AI-powered plant immunity could protect crops from bacterial diseases
AlphaFold has been used to predict how targeted immune receptor engineering could lead to increased disease resistance in plants.
Scientists at the University of California, Davis (CA, USA), have leveraged artificial intelligence (AI) to enhance plants’ ability to recognize a wider array of bacterial threats. This could lead to novel strategies to support broad-spectrum disease resistance in vital food crops such as tomatoes and potatoes.
Like animals, plants possess immune receptors to detect and defend against bacteria. One such receptor, FLS2, is responsible for recognizing flagellin, a protein found in the tiny tails bacteria use to move. FLS2 is a well-studied pattern recognition receptor present in most land plants that can detect bacteria and trigger an immune response. However, bacteria are constantly evolving to evade detection. Gitta Coaker, a professor in the Department of Plant Pathology and lead author of the study, explained, “Bacteria are in an arms race with their plant hosts, and they can change the underlying amino acids in flagellin to evade detection.”
The protein–protease duo essential to parasite immunity
A newly discovered pathway reveals potential therapeutic target against helminth infections.
Immune system 2.0
To help plants keep pace with evolving bacterial threats, Coaker’s team combined natural variation with AlphaFold, a tool designed to predict the 3D shapes of proteins. By re-engineering FLS2 to be more sensitive to bacterial markers, the researchers upgraded the plant’s immune system to identify a broader range of intruders.
The team introduced FLS2 variants into Nicotiana benthamiana plants using CRISPR-Cas9, studying how they perceived flg22 – an epitope on the bacterial flagellin monomer – via reactive oxygen species (ROS) production and mitogen-activated protein kinase (MAPK) phosphorylation assays. To help Coaker and the team understand the effects of any design modifications to FLS2, they used AlphaFold to predict the structures and possible interactions of FLS2 homologues, flg22 variants and co-receptor NbSERK3A.
The team focused on four FLS2 homologues that recognize more bacteria, even though they aren’t found in common crop species. By comparing these broadly detecting receptors with more narrowly focused ones, the researchers pinpointed the specific amino acids that needed to be altered. “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 stated.
AI tool for forecasting infectious disease risk
The first AI tool to use large language modeling to predict infectious disease risk has been developed.
Broader implications for crop protection
This breakthrough paves the way for developing broad-spectrum bacterial resistance in crops through predictive design, according to Coaker. One significant target for this research is Ralstonia solanacearum, the bacterium responsible for bacterial wilt, a major threat to crops. Certain strains of this soil-borne pathogen can infect over 200 plant species, including staple crops such as tomatoes and potatoes.
In future, the team plans to develop machine learning tools to predict which immune receptors are most suitable for editing to improve the perception and effectiveness of pattern recognition receptors. They are also working to narrow down the number of amino acids that need to be modified. Furthermore, this approach could be applied to enhance the detection capabilities of other immune receptors using a similar strategy.
