1. Introduction

Plants are exposed to a variety of biotic and abiotic stresses, including herbivore attack. During the long plant-insect coevolution, plants have evolved constitutive defenses and inducible defenses against insect herbivores (Wu and Baldwin, 2010; Erb and Reymond, 2019). Constitutive defenses are expressed irrespective of whether plants are attacked by insects, while inducible defenses are expressed in response to herbivory (Wu and Baldwin, 2010). Depending on the location, inducible defenses can be classified into local and systemic defenses. In response to wounding or herbivory, not only the attacked organs/tissues (e.g., leaves and roots) but also the distant undamaged organs/tissues increase defenses. This is important for plant survival, as insects often migrate from local to systemic organs and/or insects of the same or different species may subsequently infest on the same plants.

Although much is known about how plants respond to wounding or herbivory locally (Erb and Reymond, 2019), much research is still needed to understand systemic signaling, which regulate the deployment of defenses in systemic organs/tissues. More than half a century ago, Green and Ryan (1972) demonstrated that wounding a tomato (Solanum lycopersicum) leaf by adult Colorado potato beetles (Leptinotarsa decemlineata) rapidly induced accumulation of proteinase inhibitors, defensive metabolites that inhibit insect digestive enzyme activity, in both local and systemic leaves, indicating that certain mobile/systemic signals were transported from the wounded leaf to distal systemic leaves, where they activated defenses. Later, a small peptide systemin was proposed to be a key component of systemic signals (Pearce et al., 1991). However, systemin is specific to the genus Solanum and later studies indicated that systemin is not a systemic signal (Wang et al., 2018). Using grafting technique, Li et al. (2002) elegantly demonstrated that the phytohormone jasmonic acid (JA) or certain metabolites regulated by the JA pathway are possibly the systemic mobile signals. However, subsequent studies indicated that JA or the actual functional jasmonate JA-isoleucine conjugate (JA-Ile) is not the systemic signal per se, and the actual systemic signals are rather more upstream and faster-moving (Wu et al., 2007; Wang et al., 2008; Koo et al., 2009).

The Arabidopsis (Arabidopsis thaliana) Zat12 is a gene that is rapidly induced within minutes after wounding treatment (Davletova et al., 2005) and respiratory burst oxidase homolog D (RBOHD) plays an important role in producing reactive oxygen species (ROS) in extracellular spaces, which function as signaling molecules (Waszczak et al., 2018). Using wild-type (WT) and rbohD mutant Arabidopsis expressing the luciferase (Luc) gene driven by the Zat12 promoter, Miller et al. (2009) found that wounding-induced systemic signals travelled at a speed of up to 8.4 cm/min up- and downward along the wounded inflorescence, and importantly, RBOHD-mediated ROS signaling is required for the propagation of wounding-induced systemic signals. However, RBOHF, which also catalyzes the production of ROS, is not involved in wounding-induced systemic induction of Zat12 expression in Arabidopsis (Miller et al., 2009). The involvement of ROS pathway in systemic signaling has also been demonstrated by the findings that the H₂O₂ receptor H₂O₂-induced Ca²⁺ increases 1 (HPCA1) is important for both systemic ROS signaling and Ca²⁺ signaling in response to high light stress (Fichman et al., 2022). Other molecules and pathways are also involved in leaf-to-leaf systemic signaling. Wounding causes rapid accumulation of extracellular ATP (eATP) in a nanomolar level at the sites of wounds, and exogenously applying ATP to unwounded plants can trigger the generation of ROS waves in an purinoreceptor 2 kinase receptor-dependent manner (Myers et al., 2022). The hydraulic wave and squeeze-cell hypothesis implicate pressure as a key component of systemic signaling (Farmer et al., 2014), and the stretch-activated anion channel MSL10 is required for wounding-induced systemic electrical and Ca²⁺ signaling (Moe-Lange et al., 2021).

Increasing lines of evidence have indicated the importance of glutamate (Glu) and Ca²⁺ in systemic signaling in plant response to wounding or herbivory. Mousavi et al. (2013) demonstrated that wounding an Arabidopsis leaf led to rapid (5.8 ± 1.1 cm/min) changes of surface potentials in systemic leaves; importantly, mutations in the glutamate receptor-like genes GLR3.3 and GLR3.6 attenuated wounding-induced leaf-to-leaf systemic surface potential changes and these mutants also exhibited decreased expression of JAZ10 in the systemic leaves. Toyota et al. (2018) established a Ca²⁺ imaging system by expressing the GCaMP3s protein in Arabidopsis, and using these Ca²⁺ reporter plants, it was discovered that upon wounding, systemic Ca²⁺ signals very rapidly move from the wounded leaf to systemic leaves at a speed of about 1 mm/s; however, the glr3.3 glr3.6 mutants are unable to transport systemic Ca²⁺ signals from leaf to leaf (Toyota et al., 2018). Furthermore, by imaging Glu, Toyota et al. (2018) found rapidly and highly increased Glu contents in the wounded leaves. It was thus proposed that wounding-induced Glu activates GLR ion channels, eliciting systemic signal propagation from leaf to leaf through Ca²⁺ signals (Toyota et al., 2018).

Wounding- or herbivory-induced systemic signaling not only exists between leaves but also between leaf and root. For example, in maize, feeding of western corn rootworm Diabrotica virgifera virgifera on roots induced leaf resistance to the insect Spodoptera littoralis (Erb et al., 2009). Recent evidence also indicated that GLR3.3 and GLR3.6 also function in regulating systemic root-to-shoot Ca²⁺ waves and slow wave potentials, and during such root-to-shoot signaling, GLR3.3 and GLR3.6 displayed Ca²⁺-permeable channel activity gated by both Glu and extracellular pH (Shao et al., 2020). However, little is known about the systemic signaling between other organs. For example, during germination and young seedling stage, not only cotyledons and roots but also hypocotyls could be mechanically damaged or attacked by insects. However, little is known about whether and how wounding/herbivory-induced hypocotyls communicate with other systemic organs and activate corresponding responses.

Grafting is one of the most important techniques used in studying systemic signaling (Tsutsui and Notaguchi, 2017). Unlike normal mutants, whose local and systemic organs all uniformly have a mutation(s) and whose systemic responses may be altered by the lose-of-function or change-of-function of the mutated gene(s), grafted plants using different combinations of WT and mutant plants allow to genetically dissect the functions of genes-of-interest in production, regulation, transmission, and/or perception of systemic signals, and thus, grafting has become an important technique used in studying systemic signaling (Tsutsui and Notaguchi, 2017).

Much research is still needed for understanding the early systemic signaling events triggered by wounding and herbivory, including how systemic Ca²⁺ signaling is regulated and what its physiological role is. In this study, we focused on wounding hypocotyl-induced systemic responses in Arabidopsis leaves and its underlying mechanisms, and using micrografting and in vivo imaging of Ca²⁺ and ROS, we examined the functions of multiple signaling pathways in wounding-induced hypocotyl-to-leaf systemic signaling. It was discovered that the RBOHD-dependent reactive oxygen species pathway and the two amino acid-gated ion channels GLR3.3 and GLR3.6 cooperatively play important roles in regulating wounding-induced hypocotyl-to-leaf systemic responses, including systemic Ca²⁺ waves, JA and defensive metabolites, and insect resistance.

2. Materials and methods 2.1. Plant materials and growth conditions

Arabidopsis thaliana (Col-0) WT and mutant seeds were germinated on Petri dishes containing 1/2 Murashige and Skoog (MS) salts, 1% sucrose, and 1% agar. Plants were cultivated under the short-day conditions (8 h of light, 22 ℃/16 h of dark, 19 ℃). Mutants used in this study were aha1 (AT2G18960; SALK_065288C), msl10 (AT5G12080; SALK_076254C), pdlp5 (AT1G70690; SALK_016278C), hpca1 (AT5G49760; CS926093), dde2-2 (AT5G42650; CS65993), rbohD (AT5G47910; CS68747), rbohF (AT1G64060; CS68748), glr3.3 (AT1G42540; SALK_099757), glr3.6 (AT3G51480; SALK_091801), dorn1 (AT5G60300; SALK_042209), rbohD rbohF (Kwak et al., 2003), and glr3.3 glr3.6 (Xue et al., 2022). Through hybridization we obtained the triple mutants rbohD glr3.3 glr3.6 and dde2-2 glr3.3 glr3.6.

2.2. Preparation of Ca²⁺ sensor plants

The plasmid pGP-CMV-GCaMP6s was obtained from Addgene (https://www.addgene.org/, #40753). Bgl Ⅱ and BstE Ⅱ were used to digest the pGP-CMV-GCaMP6s plasmid and the resulted fragment of GCaMP6s was inserted it into the plant binary expression vector pCAMBIA3301, resulting in the pCAMBIA3301-GCaMP6s construct. Subsequently, pCAMBIA3301-GCaMP6s was transformed into Agrobacterium tumefaciens GV3101, which was used to transform Arabidopsis by floral dipping.

2.3. Arabidopsis micrografting

First, 5-day-old Arabidopsis seedlings were subjected to a one-day dark treatment. The hypocotyls were cut using a sterile double-sided blade. The rootstock and scion of the same or different genotypes were then placed onto the grafting medium (2.215 g/L MS (phytotechlab), 20 μg/L indole acidic acid, 40 μg/L 6-benzylaminopurine, 10 g/L agar (Solarbio)), ensuring that there were no gaps between the joints. The grafted Arabidopsis seedlings were cultivated under 8 h light/16 h dark and 22 ℃ for 8 days. Subsequently, the grafted seedlings were transferred to 1/2 MS medium and cultured under 10 h light/16 h darkness (other conditions remained unchanged) for another 10 days. These seedlings were used for all experiments, except for bioassays. For bioassays, these seedlings were transferred to soil and grown for another 10 days, before infestation of insects.

2.4. Plant treatment and imaging of Ca²⁺

To treat the seedlings, the hypocotyls were squeezed with tweezers at the positions about 3 mm below the graft junctions. A stereo fluorescence microscope (Nikon, SMZ18) equipped with a fluorescence light source (Nikon, INTENSILIGHT C-HGFI) and a Scientific CMOS camera from Teledyne Photometrics (Prime BSI) was used for real-time detection of Ca²⁺ signals in vivo. Nikon GFP-B (ex = 450−500 nm) excitation, emission, and dichroic filters were used in this experiment. Images were acquired at a constant exposure time (400 ms) and process with the NIS-Elements AR software (Nikon). The GCaMP6s signals were captured over 6.5 min at regions of interest of systemic leaves after the treatment. To calculate the fractional fluorescence changes (ΔF/F), the equation ΔF/F═(F–F0)/F0 was used, where F0 denotes the average baseline fluorescence determined by the average of F over the first 10 frames of the recording before treatment.

2.5. Determination of active translocation velocity in vascular bundles

The fluorescent tracer 8-hydroxypyrene-1, 3, 6-trisulfonic acid (HPTS) was prepared as an aqueous solution with a concentration of 0.4 mM. The hypocotyl of each grafted seedling was severed at the positions about 3 mm below the graft junctions, and immediately 5 μL of the HPTS solution was applied to the incision, and the fluorescence changes of the scion were recorded for 6.5 min. The same equipment and parameters used for Ca²⁺ imaging were employed for HPTS image acquisition.

The transportation times (Δt) of HPTS to the 5th leaf and the 9th leaf were determined from the images, respectively. For each seedling, Δx = hypocotyl length from the cut point + distance from the center of the rosette leaves to the 5th leaf and 9th leaf. The 5th leaf was defined as the old leaf, and the 9th leaf was defined as the new leaf. The calculations were based on v = Δx/Δt.

2.6. Live imaging of ROS

Imaging of ROS followed a previously published method (Fichman et al., 2019). Plants were fumigated in a seal box with 50 μM 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) in 50 mM phosphate buffer (pH 7.4) using a portable mini nebulizer for 30 min. The hypocotyls of fumigated seedlings were damaged with tweezers and fluorescence images were acquired on the same fluorescence stereo microscope at a constant exposure time (700 ms) and analyzed using the NIS-Elements AR software (Nikon). Each seedling was videoed for 6 min.

2.7. Glutamate content measurement

In brief, lyophilized samples (20 mg) were extracted with 1 mL of 80% methanol (v/v) and centrifuged at 16000 g for 15 min. The supernatant of each sample was diluted in a ratio of 1:10 (v/v) in water containing D₃-l-glutamic acid as the internal standard at a concentration of 1 μg/mL. Concentrations of Glu in the diluted extracts was analyzed using a HPLC-MS/MS (LCMS-8040 system, Shimadzu) by comparing the peak areas of glutamic acid with the isotope-labeled glutamic acid.

2.8. Quantification of phytohormones and glucosinolates

Phytohormone determination was carried out using an HPLC-MS/MS (LCMS-8040 system, Shimadzu) following a previously established method (Setotaw et al., 2024). The contents of glucosinolates in Arabidopsis were determined using HPLC (LC-20 AD, Shimadzu) according to Burow et al. (2006).

2.9. RNA-seq data acquisition and analysis

The TRIzol reagent (ThermoFisher Scientific) was used to extract total RNA. Three biological replicates were used for each group of samples. For RNA-seq data acquisition, VAHTS® Universal Plus DNA Library Prep Kit for Illumina (Vazyme) was used to construct cDNA libraries. The generated cDNA libraries were sequenced on a DNBSEQ-T7 platform (BGI) to acquire sequence reads (6 G depth). The Bioconductor DEseq2 (v.3.9) package was employed to infer differential gene expression. Differentially expressed genes (DEGs) were defined as transcripts with false discover rates (FDR) < 0.05 and absolute values of log2 (fold change) > 1. The RNA-seq data can be retrieved from the Beijing Institute of Genomics under the BioProject: PRJCA035997.

2.10. Bioassay

The hypocotyls of the grafted Arabidopsis were damaged with tweezers at the positions about 3 mm below the grafting junctions. Two days later, each seedling was infested with a three-day-old first instar Spodoptera litura larva, which was fed with artificial diet before infestation on these seedlings. The insects were each enclosed in a clip cage, which was attached to the whole scion leaves to avoid escaping, and the masses of the insects were measured one day later.

3. Results 3.1. Knocking out GLR3.3 and GLR3.6 does not affect wounding-induced hypocotyl-to-leaf Ca²⁺ waves

GLR3.3 and GLR3.6 play an important role in initiating and/or transmitting wounding-induced systemic Ca²⁺ waves between leaves (Toyota et al., 2018). Given the well-established functions of the JA, Ca²⁺, and ROS signaling pathway in wounding-induced leaf-to-leaf communications, we selected the JA synthesis mutant dde2-2 (von Malek et al., 2002), the Glu receptor double mutant glr3.3 glr3.6 (Toyota et al., 2018), the respiratory burst oxidase homolog mutants rbohD and rbohD rbohF (rbohD/F) (Miller et al., 2009; Liu et al., 2017), and the ROS receptor mutant hpca1 (Fichman et al., 2022) as rootstocks. WT and these mutants were respectively grafted with scions, which were excised from Arabidopsis seedings expressing the Ca²⁺ sensor GCaMP6s to form GCaMP6s/WT ("/" denotes grafting between scion (left) and rootstock (right)), GCaMP6s/dde2-2, GCaMP6s/rbohD rbohF, and GCaMP6s/hpca1 plants (Fig. S1A). There were no significant differences in the growth and primary root lengths of the grafted Arabidopsis with different mutants as rootstocks (Fig. S1B and S1C).

For all grafted plants, the hypocotyls were wounded with tweezers at the sites ~3 mm below the graft junctions. The Ca²⁺ signals were detected in the leaves of GCaMP6s/WT as early as 10 s after injury, and the intensity of the Ca²⁺ waves peaked at 40−50 s (Fig. 1A and Movie S1). By measuring the distances from the wound sites to the leaves, we inferred that the speed of the Ca²⁺ signals-inducing systemic signals are 0.7–0.9 mm/s. Compared to the WT rootstocks, there were no significant differences in the intensity of Ca²⁺ signals induced in the scions grafted with dde2-2, glr3.3glr3.6, rbohD, rbohD/F, or hpca1 rootstocks, and the times needed for reaching the peaks of Ca²⁺ signals were also similar among all samples (Fig. 1B−1K and Movie S2). Toyota et al. (2018) discovered that between leaves, long-distance Ca²⁺signals rely on GLR3.3 and GLR3.6 for signal propagation. However, notably, when the glr3.3glr3.6 mutant was used as rootstock, wounding hypocotyls of the rootstocks still triggered Ca²⁺ waves in the leaves (Fig. 1C and Movie S3), indicating that GLR3.3 and GLR3.6 alone do not affect the propagation of Ca²⁺ waves from hypocotyls to leaves, and importantly, there is another GLR3.3- and GLR3.6-independent signaling pathway which regulates the transmission of mechanical damage-induced Ca²⁺ signals from hypocotyl to leaf.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.pld.2025.05.004

We next selected the ATP receptor mutant dorn1 (Choi et al., 2014), the H⁺-ATPase mutant aha1 (Kumari et al., 2019), the plasmodesmata-localized protein 5 (PDLP5) mutant pdlp5 (Fichman and Mittler, 2021), and the MscS (mechanosensitive channel of small conductance)-like 10 (MSL10) mutant msl10 (Moe-Lange et al., 2021) as the rootstocks to graft with the scions expressing GCaMP6s. Among these, the pdlp5 rootstocks could not form healthy graft junctions with the WT scions (Fig. S1D), and thus no further experiments were done. When the mutants dorn1, aha1, and msl10 were used as the rootstocks, normal Ca²⁺ waves were observed in the leaves (Fig. S2A−S2I). Thus, these pathways are not necessary for wounding-induced hypocotyl-to-leaf systemic Ca²⁺ waves.

3.2. RBOHD and GLR3.3/3.6 cooperatively regulate hypocotyl-to-leaf Ca²⁺ waves

Next, we examine the possibility that some of the signaling pathways may crosstalk to regulate hypocotyl-to-leaf Ca²⁺ waves. Through hybridization, we obtained the triple mutant dde2-2 glr3.3 glr3.6 and rbohD glr3.3 glr3.6. Compared to the WT rootstocks, wounding the dde2-2 glr3.3 glr3.6 hypocotyls induced similar Ca²⁺ waves in leaves, in terms of both signal intensity and speed of propagation (Fig. 2B and D). Strikingly, after wounding the mutant hypocotyls of WT/rbohD glr3.3 glr3.6 plants, the times for the Ca²⁺ waves to reach their peaks were delayed by 30–40 s (Fig. 2C and Movie S4), although there were no obvious changes of the intensity of Ca²⁺ waves (Fig. 2E). Detailed observation indicated that the delay was a result of prolonged time required for the initiation of Ca²⁺ waves: detectable levels of Ca²⁺ signals were observed in GCaMP6s/WT leaves approximately 10 s after treatment, while in the leaves of GCaMP6s/rbohD glr3.3 glr3.6 the Ca²⁺ signals were detectable approximately 40 s after the treatment (Fig. 2A and C); however, after the initiation of Ca²⁺ signals, the speed of transfer and signal intensity were similar between GCaMP6s/WT and GCaMP6s/rbohD glr3.3 glr3.6 plants. This result illustrates the complex regulation of hypocotyl-to-leaf systemic Ca²⁺ signals: the RBOHD-mediated ROS signaling pathway and the Ca²⁺ signaling pathway mediated by GLR3.3 and GLR3.6 function independently but cooperatively in regulating the hypocotyl-to-leaf Ca²⁺ waves. Furthermore, these data suggest that there is another unknown regulatory pathway that is involved in the activation and transmission of hypocotyl-to-leaf systemic Ca²⁺ signals.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.pld.2025.05.004

3.3. RBOHD and GLR3.3/3.6 do not regulate bulk flow from hypocotyl to leaves

Bellandi et al. (2022) demonstrated that wounding a leaf triggers the release of amino acids, especially Glu, that diffuse locally through the apoplast, activating the Ca²⁺-permeable channel GLR 3.3. To test whether RBOHD and GLR3.3/3.6 control the bulk flow of vasculature, WT Arabidopsis and the mutants rbohD, glr3.3 glr3.6, and rbohD glr3.3 glr3.6 were employed as rootstocks, and WT scions were respectively grafted with these rootstocks. The leaves were harvested for free Glu content measurement after the hypocotyls from these rootstocks were wounded and those from the untreated ones were harvested as controls. Compared with the control scions, the Glu contents only slightly increased in all grafted plants with no large differences among the scions of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants (Fig. 3A). Thus, the delayed Ca²⁺ waves in the GCaMP6s/rbohD glr3.3 glr3.6 scions seem not to be a result of changes of Glu accumulation in the scions. Moreover, a stable isotope labeled Glu (D₃-L-Glu) solution was applied to wounded hypocotyls of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 rootstocks, and the contents of D₃-L-Glu was measured in the leaves. D₃-L-Glu was detected in the concentrations of about 0.1 mg/g as early as 1 min in all the leaves of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants, and no differences were detected among all these plants (Fig. 3B–S3A and S3B). Furthermore, to rule out the possibility that bulk flow was slower in the WT/rbohD glr3.3 glr3.6 plants than in the WT/WT plants, we measured the transport velocity of 8-hydroxypyrene-1, 3, 6-trisulfonic acid (HPTS) (Wright and Oparka, 1996). Compared with the WT/WT plants, we did not find any obvious changes of vascular transport velocities in the grafted plants whose root stocks were rbohD, glr3.3 glr3.6, or rbohD glr3.3 glr3.6 (Fig. 3C and Movies S5–S8). Again, all the scions showed a similar transport velocity of 0.13–0.15 mm/s in all plants (Fig. 3D). Importantly, these data indicate that the transport velocity of amino acid bulk flow is much lower than that of the Ca²⁺ wave-inducing systemic signals, and thus, RBOHD and GLR3.3/3.6 do not regulate bulk flow in hypocotyl to systemic leaf of Arabidopsis seedlings.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.pld.2025.05.004

3.4. RBOHD and GLR3.3/3.6 in hypocotyl is required for induction of ROS waves in systemic leaves

After wounding or herbivory, ROS and JA are the two key signaling pathways that are involved in activating defenses (Waszczak et al., 2018), and between leaves, ROS waves in systemic leaves are rapidly induced by different stimuli, including wounding (Miller et al., 2009). To examine whether RBOHD and GLR3.3/3.6 regulate wounding hypocotyl-induced ROS in leaves, the hypocotyls of rootstocks of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants were applied with the fluorescent ROS probe H₂DCFDA and thereafter wounded, and the ROS bursts were monitored. In the WT/WT Arabidopsis, ROS in the leaves rapidly increased after the hypocotyls of rootstocks were damaged, whose levels were detectable as early as 40 s and continued to increase over the 6 min of monitoring (Fig. 4A–B). Although we did not detect any differences between the ROS levels of the leaves of WT/WT and WT/glr3.3 glr3.6 (Fig. 4C–D), highly decreased ROS levels were similarly observed in the leaves of WT/rbohD and WT/rbohD glr3.3 glr3.6 (Fig. 4E−H), a result that is in line with the decreased but still present systemic ROS accumulation in wounded leaves of rbohD mutants (Zandalinas and Mittler, 2021). Notably, at about 5 min, the ROS levels in the leaves of WT/rbohD glr3.3 glr3.6 reached a plateau, while the leaves of other plants still exhibited increasing levels of ROS (Fig. 4B, D, 4F and 4H), suggesting that GLR3.3/3.6 could also positively influence systemic ROS production. Thus, we concluded that wounding hypocotyl actives a ROS burst in systemic leaves, and RBOHD and GLR3.3/3.6 in the hypocotyl all play roles this process.

3.5. Wounding hypocotyl-induced Ca²⁺ signals in leaves are associated with insect resistance

To determine whether the delay of Ca²⁺ waves in the leaves of GCaMP6s/rbohD glr3.3 glr3.6 affects their insect resistance, we prepared grafted plants using the WT Arabidopsis as scions and the mutants rbohD, glr3.3 glr3.6, and rbohD glr3.3 glr3.6 as rootstocks, respectively. Thirty minutes after wounding the hypocotyls of rootstocks, all leaves exhibited increased JA levels; although JA accumulation was normal in the leaves of WT/rbohD and WT/glr3.3 glr3.6 plants, in line with the delayed Ca²⁺ waves in the WT/rbohD glr3.3 glr3.6 plants, the JA levels in the WT/rbohD glr3.3 glr3.6 leaves were about 50% reduced (Fig. 5A).

Glucosinolates (GSs) are a group of important herbivore-resistant compounds in Brassicaceae (Wittstock and Halkier, 2002). Similarly, we wounded the hypocotyls of rootstocks of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants, and after 2 days, the contents of GSs in the scions were measured; the untreated plants served as respective controls. 3-methylsulfinylpropyl-glucosinolate (3MSOP), 4-methylsulfinylbutyl isothiocyanate (4MSOB), 1-methoxy-3-indolylmethyl glucosinolate (1MO-I3M), 8-methylsulfinyloctyl isothiocyanate (8MSOO), and 5-methylsulfinylpentyl glucosinolate (5MSOP) did not show obvious differences among the control plants (Fig. 4B−4F, S4A and S4B), while indole-3-ylmethyl-glucosinolate (I3M) and 4-methoxy-3-indolylmethyl glucosinolate (4MO-I3M) had lower levels in the WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants, indicating that RBOHD and GLR3.3/3.6 in the rootstock regulate specific GSs in leaves even under the control conditions (Fig. 4B–C). In the WT/WT scions, the GSs I3M, 4MO-I3M, 8MSOO, 3MSOP, 4MSOB, 1MO-I3M and 5MSOP all exhibited increased levels after the treatment (Fig. 5B−5E, S4A and S4B). Under control conditions, the contents of indole GS I3M in the leaves of WT/rbohD and WT/glr3.3 glr3.6 were about 38% and 24% lower than in the WT/WT leaves, and the leaves of WT/rbohD glr3.3 glr3.6 plants even showed 46% decrease (Fig. 5B), suggesting that RBOHD and GLR3.3/3.6 in hypocotyl regulate leaf I3M content under normal conditions; after wounding hypocotyl, the contents of I3M completely and almost restored to the levels of WT/WT in the WT/glr3.3 glr3.6 and WT/rbohD plants, respectively, but I3M levels in the leaves of WT/rbohD glr3.3 glr3.6 were still 43% lower than in the leaves of WT/WT plants (Fig. 5B). Another indole GS 4MO-I3M were decreased in the systemic leaves of WT/glr3.3 glr3.6 and WT/rbohD glr3.3 glr3.6 plants, but not in the WT/rbohD plants, indicating that GLR3.3 and GLR3.6 are necessary for regulating systemic 4MO-I3M accumulation (Fig. 5C). Compared with those in the WT/WT plants, the contents of the long-chain aliphatic GS 8MSOO were about 35% and 32% lower in leaves of WT/rbohD and WT/glr3.3 glr3.6 plants, and the WT/rbohD glr3.3 glr3.6 exhibited even more decreased levels of 8MSOO (60% decreased) (Fig. 5D). For the short-chain aliphatic GSs 3MSOP and 4MSOB, only when rbohD glr3.3 glr3.6 served as the rootstocks, their contents were respectively 50% and 74% decreased (Fig. 5E–F). All these data from quantification of GSs indicated the essential and cooperative role of RBOHD and GLR3.3/3.6 in hypocotyl in regulating accumulation of GSs in leaves after wounding hypocotyl.

Next, we aimed to further investigate the functions of these three genes in systemic herbivore resistance. A bioassay on grafted plants were performed. Given that the grafted plants used in previous experiments were too small for insect feeding, they were transferred to soil for further cultivation. After ten days of growth in soil, the hypocotyls below the grafting points were wounded and these plants were regarded as the treatment group, while the plants of control group did not undergo any treatment. After two days, Spodoptera litura larvae were infested on the grafted Arabidopsis, and the weight of each larva was measured after one day of feeding. When the WT, rbohD, and glr3.3 glr3.6 plants were used as rootstocks, the weights of larvae on the scions were about 18%−36% lower in the treatment group compared to the control group. However, the weights of larvae of the treatment and control group of WT/rbohD glr3.3 glr3.6 plants were similar (Fig. 6A–B). Thus, RBOHD, GLR3.3, and GLR3.6 in hypocotyl cooperatively control the resistance of systemic leaves to S. litura larvae.

These results indicate that both the RBOHD- and GLR3.3/3.6-controlled hypocotyl-to-leaf systemic Ca²⁺ signaling likely regulates systemic accumulation of JA and GSs and thus insect resistance.

3.6. RBOHD and GLR3.3/3.6 regulate systemic transcriptome changes after wounding hypocotyl

To better understand the role of RBOHD and GLR3.3/3.6 in regulating wounding-induced hypocotyl-to-leaf systemic signaling, we wounded the hypocotyls of rootstocks of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants, and after 30 min, all the leaves of each scion were harvested for RNA-seq analysis, and leaves from untreated plants were also harvested as controls. In the leaves of WT/WT plants, 670 genes were found to be regulated, while 315 and 216 differentially regulated genes (DEGs) were found in the leaves of WT/rbohD and WT/glr3.3 glr3.6 respectively; strikingly, there were only 58 DEGs in the leaves of WT/rbohD glr3.3 glr3.6 plants (Table S1), highlighting the critical role of both RBOHD and GLR3.3/3.6 in regulating the hypocotyl-to-leaf systemic responses.

Venn diagram analysis indicated that among the 670 DEGs in the leaves of WT/WT plants, 562 are specific for WT/WT plants (84%), and only 56 (8%), 49 (7%), and 20 (3%) are common with those in WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6, respectively (Fig. 7A). Given the important herbivore resistance phenotype of the WT/rbohD glr3.3 glr3.6 plants, the DEGs of WT/WT leaves and those of WT/rbohD glr3.3 glr3.6 were specifically compared (Fig. 7B): 650 DEGs in the WT/WT plant leaves were not regulated in the leaves of WT/rbohD glr3.3 glr3.6 plants, indicating the importance of both RBOHD and GLR3.3/3.6 in regulating these genes in systemic leaves. Gene Ontology (GO) analysis indicated that the main enriched terms of biological process included response to stress, response to chemical, response to stimulus, defense response, and response to hormone (Fig. 7C). To inspect the regulation of the common DEGs between the leaves of WT/WT plants and WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants, scatter plots indicating the fold-change values were used to visualize the regulation of these commonly regulated DEGs. Although the common DEGs between the WT/WT and WT/rbohD leaves were regulated at the similar levels (Fig. 7D), among the common DEGs between the WT/WT and WT/glr3.3 glr3.6 leaves and the common DEGs between WT/WT and WT/rbohD glr3.3 glr3.6 leaves, many of these DEGs upregulated in the WT/WT leaves were downregulated in the WT/glr3.3 glr3.6 (the nine genes in the fourth quadrant) and WT/rbohD glr3.3 glr3.6 leaves (the 18 genes in the fourth quadrant) (Fig. 7E–F), and GO analysis indicated that the former 9 genes were enriched in two terms "response to stress" and "response to stimulus" and the latter 18 genes were enriched in photosynthesis- and metabolism-related terms (Table S2).

Specifically, the expression patterns of genes associated with the ROS, Ca²⁺, and JA signaling pathway were inspected. Among the ROS-related genes, Zat12, APX1, and MAPKKK18 exhibited large increases after wounding in the leaves of WT/WT plants. However, the transcript levels of these genes showed only small changes in the leaves of WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants (Fig. S5A). The ROS scavenging enzyme-encoding genes SOD1, CAT3, and CSD2 exhibited moderate upregulation in systemic leaves of WT/WT plants following hypocotyl wounding. In contrast, WT/rbohD and WT/glr3.3 glr3.6 scions showed small transcriptional changes in these genes. Strikingly, WT/rbohD glr3.3 glr3.6 scions displayed an opposite regulatory pattern: these genes maintained elevated basal expression levels under control conditions but underwent significant transcriptional repression post-wounding (Fig. S5A). For the genes associated with the Ca²⁺ signaling pathway, we primarily focused on those encoding calcium-dependent protein kinases (CPKs), calmodulin (CaM), and cyclic nucleotide-gated channels (CNGCs). Except for CNGC2 and CNGC7, the expression levels of the Ca²⁺ signaling-related genes were highly induced by wounding in the leaves of WT/WT plants. However, most of these Ca²⁺ signaling-related genes had decreased expression levels in the leaves of WT/rbohD and WT/glr3.3 glr3.6 plants, and notably, many of these genes had very low levels in the leaves of WT/rbohD glr3.3 glr3.6 plants under both control and wounding conditions (Fig. S5B). Both JA signaling-related genes (JAZs, MYC2, BHLH10) and JA biosynthesis-related genes (LOX3, AOS, OPR3) were highly induced in the leaves of WT/WT plants after wounding the hypocotyls of rootstocks, and similar to the Ca²⁺ signaling-related genes, highly decreased levels of these genes were found in the leaves of WT/glr3.3 glr3.6 and WT/rbohD glr3.3 glr3.6 plants, and the WT/rbohD glr3.3 glr3.6 plants had the lowest levels of these JA signaling- and biosynthesis-related genes (Fig. S5C).

These transcriptome data indicate that RBOHD and GLR3.3/3.6 in hypocotyl function cooperatively in regulating the systemic transcriptional responses in leaves.

4. Discussion

Plants use systemic signals to coordinate their growth, development, and responses to environmental stresses, including wounding and herbivory. However, little is known about how different tissues and organs rapidly communicate through systemic signals during early responses to wounding and herbivory. In this study, using Arabidopsis micrografting, Ca²⁺ imaging, phytohormone and defensive metabolite analysis, we demonstrate that RBOHD and GLR3.3/3.6 cooperatively regulate hypocotyl-to-leaf systemic responses, including Ca²⁺ signaling, transcriptome reconfiguration, accumulation of JA and GSs.

Previous studies have indicated that wounding leaf and root rapidly activated leaf-to-leaf and root-to-shoot systemic Ca²⁺ signals (Toyota et al., 2018; Shao et al., 2020). In this study, we found that wounding hypocotyl rapidly activated systemic Ca²⁺ signals in leaves. The Ca²⁺ signals reached the peak values about 40–50 s after wounding the hypocotyls, at a speed of 0.7−0.9 mm/s (Fig. 1A and Movie S1). This is similar to what was found in wounding-induced leaf-to-leaf Ca²⁺signals, whose speed is about 1 mm/s (Toyota et al., 2018). Despite of the similar speed of activation of systemic Ca²⁺signals, the glr3.3 glr3.6 mutants exhibited highly diminished wound-induced leaf-to-leaf and root-to-shoot systemic Ca²⁺signals (Toyota et al., 2018; Shao et al., 2020), whereas such systemic Ca²⁺signals were normal when the glr3.3 glr3.6 mutant hypocotyls were wounded (Fig. 1C and Movie S3). Thus, GLR3.3 and GLR3.6, two putative Glu receptors and Ca²⁺-permeable channels, have different functions in regulating leaf-to-leaf and hypocotyl-to-leaf systemic Ca²⁺ signals. Moreover, although the systemic Ca²⁺ waves in leaves were still induced after wounding the hypocotyls of the rbohD glr3.3 glr3.6 mutant rootstocks, a delay of Ca²⁺ signals, but not change of signal intensity, was detected after wounding the hypocotyls of GCaMP6s/rbohD glr3.3 glr3.6 plants (Fig. 2C and E; Movie S4). The different regulation of leaf-to-leaf and hypocotyl-to-leaf systemic Ca²⁺ signaling led us to hypothesize that when leaves and hypocotyls are wounded, different mobile signals may be produced locally, which respectively require GLR3.3 and GLR3.6 and do not require GLR3.3 and GLR3.6, and these mobile signals are all able to move to systemic leaves to trigger Ca²⁺ signals, and after wounding hypocotyls, GLR3.3, GLR3.6, and ROS produced by RBOHD may regulate the dynamics of production/accumulation of the mobile signals, resulting in the delayed but otherwise normal mobile signals from the rbohD glr3.3 glr3.6 hypocotyls.

A study by Bellandi et al. (2022) suggested that apoplastic diffusion and bulk flow of amino acids play a critical role in transmission of systemic Ca²⁺ signals. We found that the free Glu contents increased similarly in the leaves of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants after wounding the hypocotyls (Fig. 3A). Given the important role of Glu in activating the GLR3.3 and GLR3.6 receptors, we also measured the contents of D₃-L-Glu in leaves after exogenous supplementation of D₃-L-Glu to hypocotyls of WT/WT, WT/rbohD, WT/glr3.3 glr3.6, and WT/rbohD glr3.3 glr3.6 plants, and similarly, the transport of exogenously supplied HPTS was examined. These experiments suggested that RBOHD and GLR3.3/3.6 do not regulate bulk flow and supported the scenario that probably not the transport of wounding-induced mobile signals, but the dynamics of production and/or accumulation of the mobile signals were altered in the rbohD glr3.3 glr3.6 mutant hypocotyls.

Ca²⁺ as a crucial second messenger playing a critical role in plant growth, development, and responses to various stresses (Tian et al., 2020). Many studies have indicated that inhibiting Ca²⁺ signaling by applying inhibitors, strongly affects the local and systemic responses. For example, tomato plants had decreased levels of simulated herbivory-induced JA and JA-Ile after applying an calmodulins (CaMs)-specific inhibitor to tomato, compared with tomato plants being treated with an inactive inhibitor analog (Hu et al., 2022). Wounding Arabidopsis leaves, whose petioles were pretreated with a Ca²⁺ channel inhibitor, showed highly compromised systemic elevation of both Ca²⁺ signal and induction of wound-related marker genes (Toyota et al., 2018). In this study, we found that wounding rbohD glr3.3glr3.6 hypocotyls resulted in a delay in the systemic Ca²⁺ wave in the leaves. Concomitantly, decreased JA and GSs were detected in these systemic leaves (Fig. 5A−5F), leading to compromised systemic herbivore resistance (Fig. 6A). These data imply the critical role of Ca²⁺ signals, given that such a delay of systemic Ca²⁺ signals, which had normal intensity and duration, had such a profound impact on the defenses of systemic leaves. This point was also supported by our transcriptome analysis: wounding the hypocotyls of WT/WT, WT/rbohD, and WT/glr3.3 glr3.6 respectively led to 670, 315, and 216 DEGs in the leaves, while only 58 DEGs were found in the leaves of WT/rbohD glr3.3 glr3.6 plants (Table S1). How the Ca²⁺ waves, including the signal intensity and speed of transport, are accurately decoded to into downstream responses requires further study. Of course, it is very likely that the systemic regulation transcriptome, hormone JA, and GSs is a result of the crosstalk between systemic Ca²⁺ signals and other ROS- and GLR3.3/3.6-regulated pathways.

ROS burst is also one of the most rapid local and systemic responses when plants are challenged by stresses (Mittler et al., 2022). We found that wounding hypocotyl also very rapidly activated a burst of ROS in systemic leaves, but the speed of was slightly slower than that of Ca²⁺ signals (Fig. 1A–4A), suggesting that ROS may not have an impact on the Ca²⁺ signals in systemic leaves; consistently, wounding the hypocotyls of rootstocks of WT/rbohD plants induced normal systemic Ca²⁺ signals in leaves (Fig. 1D and Movie S2). The WT/rbohD plants showed a compromised systemic ROS phenotype, indicating that RBOHD in hypocotyl controls ROS in leaves (Fig. 4A–E). Moreover, in line with the finding that during salt stress, ROS and Ca²⁺ signals are interdependent in Arabidopsis plants (Evans et al., 2016), our data suggest that GLR3.3 and GLR3.6 in hypocotyl likely also play a role in regulating systemic ROS in leaves (Fig. 4G–H). We used young seedlings for imaging analysis, which could not be monitored under the fluorescence stereomicroscope for more than 6 min, as they started to wither under the extensive excitation light. It is worth to examine the ROS levels in longer time frames, such as at least 30 min. We propose that after wounding hypocotyls, ROS is produced locally near the wounds, and a certain unknown mobile signal, which is likely different from the mobile signals activating Ca²⁺ signals in leaves, is induced by ROS produced by RBOHD; this mobile signal is transported rapidly to systemic leaves where it initiates ROS production. These findings add another layer of complexity to the systemic wound signaling.

In addition, our micrografting experiments also indicated that the JA pathway, H⁺-ATPase AHA1, extracellular ATP receptor DORN1, H₂O₂ receptor HPCA1, and stretch-activated anion channel MSL10, which are related to different stress-induced systemic responses (von Malek et al., 2002; Choi et al., 2014; Kumari et al., 2019; Moe-Lange et al., 2021; Fichman et al., 2022; Myers et al., 2022), do not regulate wounding-induced hypocotyl-to-leaf systemic Ca²⁺ signals. Among these, MSL10 was found to be important for wounding-induced leaf-to-leaf systemic Ca²⁺ signals (Moe-Lange et al., 2021). It is likely that different organs may use overlapping but different signaling molecules, which are regulated by complex interactive signaling pathways, to communicate during responses to different stresses.

Acknowledgements

This work was supported by National Natural Science Foundation of China (U23A20199), Yunnan Revitalization Talent Support Program "Yunling Scholar" and Yunnan Fundamental Research Projects (202201AS070056). We are grateful to the Service Center for Experimental Biotechnology and the High-Performance Computing Facility at the Kunming Institute of Botany, CAS, for providing plant cultivation and computing services. We thank Dr. Yuxing Xu (Kunming Institute of Botany, CAS) for discussions.

CRediT authorship contribution statement

Che Zhan: Writing – original draft, Visualization, Formal analysis, Data curation. Na Xue: Methodology. Zhongxiang Su: Software. Tianyin Zheng: Visualization, Data curation. Jianqiang Wu: Writing – review & editing, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2025.05.004.