A helper NLR channels organellar calcium to trigger plant immunity
The helper NLR NRG1 works at the chloroplast envelope rather than the plasma membrane, draining stromal calcium into the cytosol through a coiled-coil funnel long enough to span a double membrane.
The standing model has activated NLRs inserting into the plasma membrane and pulling calcium in from outside the cell. NbNRG1 does something else. On activation by XopQ it forms puncta on the chloroplast outer envelope, and stromal calcium falls as those puncta appear. Time-lapse imaging ties the two events together in the same chloroplast. AlphaFold 3 models put the NRG1 coiled-coil funnel about 50 A longer than those of ZAR1, Sr35 or NRC4, close to the thickness of the chloroplast outer envelope. The specificity is real, not incidental: trapping NRC4 at the chloroplast with a nanobody kills its cell-death activity, while NRG1 trapped there works fine and still drains the stroma. Five CCR-NLRs spanning roughly 360 million years, back to a fern, all target chloroplasts.
Key findings (10)
Activated NbNRG1 accumulates at multiple subcellular membranes, including the ER, mitochondria, Golgi and chloroplast, whereas the canonical CC-NLR helper NbNRC4 (D478V/L9E autoactive) accumulates exclusively at the plasma membrane under the same conditions.
Evidence confocal colocalisation with organelle markers (SP:RFP:HDEL, ScCOX4, GmMan1, RFP:Remorin1.3)
Source Figures 1A-1F; figure S4
Effector-dependent NRG1 activation and chloroplast localisation require the canonical EDS1-SAG101 signalling module. XopQ induced puncta around chloroplasts in nrg1/adr1 knockout plants but not in eds1/pad4/sag101a/sag101b (epss) plants.
Evidence confocal imaging in N. benthamiana knockout backgrounds
Source Figure S3
Activated NbNRG1 sits on the chloroplast outer envelope, not in the stroma. NbNRG1 puncta encircle the stromal marker CTP1:RFP and colocalise with the outer-membrane marker TOC64:GFP; puncta also appear on stromules. The punctate distribution was reproduced on chloroplasts purified from infiltrated tissue.
Evidence confocal colocalisation; line intensity profiles; imaging of purified chloroplasts
Source Figures 1G and 1H; figure S6
NbNRG1 drives calcium out of the chloroplast stroma. Chloroplasts bearing NbNRG1:mScarlet3 puncta showed significantly reduced stromal GCaMP6s signal against an empty-vector control (Wilcoxon, W = 5511, P < 0.005), while the pore-deficient NbNRG1delta20 did not (W = 3446, P = 0.208). On XopQ activation the cytosolic-to-stromal sensor ratio (RCaMP1h:GCaMP6s) rose sharply (W = 2, P < 0.005).
Evidence dual genetically encoded calcium sensors (RBCS1A-GCaMP6s stromal, RCaMP1h cytosolic); 90 and 80 chloroplasts quantified
Source Figures 2A-2D; figures S8, S9
Time-lapse imaging over 60 minutes links the two events directly: as NbNRG1:mScarlet3 shifts from diffuse cytoplasmic to punctate around a chloroplast, stromal GCaMP6s fluorescence declines in that same chloroplast.
Evidence 1-hour live-cell time series, replicated in an independent experiment
Source Figure 2E; figure S11; raw data S2
AlphaFold 3 models give CCR-NLR resistosomes a markedly elongated CC funnel. The NbNRG1 funnel extends roughly 50 A beyond those of ZAR1, Sr35 and NRC4 (CC lengths ~117 A for NbNRG1 and ~99 A for NbADR1, against ~66-69 A for the canonical CC-NLRs), comparable to the ~50-60 A thickness of the chloroplast outer envelope.
Evidence AlphaFold 3 oligomer prediction; NbNRG1 and NbADR1 modelled as pentamers on pTM/ipTM grounds
Source Figure 3A; figures S13-S15; data S13
Membrane specificity is functional, not incidental. Nanobody trapping at the chloroplast outer envelope (CP:LaG-24) abolished the HR activity of NbNRC4D478V:GFP but left NbNRG1:GFP activity intact (NRG1 t test P = 0.7089, n.s.; NRC4 Wilcoxon W = 7, P < 0.005). Chloroplast-trapped NbNRG1 depleted stromal calcium; trapped NbNRC4 did not. Trapping either protein at the plasma membrane affected neither.
Evidence GFP-binding nanobody relocalisation (CHUP1-anchored CP:LaG-24; Lti6b:LaG-24 and PIP2a:LaG-24 for plasma membrane); HR quantification at 800 nm across 31 leaves
Source Figures 3B-3D; figures S16-S18; raw data S3
The extended fourth alpha helix directs the targeting. The NbNRG1 K126E/L129E double mutant was fully retained at the plasma membrane with markedly reduced HR, while each single mutant retained partial organelle localisation and near-wild-type HR. Both residues face the outside of the funnel and do not alter predicted pore architecture, and both are distal to the EDS1-SAG101 interface, so upstream activation is likely preserved.
Evidence site-directed mutagenesis; confocal localisation; HR assay; immunoblot
Source Figures 3E and 3F; figure S19
Chloroplast targeting is conserved across CCR-NLRs spanning roughly 360 million years of land-plant divergence. Five CCR-NLRs, from Daucus carota, Vicia faba, Amborella trichopoda (two) and the fern Ceratopteris richardii, all formed discrete puncta on the chloroplast outer membrane and all had extended CC domains relative to NbNRC4.
Evidence phylogenetic analysis of 56,280 CCR-NLR NB-ARC sequences; heterologous expression and confocal imaging of five orthologs
Source Figures 4A and 4B; figure S20
CCR-NLRs lack the MADA motif shared by canonical CC-NLRs and instead carry several distinct N-terminal motifs, with an initial split into two CCR-NLR groups in the monilophytes and gymnosperms, and ADR1- and NRG1-specific motifs appearing during angiosperm divergence.
Evidence N-terminal motif search across land-plant CCR-NLRs with canonical CC-NLRs as outgroup
Source Figure S12; data S4-S12
NLRs
| Name | Class | Role in this paper |
|---|---|---|
| NRG1 N requirement gene 1, NbNRG1, AtNRG1.1, NbNRG1delta14, NbNRG1delta20 | helper subject | central subject; shown to target the chloroplast envelope and channel stromal calcium into the cytosol |
| NRC4 NbNRC4, NbNRC4D478V, NbNRC4D478V/L9E | helper comparator | plasma-membrane-exclusive counterpart; loses HR activity when trapped at the chloroplast, establishing that membrane specificity is functional |
| ADR1 NbADR1 | helper comparator | second CCR-NLR subclass; modelled alongside NRG1 and shown to have an intermediate extended funnel |
| ZAR1 AtZAR1, NbZAR1 | singleton comparator | canonical pentameric reference for CC funnel length; outgroup in the CCR-NLR phylogeny |
| Sr35 TmSr35 | singleton comparator | canonical reference for CC funnel length |
| Roq1 recognition of XopQ 1 | sensor cited-only | the sensor upstream of XopQ recognition that activates the EDS1-SAG101-NRG1 module |
| DcCCR-NLR | helper subject | one of five orthologs tested for conserved chloroplast targeting |
| VfCCR-NLR | helper subject | ortholog tested for conserved chloroplast targeting |
| AmCCR-NLR1/2 | helper subject | basal angiosperm orthologs tested for conserved chloroplast targeting |
| CrCCR-NLR | helper subject | fern ortholog; anchors the ~360-million-year conservation claim |
Effectors
| Name | Pathogen | Role |
|---|---|---|
| XopQ | Xanthomonas spp. | the activating trigger throughout; used to switch NbNRG1 from cytoplasmic to chloroplast-punctate and to drive stromal calcium efflux |
Structures
| Entry | State | Description |
|---|---|---|
| AlphaFold3 predicted | pentamer (predicted) | NbNRG1 and NbADR1 CCR-NLR resistosomes modelled as pentamers (chosen over hexamers on pTM/ipTM grounds); NbNRG1 ipTM 0.43, pTM 0.42; NbADR1 ipTM 0.40, pTM 0.44 |
| AlphaFold3 predicted | pentamer / hexamer (predicted) | Comparator CC-NLR resistosome models: AtZAR1 (ipTM 0.68, pTM 0.70), TmSr35 (ipTM 0.71, pTM 0.73), NbNRC4 (ipTM 0.67, pTM 0.69) |
| AlphaFold3 predicted | monomer and oligomer | NbNRG1delta14 and NbNRG1delta20 truncations, and resting-state monomer models showing the extended fourth alpha helix |
Open questions
- Whether Arabidopsis NRG1 also reaches organellar membranes is untested. Prior work reported plasma-membrane NRG1 resistosomes, but used autoactive AtNRG1 alleles expressed in N. benthamiana rather than native NRG1 in Arabidopsis.
- No experimental structure of an activated CCR-NLR resistosome exists; the elongated funnel is an AlphaFold 3 prediction, and the NbNRG1 model scores are low (ipTM 0.43).
- Whether NRG1 conducts calcium across both chloroplast envelope membranes, or only the outer one, is not resolved.
- The functional consequence of NRG1 targeting mitochondria, ER and Golgi is untested; only the chloroplast was followed up.
- Whether multicompartment targeting is an evolved counter-strategy against effectors that suppress plasma-membrane NLR function is proposed, not tested.
Connections in this corpus
| Paper | Why |
|---|---|
guo-2026-ccg10-resistosome | Direct tension over NRG1. Guo groups NRG1 with EDVID-lacking receptors adopting a WAI3-like CC arrangement, and predicts a pentamer for it from the HD1-WHD angle; Ibrahim independently models it as a pentamer but shows it acting at the chloroplast rather than the plasma membrane. Guo also uses NRG1.1DV as the positive control that works in HEK293T where WAI3 does not, which sits oddly beside NRG1's organellar role here. |
gao-2026-vnt1-splicing | Six co-authors in common (Ibrahim, Toghani, Yuen, Eilmann, King, Wang) and a shared Bozkurt-lab background. Both papers concern how and where NLR activation is controlled rather than how effectors are recognised. |
toghani-2026-sni-undecamer | Both use AlphaFold 3 to reason about resistosome architecture where no experimental structure exists, and both accept predicted oligomeric state as a working hypothesis to test biochemically. |
toghani-2026-sensor-helper-interface | Shared reliance on AlphaFold 3 for NLR complexes that resist experimental capture; both papers then validate predictions by targeted mutagenesis rather than by structure determination. |
seager-2026-avrcap1b-nrc3 | Both address helper NLR function in N. benthamiana; Seager works on how an effector blocks NRC assembly at the plasma membrane, which is the compartment Ibrahim shows NRG1 avoiding. Together they bear on whether compartment choice is a route around effector suppression. |
madhuprakash-2026-avrcap1b-tol9a | Madhuprakash shows an effector hijacking an activated helper NLR pathway; Ibrahim proposes organellar targeting as a possible counter to effectors that suppress plasma-membrane NLR function. Same problem from opposite sides. |
pai-2026-lettuce-nrc | Both use large-scale NLR phylogenetics to place a helper clade in evolutionary context, this paper across 56,280 CCR-NLRs in land plants and Pai across the NRC network in asterids. Pai counts only 9 CCR-NLRs in the lettuce NLRome, a useful data point on how differently the two helper systems have expanded within one genome. |
marques-2026-legume-tissue | Both survey the CCR-NLR clade at scale and reach complementary kinds of conclusion. This paper places NRG1 across 56,280 CCR-NLRs and shows what it does subcellularly; Marques shows where in the plant CCR-NLRs are expressed, finding the class leaf-biased overall with every root-specific member belonging to the NRG1 subfamily. Compartmentalisation at two scales, organelle and organ. |
gomezdelacruz-2026-mla3-mimicry | Both use AlphaFold modelling to explain a functional property and then test it by targeted substitution, this paper on membrane targeting via the extended fourth helix, Gomez De La Cruz on effector binding via three C-terminal LRR residues. They differ sharply in model confidence (ipTM 0.43 here against pTM 0.859 there) and in AlphaFold version, so they are a useful contrast on evidential weight. |
Provenance and caveats
CC domain lengths (~117 A NbNRG1, ~99 A NbADR1, ~66-69 A canonical) are read from the Figure 3A annotation. The ~50 A difference quoted in the Discussion is the paper's own rounding of NbNRG1 against NbNRC4.
The AlphaFold 3 model underpinning the elongated-funnel claim has ipTM 0.43 and pTM 0.42, which is low. The paper is explicit that no experimental CCR-NLR resistosome structure exists. Any wiki statement about the NRG1 funnel must carry that caveat.
Reference numbering in in_corpus_or_lab_lineage is taken from the printed Science reference list where legible; two entries could not be assigned a number from the two-column extraction and are recorded with n: null. [UNVERIFIED - confirm at Stage 3]
Sources
- Full citation: Ibrahim T, King FJ, Toghani A, Wang L, Jenkins S, Yuen ELH, Wang H-Y, Vuolo C, Eilmann N, Adamkova V, Chia K-S, Castel B, Jones JDG, Carella P, Wu C-H, Kourelis J, Kamoun S, Bozkurt TO (2026) A helper NLR channels organellar calcium to trigger plant immunity. Science 392:499-505.
- DOI: 10.1126/science.aeb6690
- PDF on file:
papers/Science_2026.pdf - Extraction:
data/papers/ibrahim-2026-nrg1-chloroplast.json(schema v1.0) - References: 79 works, retrieved from OpenAlex
- Licence: AAM available under CC BY (cOAlition S)