Effector suppression of NLR function

Two papers here describe the same effector from opposite ends, and they should be read as one result. AVRcap1b, from the potato late blight pathogen, binds a host trafficking protein through one end of itself and an activated helper NLR through the other. Neither contact alone suppresses immunity (madhuprakash-2026-avrcap1b-tol9a, seager-2026-avrcap1b-nrc3).

The host-protein contact

AVRcap1b binds the ENTH domain of NbTOL9a, a member of the TOL family that acts in ESCRT vesicle trafficking. Domain chimeras localise the interaction to that domain, and the complex reconstitutes in vitro from purified components. The co-crystal structure shows the effector's WY1 and LWY2 regions gripping the ENTH domain across about 800 A2, with the alpha1 and alpha2 helices of ENTH inserting into a shallow concave pocket (madhuprakash-2026-avrcap1b-tol9a).

The effector fold itself was new: L-shaped, where the only previously solved RXLR-LWY structure is linear. The bend comes from one module, LWY5, and not from loss of the inter-repeat features thought to enforce linearity, which are all conserved. L-shaped effectors turn out to form a family of 152 sequences across Phytophthora (madhuprakash-2026-avrcap1b-tol9a).

Mutating the interface costs both binding and suppression. But binding is not sufficient: orthologs from three closely related species preserve all six key interface residues, bind NbTOL9a as well as the original, and suppress nothing (madhuprakash-2026-avrcap1b-tol9a).

The receptor contact

The effector associates with sensor-activated NbNRC2 and never with the resting state, through an interface distinct from the ENTH one. The P92E variant, which cannot bind NbTOL9a, retains that association (madhuprakash-2026-avrcap1b-tol9a) and retains partial suppression (seager-2026-avrcap1b-nrc3). That variant is the hinge between the two papers.

Cryo-EM of the receptor side shows what the contact does. The effector wraps around three activated NRC3 protomers on the face that would be buried inside a completed hexamer, and assembly stops there. LWY6 and LWY7 account for 61% of the buried surface, and a four-residue motif at the periphery of the NB domain, YEFF, is engaged by both. Mutate that motif and the receptor still builds a resistosome but no longer binds the effector — binding and assembly separated cleanly (seager-2026-avrcap1b-nrc3).

Removing the effector's N-terminal domains lets a fourth protomer in, which is how the block is shown to be steric rather than allosteric, and how the stepwise character of assembly is demonstrated at all (seager-2026-avrcap1b-nrc3).

Two ways to stop a receptor

The comparison the papers draw is with SS15, a potato cyst nematode effector that locks resting NRCs before they can activate. AVRcap1b acts only after activation. A metazoan and an oomycete have converged on the same network node from opposite sides of the activation step, and the trade-offs differ: SS15 must arrive first and in favourable stoichiometry, AVRcap1b can arrive late (seager-2026-avrcap1b-nrc3).

The proposed model is that AVRcap1b is not an inhibitor at all but a bridge, connecting an activated helper to the ESCRT pathway (madhuprakash-2026-avrcap1b-tol9a). No ternary complex has been resolved. The model is built from two binding surfaces and one mutant that separates them.

The NRC2 complex, the effector's originally described target, still could not be purified. The structure is of NRC3, which is where the work went instead (seager-2026-avrcap1b-nrc3).