-world conditions,” says
Dr. Pan. “By combining computational power with biochemical expertise, we can engineer enzymes with properties previously thought unattainable.”
Future Directions: Expanding the Scope of AI-Driven Molecular Engineering
Dr. Pan’s team is in the early stages of validating their technology, but they envision a future where
AI-generated protein machines can perform a variety of critical tasks. Potential applications include:
- Custom-designed enzymes for targeted disease monitoring
- Therapeutic proteins optimized for human physiology control
- Enzymatic solutions for industrial and environmental challenges
While the current focus is on dehalogenase, the principles developed in this study could be adapted to engineer a broad range of functional enzymes. By bridging the gap between AI-driven predictions and real-world biochemical functionality, this research marks an important step toward the future of molecular engineering.
Story & Editing & Design by: Christi Swiers
Marketing & Communications Specialist
UF Mechanical & Aerospace Engineering
March 5, 2025