For many of us, a cup of hot coffee is how we start our day. For Texas A&M Health researchers, it could also provide new ways to control engineered cells in future drugs.
A team at Texas A&M Health Biosciences Institute has developed an artificial intelligence-designed molecular switch that uses caffeine to rapidly isolate engineered proteins in living cells and trigger a cellular response on demand. The platform, called CODS (short for caffeinergic dissociation system), could help scientists build safer and more controllable gene and cell therapies.
This study Journal of the American Chemical SocietyThe study was led by Yubin Zhou, PhD, FAAAS, FAIMBE, FRSC, director of the Center for Translational Cancer Research at the Institute of Biological Sciences and Professor at the Texas A&M Naresh K. Vashish School of Medicine, along with Tianlu Wang, PhD, and colleagues. Graduate students Brendan McKee and Itsuki Nonomura played central roles in this research, with McKee driving AI-guided protein design and computational modeling efforts and Nonomura leading key molecular engineering and live-cell validation studies.
AI is changing the way biology is designed. Instead of relying solely on parts of proteins that already exist in nature, it is now possible to design new mini-proteins with specific behaviors. Here, AI was used to turn caffeine into a precise trigger to control engineered cells. ”
Yubin Zhou, MD, Texas A&M Naresh K. Vashisht Professor of Medicine
AI as a molecular designer
The new research builds on Zhou’s previous caffeine reaction technology, but goes in a decidedly different direction.
Previous systems showed that caffeine helps hold engineered proteins together. However, CODS does the opposite. Use caffeine to wean out proteins. This distinction is important because future treatments may require ways to not only activate cells, but also to halt, quiescent, or reset cells as needed.
To build CODS, the team used AI-guided protein design to create a small synthetic binder that recognizes a caffeine-responsive protein module. In the absence of caffeine, the binder holds the system together, but when caffeine is added, the proteins separate.
In this way, CODS acts like a molecular fastener. Without caffeine, the clasp will remain closed. The clasp opens when you ingest caffeine.
“Many genetically encoded molecular tools act like accelerators,” Wang says. “CODS provides something more like a brake or pause button.”
high performance computing
The AI-powered design process required significant computational power. The research team used protein design algorithms and molecular simulations to identify, evaluate, and purify synthetic binders before testing the most promising candidates in living cells.
This research was made possible by the Texas A&M High Performance Research Computing (HPRC) service, which provides the computing power needed to run advanced AI-driven protein design workflows at scale.
“High-performance computing was essential to this project,” says Chou. “Designing AI proteins requires computational power, and Texas A&M HPRC services helped us move from conceptual idea to functional molecule switch more quickly.”
The resulting system was responsive to very low caffeine concentrations, worked within minutes, and could be undone repeatedly by adding or removing caffeine.
Regulation of genes, cell death, and immune cells
Researchers demonstrated CODS in three main ways.
First, they used it to control gene activity. Without caffeine, the engineered genetic circuit remained active. When caffeine is added, CODS sequesters the target protein needed to keep the gene switched on, causing a sharp drop in gene activity. Removal of caffeine restored the system.
Second, the research team used CODS to control programmed cell death. By rewiring cell death proteins with a caffeine-responsive switch, they created a system in which caffeine triggers an inflammatory cell death known as pyroptosis. This will help scientists study inflammation and may one day support the design of therapeutic cells that can be removed when needed.
Finally, the most translational demonstration involved CAR T cells, immune cells designed to recognize and attack cancer. CAR T-cell therapy has shown remarkable success in some blood cancers, but it can also cause serious side effects if immune cells become too active. A caffeine-induced safety switch may offer clinicians a way to temporarily reduce CAR T cell activity without permanently destroying therapeutic cells.
Using CODS, this Texas A&M team built a split CAR system that remains active when caffeine is not present and remains passive when caffeine is added. In clinical tests, caffeine strongly reduced activation of CAR T cells, suggesting that CODS may be a practical safety off switch for genetically engineered immune cells.
Beyond coffee: Towards programmable medicine
Chou emphasized that caffeine itself is not a cancer treatment drug. Instead, caffeine acts as a safe, familiar signal that can communicate with specially engineered cells.
“Coffee is not a substitute for medicine,” says Chou. “But caffeine can help us imagine drugs that are more controllable, more responsive, and safer for patients.”
A broader advance is the use of AI to design new proteins that behave in ways that nature does not readily provide. Similar strategies could eventually be used to construct switches controlled by other well-known molecules, commercially available drugs, or clinically approved drugs.
Before CODS can move into clinical use, the system will require further testing in therapeutic cells, animal models, and disease-relevant settings. Still, this study represents an important step toward programmable medicine by providing a framework for designing treatments that can be adjusted after implementation.
“Strong treatments require strong controls,” Zhou says. “By combining AI-designed proteins, high-performance computing, and familiar small molecules, we are building a new language to communicate with engineered cells.”
sauce:
Reference magazines:
Tetsuya Nonomura Others. (2026). AI guide Also Design of a caffeine-induced protein dissociation system. Journal of the American Chemical Society. DOI: 10.1021/jacs.6c02343. https://pubs.acs.org/doi/10.1021/jacs.6c02343
