AI modelling of protein structures can distinguish between sensor and helper NLR immune receptors

NLR proteins are intracellular immune receptors present across all kingdoms of life, with exceptional diversification in plants. NLRs can function in diverse configurations, including singletons, pairs, and intricate hierarchical networks, to effectively recognize and combat invading threats¹. In NLR pairs and networks, sensor NLRs detect pathogen secreted effectors but require helper NLRs for immune signaling. Upon activation, singleton and helper NLRs oligomerize, insert into the plasma membrane, and act as calcium channels. Here, we propose that the AI system AlphaFold 3 can classify NLR pairs and networks proteins into sensor or helper categories based on predicted structural characteristics. Helper NLRs showed higher AlphaFold 3 confidence scores than sensors when modeled in oligomeric configurations. Furthermore, we could predict funnel-shaped structures—essential for activating immune responses—in helpers but not in sensors. These approaches could overcome the limitations of sequence-based annotation methods and provide new insights into the structural and functional organization of immune receptors.