Effector recognition and its evolution

How a receptor acquires a new specificity has one standard answer: it swallows the effector's host target and keeps it as an integrated decoy, so that anything binding the target binds the receptor. One paper here describes the inverse, and shows it is transferable.

Mimicry instead of integration

The barley receptor MLA3 recognises Pwl2, an effector of the blast fungus, and binds it directly through the last 85 residues of its LRR domain. Pwl2's actual virulence target is HIPP43, a heavy metal-associated protein. Modelling the MLA3-Pwl2 complex and comparing it with the experimental Pwl2-OsHIPP43 structure shows the effector using an overlapping surface in both cases, with the two host proteins engaging it in related orientations despite no sequence similarity (gomezdelacruz-2026-mla3-mimicry).

The mimicry is of an interface, not of a domain. The HIPP43 HMA fold is four antiparallel beta strands and two helices; the corresponding MLA3 region is three parallel strands and one helix, and is simply the last three LRR repeats. So this is not a misannotated integrated domain (gomezdelacruz-2026-mla3-mimicry).

Two experiments make the case rather than the model. Expressing OsHIPP43 weakens MLA3-mediated cell death, so receptor and host target compete for the same effector. And the interface residues are interchangeable in both directions: MLA3 residues put into OsHIPP43 preserve Pwl2 binding, and OsHIPP43 residues put into MLA3 preserve recognition (gomezdelacruz-2026-mla3-mimicry).

Specificity is three residues, and often one

MLA3 belongs to an allelic series of more than thirty alleles sharing over 90% identity, of which only MLA3 recognises Pwl2. Of twelve predicted contact residues, three are unique to MLA3. Lys926 sits in a negatively charged pocket of the effector; the other alleles carry a glutamate or aspartate there. Val931 and Tyr932 hydrogen bond to Pwl2 Arg44 where other alleles carry bulky residues that clash (gomezdelacruz-2026-mla3-mimicry).

A single D926K substitution gives the non-recognising MLA23 the ability, and the reciprocal K926D removes it. Of all nineteen alternatives at that position, only arginine preserves recognition, so what matters is a positive charge (gomezdelacruz-2026-mla3-mimicry).

Note the numbering trap: MLA3 Lys926 is MLA23 Asp926 and MLA1 Glu926. The same position, three names (gomezdelacruz-2026-mla3-mimicry).

Recognition that has to be unlocked

A different constraint appears in the potato late blight system. Rpi-vnt1.1 can recognise its effector AVRvnt1, but only the spliced isoform of its own mRNA assembles into a resistosome. Surface immune signalling shifts the isoform balance, so pattern recognition licenses the receptor before the effector arrives — and the shift happens whether or not the effector is present (gao-2026-vnt1-splicing). Recognition capacity and recognition competence are separable.

The receptors with nothing to recognise

Across the lettuce NRC network, nine sensors were shown to be functional by activating them genetically. No cognate effector or pathogen ligand is known for any of them (pai-2026-lettuce-nrc). The network's wiring is mapped; what it detects is not.

That gap is worth holding next to the mimicry result. If a receptor can acquire specificity by copying a host protein's binding surface, then the relevant question for an uncharacterised sensor is not only which effector it sees but which host protein it resembles.