Disentangling cause and consequence: genetic dissection of the DANGEROUS MIX2 risk locus, and activation of the DM2h NLR in autoimmunity

Nucleotide-binding domain-leucine-rich repeat-type immune receptors (NLRs) protect plants against pathogenic microbes through intracellular detection of effector proteins. However, this comes at a cost, as NLRs can also induce detrimental autoimmunity in genetic interactions with foreign alleles. This may occur when independently evolved genomes are combined in inter- or intraspecific crosses, or when foreign alleles are introduced by mutagenesis or transgenesis. Most autoimmunity-inducing NLRs are encoded within highly variable NLR gene clusters with no known immune functions, which were termed autoimmune risk loci. Whether risk NLRs differ from sensor NLRs operating in natural pathogen resistance and how risk NLRs are activated in autoimmunity is unknown. Here, we analyzed the DANGEROUS MIX2 risk locus, a major autoimmunity hotspot in Arabidopsis thaliana. By gene editing and heterologous expression, we show that a single gene, DM2h, is necessary and sufficient for autoimmune induction in three independent cases of autoimmunity in accession Landsberg erecta. We focus on autoimmunity provoked by an EDS1-yellow fluorescent protein (YFP)NLS fusion protein to characterize DM2h functionally and determine features of EDS1-YFPNLS activating the immune receptor. Our data suggest that risk NLRs function in a manner reminiscent of sensor NLRs, while autoimmunity-inducing properties of EDS1-YFPNLS in this context are unrelated to the protein's functions as an immune regulator. We propose that autoimmunity, at least in some cases, may be caused by spurious, stochastic interactions of foreign alleles with coincidentally matching risk NLRs.