ANNEXIN1 mediates calcium-dependent systemic defense in Arabidopsis plants upon herbivory and wounding

and can exhibit Ca channel-like activity. Arabidopsis ANNEXIN1 (ANN1) is sug-gested to contribute to Ca transport. (cid:1) Here, we report that wounding and simulated-herbivory-induced cytosolic free Ca elevation was impaired in systemic leaves in ann1 loss-of-function plants. We provide evidence for a role of ANN1 in local and systemic defense of plants attacked by herbivorous Spodoptera littoralis larvae. (cid:1) Bioassays identiﬁed ANN1 as a positive defense regulator. Spodoptera littoralis feeding on ann1 gained signiﬁcantly more weight than larvae feeding on wild-type, whereas those feeding on ANN1-overexpressing lines gained less weight. Herbivory and wounding both induced defense-related responses on treated leaves, such as jasmonate accumulation and defense gene expression. These responses remained local and were strongly reduced in systemic leaves in ann1 plants. (cid:1) Our results indicate that ANN1 plays an important role in activation of systemic rather than local defense in plants attacked by herbivorous insects.


Introduction
Plants are challenged throughout their life by various abiotic and biotic stress factors. These changes in the environment require fast adaptation. Consequently, plants evolved a multilayered metabolic barrier, composed of mechanical and chemical defenses (Maffei et al., 2012;Mith€ ofer & Boland, 2012). An attack by herbivorous insects represents a major threat to the plant's survival. In particular, an attack by chewing insects is a combination of plant tissue wounding and application of insect-specific herbivore-associated molecular patterns, mainly present in their oral secretions (OSs;Mith€ ofer & Boland, 2008;Vadassery et al., 2012b;Kiep et al., 2015). The establishment of chemical defenses to such an insect herbivory is mediated by a network of signaling pathways (including calcium (Ca 2+ ) ions, protein phosphorylation, phytohormones, and reactive oxygen species (ROS) and reactive nitrogen (N) species) that finally initializes synthesis and accumulation of a plethora of defensive metabolites (Seybold et al., 2014;Zebelo et al., 2014). The elevation in cytosolic free Ca, [Ca 2+ ] cyt , is one of the earliest signaling events initiated upon the plant's interaction with feeding insects (Maffei et al., 2004;Kiep et al., 2015;Toyota et al., 2018). Jasmonates represent the most important class of wound-induced phytohormones to be activated, with the main components being jasmonic acid (JA) and its biologically active isoleucine conjugate (+)-7-iso-jasmonoyl-L-isoleucine (JA-Ile) (Wasternack, 2007;Mith€ ofer & Boland, 2008. The connection between jasmonates and [Ca 2+ ] cyt is likely mediated by Ca 2+ -sensing proteins. In plants, canonical calmodulins, calmodulin-like proteins (CMLs), calcineurin B-like proteins, and Ca 2+ -dependent protein kinases are good candidates (Swarbreck et al., 2013;Yan et al., 2018;Mohanta et al., 2019;Tai et al., 2019). In particular, a connection between [Ca 2+ ] cyt and jasmonate signaling has been shown for CML42 and CML37 (DeFalco et al., 2010;Vadassery et al., 2012a,b;Scholz et al., 2014Scholz et al., , 2016Heyer et al., 2018b).
Strikingly, Arabidopsis senses local herbivore attack and transmits this information to unwounded vascular-connected systemic leaves through a long-distance signaling system Kiep et al., 2015). Systemic signaling has also been identified leading to activation of jasmonate accumulation and signaling in distal leaves Heyer et al., 2018b). Recently, it has been shown that wound-induced electrical signals precede vascular Ca 2+ fluxes as xylem contact cells and phloem sieve elements function together for leaf-to-leaf electrical signaling (Nguyen et al., 2018). Probably, the systemic electrical signaling is mediated by GLR-type cation channels, because in glr3.3 glr3.6 double mutants the wound-activated electrical signal propagation, as well as propagation of [Ca 2+ ] cyt signals between leaves, is attenuated (Toyota et al., 2018). In addition, TPC1 was shown to be involved in systemic [Ca 2+ ] cyt elevations (Kiep et al., 2015). This channel trio of GLR3.3, GLR3.6 and TPC1 also operates in local [Ca 2+ ] cyt elevations induced by aphid feeding (Vincent et al., 2017). Very recently, it was demonstrated that a rapidly activated cyclic nucleotide-gated Ca 2+ channel (CNGC19) also plays a partial role in wounding-induced Ca 2+ influx (Meena et al., 2019). Thus, it becomes increasingly evident that herbivory-induced [Ca 2+ ] cyt elevation involves multiple channels and pathways regulating local and long-distance [Ca 2+ ] cyt signals.
Nevertheless, some studies showed that conventional channels might not always be responsible for Ca 2+ influx pathways. Thus, the involvement of other passive Ca 2+ transport-mediating proteins, such as annexins, becomes an interesting possibility , 2011Davies, 2014;Ma et al., 2019).
Annexins are found in eukaryotic organisms and form a diverse multigene superfamily of Ca 2+ -dependent membrane-binding proteins that serve as targets for Ca 2+ in most eukaryotic cells. In angiosperms, annexins are found in vegetative and generative organs (Laohavisit & Davies, 2011;Clark et al., 2012). They are composed of motifs 60-70 amino acids long, repeated four times. The ability of annexins to conduct Ca 2+ has become evident from in vivo and in vitro assays (Demidchik & Maathuis, 2007;Laohavisit et al., , 2012Richards et al., 2014;Ma et al., 2019). Unlike conventional Ca 2+ channels, which are routed from the Golgi complex to reside in a specific membrane, annexin proteins are able to occupy multiple cellular locations simultaneously. This characteristic makes annexins capable of a fast-recruitment response that can be driven by localized stimulation of membrane regions and might be independent of vesicular deliveryreviewed by , 2011 and Clark et al. (2012).
Among the eight annexins described to date in Arabidopsis thaliana, ANNEXIN1 (ANN1) is the best-studied one. It was initially detected in the cytosol of cells, and later in the plasma membrane, endoplasmic reticulum, vacuole, mitochondria, chloroplast, and in the cell wall (Laohavisit & Davies, 2011). ANN1 overexpression has a protective effect on plant survival under drought conditions, whereas lack of expression increases stress sensitivity (Konopka-Postupolska et al., 2009). Studies on Arabidopsis roots have correlated its localization in the plasma membrane with the presence of a Ca 2+ conductance, which is activated by voltage hyperpolarization and extracellular hydroxyl radicals and is involved in the elongation of root hair cells (Foreman et al., 2003;Laohavisit et al., 2012). The Arabidopsis ann1 knockout mutant lacks this Ca 2+ -channel-like activity in the plasma membranes of root epidermal cells and root hairs. Furthermore, ann1 mutants also have shorter roots compared to wild-type (Columbia-0, Col-0) plants (Laohavisit et al., 2012). More recent studies have shown that ANN1 is involved in root and seedling [Ca 2+ ] cyt elevation in response to hydrogen peroxide (Richards et al., 2014;Zhao et al., 2019).
As ANN1 is firmly implicated in [Ca 2+ ] cyt elevation and this occurs during insect feeding, our aim was to elucidate a putative role for this annexin in plant responses to herbivory-related cues. Therefore, we performed a set of assays to characterize ann1 mutants. The effect of the lack of ANN1 was analyzed by observing larval growth of the crop-pest moth Spodoptera littoralis on two different ann1 knockout and two ANN1-overexpressing lines. Further, the plant's response to mechanical injuries (i.e. mechanical wounding with and without the addition of larval OS) and after lesions caused by S. littoralis feeding on leaves was investigated. We found a role for ANN1 in Arabidopsis for both local and systemic defense responses against S. littoralis attack by mediating [Ca 2+ ] cyt elevations, jasmonate level, and defense-related gene expression. This study contributes to our understanding of the molecular identity of Ca 2+ channels involved in the plant response to wounding and herbivory.
Plants were kept in short-day conditions after stratification for 2 d at 4°C. Four to five-week-old plants grown in 10 cm round pots were used for all experiments. The growth chamber was adjusted to 50-60% relative humidity and 21°C with a 10 h : 14 h, light : dark photoperiod and a light intensity of 100 lmol m À2 s À1 . For experiments investigating the systemic response and translocation of metabolites, the leaves of each plant were counted according to their age (Dengler, 2006;Farmer et al., 2013;Kiep et al., 2015).
MecWorm (Mith€ ofer et al., 2005) treatment was used for mechanical wounding of the plant with punches every 5 s (12 punches per minute) on treated leaf 8. To investigate the systemic response upon treatment of leaf 8, the local and systemic leaves 5, 8, 9 and 13 were analyzed. Untreated plants were used as control and had the same growth and handling conditions as the treated ones.
To study the mechanical-wounding-induced systemic response, wounds were generated on leaf 8 with a pattern wheel (six vertical movements on each side of the midrib), and 20 µl of water (MW + W: mechanical wounding + water) or of S. littoralis OS, diluted 1 : 1 in water (MW + OS) was applied to the wounds (Vadassery et al., 2012a). Treated plants were kept in the growth chamber with a cover to prevent evaporation. Samples of leaf 8 and selected systemic leaves were harvested in liquid N 2 and kept at À80°C till further analysis.

Analysis of [Ca 2+ ] cyt elevations
For the analysis of [Ca 2+ ] cyt in whole plants, leaves of 4-wk-old Arabidopsis rosettes were numbered according to their phyllotactic sequence (Dengler, 2006). The day before the experiment, plants were sprayed with 10 µM coelenterazine in 0.01% (v/v) Tween 20 and incubated in the dark for 16 h for aequorin reconstitution. Aequorin imaging was performed according to Kiep et al. (2015) using a high-resolution photon-counting camera system (HRPCS218; Photek, St Leonards-on-Sea, UK) comprising an intensified CCD camera (ICCD218; Photek) and a camera controller (HRPCS4; Photek). The camera was mounted on a darkbox (DB-2; Photek). Signal acquisition and processing were performed with the IFS32 software (Photek). Photons were captured in photon-counting mode with a 200 ms frame rate, and cumulative images were integrated offline after the experiments as indicated in the figure legends. At the end of each treatment, the rosettes were flooded with 40 ml discharge solution (1 M calcium chloride, 10% (v/v) ethanol) to achieve a complete discharge of aequorin to enable calibration of the data obtained and determine the cytosolic Ca 2+ concentration according to Knight et al. (1996). The identical regions of interest (ROIs) found in treatment and discharge images were identified, and the average signal intensity in the ROIs at a given time point, as well as the cumulative counts in the ROIs, were determined by using the IFS32 software. The wounding treatment consisted of mechanically wounding the midrib of leaf 8 with a pattern wheel and adding 20 µl of water (MW + W) or 20 µl of OS (MW + OS) across all holes of the mechanically wounded leaf.

Insect material and feeding assays
Larvae of the generalist herbivore S. littoralis were hatched and reared on artificial diet at 23-25°C with 10 h : 14 h, light : dark cycles (Bergomaz & Boppr e, 1986). The OS was collected from S. littoralis larvae fed on Arabidopsis Col-0 plants and stored on ice. The OS was centrifuged at 10 000 g and 4°C to remove residual plant tissue pieces and then diluted with water (1 : 1) as previously described (Vadassery et al., 2012a).
For short-term feeding assays, third instar S. littoralis larvae were used after being kept separately without food overnight. This treatment ensured an immediate start of feeding after placement on the plant. The locally fed leaves were collected in liquid N 2 after the indicated time points and kept at À80°C until further analysis.
For local larval feeding assays, overnight-starved third-instar larvae were placed on leaf 8 for direct feeding. Each plant received one larva. After feeding on c. 40% of the leaf, which took between 5 and 10 min, the larva was removed. After 90 min, the local and systemic leaves were harvested for phytohormone extraction and gene expression analysis.
For 1 wk feeding assays, 30 first-instar larvae were placed on each of 10 Col-0, ann1 and ANN1-OE plants (three larvae per plant). To achieve similar starting conditions, all larvae determined for one plant genotype were pooled and weighed before the experiment. The minimal starting weight of 30 larvae was set to 60 mg. After 1 wk, the weight of all larvae found again was recorded separately. Owing to a limited number of first instar larvae available, the experiment was carried out several times and the weight data of each genotype were combined.

Gene expression
Total RNA was extracted from frozen material (c. 50-100 mg) using the Trizol method according to the manufacturer's protocol (Thermo Fisher, Darmstadt, Germany). Genomic DNA in total RNA samples was removed using the Turbo DNAse-free kit (Ambion, Thermo Fisher) according to the manufacturer's protocol. The integrity and amount of RNA were monitored by agarose gel electrophoresis and spectrophotometric quantification, respectively. Complementary DNAs were synthesized using the GeneAmp Core PCR RNA Kit (Applied Biosystems) according to the manufacturer's instructions. The pair of primers specific for ANN1 (AT1G35720: forward 5 0 -ATGGCGACTCTTAA GGTTTCTGAT-3 0 according to Clark et al. (2001) and reverse 5 0 -GCCTGATGACTTTCCTCTGTTCAG-3 0 ) was used, producing a product size of 151 bp. For VEGETATIVE STORAGE PROTEIN2 (VSP2; AT5G24770) we used forward 5 0 -ACGACT CCAAAACCGTGTGCAA-3 0 and reverse 5 0 -CGGGTCGGT CTTCTCTGTTCCGT-3 0 (Vadassery et al., 2012b), and for JASMONATE-ZIM-DOMAIN PROTEIN 10 (JAZ10; AT5G13220) we used forward 5 0 -TCGAGAAGCGCAAGGA GAGATTAGT-3 0 and reverse 5 0 -AGCAACGACGAAGAAGG CTTCAA-3 0 (Scholz et al., 2014). Quantitative reverse transcription PCR was performed on a CFX96 Real Time System (Bio-Rad). Brilliant II QPCR SYBR green Mix (Agilent, B€ oblingen, Germany) was used to monitor the synthesis of double-stranded DNA. Each biological sample was analyzed in technical triplicates. The cycle protocol consisted of 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min, and ended by a dissociation curve determined between 60°C and 95°C. The specificity of PCR amplifications was evaluated by the presence of a single peak in denaturation curves and by visualization of simple amplification products of expected size in ethidium bromide gel electrophoresis. The primer efficiencies were calculated using LINREGPCR (v.11.0, (Ruijter et al., 2009). The mean relative expression of the gene was calculated according to Pfaffl (2001), using DDC t with the ribosomal protein S18 gene (AT1G34030) as reference (Scholz et al., 2014)

Extraction and quantification of phytohormones
A total of 250 mg of leaf material was used for phytohormone analyses. The extraction procedure and the determination of JA and JA-Ile were performed as previously described (Vadassery et al., 2012a) with some modifications. An API5000 triple quadrupole mass spectrometer (AB Sciex, Darmstadt, Germany) was used for detection (Heyer et al., 2018a). Moreover, in this study, a different mixture of labeled jasmonates was used as internal standard. Instead of 15 ng of JA-[ 13 C 6 ]-conjugate used in the previous study, 12 ng of D 6 -JA-Ile (HPC Standards GmbH, Cunnersdorf, Germany) was used. In addition, the 60 ng of 9,10-D 2 -9,10-dihydrojasmonic acid was replaced by 60 ng of D 6 -JA (HPC Standards GmbH) as previously reported by Scholz et al. (2017).

Statistics
To ensure reproducibility, all experiments were repeated with independent biological replicates. The exact number of replicates is indicated in the particular figure legends. For statistical analyses, one or two-way ANOVA followed by post hoc (Student-Newman-Keuls; Sid ak; Tukey) tests or Student's t-tests were used as indicated in the figure legends. Different letters indicate significant differences between treatments or leaves. GraphPad PRISM 6 and ORIGINPRO 9.3 software were used for data analysis and graph composition.

ANNEXIN1 is induced by mechanical wounding and herbivory
Initially, in order to learn whether ANN1 was involved in the plant's defense response against wounding and herbivory, we implemented different wounding-related treatments on Col-0 and ann1-1 knockout plants and then analyzed ANN1 expression levels. Different treatments were performed on leaf number 8: first, mechanical wounding (MW) with a pattern wheel and applying water (MW + W) or OS of S. littoralis (MW + OS) simulating herbivory; and second, direct feeding of S. littoralis larvae on one leaf. Col-0 plants showed a high accumulation of ANN1 transcripts after all treatments, whereas in the ann1-1 mutant no induction of ANN1 was detected (Fig. 1). No significant difference between the different treatments was detected (Fig. 1a), suggesting a universal response of ANN1 expression in response to mechanical wounding, with or without the presence of larval OS, and to herbivore attack.

Local and systemic calcium signaling is affected in ann1 plants
Since, on the one hand, a [Ca 2+ ] cyt signal precedes jasmonate accumulation and subsequent defense-related responses upon wounding and herbivory (Fisahn et al., 2004;Maffei et al., 2007;Bricchi et al., 2010;Toyota et al., 2018;Meena et al., 2019) and, on the other hand, annexins are components of [Ca 2+ ] cyt signal generation, we aimed to investigate how the elevation of [Ca 2+ ] cyt upon stress is influenced by the presence or absence of the ANN1 protein. Therefore, we used Col-0 and ann1-1 plants, both containing the [Ca 2+ ] cyt reporter (apo)aequorin. The [Ca 2+ ] cyt elevation was analyzed in whole-plant rosettes according to Kiep et al. (2015). Leaf 8 was wounded with a pattern wheel before adding Fig. 1 Response of ANNEXIN1 (ANN1) in Arabidopsis thaliana after treatment with different stresses. Levels of ANN1 transcripts (AE SE) were determined on (a) the wild-type (Col-0) and (b) the ann1-1 mutant genotype after injuring leaf 8 with a pattern wheel, applying either water (MW + W) or oral secretion (MW + OS) to the wounds, or promoting Spodoptera littoralis larvae feeding (Larvae, third instar) feeding on leaf 8 until c. 40% of the leaf was eaten. Per genotype and treatment, nine replicates (n = 9) were done. All plants were incubated for 90 min before sampling leaves. Leaves of untreated plants were used as controls (Ctrl). Differences between treatments within the same genotype were analyzed using two-way ANOVA (Student-Newman-Keuls post hoc test); significant differences are indicated by different letters (P < 0.05); ns, not significant New Phytologist (2021)

Research
New Phytologist 20 µl water or OS (1 : 1 diluted) to the wounds. We observed an immediate and monophasic local elevation of [Ca 2+ ] cyt , with a peak after c. 30-45 s, and a return to background levels after c. 4 min ( Fig. 2; Supporting Information Fig. S1; Videos S1-S4). The local response was similar to those described in previous studies, but presenting a slightly faster [Ca 2+ ] cyt peak response (cf. 1 min: Verrillo et al., 2014;Kiep et al., 2015). In the case of MW + OS treatment, the Col-0 [Ca 2+ ] cyt elevation lasted longer than with the MW + W (water) treatment. Although slightly weaker than in the Col-0 plants, the local [Ca 2+ ] cyt response was clearly detectable in the wounded ann1-1 leaf and moved into the petiole. Strikingly, neither water nor OS induced a systemic [Ca 2+ ] cyt response in the ann1-1 genotype (Figs 2, S1). By contrast, upon MW + W treatment, a systemic [Ca 2+ ] cyt response (also monophasic) was observed in Col-0 plants (systemic leaf 5), which started after 4 min and reached a maximum at 4.5 min before decreasing to background levels (Fig. S1). Upon MW + OS treatment, the systemic response was more pronounced. Here, leaves 5, 6, 7, 10 and 11 responded in Col-0 (Fig. 2c). This response was highly significantly different when compared with ann1. It was possible to observe a [Ca 2+ ] cyt wave running from the treated leaf to the connected leaves in Col-0 (Videos S1, S3), whereas only the local response was detected in treated ann1 plants (Videos S2, S4). Therefore, ANN1 seems to be an important player in the systemic [Ca 2+ ] cyt wave upon wounding and herbivory challenge.
ann1 plants are more susceptible to herbivore feeding, whereas ANN1-overexpressing plants are more resistant To further evaluate the impact of ANN1 on the defense against chewing herbivores, we carried out feeding assays using first instar S. littoralis larvae on Col-0 wild-type plants and two different knockout lines (ann1-1, ann1-2) and two different ANN1 overexpressor lines (ANN1-OE12, ANN-OE10). We performed two independent sets of experiments, each with Col-0 as wild-type control, one knockout, and one overexpressor line. Our results show that larvae feeding on ann1 plants gained significantly more weight than wild-type-fed larvae (Fig. 3). This effect was also observed when we evaluated the larval growth on the APOAEQUORIN-containing ann1 mutant (ann1-1/AEQ) (Fig. S2). The opposite happened when the larvae were feeding on the ANN1-OE plant lines, where they gained significantly less weight than those feeding on Col-0 plants (Fig. 3).
Spodoptera littoralis feeding-induced jasmonate accumulation is affected in ANN1 mutant plants Jasmonates are rapidly induced upon wounding and herbivory. To test if these were differentially induced in ann1 or ANN1-OE10 lines compared with the Col-0 wild-type, we analyzed jasmonate concentrations after herbivore feeding. In independent short-term (30, 90 min) feeding assays, the levels of both JA and JA-Ile increased significantly in the treated leaves of all genotypes at both time points (Fig. 4). Even fed leaves of ann1-1 and ann1-2 plants still accumulated both JA and JA-Ile; however, their levels were significantly lower than in Col-0 plants (Fig. 4). By contrast, in the overexpressing line ANN1-OE10 the level of JA-Ile was significantly higher than in Col-0, although the JA level was not (Fig. 4c,d).
Systemic transcriptional responses to S. littoralis attack are impaired in ann1-1 plants responses was analyzed in parallel, focusing on the full knockout line ann1-1 (Fig. 1b). We performed an experiment in which S. littoralis was allowed to feed on one defined local leaf (leaf 8) followed by leaf sampling after 90 minleaf numbering according to Dengler (2006), Farmer et al. (2013), and Kiep et al. (2015). In addition to the treated leaf 8, unwounded systemic leaves 5 and 13 (vascularly connected to leaf 8) and leaf 9 (unconnected to leaf 8) were sampled and analyzed for the expression of two jasmonate-responsive genes, VSP2 and JAZ10. Compared with the nontreated Col-0 plants, VSP2 was induced 10-fold in the local leaf 8 and even more strongly in the directly vascularly connected leaf 13, whereas no change in expression was  Fig. 3 Weight of Spodoptera littoralis larvae after feeding on Arabidopsis thaliana Col-0 plants, two annexin1 (ann1) mutants, and two ANN1overexpression lines. Thirty first-instar larvae of S. littoralis were preweighed, and three larvae were placed on each plant. In each independent experiment, 10 plants per genotype were used. Larva weight was measured individually after 7 d of feeding. Each set of experiments was independently repeated: (a) n = 3; (b) n = 4. The combined total number of larvae n recovered after 1 wk is indicated. The box indicates the middle 50% of the data points; the black line within the box is the median. Whiskers are defined as 1.5-fold interquartile range; dots represent outliers. Statistical differences between the genotypes after feeding were analyzed using one-way ANOVA (Tukey's post hoc test) and indicated by different letters (P < 0.001).

Research
New Phytologist detectable in the ann1-1 mutant (Fig. 5a). Also, JAZ10 was significantly induced in the treated leaf 8 in Col-0 as well as in all systemic leaves, again with the highest expression in leaf 13. In addition, in ann1-1 plants, JAZ10 showed a significant induction in systemic leaves 5 and 13 (Fig. 5b).

Herbivory-like feeding-induced systemic jasmonate accumulation is abolished in ann1-1 plants
Considering that ANN1 is involved in jasmonate-related defense induction in systemic leaves, we wanted to study whether mechanical wounding alone employs ANN1 or if a chemical signal of the OS of the larvae is necessary. Thus, we tested the systemic jasmonate response after local leaf wounding with a pattern wheel followed by either water or OS application to the small wounds. The results show a clear local jasmonate response in Col-0 and in ann1-1 plants upon wounding and water (Fig. 6). This local response was significantly higher when wounded sites were treated with OS. Interestingly, only OS treatment induced a systemic response in the vascularly connected leaves 5 and 13 in Col-0, represented by a strong increase of JA and JA-Ile. This was not the case in ann1-1 plants, where the systemic effect was completely absent (Fig. 6). These results were supported by other measured jasmonates. The biosynthetic precursor cis-12-oxo-phytodienoic acid and the catabolite hydroxy-JA both showed very similar results to those found for JA and JA-Ile (Fig. S3).
Previous work has shown that pattern-wheel wounding does not fully represent the insect feeding-like wounding such as, for example, the use of MecWorm does (Mith€ ofer et al., 2005). Therefore, MecWorm-mediated wounding was applied in an additional experiment. Based on the studies of Heyer et al. (2018b), we wounded leaf 8, including the midrib. As shown in Fig. 7, the continuous mechanical wounding of leaf 8 inflicted by MecWorm significantly elevated the local levels of JA-Ile, which we focused on as the bioactive form of the jasmonates, in both Col-0 and ann1-1 plants. Strikingly, in contrast to Col-0, in ann1-1 plants, local accumulation of jasmonates in leaf 8 was not accompanied by a comparable systemic increase of jasmonates. Moreover, a highly significant difference in JA-Ile accumulation in directly connected leaf 13 was observed when Col-0 and ann1-1 plants were compared (Fig. 7).

Discussion
Various studies have demonstrated that a rapid, early and transient increase of [Ca 2+ ] cyt is involved and essential for the successful induction and regulation of jasmonate accumulation and further downstream plant defense strategies upon wounding and insect herbivory (Fisahn et al., 2004;Maffei et al., 2004Maffei et al., , 2007Arimura et al., 2008Arimura et al., , 2011Scholz et al., 2014;Toyota et al., 2018;Yan et al., 2018;Kumari et al., 2019;Meena et al., 2019). Such stress-induced [Ca 2+ ] cyt elevation occurs both locally and systemically (Kiep et al., 2015). There are various channels that have been shown to be involved in wounding or herbivory-related Ca 2+ influx into the cytosol such as the TPC1 (Kiep et al., 2015;Vincent et al., 2017), glutamate receptor-like channels (GLRs; Mousavi et al., 2013;Toyota et al., 2018), and cyclic nucleotidegated channel 19 (CNGC19; Meena et al., 2019). However, Fig. 5 Local and systemic transcriptional responses to Spodoptera littoralis feeding. (a) Levels of VEGETATIVE STORAGE PROTEIN2 (VSP2) and (b) JASMONATE-ZIM-DOMAIN PROTEIN 10 (JAZ10) transcripts were determined in Arabidopsis thaliana Col-0 and annexin1 (ann1-1) genotypes in different leaves after local S. littoralis (Larvae) feeding for 90 min on leaf 8 (t). Treated leaf 8 and untreated leaves 5, 9 and 13 were analyzed (mean AE SE). Per genotype and treatment, nine replicates (n = 9) were done. Leaves of untreated plants were used as controls. RPS18 was used as a reference gene to normalize the data. Statistical differences between treatments were analyzed using two-way ANOVA (Student-Newman-Keuls post hoc test); significant differences are indicated by different letters (P < 0.05). Ca 2+ influx may not be mediated only by conventional channels additional, unconventional ones such as annexins might also contribute (Laohavisit & Davies, 2011;Laohavisit et al., 2012;Davies, 2014;Ma et al., 2019). Diverse studies in plants have gathered evidence of annexins' ability to influence Ca 2+ transport. A growing body of data suggests that annexins play a role in plant response to nematode parasitism. Some cyst-secreted effectors are annexin-like and able to affect plant defense possibly by mimicking the endogenous annexin functions and impairing H 2 O 2 -induced [Ca 2+ ] cyt transients (Patel et al., 2010;Zhao et al., 2019). Moreover, it was shown that compared with wild-type, overexpression of ANN1 and ANN4 decreases susceptibility against Meloidogyne incognita nematode infection of roots while the ann1 and ann4 lines were more susceptible (Zhao et al., 2019). Our study demonstrates a role for ANN1 in systemic leaf defense responses against herbivore attack and mechanical wounding.

ANNEXIN1 is induced upon insect herbivory
Biotic interaction has been shown to influence annexin transcription in crops such as alfalfa (Medicago sativa), Indian mustard (Brassica juncea), tomato (Solanum lycopersicum), and wheat (Triticum aestivum) (Kov acs et al., 1998;Jami et al., 2009;Lu et al., 2012;Xu et al., 2016). It was shown in multiple studies that Arabidopsis ANN1 gene expression is influenced by diverse environmental signals (Konopka-Postupolska et al., 2009;Clark et al., 2010;Guelette et al., 2012). Here, we demonstrate the expression response of Arabidopsis ANN1 to herbivory. The assays were designed to understand the attack in a mechanistic and holistic way, which we achieved by dissecting the insect . Per genotype and treatment, seven replicates (n = 7) were done. Untreated leaves 5, 9 and 13 and treated leaf 8 (t) were analyzed. Leaves of untreated plants were used as controls. Statistical differences found in leaves between treatments, or between genotypes were analyzed using two-way ANOVA (*, P < 0.05; **, P < 0.005; ***, P < 0.001; Student-Newman-Keuls test). No indication, no significant difference. Fig. 7 Accumulation of (+)-7-iso-jasmonoyl-L-isoleucine (JA-Ile) in leaves of Arabidopsis thaliana after mechanical wounding with MecWorm. JA-Ile levels were analyzed in Col-0 and annexin1 (ann1-1) genotypes in different leaves after mechanical wounding for 90 min, including the midrib. In treated plants, leaf 8 (t) was subjected to mechanical damage; untreated leaves 5, 9 and 13 and treated leaf 8 were analyzed (mean AE SE). Per genotype and treatment, six to eight replicates (n = 6-8) were done. Leaves of untreated plants were used as controls. Statistical differences between leaves of control and treated plants on the same genotype and between Col-0 and ann1-1 leaves upon MecWorm treatment were analyzed using an unpaired Student's t-test (*, P < 0.05; **, P < 0.005

Research
New Phytologist attack into different modules of stress: mechanical wounding alone (MW + W), mechanical wounding plus OS (MW + OS), or the complex stress of larval feeding (Fig. 1). The high expression of ANN1 upon the different but complementary stresses showed that ANN1 transcription activation is triggered quickly (here, after 90 min) by insect-feeding-related damage, but also by wounding alone. The latter confirms earlier results showing ANN1 induction 24 and 48 h after wounding (Konopka-Postupolska et al., 2009).

ANNEXIN1 is involved in the defense against herbivores
As with other cellular components that are involved in [Ca 2+ ] cyt signaling (Vadassery et al., 2012b;Scholz et al., 2014;Meena et al., 2019), we show here that ANN1 is also an important player in the regulation of insect-feeding-induced defense. Using two different knockout and two different overexpression lines, we showed that S. littoralis larvae feeding on ann1 plants gained substantially more weight (in total, +27.6%), whereas those feeding on ANN1-OE plants were much smaller (À26.9%) than larvae feeding on Col-0 plants (Figs 3, S2). Thus, ANN1 is a positive regulator in herbivory-induced defense in Arabidopsis and contributes to resistance against S. littoralis, comparable to CNGC19 and GLRs.
To gain further insight into the putative role of ANN1 in systemic defense-related signaling, we performed a series of experiments in which a defined local leaf was treated and systemic leaves were analyzed Kiep et al., 2015). Using (apo)aequorin-expressing ann1-1 and wild-type plants, we demonstrated that ANN1 is indispensable for the systemic [Ca 2+ ] cyt response; this is comparable to what has been found previously for TPC1 and GLRs (Kiep et al., 2015;Toyota et al., 2018). By contrast, the local response was slightly but not significantly reduced in the ann1-1 mutant (Figs 2, S1). However, compared with other studies (Nguyen et al., 2018;Toyota et al., 2018), the systemic [Ca 2+ ] cyt response was found to be weaker. This can be explained by different experimental conditions; first, in the other studies, the GCaMP3 fluorescent-protein-based [Ca 2+ ] sensor was used, a highly sensitive calcium fluorescence reporter, whereas we used bioluminescent aequorin. Second, their mode of wounding was much harsher. Whereas Nguyen et al. (2018) destroyed half of the leaf tissue, we used a pattern wheel that caused only a few small holes in the tissue. This supports the view that the intensity of wounding correlates with the intensity of the response (Nguyen et al., 2018).
As a readout for the defense response, the expression of jasmonate-responsive genes VSP2 and JAZ10 was examined. Upon larval feeding, a local increase was detected in Col-0 for both genes, as well as a strong increase in the vascularly connected leaf 13 that was even higher than the local response (Fig. 5). Such a strong systemic increase was not found for either VSP2 or JAZ10 in ann1-1 plants, suggesting that the feeding-related signals necessary to induce the systemic gene activation do not reach the distal leaves. The fact that in this and other experiments leaf 13 and (to a lesser extent) leaf 5 respond more strongly can be explained by the direct (leaf 13) and indirect (leaf 5) vascular connections to leaf 8 (Dengler, 2006).
As [Ca 2+ ] cyt elevations initiate Ca 2+ signaling Mith€ ofer & Boland, 2012;Scholz et al., 2014;Vadassery et al., 2014) and precede downstream signals, such as phytohormone accumulation, we further investigated the concentrations of the jasmonates JA and JA-Ile at two time points (30 and 90 min). These jasmonates were found to be strongly induced locally in both Col-0 and ann1 upon herbivore attack; but accumulation was significantly lower in ann1 plants, whereas in the ANN1-OE10 at least the JA-Ile level was significantly higher (Fig. 4). This is in accordance with the finding that [Ca 2+ ] cyt signals in ann1 plants are affected in local leaves to a certain extent and suggests that a full induction of jasmonates in response to insect herbivores is not possible in those plants. Strikingly, the induction of ANN1 upon herbivory or wounding (Fig. 1) is high if compared with downstream responses such as jasmonate or gene induction. Very likely, the reason behind this finding is that gene expression does not always reflect the corresponding protein expression. Lee et al. (2004) already noted that ann1-1 transcript level does not necessarily correspond to the ANN1 protein.
Besides using larvae, we further evaluated the phytohormonal responses in local and systemic leaves of plants treated with mechanical wounding. We chose this approach because such experiments can be better standardized than exposure to larvae that might feed or not, particularly when only a short time period is investigated. Pattern-wheel-mediated mechanical wounding, as well as wounding by MecWorm, supported the finding of local jasmonate induction in wild-type and mutant (Figs 6, 7). Strikingly, after MW + OS treatment, the response of JA-Ile in ann1-1 plants was significantly reduced compared with wild-type (Fig. 6). In addition, an elevation of jasmonates was also observed in systemic leaves of OS-treated plants in wild-type but not in ann1-1 mutant plants. This further indicates that ANN1 is involved in systemic [Ca 2+ ] cyt -dependent jasmonate elevation. This notion was supported by results obtained by mechanical wounding alone, using MecWorm treatment. In ann1-1 plants, leaf 13 showed significantly lower accumulation of JA-Ile than in Col-0 plants (Fig. 7), suggesting that ANN1 might not only be involved in systemic OS-specific signaling but also in systemic wound-induced jasmonate accumulation.
Conflating the data obtained, we propose that ANN1 is a positive factor of the [Ca 2+ ] cyt -dependent systemic defense response against wounding and larval feeding. Thus, the absence of this unconventional Ca 2+ channel causes an impaired systemic response. ANN1 can exist as an integral plasma membrane protein (Alexandersson et al., 2004;Marmagne et al., 2007). Nevertheless, as small amphipathic proteins, annexins are distributed throughout cells and can be transported within the plant via the phloem (Guelette et al., 2012). It is possible that annexins may be recruited directly to membranes, independently of vesicle delivery, to operate in stimulus-specific signaling (Laohavisit & Davies, 2011;Laohavisit et al., 2012;Davies, 2014;Espinoza et al., 2017). Therefore, a plant that contains the conventional and unconventional Ca 2+ channels should be able to recruit annexins to the tissues under stress when needed, whereas the ann1 genotype can only launch reduced [Ca 2+ ] cyt -mediated defense responses. It should be kept in mind that annexins might also act via the regulation of channel activities, for example by selective channel delivery to or retraction from membranes, in a similar way as the KAT1 plasma membrane potassium ion (K + ) channel is cycled during abscisic acid induced stomatal closure (Sutter et al., 2007). However, the observed effect of abolishing systemic [Ca 2+ ] cyt -induced responses in ann1 plants may be specific for certain stresses, such as herbivory, and is not necessarily involved in all types of biotic and abiotic stress responses. An example is a study showing that ANN1 was not necessary for systemic signaling and development of acquired resistance in uninfected leaves during challenge with avirulent bacteria Pseudomonas syringae pv tomato (Carella et al., 2016).
The success of the plant's defense response against stresses does not depend on one single signaling component alone. Instead, it is a coordinated local and systemic communication between cells and distant organs. For that, various signals, such as ROS, hydraulic pressure, as well as electropotential waves, [Ca 2+ ] cyt and others, are employed in a tightly linked manner (Foreman et al., 2003;Zimmermann et al., 2009;Maischak et al., 2010;Farmer et al., 2013;Mousavi et al., 2013;Davies, 2014;Seybold et al., 2014;Ranjan et al., 2015;Peiter, 2016;Alonso et al., 2019;Gully et al., 2019;Saijo & Loo 2019). In previous studies using epidermal root tissue, hydroxyl-radical-activated plasma membrane conductance of Ca 2+ and K + were absent in the ann1-1 mutant (Laohavisit et al., 2012). Expression and protein levels of Arabidopsis ANN1 also correlated with the occurrence of the radical-activated plasma membrane Ca 2+ conductance in the root epidermis and at the apex of root hairs (Clark et al., 2001;Dinneny et al., 2008). These results strongly suggest that ANN1 is very likely a Ca 2+ -permeable protein in Arabidopsis and might provide a molecular link between ROS and [Ca 2+ ] cyt in the systemic defense-related signaling in plants. Further studies will address this hypothesis.
In conclusion, we investigated the role of ANN1 in local and systemic plant defense against wounding and herbivorous insects in Arabidopsis. Plant tissue wounding and cell disruption caused by feeding insects strongly induced ANN1 expression, demonstrating that it is part of the rapid defense response against invertebrate pests; neither jasmonates nor defense-related genes were upregulated systemically in ann1 mutants. ANN1 mediates plant defense, affecting larval growth, and is crucial for the induction of signaling upon herbivory within the whole plant. ANN1 is an important part of systemic [Ca 2+ ] cyt signaling, thereby connecting [Ca 2+ ] cyt to subsequent downstream signals and defense responses against herbivores.

Supporting Information
Additional Supporting Information may be found online in the Supporting Information section at the end of the article.   Video S4 Time-lapse of [Ca 2+ ] cyt response in Arabidopsis thaliana ann1-1 rosette induced by mechanical wounding + water of the leaf lamina.
Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.