Rapid local and systemic jasmonate signalling drives the initiation and establishment of plant systemic immunity

Abstract

Successful recognition of pathogen effectors by plant disease resistance proteins, or effector-triggered immunity (ETI), contains the invading pathogen through localized hypersensitive cell death. ETI also activates long-range signalling to establish broad-spectrum systemic acquired resistance (SAR). Here we describe a sensitive luciferase (LUC) reporter that captures the spatial–temporal dynamics of SAR signal generation, propagation and establishment in systemic responding leaves following ETI. JASMONATE-INDUCED SYSTEMIC SIGNAL 1 (JISS1) encodes an endoplasmic-reticulum-localized protein of unknown function. JISS1::LUC captured very early ETI-elicited SAR signalling, which surprisingly was not affected by classical SAR mutants but was dependent on calcium and was also wound responsive. Both jasmonate biosynthesis and perception mutants abolished JISS1::LUC signalling and SAR to Pseudomonas syringae. Furthermore, we discovered that ETI initiated jasmonate-dependent systemic surface electrical potentials. These surface potentials were dependent on both glutamate receptors and JISS1, despite neither JISS1 loss-of-function nor glutamate receptor mutants altering SAR to Pseudomonas syringae. We thus demonstrate that jasmonate signalling, usually associated with antagonism of defence against biotrophs, is crucial to the rapid initiation and establishment of SAR systemic defence responses (including the activation of systemic surface potentials) and that JISS1::LUC serves as a reporter to further dissect these pathways.

A novel reporter captured spatial temporal dynamics of effector-triggered-immunity (ETI)-induced systemic immunity, revealing that signal propagation and establishment in systemic acquired resistance depend on jasmonates. Furthermore, ETI initiates jasmonate-dependent systemic induced surface electrical potentials.

Main

Despite the discovery of plant systemic acquired resistance (SAR) over a century ago, our knowledge of the signalling processes underlying the establishment, propagation and especially initiation of this response remains fragmentary. Classically, SAR is established following effector-triggered immunity (ETI) leading to the hypersensitive response (HR). SAR has also been reported to be activated via pathogen-associated molecular pattern recognition and virulent bacterial phytopathogens, although the latter has also been reported to trigger systemic induced susceptibility^1,2.

Multiple molecules are implicated in SAR induction, including salicylic acid (SA) and its volatile derivative methyl salicylate, azelaic acid (AZA), glycerol-3-phosphate, dehydroabietinal, pipecolic acid (Pip) and N-hydroxy-pipecolic acid (NHP). More recently, extracellular NAD(P), the volatile monoterpenes α- and β-pinene, vitamin B6 and small RNAs derived from TAS3a were shown to induce SAR^3–6. HR-generated reactive oxygen species (ROS) and nitric oxide (NO) are integral to ETI-initiated SAR, most likely via C₁₈ unsaturated fatty acid oxidation of chloroplast lipids^2,7. Hydrolysis of C₁₈ fatty acids released from thylakoid membrane monogalactosyldiacylglycerol and digalactosyldiacylglycerol generates AZA^8,9. The importance of lipid signalling in SAR is highlighted by the involvement of lipid transfer proteins, AZELAIC ACID INDUCED1 (AZI1) and DEFECTIVE IN INDUCED RESISTANCE1 (DIR1)⁸. Plants defective in SA, glycerol-3-phosphate, NO or ROS biosynthesis have reduced levels of Pip in distal tissues, reinforcing the complex metabolic interplay in the establishment of SAR¹⁰. Airborne defence cues also activate SAR¹¹; thus, one can conclude that multiple signals translocating apoplastically, symplastically^12,13 and as volatiles can collectively confer broad-spectrum systemic resistance against diverse pathogens, including viral, bacterial, oomycete, fungal and insect pests². The synthesis, activities and interactions of these SAR inducers have been extensively reviewed^10,14–16.

Despite progress in understanding the individual signalling networks leading to SAR, the spatial–temporal dynamics and interactions of various chemical signals in the SAR pathway remain unclear. Recognition of Pseudomonas syringae pv. tomato DC3000 (DC) carrying avrRpm1 (DCavrRpm1)¹⁷ by the Resistance to P. maculicola 1 (RPM1) disease resistance protein provides a robust ETI model to dissect signal generation and transduction dynamics underlying SAR. We previously demonstrated that RPM1 activation triggers early increases in cytosolic calcium, beginning ~1.5–2 h post-infection (hpi)^18,19, followed by lipid-peroxidation-triggered biophoton generation ~3 hpi^20,21 and visible leaf collapse ~6 hpi. RPM1 activation elicits rapid transcriptional reprogramming 4 hpi in systemic leaves, which strongly overlaps with jasmonate-triggered systemic wound responses²². Here we report JASMONATE-INDUCED SYSTEMIC SIGNAL 1 (JISS1), a jasmonate-responsive SAR reporter that captures unexpectedly rapid temporal–spatial dynamics following ETI. We show that SAR requires enzymatic production of a local jasmonate signal that propagates via the vasculature and epidermal cells to systemic leaves and is coupled to calcium- and jasmonate-dependent systemic surface electrical potentials.

Results

JISS1 expression reveals temporal and spatial dynamics of early effector–resistance gene interactions

JISS1 (At5g56980; previously known as A70 (ref. ²²)), a protein of unknown function, is an early SAR marker²². To monitor SAR transcriptional dynamics, we fused the promoter of JISS1 and the sequence encoding the first 84 amino acids of JISS1 to luciferase (Extended Data Fig. 1). Homozygous JISS1 promoter::luciferase (JISS1::LUC) lines showed rapid systemic luciferase activity following challenge with DCavrRpm1, but not with virulent DC; the type-III-secretion-system-deficient DChrpA, which elicits pathogen-associated-molecular-pattern-triggered immunity (PTI) responses; or mock challenge (MgCl₂) (Fig. 1a). SAR signal propagation was remarkably rapid, with strong luciferase activity first evident in the petiole of the challenged leaf ~3 hpi (Fig. 1b), and within 30 min JISS1::LUC activity was established^23,24. This activity spread to adjacent leaves (~4 hpi, Fig. 1b), reaching maximal intensity ~4.5 hpi, ~1 h prior to any visible collapse of the challenged leaf.

Extended Data Fig. 1. Schematic of the JISS1:LUC construct and its T-DNA integration location in the Arabidopsis genome.

(A). The Photinus pyralis LUC2P reporter gene (from Promega pGL4-11) was PCR amplified and cloned into pCAMBIA1302 digested with Kpn1 and Pml1 to generate pC1LUCP with a NOS terminator. A 1651 bp JISS1 fragment containing the promoter and encoding the first 84aa of JISS1 was PCR amplified as described⁶⁴ generating a hygromycin selectable JISS1:LUC construct. (B) Eight independent homozygous lines A. thaliana Col-5 JISS1:LUC expressing transgenic lines were generated and tested for luciferase expression in systemic leaves following an immunising challenge with DCavrRpm1. A homozygous line showing strong JISS1:LUC expression was chosen and the genomic location of the JISS1:LUC T-DNA construct was determined at position 18270147 in the coding region of AT4G39240.1 sequence analysis as per⁷⁰.

Fig. 1. JISS1 expression is induced systemically by ETI.

a, Luciferase activity in JISS1::LUC plants following DCavrRpm1, DC, DChrpA or mock (MgCl₂) challenges at 4:30 hpi. Throughout the figure, white asterisks indicate infiltrated leaves, and red carets indicate leaves used for signal intensity analysis (d). The images are false-coloured by signal intensity, as indicated by individual calibration bars. b, Temporal spatial dynamics of luciferase activity in JISS1::LUC plants following DCavrRpm1 challenge, initiating at 3 hpi. The 3:20 hpi, 3:50 hpi and 4:30 hpi images capture the systemic spread of the signal over time. c, Different Avr genes display temporal specificity in the activation of systemic JISS1::LUC: DCavrRpm1 (4 hpi), DCavrRps4 (13:20 hpi) and DCavrRpt2 (15:20 hpi), compared with the DChrpA control. d, JISS1::LUC signal intensity in leaves adjacent to infiltration (red carets in c) plotted over time (h). e, F_v/F_m is strongly suppressed during ETI following DCavrRpm1 (3 hpi), DCavrRpt2 (10:00 hpi) or DCavrRps4 (16:00 hpi) challenge compared with DChrpA. f, Visualization of F_v/F_m suppression wherein orange indicates F_v/F_m of a healthy leaf (~0.8), green represents a reduction in F_v/F_m as ETI progresses (~0.6) and blue represents a strong impact of ETI on F_v/F_m (~0.3). Representative images of over ten repeats are shown.

Source data

Challenge with DCavrRpt2 or DCavrRps4 also induced systemic luciferase activity following recognition by Resistance to P. syringae 2 (RPS2)²⁵ and RPS4 (ref. ²⁶), respectively (Fig. 1c). The spatial pattern of systemic luciferase reporter activity was identical for all ETI responses, but initiation timing differed for each resistance (R) protein, consistent with rapid translocation of an ETI-induced signal (Fig. 1c,d and Supplementary Video 1). To understand the temporal context of R protein elicitation, we investigated local ETI dynamics using two non-destructive physiological readouts, chlorophyll fluorescence and biophoton generation. The chloroplast senses and responds to biotic stress, best exemplified by decreases in the quantum efficiency of photosystem II (F_v/F_m) following ETI elicitation²⁷. Quantitative and spatial F_v/F_m parameters were determined following challenge with DCavrRpm1, DCavrRpt2, DCavrRps4 or the DChrpA control (Fig. 1e,f and Extended Data Fig. 2a,b). Biophotons, generated from chloroplast lipid peroxidation²¹ and associated with the initiation of the HR²⁰, were additionally assayed (Extended Data Fig. 2c and Supplementary Video 2). Systemic signal initiation elicited by these three R proteins is preceded in the local challenged leaf initially by strong suppression of F_v/F_m and subsequently by biophoton generation. Although DCavrRpt2 challenge led to earlier suppression of F_v/F_m and biophoton generation than DCavrRps4, JISS1::LUC activation was faster following DCavrRps4 challenge. These data imply that SAR signal generation is not definitively linked to biophoton generation and F_v/F_m suppression.

Extended Data Fig. 2. DCavrRpt2 F_v/F_m dynamics and biophoton generation is earlier than DCavrRps4 although its systemic JISS1:LUC expression is considerably later.

(A, B) Visualisation (A) and quantitation (B) of F_v/F_m suppression within a plant over infection time showing, like with biophoton generation (C), ETI elicited by DCavrRpt2 affects chloroplast physiology (PSII) faster than DCavrRps4 challenge. In (A) orange represents a healthy PSII signature whereas blue signifies strongly disrupted PSII due to ETI induced ROS. (C) The sequence of biophoton generation, matches F_v/F_m suppression, during ETI elicited by DCavrRpm1, DCavrRpt2 or DCavrRps4 challenge (times indicated). Stills taken from Supplementary Video 2.

Jasmonates are involved in ETI-induced systemic signalling

To examine the regulation of JISS1 in SAR, we crossed the JISS1::LUC reporter into classical SAR mutant lines expediated by identifying the JISS1::LUC transfer DNA (T-DNA) insertion position (Extended Data Fig. 1b): npr1, where NONEXPRESSOR OF PATHOGENESIS-RELATED 1 (NPR1) encodes a repressor of ETI but is important for SAR^28–30; npr1 npr3 npr4 (ref. ²⁸), impaired in SA signal transduction (notably, NPR3/4 is required for full RPS2 ETI)³¹; nac19 nac55 nac72, altered in the regulation of SA accumulation³²; and SA INDUCTION DEFICIENT 2 (sid2), deficient in the accumulation of SA³³. Surprisingly, when challenged with DCavrRpm1, all SAR-compromised lines showed wild-type reporter dynamics (Fig. 2a and Extended Data Fig. 3a). Furthermore, SAR elicitors, AZA (or nonanoic acid (NA), its precursor), Pip and NHP did not significantly increase local JISS1–LUC activity (Fig. 2b), and JISS1–LUC activity in the NHP biosynthetic FLAVIN-DEPENDENT MONOOXYGENASE 1 (fmo1) mutant³⁴ was wild-type-like in response to DCavrRpm1 (Fig. 2c). Together, these data indicate that the signal inducing JISS1 occurs upstream of or in parallel to previously characterized SAR elicitors.

Fig. 2. JISS1::LUC is activated by the jasmonate signalling pathway but not classical SAR elicitors.

a, The classical SAR mutants sid2, nac19 nac55 nac72 and npr1 npr3 npr4 do not alter JISS1::LUC luciferase signatures following DCavrRpm1 challenge (4 hpi). Throughout the figure, white asterisks indicate infiltrated leaves, and the images are false-coloured by signal intensity, as indicated by individual calibration bars. b, Infiltration of 1 mM Pip, NHP, NA or AZA into leaves of JISS1::LUC plants does not activate luciferase signal (1 hpi). c, No attenuation of luciferase activity is observed in a JISS1::LUC fmo1 mutant line following DCavrRpm1 challenge (4:30 hpi). d, JA (250 µM), but not SA (1 mM), ABA (1 mM) or mock (10 mM MgCl₂) challenge induces local JISS1::LUC signal propagation (1 hpi). e, Luciferase activity in JISS1::LUC leaves infiltrated with JA (250 µM) or COR (500 nM) (2 hpi). f, JISS1::LUC activity following challenge with DC or the COR-deficient DC mutant DB4G3 (cor⁻) with or without avrRpm1. DCavrRpm1 cor⁻ induced comparable JISS1::LUC systemic activity to DCavrRpm1, whereas no luciferase activity was detected in DCcor⁻ challenged leaves (3:50 hpi). Representative images of at least three repeats are shown.

Extended Data Fig. 3. JISS1 signal propagation is not impaired in npr1, but responds to wounding and the JA elicitor coronatine.

A white asterisk denotes treated leaves and images are false coloured by signal intensity, indicated by the calibration bar. (A) A JISS1:LUC/npr1 line exhibited comparable systemic luciferase activity to the wild-type Col-0 JISS1:LUC line following DCavrRpm1 challenge (4:20 hpi). (B, C) Strong and transient luciferase activity was observed in response to 250 nM coronatine (COR), a milder response was triggered by wounding and no response was detected following mock treatment (10 mM MgCl₂) 2 hpi.

We next tested key immunity-associated phytohormones for JISS1 induction. Even at high concentrations, leaves infiltrated with SA (1 mM) or abscisic acid (ABA; 1 mM) failed to elicit JISS1–LUC activity. However, jasmonic acid (JA; 250 µM), a key elicitor of systemic wound signalling, induced luciferase locally (Fig. 2d). Collectively, jasmonates comprise JA and its derivatives, with bioactive jasmonoyl-isoleucine (JA-Ile) binding the Skp/Cullin/F-box SCF^COI1 (CORONATINE-INSENSITIVE PROTEIN 1)–JAZ1 E3 ubiquitin ligase jasmonate co-receptor complex^35,36. DC does produce the highly active JA-Ile mimic coronatine (COR), but not until ~10 hpi^36,37. Strikingly, 500 nM COR strongly induced JISS1–LUC locally within 2 h (relative to MgCl₂ or wounding; Fig. 2e and Extended Data Fig. 3b,c). JISS1–LUC was also induced in plants inoculated with both DC and the DCcor⁻ mutant³⁸ bacteria expressing avrRpm1 (Fig. 2f), excluding COR as the systemic elicitor.

ETI elicits a rapid and propagative jasmonate-dependent signal essential for effective SAR

We reported significant transcriptional overlap between the JA-regulated wounding response and the SAR response at 4 hpi with DCavrRpm1 compared with mock challenges²². Re-examination of these data identified nine JASMONATE ZIM DOMAIN (JAZ) family members (including JAZ10), which act to reimpose the repression of jasmonate signalling^36,39, induced systemically by DCavrRpm1 (Supplementary Table 2). As for JISS1, the wound-responsive JAZ10–GUS reporter⁴⁰ is also activated in systemic leaves challenged with DCavrRpm1, but not DC or DChrpA (Fig. 3a). This systemic JAZ10–GUS expression is abolished in the jasmonate receptor mutant coi1-16 (Fig. 3a)⁴¹. Following DCavrRpm1 challenge, JISS1–LUC systemic expression was not detected in the JA biosynthetic mutant (ALLENE OXIDE SYNTHASE) (aos)⁴² (Fig. 3b), and JISS1 expression was reduced compared with that in Col-0 (Extended Data Fig. 4a). While local application of JA or COR to JISS1::LUC coi1-16 failed to elicit local or systemic signals (Fig. 3c), it restored signalling in the JISS1::LUC aos plants (Fig. 3d).

Fig. 3. JISS1::LUC signal propagation is dependent on JA biosynthesis and perception.

a, DCavrRpm1 GUS activity in representative systemic leaves of a JAZ10::GUS reporter line at the times indicated compared to a JAZ10::GUS coi1-16 line 8 hpi (n = 4 biological replicates). Throughout the figure, white asterisks indicate infiltrated leaves, and red carets indicate leaves used for signal intensity analysis. The images are false-coloured by signal intensity, as indicated by individual calibration bars. b, Luciferase activity is not induced in DCavrRpm1-treated JISS1::LUC aos or JISS1::LUC coi1-16 mutant lines (5 hpi). c, Luciferase activity is absent in a JISS1::LUC coi1-16 mutant but not JISS1::LUC leaves 1 hpi following infiltration with 250 µM JA or 500 nM COR (a JA-Ile mimic). d, Luciferase activity is restored in JISS1::LUC aos treated with 250 µM JA or 500 nM COR 1 hpi. e, Schematic of the JA biosynthetic pathway highlighting the positions of jasmonate inhibitor activity. 13-HPOT, 13-hydroperoxy-9,11,15-octadecatrienoic acid. f–h, Treatment with phenidone (Phen) (2 mM; 4:40 hpi) (f), DIECA (2.5 mM; 5:30 hpi) (g) or jarin-1 (25 µM; 5:10 hpi) (h) inhibits DCavrRpm1-induced JISS1::LUC activity. In f–h, the top left image shows the inhibitor pre-infiltrated into the right-hand systemic leaf immediately adjacent to the local challenged leaf, the top right image shows the challenged leaf co-infiltrated with the inhibitor, the bottom right image shows the petiole of the immunized leaf treated with the inhibitor and the bottom left image shows the leaves with no inhibitor. i, JISS1::LUC signal intensity in two leaves adjacent to DCavrRpm1 infiltration, one treated with inhibitor and one without inhibitor (red carets in f–h) plotted over time (min). j, SAR growth curve of DC following DCavrRpm1 or mock immunizing challenge on Col-0 (*P = 0.0219, t = 3.068, d.f. = 6 (unpaired two-tailed t-test)), jiss1 (***P < 0.0001, t = 9.162, d.f. = 6) and aos (not significant (NS), P = 0.1415, t = 1.692, d.f. = 6). (n = 4 biological replicates). Box plots represent the minimum, maximum and median values for the statistical analysis of each treatment. Col-0 mock: min = 5.47, max = 5.88, median = 5.64, Q1 = 5.53, Q3 = 5.76; Col-0 DCavrRpm1: min = 4.20, max = 5.29, median = 4.85, Q1 = 4.43, Q3 = 5.22; jiss1 mock: min = 5.26, max = 5.73, median = 5.55, Q1 = 5.45, Q3 = 5.63; jiss1 DCavrRpm1: min = 3.64 max = 4.22, median = 3.89, Q1 = 3.67, Q3 = 4.13; aos mock: min = 5.28, max = 5.44, median = 5.32, Q1 = 5.31, Q3 = 5.35; aos DCavrRpm1: min = 4.81, max = 5.45, median = 5.05, Q1 = 4.88, Q3 = 5.26. The images are representative of at least three repeats.

Extended Data Fig. 4. HR development is not affected in mutants with impaired JA biosynthesis (aos) or perception (coi1-16) but coi1-16 has reduced disease susceptibility.

(A) RT-PCR verification loss of JISS1 expression in the aos T-DNA insertion line. Actin2 expression was used as a loading control (B) Phenotyping of Col-0 and aos and coi1-16 mutants showing that loss of JA biosynthesis and signalling does not affect HR development. (C) SAR growth curve of Psm4 following DCavrRpm1 or mock pre-treatment on Col-0 (**** p = <0.0001, t = 15.36, df=10 [unpaired t test]) and the coi1-16 mutant (p = 0.9421, t = 0.07450, df=10). coi1-16 is more resistant to Psm4 than wild-type Col-0 (*** p = 0.0004, t = 5.154, df=10) (n = 6 biological replicates). Col-0 mock: min=7.88, max=7.96, median=7.92, Q1 = 7.91, Q3 = 7.94; Col-0 DCavrRpm1: min=7.15, max=7.42, median=7.22, Q1 = 7.19, Q3 = 7.31; coi1-16 mock: min=7.03, max=7.71, median=7.28, Q1 = 7.07, Q3 = 7.53; coi1-16 DCavrRpm1: min=7.19, max=7.36, median=7.33, Q1 = 7.31, Q3 = 7.35.

Source data

We next determined whether the HR was impacted in aos or coi1-16 mutants. Consistent with biophoton and chlorophyll fluorescence data, collapse of the wild-type Col-0 leaf occurred first with DCavrRpm1 (~5.5 hpi), followed by DCavrRpt2 (~14 hpi) and then DCavrRps4 (~18 hpi). Crucially, timing of the HR was comparable between Col-0 and the aos and coi1-16 mutants (Extended Data Fig. 4b), indicating that the loss of jasmonate biosynthesis or perception does not attenuate local HR but does attenuate JISS1–LUC local and systemic signalling and abolishes SAR. We thus concluded that systemic signalling is both jasmonate and HR dependent.

Aside from JA-Ile, other jasmonate molecules have biological activity in planta⁴³, including 12-oxophytodienoic acid (OPDA), although OPDA is not perceived by the SCF^COI1 receptor complex⁴⁴. We tested the JA biosynthetic inhibitors phenidone, a lipoxygenase inhibitor⁴⁵; diethyldithiocarbamic acid (DIECA), which interferes with octadecanoid signalling⁴⁶; and jarin-1, an inhibitor of JA-Ile synthetase⁴⁷ (Fig. 3e). Concurrent with DCavrRpm1 challenge, treatment with phenidone (2 mM), DIECA (2.5 mM) or jarin (25 µM), via co-infiltration, petiole application or infiltration into the adjacent systemic responding leaves, abolished or markedly attenuated DCavrRpm1-elicited JISS1–LUC activity (Fig. 3f–i). These data independently reinforce a role for jasmonates in SAR, in both local signal generation and establishment in naive responding leaves, and specifically implicate JA-Ile biosynthesis as being necessary for both effective signal transduction and the establishment of SAR.

The coi1 mutant is known to be more resistant to local DC challenge than wild-type Arabidopsis⁴¹. We assessed SAR to different P. syringae pathovars (less virulent P. syringae pv. maculicola race 4 (Psm4) or more virulent DC) in the aos and coi1-16 mutants alongside a JISS1 T-DNA insertion loss-of-function line (jiss1; Fig. 3j and Extended Data Figs. 4c and 5a,b). Crucially, neither coi1-16 nor aos exhibited SAR relative to Col-0 across multiple independent assays (n > 4). Interestingly, jiss1 lines still induced SAR following DCavrRpm1 challenge (Fig. 3j and Extended Data Fig. 5c).

Extended Data Fig. 5. jiss1 mutants do not impact SAR and the JISS1 protein is not required for systemic signalling.

(A) RT-PCR verification loss of JISS1 expression in the jiss1 T-DNA insertion line. Actin2 expression was used as a loading control. (B) T-DNA insertion in the coding sequence of the JISS1 gene (AT5G56980) is located at position Chr5: 23056664–23056777. Relative positions of forward and reverse gene specific RT-PCR primers are indicated, and sequences are given in Methods. (C) SAR growth curve of DC following DCavrRpm1 or mock pre-treatment on wild-type (Col-0) and the jiss1 mutant (n = 4 biological replicates). Replicate 1: Col-0 ** p = 0.0094, t = 3.759, df=6 [unpaired two-tailed t test], jiss1 ** p = 0.0010, t = 5.943, df=6 [unpaired two-tailed t test]. Col-0 mock: min=5.06, max=6.05, median=5.37, Q1 = 5.12, Q3 = 5.71; Col-0 DCavrRpm1: min=3.09, max=4.73, median=3.64, Q1 = 3.18, Q3 = 4.24; jiss 1 mock: min=5.86, max=6.16, median=5.94, Q1 = 5.92, Q3 = 6.00; jiss1 DCavrRpm1: min=4.56, max=5.31, median=4.92, Q1 = 4.72, Q3 = 5.12. Replicate 2: Col-0 ** p = 0.0076, t = 3.947, df=6 [unpaired two-tailed t test], jiss1 ** p = 0.0015, t = 5.498, df=6 [unpaired two-tailed t test]) Col-0 mock: min=4.78, max=5.65, median=5.15, Q1 = 4.94, Q3 = 5.39; Col-0 DCavrRpm1: min=4.02, max=4.66, median=4.16, Q1 = 4.11, Q3 = 4.31; jiss 1 mock: min=5.26, max=5.72, median=5.38, Q1 = 5.27, Q3 = 5.53; jiss1 DCavrRpm1: min=4.02, max=4.81, median=4.28, Q1 = 4.17, Q3 = 4.45. Replicate 3: Col-0 *** p = 0.0001, t = 8.699, df=6 [unpaired two-tailed t test], jiss1 *** p = 0.0002, t = 7.951, df=6 [unpaired two-tailed t test]. Col-0 mock: min=5.97, max=6.42, median=6.35, Q1 = 6.25, Q3 = 6.37; Col-0 DCavrRpm1: min=3.70, max=4.63, median=4.41, Q1 = 4.21, Q3 = 4.49; jiss 1 mock: min=5.64, max=6.40, median=6.28, Q1 = 6.09, Q3 = 6.34; jiss1 DCavrRpm1: min=3.45, max=4.52, median=3.72, Q1 = 3.65, Q3 = 3.93. (D-G) The JISS1:LUC reporter in the jiss1 mutant responds as wild type following challenge with either DCavrRpm1 (D-E) or DCavrRps4 (F-G). A white asterisk denotes infiltrated leaves and images are false coloured by signal intensity, as indicated by calibration bar. Luciferase activity shows similar dynamics and signal intensity in response to DCavrRpm1 (D-E) and DCavrRps4 (F-G) in JISS1:LUC and JISS1:LUC/jiss1 lines.

Source data

These data collectively imply that ETI elicits rapid de novo synthesis of a jasmonate-dependent signal that propagates systemically and is essential for the effective establishment of SAR. JISS1 expression is intimately linked to this signal, but JISS1 itself is not critical for SAR or SAR signalling (Extended Data Fig. 5c–f), and its precise biological function remains unclear.

JISS1 signal localizes to the vasculature and epidermal endoplasmic reticulum

We next investigated JISS1 subcellular localization and expression following ETI elicitation. JISS1 was annotated as chloroplast localized (TAIR10; AtSubP). We thus generated both a full-length (JISS1_pro::JISS1-GFP) and a truncated JISS1–GFP fusion (JISS1_pro::JISS1^1–84-GFP); the latter included the predicted chloroplast transit peptide. In both lines, GFP signal was detected in the systemic leaf within 4.5 hpi, predominantly in both the vasculature and epidermal cells of the petiole and lamina (Fig. 4a,b and Extended Data Fig. 6a). Within the epidermal cells, both versions of the JISS1–GFP fusion proteins surprisingly localized to the endoplasmic reticulum (ER) network, colocalizing with the ER luminal marker, RFP–HDEL (Fig. 4c and Extended Data Fig. 6b). These data are suggestive of a SAR signal moving symplastically and are consistent with the abolition of SAR in plasmodesmata-permeability-restricted Arabidopsis overexpressing PLASMODESMATA-LOCATED PROTEIN 5 (ref. ¹³). As observed with JISS1–LUC activity, GFP expression in systemic leaves was markedly suppressed by pretreatment with the JA biosynthetic inhibitors phenidone and DIECA (Extended Data Fig. 6c).

Fig. 4. JISS1 signal propagates symplastically through the epidermis and vasculature and localizes to the ER.

a–c, Confocal images of global JISS1::JISS1-GFP expression in representative systemic leaves following DCavrRpm1 challenge. Panel a shows JISS1::JISS1^1–84-GFP expression in the petiole of a challenged leaf showing JISS1–GFP signal predominantly in the vasculature and epidermal cells. Ab, abaxial surface; Ad, adaxial surface. Scale bar, 100 µm. Panel b shows a still from Supplementary Video 3 of JISS1(1–84)–GFP highlighting that the GFP signal is predominately restricted to the central vasculature and abaxial epidermal cell layer. Panel c shows representative confocal images of Arabidopsis epidermal leaf cells stably expressing JISS1::JISS1-GFP (green) and the ER luminal marker RFP–HDEL (magenta), which shows that JISS1 localizes strictly to the ER. Scale bars, 10 µm. The images are representative of at least three biological replicates.

Extended Data Fig. 6. JA dependent localisation of JISS1:JISS1^1–84-GFP and JISS1:JISS1-GFP in systemic tissue.

(A) Petiole section from systemic leaf showing propidium iodide staining (magenta) and JISS1:JISS1^1–84-GFP expression (green) with merged image. Scale bar, 200 µm (B) Representative confocal images of JISS1:JISS1-GFP and JISS1:JISS1^1–84-GFP plants following DCavrRpm1 challenge, showing identical subcellular localisation of GFP in ER network of epidermal cells. Scale bars, 5 µm and 10 µm, respectively. (C) Representative confocal images of Arabidopsis epidermal leaf cells of JISS1:JISS1^1–84-GFP systemic leaves pre-treated with JA inhibitors, phenidone (2 mM) and DIECA (2.5 mM) prior to DCavrRpm1 challenge or mock treated. Mock treatment of systemic leaves shows JISS1^1–84-GFP labelling of the entire ER network 4 hpi whereas Systemic leaves infiltrated with phenidone or DIECA exhibit strongly reduced GFP signal in the ER 4 hpi compared to signal across the entire ER network after mock treatment. Scale bar, 10 µm. Images are representative of at least 3 biological replicates.

Systemic electric signal propagation is a general feature of ETI activation

Wound-induced accumulation of JA and JA-Ile in undamaged distal leaves, together with altered expression of jasmonate-responsive genes, is preceded by the rapid generation of wound-activated surface electrical potentials (WASPs), caused by plasma membrane depolarization⁴⁰. WASPs are mediated by glutamate-triggered systemic activation of cytosolic Ca²⁺ signalling via specific vasculature-localized members of the glutamate-like receptor family, GLR3.3 and GLR3.6 (refs. ^40,48). These cation-permeable ion channels also contribute to systemic defence against herbivory⁴⁸. To confirm the link between JISS1 and JA signalling, we assayed the activity of JISS1::LUC in response to wound stimuli. Either crushing with forceps or severing the leaf at the petiole triggered rapid, transient induction of JISS1::LUC (Fig. 5a). Wounding induced a much weaker signal than COR application (Extended Data Fig. 3b,c). Given the similar spatial distribution of WASPs and JISS1–LUC signals, and the transcriptional parallels between wounding and early SAR responses²², we measured leaf surface potentials between electrodes attached to the midrib/petiole junction of a local challenged leaf (‘infiltrated’) and to fully expanded leaves immediately adjacent to (‘adjacent’) and directly opposite (‘distal’) the challenged leaf (Fig. 5b).

Fig. 5. Systemic electric signal propagation is a general feature of ETI activation.

a, White asterisks indicate wounded JISS1::LUC leaves, and the images are false-coloured by signal intensity, as indicated by individual calibration bars. Luciferase activity in severed or crushed leaves (n = 3) initiated and peaked at ~40 min and ~2 h post-wounding, respectively. b, Plant electrophysiology experimental set-up with a cartoon showing spatial sampling and colour coding of leaves for working electrodes (W). The inset illustrates electrode positioning. Ref, reference electrode. c, Leaves challenged by DCavrRpm1 but not DChrpA (red) show an initial depolarization ~2 hpi and subsequent repolarization. From 4 to 7 hpi, SISPs are propagated in the two systemic leaves (blue and green) adjacent to the DCavrRpm1-immunized leaf, with the distal leaf (black) responding later (from ~7 hpi). No SISPs are observed in the DChrpA treatments. The coi1-16 mutant shows depolarization of the DCavrRpm1-challenged leaf but no SISP initiation. d, Dex-induced avrRpm1 expression does not replicate DCavrRpm1-induced SISPs; however, infiltration of DChrpA 1 h after Dex application re-instigates SISPs. The experiments were repeated at least twice (see Extended Data Fig. 7 for further details).

No specific changes in leaf surface potentials were induced in untreated or DChrpA-challenged leaves, whereas DC induced a small depolarization of the local challenged leaf (Fig. 5c and Extended Data Fig. 7a,b). By contrast, challenge with DCavrRpm1 initially induced strong depolarization with an amplitude of about −100 mV over a duration of ~2 h, followed by repolarization (Fig. 5c and Extended Data Fig. 7c). The timing of depolarization strongly correlated with biophoton generation, the suppression of F_v/F_m and JISS1–LUC systemic signal initiation (Fig. 1 and Extended Data Fig. 2). Following repolarization of the challenged leaf, systemic immunity surface potentials (SISPs) were detected in adjacent leaves with maximal depolarization ~7 hpi. SISPs were also detected in distal leaves, but maximal depolarization occurred later (~10 hpi; Fig. 5c and Extended Data Fig. 7c). These SISPs, unlike WASPs and herbivory responses, which travel at speeds in excess of millimetres per second⁴⁹, represented slower variation potentials, which largely mirrored the spatial dynamics of systemic JISS1–LUC activity, albeit with delayed propagation. As both DCavrRpt2 and DCavrRps4 also trigger SISPs (Extended Data Fig. 7d,e) with the timing of initiation largely consistent with the initiation of suppression of F_v/F_m (Fig. 1e,f, and Extended Data Fig. 2a,b), we concluded that SISPs are specifically elicited by ETI.

Extended Data Fig. 7. ETI activation of systemic electric signalling is consistent.

Replicated data displayed consistent with Fig. 5. (A) Col-0 plants do not show development of SISPs in non-treated plants. (B) Col-0 plants do not show development of SISPs following DC challenge. (C-E) Challenge with DCavrRpm1, DCavrRpt2, and DCavrRps4 exhibits depolarisation of the infected leaf from 2 hpi, followed by induction of SISPs from 6-10 hpi. (F) Infiltration of DCavrRpm1 in the coi1-16 mutant causes transient depolarization (3-4 hpi) in the challenged but not distal nor adjacent leaves. (G) Dex-avrRpm1 plants do not show SISPs in non-treated plants. (H-I) Infiltration of DChrpA or DC 1 h after Dex induction of avrRpm1 induces depolarisation and repolarisation of the infiltrated leaf and initiation of SISPs in some systemic responding leaves, albeit weaker than DCavrRpm1.

Consistent with jasmonate dependency, SISPs were abolished in coi1-16, with only the DCavrRpm1-challenged leaf (red trace) undergoing initial depolarization (Fig. 5c and Extended Data Fig. 7f) Surprisingly, despite eliciting a local visible HR, SISPs were not induced by conditional (dexamethasone (Dex)) induction of avrRpm1. Instead, a steady depolarization of the induced leaf with no subsequent repolarization was observed (Fig. 5d and Extended Data Fig. 7g). However, co-infiltration of Dex-induced leaves with DChrpA or DC (Fig. 5d and Extended Data Fig. 7h,i) broadly recapitulated SISP changes, implying that PTI is required for SISP generation.

We also measured SISPs in glr3.3a, glr3.6a, jiss1 and the glr3.3a glr3.6a mutant, the latter of whose functions are necessary for full wound and herbivory responses^40,50. Mirroring WASP attenuation⁴⁰, SISPs were abolished in the glr3.3a, glr3.6a and glr3.3a glr3.6a mutants (Fig. 6a and Extended Data Fig. 8a–c). glr3.6a showed depolarization of the local challenged leaf (Fig. 6a and Extended Data Fig. 8b)⁵⁰, which mirrored jiss1 and coi1-16 responses in which SISPs were abolished and depolarization was only seen in the local challenged leaf (Figs. 5c and 6a and Extended Data Figs. 7f and 8d).

Fig. 6. JISS1 systemic signal is calcium dependent.

a, Glutamate receptor mutants glr3.3a, glr3.6a and glr3.3a glr3.6a do not propagate SISPs following DCavrRpm1 challenge. glr3.6a, like jiss1, shows limited depolarization of the DCavrRpm1-challenged leaf. b, DCavrRpm1-challenged JISS1::LUC glr3.3a glr3.6a, JISS1::LUC glr3.3a and JISS1::LUC glr3.6a mutants all exhibited comparable systemic luciferase activity to that of JISS1::LUC (4:10 hpi). In b and d, white asterisks indicate infiltrated leaves, and luciferase images are false-coloured by signal intensity, as indicated by the calibration bars. c, SAR growth curve of Psm4 following DCavrRpm1 or mock pretreatment on Col-0 (*P = 0.0102, t = 3.343, d.f. = 8 (unpaired two-tailed t-test)) and the glr3.3a glr3.6a mutant (***P = 0.0005, t = 5.550, d.f. = 8). glr3.3a glr3.6 mutant mock and the Col-0 mock were significantly different (**P = 0.0076, t = 3.593, d.f. = 8) (n = 5 biological replicates). Box plots represent the minimum, maximum and median values for the statistical analysis of each treatment. Col-0 mock: min = 7.16, max = 7.20, median = 7.19, Q1 = 7.17, Q3 = 7.20; Col-0 DCavrRpm1: min = 6.43, max = 7.13, median = 6.86, Q1 = 6.71, Q3 = 6.88; glr3.3a glr3.6a mock: min = 7.41, max = 8.23, median = 7.63, Q1 = 7.52, Q3 = 8.18; glr3.3a glr3.6a DCavrRpm1: min = 5.88, max = 6.94, median = 6.29, Q1 = 6.17, Q3 = 6.67. d, Treatment with the calcium inhibitor LaCl₃ (1 mM) inhibits DCavrRpm1-induced JISS1::LUC activity compared with the control (4:50 hpi). Top left, two systemic leaves below the challenged leaf were pre-infiltrated with LaCl₃. Top right, local leaf co-infiltrated with LaCl₃ and DCavrRpm1. Bottom right, petiole of an immunized leaf treated with LaCl₃. Bottom left, no inhibitor. All experiments were repeated at least three times with similar results. SAR (c) was evident with either DC and Psm4 challenge or secondary challenges.

Extended Data Fig. 8. ETI activation of systemic electric signalling is consistent.

Replicated data (n = 2) displayed consistent SISP signatures to Fig. 6. (A) Challenge of glr3.3a with DCavrRpm1 does not show SISPs. (B) Challenge of glr3.6a with DCavrRpm1 shows limited depolarization in the infiltrated but not distal leaves. (C) Challenge of glr3.3a glr3.6a with DCavrRpm1 does not show SISPs. (D) Challenge of jiss1 with DCavrRpm1 shows limited depolarization in the infiltrated but not distal leaves.

Calcium signalling is necessary for systemic immunity

Given the finding that ETI alone is not sufficient to induce SISPs, we crossed JISS1::LUC into the glr3.3a, glr3.6a and glr3.3a glr3.6a mutants. Interestingly, JISS1–LUC dynamics were unaltered in all loss-of-function glr lines tested (Fig. 6b), though both jiss1 and glr mutants lost the ability to generate SISPs.

AvrRpm1–RPM1 mediated SISP generation thus requires functional JISS1, COI1 and vasculature-specific GLR Ca²⁺ channels, which link systemic electrical signal propagation in response to wounding⁴⁰ and herbivory⁵⁰. As the loss of GLRs does not abolish JISS1–LUC activity or SAR to Psm4 (Fig. 6c), SISPs most likely encode another SAR signal, for example for herbivory responses. Since SISPs, like WASPs, are dependent on GLR3.3, which triggers long-distance calcium signalling⁵¹, we investigated the role of calcium in JISS1–LUC activity. The calcium channel blocker LaCl₃ (1 mM) abolished DCavrRpm1-induced JISS1–LUC activity in systemic leaves, as did co-infiltration of LaCl₃ with DCavrRpm1. Also, the application of LaCl₃ to the petiole surface markedly attenuated JISS1–LUC activity (Fig. 6d). Therefore, Ca²⁺ signalling, which is intimately linked to ETI⁵², is necessary for JISS1–LUC SAR activity and SISP propagation.

Discussion

SAR confers broad-spectrum resistance to viral, fungal, oomycete and bacterial pathogens and insect pests despite their diverse lifestyles and virulence strategies. A variety of signalling molecules have been shown to be important for SAR. Many of these are synthesized de novo, implying upstream inductive signals. Our current understanding of signal generation, translocation and establishment in systemic responding leaves is constrained by a lack of information on the spatial–temporal dynamics of SAR activation. Here we used JISS1 (previously A70)²², a jasmonate-responsive gene, to develop a reporter that overcomes these constraints to faithfully report ETI-elicited SAR. Using a JISS1::LUC reporter, chlorophyll fluorescence^27,53 and biophoton²⁰ imaging, we characterized SAR dynamics during HR elicited by two coiled-coil nucleotide-binding leucine-rich repeats (CNLs,RPM1 and RPS2) and a Toll-like/Interleukin 1 nucleotide-binding leucine-rich repeat receptor (TNL,RPS4). We demonstrate that jasmonate signalling is essential for SAR and further show that the ER-localized JISS1 is essential for the generation of electrical surface potentials in systemic responding leaves (SISPs), like those reported to underpin resistance to herbivory⁵⁰.

JISS1::LUC signal generation was unaffected in the mutants sid2, npr1 npr3 npr4, fmo1 and nac19 nac55 nac72 (Fig. 2 and Extended Data Fig. 3a). Infiltration of the SAR inducers NA, AZA, Pip and NHP also failed to induce JISS1::LUC. Of the classic plant immunity-associated hormones, neither ABA nor, unexpectedly, SA activated JISS1::LUC. JA and COR elicited local but not systemic luciferase activity, and COR-deficient DCavrRpm1 still replicated JISS1::LUC temporal and spatial activation dynamics (Fig. 2).

Consistent with a key role for jasmonates in establishing SAR, both jasmonate biosynthetic (aos) and signalling (coi1-16) mutants were SAR deficient and failed to activate JISS1–LUC following DCavrRpm1 challenge (Fig. 3b). These data expand our previous results²² demonstrating compromised SAR in the JA biosynthetic mutant opr3 (12-OXOPHYTODIENOATE REDUCTASE 3) and the JA signalling mutant jin1 (JASMONATE-INSENSITIVE 1; myc2)²². Neither COR nor JA infiltration activated luciferase in coi1-16 JISS1::LUC plants (Fig. 3c), but their exogenous application restored activity in JISS1::LUC aos (Fig. 3d). Reinforcing the importance of jasmonate signalling, the JA-Ile synthetase inhibitor jarin (like phenidione and DIECA) significantly attenuated JISS1::LUC signal generation, propagation and distal activation (Fig. 3f–i). Interestingly, jiss1 was SAR competent, and jiss1 JISS1::LUC lines were activated by ETI, indicative of a mobile jasmonate-dependent SAR signal propagating through the systemic tissue (Fig. 3j and Extended Data Fig. 5c–g). Thus, genetically, pharmacologically and spatially, via real-time transcription monitoring, we conclusively show a key role for jasmonates in SAR, independent of the classical SAR mutant lines npr1, sid2, npr1 npr3 npr4 and nac19 nac55 nac72.

It may seem counterintuitive that jasmonates underpin SAR, given well-documented JA/SA antagonism in biotrophic immunity; however, SA/JA synergism has been reported⁵⁴. Spatial separation of JA and SA signalling during ETI would be a parsimonious explanation for SA/JA dependency and reinforce the importance of short-distance local cellular signalling in ETI responses⁵⁵. Indeed, dual-expressed SA and JA reporters exhibit temporally separated distinct concentric domains (inner/early for SA and outer/later for JA) between the ETI-responding cells⁵⁶, elegantly explaining the COI1 dependency of the JISS1 reporter and the need to consider spatial cellular context during ETI signalling.

Our data support the model that early ETI leads to chloroplast ROS generation, indicative of the strong suppression of F_v/F_m²⁷ and subsequent biophoton generation^20,21. Enzymatic and non-enzymatic chloroplast galactolipid-derived oxylipins^57,58 are thus potential substrates for the generation of jasmonate-based mobile SAR signals and JISS1–LUC activation. Both fatty acid desaturase and chloroplast galactolipid mutants necessary for JA synthesis are SAR deficient^7,9, whereas JA levels increase ~75-fold within 5–10 hpi of DCavrRpm1 challenge⁵⁷.

The JISS1 reporter also responded to wounding stimuli, consistent with our previous microarray data²². The spatial expression pattern of JISS1–LUC activity mirrored jasmonate-dependent WASPs⁴⁰. ETI-responding leaves triggered surface depolarizations (~100 mV) comparable to WASPs, but SISP generation and propagation dynamics were markedly slower. Conditional expression of avrRpm1 in planta was insufficient to generate SISPs and required pathogen-associated molecular patterns (Fig. 5d and Extended Data Fig. 7d). Interestingly, wounding and herbivory release damage-associated molecular pattern molecules^59,60. Our results suggest that PTI/ETI mutual potentiation⁶¹ may extend to SISP signalling.

Glutamate receptor-like mutants abolish WASPs^40,48 and were required for SISP generation. They have previously been implicated in PTI⁶², resistance to pathogens⁶³ and anti-herbivory defence^50,51. glr3.3a exhibited markedly less depolarization of the challenged leaf (Fig. 6a, red trace) than glr3.6a. Notably, jiss1 failed to induce SISPs; also, like jiss1, no tested glr mutants abolished ETI JISS1::LUC activity, and glr3.3 glr3.6 maintained SAR to Psm4 (Fig. 6). Collectively, the loss of SISPs in coi1-16 and jiss1 (Figs. 5c and 6a) supports an additional role for jasmonates and JISS1 in SISP generation. SISPs may thus represent one of a suite of signals that collectively drive reprogramming of distal leaves to confer broad-spectrum resistance. Indeed, we previously showed that the generalist insect Helicoverpa armigera modified its feeding behaviour on Arabidopsis leaves, moving away from regions where JISS1::LUC was induced⁶⁴.

Unlike JISS1, GLRs are believed to be ligand-gated channels. It remains unclear how either is activated to induce or maintain SISPs. More research is therefore necessary to understand the GLR mechanism of action in SISP generation. Localization of JISS1::JISS1–GFP predominantly to the epidermal ER and vasculature was unexpected. It is notable that both GLR3.3a and GLR3.6a localize to the leaf vasculature. Interestingly, GLR3.3a is primarily ER localized in phloem sieve elements, whereas GLR3.6a was predominantly in xylem contact cell tonoplast membranes^49,65. Since distinguishing vascular cell types by cross-section is challenging, we examined single-cell transcriptomics data focusing on the vasculature⁶⁶ and cell layers on the adaxial leaf surface⁶⁷. The majority of JISS1 vascular expression was in a discrete xylem parenchyma cluster, which also showed strong representation of genes encoding hormone pathways including JA as well as specific amino acid biosynthetic and degradation pathways⁶⁶. JISS1 expression was also evident in bundle sheath, adaxial procambium stem cells and a phloem parenchyma cluster, whereas epidermal and guard cell JISS1 expression was seen in general cell-type clusters. Procko et al.⁶⁸ identified JISS1 in only 2 of 16 clusters corresponding to epidermal pavement and mesophyll/photosynthesis-related genes, both highly enriched for plant immune response transcripts.

In summary, JISS1 encodes an ER-localized protein. The JISS1 promoter functions as a dynamic spatial transcriptional SAR reporter, capturing propagative jasmonate-dependent signals rapidly moving through the vasculature, and symplastically via epidermal cells to systemic responding leaves. Our study reveals remarkable parallels between systemic wound signalling and the elicitation of SAR. Vascular jasmonate synthesis, such as in wounding, is well documented and has been attributed to electrical-signal-dependent remodelling of primary vein chloroplast galactolipids⁶⁵. Here, coi1, glr3.3a glr3.6a and jiss1 mutants all abolished SISPs, but the glr mutants (like jiss1) showed wild-type systemic JISS1–LUC signalling. This may represent an example of an amplification loop, considered widespread in systemic signalling⁶⁶. Whereas ETI-mediated rapid jasmonate induction establishes defence against biotrophs, SISPs may propagate mobile information decoded systemically for anti-herbivory responses^50,51. This study lays the foundation, and provides spatial–temporal tools, for dissecting the complex molecular mechanisms underpinning long-distance immune signalling.

Methods

Arabidopsis growth conditions

Arabidopsis thaliana were grown for four to five weeks in compost (Levingtons F2) in a controlled-environment growth chamber programmed at 60% relative humidity with 10 h days (21 °C; 120 μmol m⁻² s⁻¹) and 14 h nights (21 °C) as previously described^37,67.

Generating Arabidopsis transgenic lines

The JISS1::LUC plants were generated as previously described⁶⁴. In brief, the Photinus pyralis LUC2P reporter gene containing the hPEST protein destabilization sequence (Promega pGL4.11) was cloned into pCAMBIA1302 via the Kpn1 and Pml1 restriction sites, creating pC1LUCP. A 1,631-bp promoter fragment of JISS1 was PCR amplified from Col-0 genomic DNA and cloned into pC1LUCP using the primers detailed in Supplementary Table 2. Transgenic homozygous JISS1::LUC lines were generated in the Col-5 background via floral dipping⁶⁹. Ten lines were screened via DCavrRpm1 challenge, and the genomic location of JISS1::LUC in the strongest-expressing line was identified within At4g39240 using adapter ligation-based PCR⁷⁰.

Subsequently, JISS1::LUC crosses into mutant plants were determined as homozygous via diagnostic PCR for JISS1::LUC insertion using the primers detailed in Supplementary Table 2. All other primers for validating mutants following crossing are shown in Supplementary Table 3. All homozygous JISS1::LUC mutant crosses were generated in the mutant parental Col-0 background.

The JISS1^1–84-GFP line was generated using restriction enzyme cloning. The JISS1 promoter and initial coding sequence were amplified from genomic DNA using the primers detailed in Supplementary Table 2. The PCR product was cloned into pCAMBIA1305 as a carboxy-terminal fusion with GFP. The JISS1 promoter full-length JISS1::GFP line was generated using the Golden Gate assembly system⁷¹ as follows: the JISS1 promoter was amplified as above and cloned into the level 0 plasmid pICH41295, while the full-length JISS1 coding sequence was cloned into pAGM41287. The JISS1 promoter and coding sequence were combined in the FL1P1 level 1 vector to create JISS1_pro::JISS1-GFP::Ocs. Finally, the level 2 vector FL1P2 with BASTA selection was used to allow selection in planta after transformation into the Col-0 background via standard methods.

The homozygous JISS1_pro::JISS1-GFP line was subsequently crossed with an RFP-HDEL (ER luminal marker) line (a gift from the late C. Hawes, Oxford Brookes University, UK).

The JAZ10::GUS line was a gift from E. Farmer (University of Lausanne), and the npr1, npr1 npr3 npr4 and nac19 nac55 nac72 mutants were a gift from X. Dong (Duke University). All other loss-of-function mutants and the Dex-avrRpm1 line were obtained from the Nottingham Arabidopsis Stock Centre.

Bacterial growth, maintenance and inoculation

Bacterial cultures (P. syringae pv. tomato strains DC containing the empty cloning vector (pVSP61), DChrpA and DC containing the avirulence gene avrRpm1, avrRps4 or avrRpt2) were grown in Kings B medium⁷² overnight with shaking (200 rpm) at 28 °C. Cells were harvested (2,000 g for 8 min), washed and resuspended in 10 mM MgCl₂ (ref. ⁶⁷). avrRpm1 (in pVSP61) was introduced into the DC COR-deficient mutant DB4G3 (cor⁻¹/cor⁻²)³⁸. For luciferase, GUS, phenotyping and GFP assays, selected leaves were inoculated with a 1-ml needleless syringe on their abaxial surface with the appropriate bacterial suspension adjusted to a final optical density at 600 nm (OD₆₀₀) of 0.15 in 10 mM MgCl₂ (or as otherwise indicated in the figure legends).

For SAR growth assays, the immunizing inoculation comprised either 10 mM MgCl₂ (mock) or OD₆₀₀ 0.005 DCavrRpm1. Two days later, either Psm4 or DC (see the figure legends) was inoculated at OD₆₀₀ 0.001 or OD₆₀₀ 0.002, respectively, using a similar protocol to ref. ⁷³. Three days after DC or four days after Psm4 challenges, bacterial growth measurements were determined from three inoculated leaves per plant and a minimum of four independent replicates. Significant growth differences between treatments were determined using Student’s t-test (unpaired two-tailed). All experiments were repeated at least three times.

Chlorophyll fluorescence

Photosystem II chlorophyll fluorescence imaging of challenged leaves was performed with CF Imager software V2.305 (Technologica Ltd). Plants were placed in the chamber for 40 min post-inoculation and then dark adapted for 20 min. This was followed by a saturating light pulse (6,349 µmol m⁻² s⁻¹ for 0.8 s) to obtain maximum dark-adapted fluorescence (F_m). Actinic light (120 µmol m⁻²s ⁻¹) was then applied for 15 min, followed by a saturating pulse to obtain maximum light adapted fluorescence (F_m′). The plants remained in actinic light for a further 24 min and were then returned to a dark period of 20 min. This cycle (59 min long) was repeated 23 times. F_m, F_m′ and F_o (minimal fluorescence with fully oxidized PSII centres) were used to calculate chlorophyll fluorescence parameters related to photosystem II: F_v/F_m (maximum dark-adapted quantum efficiency) and non-photochemical quenching. Data were extracted using CF Imager software V2.305 (Technologica Ltd).

Luciferase visualization

A solution of 1 mM luciferin (Promega) in 0.02% Silwet L77 (Loveland Industries, Ltd) was sprayed onto JISS1::LUC plants. The plants were kept in the dark for 30 min prior to inoculation. The petioles of the treated and adjacent rosette leaves were secured by folded paperclips to minimize epinastic movement. The plants were placed in a dark box, and images were captured using either an ORCAII ER CCD camera (Hamamatus Photonics) with a 35-mm f2.8 micro Nikkor lens or a Retiga R6 Scientific CCD camera (Qimaging) fitted with a Schneider STD XENON 25-mm lens. Photons were counted for 10 min at 2 × 2 binning mode, and data were acquired with either Wasabi (Hamamatsu) or Micro-Manager v.1.4 (Qimaging) software.

Biophoton visualization

Pathogen-challenged plants were placed inside a dark box mounted with a Retiga R6 camera with a 25-mm f1.4 Navitar lens. Digital monochrome images captured photons for 20 min at 2 × 2 binning mode using MicroManager v.1.4. False-colouring, brightness adjusting and annotation were performed using Fiji (ImageJ2 v.2.9.0/1.53t)⁷⁴, as described previously²⁰.

Chemicals

JA, SA, ABA, AZA, NA, Pip, DIECA, phenidone (1-phenyl-3-pyrazolidinone), lanthanum (III) chloride (LaCl₃) and COR were all from Sigma. NHP was synthesized by Accel Pharmtech, and jarin-1 was sourced from AOBIOUS Inc. All chemicals were used at the concentrations described in the figures. DIECA (2.5 mM), phenidone (2 mM), jarin-1 (25 µM) or LaCl₃ (1 mM) was pre-infiltrated into systemic leaves or co-infiltrated with DCavrRpm1, or the petiole of the challenged leaf was treated with the inhibitor soaked in cotton wool and secured in place with cling film.

Wounding

Leaf tissue was wounded by crushing either side of the midrib with flat tweezers or by severing the petiole with scissors. Three leaves were wounded per plant, and three biological replicated were imaged.

JAZ10–GUS expression

GUS activity in systemic leaves was assessed at 4, 8 and 24 hpi via GUS staining (1 mM X-Gluc, 100 mM NaPO₄ buffer pH 7.0, 10 mM EDTA and 0.1% v/v Triton X-100) using four plants per treatment and challenging four leaves per plant. Samples were incubated at 37 °C. The leaves were de-stained by repeated washes with 70% ethanol, and representative leaf images are shown.

Phenotyping

Infected leaves were removed and imaged at 5.5, 14 and 18 hpi to assess leaf collapse.

Reverse transcription PCR

Total RNA was extracted from mature green leaves of aos, jiss1 and Col-0 (two leaves per plant) using Trizol (Invitrogen, Thermo Fisher Scientific) and treated with DNase I (Merck), according to the manufacturers’ instructions. cDNA was synthesized using ReadyScript cDNA Synthesis Mix (Sigma-Aldrich) with both oligo d(T) and random primers and diluted 1:10 for PCR. PCR was performed using BioMix Red 2xPCR mix (Meridian Bioscience) with primers specific to JISS1 (At5g56980; (LP) CAAGCATGTACGAAGCAAAT; (RP) CTCCTTGTACCTCAGAATCG; amplicon, 301 bp) or Actin2 (At3g18780; (LP) GCCATCCAAGCTGTTCTCTC; (RP) CAGTAAGGTCACGTCCAGCA; amplicon, 156 bp).

Confocal imaging

Freshly excised leaf samples were mounted in water and imaged on a Zeiss LSM 880 confocal microscope with a ×63 oil-immersion objective (ER images) or a ×10 air objective (whole-leaf images). GFP was excited at 488 nm, and emission was detected in the 498–559 nm range; RFP and chlorophyll A were excited at 561 nm and detected in the 605–649 nm range. For cell wall imaging, petiole sections were stained with an aqueous solution of propidium iodide (25 µM), with excitation at 561 nm and emission detection at 614–659 nm. All image analysis was performed in Fiji (ImageJ2 v.2.9.0/1.53t)⁷⁴. Median pixel (signal) intensity for individual leaves was measured across the defined region of interest with corresponding background subtraction using Fiji.

Electrophysiology experiments

Surface electrical potentials of challenged and systemic leaves were measured following infection using four electrodes (Fig. 5a) adopting a similar approach to ref. ⁴⁰. The working electrode (W1, red) was always placed on the lamina immediately above the petiole of the challenged leaf. Electrodes W2 (green) and W3 (blue) were placed similarly on adjacent systemic leaves, and W4 (black) on the distal systemic leaf. The reference electrode was placed in the soil. Surface potential changes were measured from the reference electrode to the working electrode. Plants were covered with a propagator lid, electrical recordings were captured using a data logger (PicoLog 1000) and signal amplitude and duration were plotted for each leaf over ~24 h. Control recordings over an extended time predominantly showed a constant surface potential.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

M.G., M.d.T.-Z., T.G., S.B., E.B. and E.S. conceptualized and designed the experiments. T.G., S.B., E.B., R.H., N.K., F.B. and E.S. performed the experiments and analysed the data. T.G., M.d.T.-Z., F.B., R.H. and S.K. generated the material and resources. D.H. designed the equipment and L.F., P.K. and D.H. provided experimental insights. M.G., S.B., E.B., E.S. and P.K. wrote the manuscript. M.G. and P.K. provided financial support.

Peer review

Peer review information

Nature Plants thanks the anonymous reviewers for their contribution to the peer review of this work.

Data availability

The dataset for Extended Data Fig. 1 has been deposited at http://affymetrix.arabidopsis.info/narrays under identifier NASCARRAYS-403. The A. thaliana reporter lines and all raw data are available from the corresponding author. Source data are provided with this paper. All other data are available in the main text or supplementary materials.

Competing interests

The authors declare no competing interests.

Extended data

is available for this paper at 10.1038/s41477-025-02178-4.

Supplementary information

The online version contains supplementary material available at 10.1038/s41477-025-02178-4.