Autoinhibition and priming

An immune receptor that fires without cause kills the cell for nothing. Five papers here bear on what holds NLRs off, and they answer at three different levels: conformation, transcript, and the interface itself.

Transcript-level licensing

The clearest case is a potato late blight receptor transcribed in two isoforms. One retains an intron and carries an N-terminal extension ahead of the MADA motif; the other is spliced and is a conventional MADA-NLR. Only the spliced form assembles, conducts calcium and triggers cell death. Surface immune signalling shifts the balance towards it, at transcript and protein level, without changing total transcript (gao-2026-vnt1-splicing).

Two things make this more than a regulatory curiosity. Locking the isoform in either direction makes the receptor insensitive to pattern-triggered signalling, which places surface signalling upstream of splicing rather than alongside it. And during real infection the shift happens whether or not the effector is present, and is lost when BAK1 is silenced. Pattern recognition licenses the intracellular receptor in advance (gao-2026-vnt1-splicing).

The inhibition is a property of length rather than sequence: extensions under 16 residues have little effect, those over 16 suppress. Modelling across 85 N-extended receptors puts the extension in the pore-forming region as a low-confidence, flexible segment that displaces the alpha1 helix. Some 619 Solanaceae receptors carry the feature, and none of the others has been tested (gao-2026-vnt1-splicing).

Conformational autoinhibition

Elsewhere autoinhibition is inferred from what breaks it. A single L619F substitution in an LRR domain renders a wheat receptor constitutively active and semi-dominant, but why is unknown: modelling the wild-type protein always returns an activated-like conformation, and no CCG10 resting-state structure exists (guo-2026-ccg10-resistosome). Only ZAR1 and NRC2 have been resolved at rest at all.

MHD-motif substitutions are used throughout the corpus as engineered autoactivity, a tool rather than a phenomenon (toghani-2026-sni-undecamer, pai-2026-lettuce-nrc). MADA-motif mutations do the complementary job, allowing oligomerisation without cell death so that complexes can be purified.

Priming as a distinct state

Sensor binding is modelled as moving the helper part-way, not all the way: the predicted sensor-bound conformation sits between the resting homodimer and the active hexamer in NB-HD1 rotation (toghani-2026-sensor-helper-interface). An effector that binds only the activated helper and never the resting one is consistent with that intermediate being real and recognisable (madhuprakash-2026-avrcap1b-tol9a).

The trigger-happy problem

Interface mutants make the cost of loosening a receptor visible. Of ten NRC3 mutations that reduce effector suppression, four fire with the sensor alone — insensitive to the effector, but no longer properly restrained. Only one mutation was both suppression-insensitive and not trigger-happy (seager-2026-avrcap1b-nrc3).

The same tension shows in the field. Transgenic lines locked in the assembly-competent isoform all expressed at low levels and gave only quantitative resistance, which the authors read as counterselection against autoimmunity, supported by that isoform triggering cell death on its own when overexpressed (gao-2026-vnt1-splicing).

A receptor that is easier to activate is not a better receptor. Every mechanism on this page is a plant paying for the ability to respond by making it difficult.