Ten Papers on NLR Biology Published in 2026

Ten Papers on NLR Biology Published in 2026

Read as one body of work rather than ten separate results. Every statement carries a source locator back to a figure, table or accession in the paper it came from.

Resistosome architecture and oligomeric state

What activated NLRs build, and how much that varies. Pentamer, hexamer, octamer, undecamer and stalled sub-hexameric i… · 9 papers

Sensor-helper communication in NLR networks

How expanded sensor NLRs signal through a small number of helper executors, and what holds the pairing specificity in … · 8 papers

Calcium signalling and where immunity is executed

Resistosomes as ion channels, and the growing evidence that which membrane they act on is itself a functional variable… · 5 papers

Effector suppression of NLR function

Pathogen effectors that target the receptors themselves rather than downstream immunity, and the structural basis of h… · 2 papers

Effector recognition and its evolution

How receptors acquire and change what they detect, including integrated decoys and the mimicry alternative.… · 3 papers

Autoinhibition and priming

What keeps a receptor off until it is wanted: resting-state conformation, N-terminal extensions, transcript isoform, a… · 5 papers

Evolution of NLR repertoires

How NLR families expand, contract, subfunctionalise and specialise across lineages and tissues.… · 9 papers

Engineering receptors

Moving interfaces, swapping domains and reprogramming specificity, in every case guided by a structure.… · 4 papers

Nine of the ten papers here cite the same 2023 review, NLR receptors in plant immunity: making sense of the alphabet soup (10.15252/embr.202357495) — the highest overlap among 424 works these papers reference between them. It locates the corpus: a set of answers to questions that review left open. Three years on, several of its framing assumptions have not survived.

Oligomeric state: no common principles

Until recently the plant resistosome came in two sizes: a pentamer for ZAR1 and Sr35, a hexamer for the NRC helpers. Both numbers now look like samples rather than rules. An autoactive wheat CCG10-NLR assembles into an octamer, and a second member of that clade from Arabidopsis does the same, so the architecture travels across the monocot-dicot split (guo-2026-ccg10-resistosome). Screening 637 NRC proteins by a structural novelty score flagged the NRC7 clade as anomalous, and purified NRC7 from three species forms an eleven-protomer ring roughly 250 A across, against the 150 A a hexamer would give (toghani-2026-sni-undecamer). Between those extremes sit partial assemblies: an effector that stalls tomato NRC3 at three protomers, and, when truncated, at four, which is what shows the hexamer builds one subunit at a time (seager-2026-avrcap1b-nrc3).

The 2022 structure that established the Sr35 pentamer is titled A wheat resistosome defines common principles of immune receptor channels (10.1038/s41586-022-05231-w). Five of the ten papers here cite it. When it was written, CC-type plant resistosomes came in one size: pentamers. NRC hexamers followed in 2024 (10.1126/sciadv.adr2594, 10.1016/j.cell.2024.07.013), then an octamer, an undecamer, and a trimer caught mid-assembly. Two structures of one architecture had been enough to license the word principles. The structures were not wrong. The generalisation was premature, which in a protein family this large and this old is the more common error: NLRs have been diversifying since before plants evolved flowers, and a sample of two is a sample of two.

Two independent attempts to predict stoichiometry from a monomer appear here, one using the HD1-WHD hinge angle (guo-2026-ccg10-resistosome), one a composite structural index (toghani-2026-sni-undecamer). Neither is validated beyond the cases that motivated it. Structures of the resting states of plant NLR remain limited.

Where a resistosome acts is also variable

The standing model has activated receptors inserting into the plasma membrane and drawing calcium in from outside. NRG1 does not. It accumulates on the chloroplast outer envelope and drains stromal calcium into the cytosol, and the specificity is functional rather than incidental: trapping NRC4 at the chloroplast with a nanobody abolishes its cell-death activity, while NRG1 trapped in the same place works normally (ibrahim-2026-nrg1-chloroplast). Five CCR-NLRs spanning roughly 360 million years, back to a fern, all target chloroplasts. Calcium remains the shared output across every functional paper here (guo-2026-ccg10-resistosome, ibrahim-2026-nrg1-chloroplast, gao-2026-vnt1-splicing); the compartment it comes from does not.

Sensor-helper communication, reached from opposite directions

Two papers converge on activation-and-release from incompatible evidence. One finds no Rx sensor in any purified NRC3 complex, by two methods, and concludes the engagement must be transient (seager-2026-avrcap1b-nrc3). The other predicts the Rx-NRC2 interface that transience requires and tests it by reciprocal charge swap: each mutation dead alone, both together restoring cell death (toghani-2026-sensor-helper-interface). The same binding mode appears across five asterid orders, and two substitutions move a lettuce sensor onto a helper it could not previously use — a boundary that phylogeny and selection analysis had independently placed there (pai-2026-lettuce-nrc).

Control before activation

Three papers converge on the same point from unrelated systems: what a receptor can do matters less than what stops it doing so. A potato late blight receptor is transcribed in two isoforms, and only the spliced form assembles; surface immune signalling shifts the balance, so pattern recognition licenses the intracellular receptor before any effector arrives (gao-2026-vnt1-splicing). Autoactivity elsewhere is engineered rather than regulatory, through MHD or LRR substitutions (guo-2026-ccg10-resistosome, toghani-2026-sni-undecamer), and interface mutants that become insensitive to effector suppression often turn "trigger-happy" instead (seager-2026-avrcap1b-nrc3). At organ scale, legume NLRs are more than twice as tissue-specific as genes generally, and lean towards roots while the transcriptome leans towards leaves (marques-2026-legume-tissue).

Effectors, from both sides

An oomycete effector binds a host trafficking protein through one end and an activated helper NLR through another, needing both to suppress immunity (madhuprakash-2026-avrcap1b-tol9a); its companion shows what the receptor-side contact does (seager-2026-avrcap1b-nrc3). Running the other way, a barley receptor detects a blast effector by copying the binding surface of the host protein that effector normally targets — the inverse of an integrated decoy, and transferable: moving that interface into a rye receptor produced dual specificity and resistance to two pathogens (gomezdelacruz-2026-mla3-mimicry).

Where they disagree

NRG1 is the sharpest case. One paper groups it with EDVID-lacking receptors and predicts a ZAR1-like pentamer from hinge geometry, using it as the control that works in human cells where the wheat receptor does not (guo-2026-ccg10-resistosome); the other puts it at the chloroplast envelope (ibrahim-2026-nrg1-chloroplast). Both model it as a pentamer, from different evidence. Separately, TIR-NLRs in legumes are predominantly root-expressed while the CCR-NLR helpers they signal through sit in leaves, which is difficult to reconcile with a simple dependency and is left unexplained (marques-2026-legume-tissue).

What none of them resolves

No experimental structure of any sensor-helper complex exists. The N-terminal alpha1 helix, the part that actually enters the membrane, is unresolved in every structure here. The plant-specific factors that NRC4 and the CCG10 receptors need in order to conduct calcium remain unidentified (guo-2026-ccg10-resistosome). And confidence in predicted models varies by almost a factor of two across these papers without any shared standard for what a given score licenses.

The alphabet soup has more letters in it than it did.

The papers

Ordered to follow the topics above, not by date or venue.

Sources

  • Corpus overview: data/pages/overview.md
  • Paper metadata: data/manifest.json
  • Shared references: data/citation-graph.json, retrieved from OpenAlex
  • Site generated by scripts/build_site.py