Researchers use AI to enhance plants' immune systems and fight bacterial diseases

Scientists at the University of California, Davis, have used artificial intelligence to enhance plants' ability to detect and defend against bacterial pathogens, potentially offering new strategies to protect crops like tomatoes and potatoes from diseases. The research, published in Nature Plants, focuses on improving the plants' internal immune systems by reengineering key receptors. Plants, like animals, have immune systems that help them detect and respond to threats. A critical part of this defense is the presence of immune receptors, which identify harmful bacteria. One such receptor, called FLS2, allows plants to recognize flagellin, a protein found in the tails of bacteria that help them move. However, bacteria are constantly evolving, and some can alter the amino acids in flagellin to avoid detection by these receptors. Gitta Coaker, a professor in the Department of Plant Pathology and the study’s lead author, explained that bacteria are engaged in an ongoing "arms race" with their plant hosts, making it challenging for plants to stay ahead. To address this, her team combined natural genetic variation with artificial intelligence, specifically using AlphaFold, a tool designed to predict protein structures. By reengineering the FLS2 receptor, they effectively upgraded the plant's immune system to detect a broader range of bacterial threats. The researchers studied receptors that were already known to recognize a wider range of bacteria, even if they weren’t found in major crop species. By comparing these with more specialized receptors, they identified which specific amino acids needed to be modified to enhance detection. "We were able to resurrect a defeated receptor, one where the pathogen has won, and enable the plant to have a chance to resist infection in a much more targeted and precise way," Coaker said. This breakthrough could pave the way for creating broad-spectrum disease resistance in crops through predictive design. One of the team's key targets is Ralstonia solanacearum, the bacterium responsible for bacterial wilt. This pathogen affects over 200 plant species and is a major threat to staple crops such as tomatoes and potatoes. Looking forward, the researchers are working on developing machine learning tools to identify which immune receptors are most promising for future editing. They are also aiming to reduce the number of amino acid changes needed to improve receptor function. The approach could be applied to other immune receptors, helping plants better detect and respond to a range of pathogens. The study was conducted in collaboration with researchers from UC Davis and the Lawrence Berkeley National Laboratory, including Tianrun Li, Esteban Jarquin Bolaños, Danielle M. Stevens, Hanxu Sha, and Daniil M. Prigozhin.