Subsequently we used this knowledge to generate engineered Pm3 NLRs that fully incorporate Pm3d- and Pm3e specificities and verified their further extended recognition spectra, now including three mildew AVRs, in Nicotiana and wheat.

Using chimeric NLRs between Pm3d, Pm3e and the ancestral Pm3CS (completely overcome by powdery mildew) we then identified individual amino acid polymorphisms defining the recognition specificity of Pm3d and Pm3e towards their various effector targets.

We found that Pm3e, like Pm3d, has the ability to recognize variants of the BgtE-20069b effector (now AvrPm3d_1/AvrPm3e_2). However, its second AVR, BgtE-5754 (AvrPm3e_1), encodes a much larger effector protein predicted to exhibit a completely different structure!

But what about Pm3e? Here our mapping approach identified two genetically unlinked AVR loci on chromosomes 4 and 9, with the later one perfectly overlapping with the AvrPm3d locus. Identifying the underlying AVRs held an interesting surprise!

However, the locus contained two AVR genes: the previously described BgtE-20069b effector (AvrPm3d_1) as well as its neighboring gene BgtE-20069a (AvrPm3d_2), a close paralog that was missed in our previous study. Only isolates compromised in both AVRs can break the Pm3d resistance.

Using our recently developed avirulence depletion mapping strategy (doi.org/10.1371/jour...), we found that there is a single AvrPm3d locus in wheat powdery mildew.

Interestingly the Pm3d and Pm3e NLRs differ by only three and two amino acids, respectively, from the ancestral Pm3CS which has been completely overcome by the powdery mildew pathogen. So what makes Pm3d and Pm3e so successful? To answer this we aimed at identifying the corresponding AVR effectors.

The wheat Pm3 NLR locus is highly diverse with at least 17 alleles providing resistance against wheat powdery mildew. We challenged transgenic Pm3 wheat lines with >90 mildew isolates collected worldwide and identified two Pm3 alleles, Pm3d and Pm3e, that confer broad-spectrum resistance.