Plants also have immune systems, just like animals. They use special receptors to spot harmful bacteria and defend themselves. One key receptor, called FLS2, helps plants detect flagellin, a protein in the tails that bacteria use to move.
But bacteria are clever. They continually modify the protein to evade the plant’s defenses.
To help plants fight back, scientists at UC Davis used artificial intelligence, specifically AlphaFold, a tool that predicts protein shapes. With it, they redesigned FLS2 to recognize more variations of flagellin, thereby strengthening the plant’s immune system and making it harder for bacteria to trick it.
The researchers studied plant receptors that could detect a wide range of bacteria, even if those receptors came from non-crop plants. By comparing these with less effective receptors, they determined which amino acids needed to be modified.
Plants use “trojan horse” to fight mold invasions
With this knowledge, they revived a receptor that had previously failed to protect the plant, making it stronger and more precise in fighting infections. This could lead to crops that are resistant to many different diseases, thanks to this innovative, predictive approach.
The researchers are focusing on a severe plant disease known as bacterial wilt, caused by the bacterium Ralstonia solanacearum. These harmful bacteria can infect over 200 different plant species, including important crops like tomatoes and potatoes.
To combat it, the team is utilizing machine learning to identify which plant immune receptors are the most effective targets for future editing. They’re also working on identifying the specific amino acids that need to be modified to improve the receptors’ ability to detect infections.
This method could enhance the ability of many plant receptors to detect harmful bacteria more effectively, thereby making crops stronger and more disease-resistant.
Journal Reference
- Li, T., Jarquin Bolaños, E., Stevens, D.M. et al. Unlocking expanded flagellin perception through rational receptor engineering. Nat. Plants (2025). DOI: 10.1038/s41477-025-02049-y
