Genetic requirements for signaling from an autoactive plant NB-LRR intracellular innate immune receptor

Plants react to pathogen attack via recognition of, and response to, pathogen-specific molecules at the cell surface and inside the cell. Pathogen effectors (virulence factors) are monitored by intracellular nucleotide-binding leucine-rich repeat (NB-LRR) sensor proteins in plants and mammals. Here, we study the genetic requirements for defense responses of an autoactive mutant of ADR1-L2, an Arabidopsis coiled-coil (CC)-NB-LRR protein. ADR1-L2 functions upstream of salicylic acid (SA) accumulation in several defense contexts, and it can act in this context as a "helper" to transduce specific microbial activation signals from "sensor" NB-LRRs. This helper activity does not require an intact P-loop. ADR1-L2 and another of two closely related members of this small NB-LRR family are also required for propagation of unregulated runaway cell death (rcd) in an lsd1 mutant. We demonstrate here that, in this particular context, ADR1-L2 function is P-loop dependent. We generated an autoactive missense mutation, ADR1-L2D484V, in a small homology motif termed MHD. Expression of ADR1-L2D848V leads to dwarfed plants that exhibit increased disease resistance and constitutively high SA levels. The morphological phenotype also requires an intact P-loop, suggesting that these ADR1-L2D484V phenotypes reflect canonical activation of this NB-LRR protein. We used ADR1-L2D484V to define genetic requirements for signaling. Signaling from ADR1-L2D484V does not require NADPH oxidase and is negatively regulated by EDS1 and AtMC1. Transcriptional regulation of ADR1-L2D484V is correlated with its phenotypic outputs; these outputs are both SA-dependent and -independent. The genetic requirements for ADR1-L2D484V activity resemble those that regulate an SA-gradient-dependent signal amplification of defense and cell death signaling initially observed in the absence of LSD1. Importantly, ADR1-L2D484V autoactivation signaling is controlled by both EDS1 and SA in separable, but linked pathways. These data allows us to propose a genetic model that provides insight into an SA-dependent feedback regulation loop, which, surprisingly, includes ADR1-L2.