Methanol induces cytosolic calcium variations, membrane depolarization and ethylene production in arabidopsis and tobacco

Abstract

Background and Aims

Methanol is a volatile organic compound released from plants through the action of pectin methylesterases (PMEs), which demethylesterify cell wall pectins. Plant PMEs play a role in developmental processes but also in responses to herbivory and infection by fungal or bacterial pathogens. However, molecular mechanisms that explain how methanol could affect plant defences remain poorly understood.

Methods

Using cultured cells and seedlings from Arabidopsis thaliana and tobacco BY2 expressing the apoaequorin gene, allowing quantification of cytosolic Ca²⁺, a reactive oxygen species (ROS) probe (CLA, Cypridina luciferin analogue) and electrophysiological techniques, we followed early plant cell responses to exogenously supplied methanol applied as a liquid or as volatile.

Key Results

Methanol induces cytosolic Ca²⁺ variations that involve Ca²⁺ influx through the plasma membrane and Ca²⁺ release from internal stores. Our data further suggest that these Ca²⁺ variations could interact with different ROS and support a signalling pathway leading to well known plant responses to pathogens such as plasma membrane depolarization through anion channel regulation and ethylene synthesis.

Conclusions

Methanol is not only a by-product of PME activities, and our data suggest that [Ca²⁺]_cyt variations could participate in signalling processes induced by methanol upstream of plant defence responses.

INTRODUCTION

Plant release of airborne volatile organic compounds (VOCs) in response to wounding due to pathogenic attack is well established (Engelberth et al., 2004; Baldwin et al., 2006; Beckers and Conrath, 2007) and plant volatile-mediated signalling is now considered for application in agriculture (Pickett and Khan, 2016). Wounding and herbivore attacks were notably shown to increase methanol emission levels, and methanol is considered as a signalling molecule within the plant and for plant–plant communication (de Gouw et al., 2000; Peñuelas et al., 2005; von Dahl et al., 2007; Körner et al., 2009; Dorokhov et al., 2012). The pectin demethylation directed by pectin methylesterase (PME) is probably the main source of the methanol released in the cell wall (Nemecek-Marshall et al., 1995; Fall and Benson, 1996). Cell walls are mainly composed of interacting polysaccharides, structural glycoproteins and proteoglycans, as well as proteins with multiple enzymatic activities which modify cell wall polysaccharides (Wolf and Greiner, 2012; Cosgrove, 2016). Polysaccharides are composed of three main polymers: cellulose, hemicelluloses and pectins. In the primary cell wall, cellulose is assembled into microfibrils, coated with xyloglucan interacting with other cell wall components in a complex molecular network (Carpita and Gibeaut, 1993). Thus, the plant cell wall represents a primary physical barrier against pathogens such as bacteria and fungi (Vorwerk et al., 2004; Xia et al., 2014). In the cell wall, the tightly packed arrangement of cellulose microfibrils makes it difficult to penetrate, leaving pectin as the prime target of pathogens (Vorwerk et al., 2004). Homogalacturonan, the main pectin component, is synthesized under a highly methylesterified form which can be demethylesterified by PMEs (Wolf and Greiner, 2012). The activity of PME has a significant role in the status of pectin methylesterification, which deeply affects the pectin’s properties, and could also be critical during many stages of plant development (Wolf and Greiner, 2012; Kohli et al., 2015; Li et al., 2016; Huang et al., 2017; Nguyen et al., 2017). The pectin network is effectively disassembled at many stages during plant development, such as organ abscission and fruit ripening, and PMEs play a central role in both pectin remodelling and disassembly, and in the firming and softening of the cell wall (Duan et al., 2016). Transgenic plants overexpressing a rice PME (OsPMEI28) had an increased level of cell wall-bound methylester groups and differential changes in the composition of cell wall neutral monosaccharides and lignin content leading to dwarf phenotypes (Nguyen et al., 2017). However, wounding such as herbivore attack, results in drastic de novo PME expression, correlated to a large increase in the emission of gaseous methanol after injury (Körner et al., 2009; Dorokhov et al., 2012). PMEs were shown to be of critical importance for virulence, and a higher degree of cell wall methylation correlates with disease resistance in multiple plant species (Hasunuma et al., 2003; Lionetti et al., 2007, 2017; Reignault et al., 2008; Raiola et al., 2011; Volpi et al., 2011). Pathogens could also regulate the expression of certain plant PMEs that results in increasing both susceptibility to the pathogen and methanol release (Peñuelas et al., 2005; von Dahl et al., 2006; Körner et al., 2009; Raiola et al., 2011; Lionetti et al., 2012). This is consistent with the increased resistance against herbivores of transgenic tobacco plants overexpressing PMEs and overproducing methanol (Dixit et al., 2013). PME-generated methanol release from wounded plants was shown to upregulate defence genes and active defensive reactions of neighbouring plants, allowing resistance to the bacterial pathogen (Dorokhov et al., 2012; Komarova et al., 2014).

Direct spraying of methanol onto leaves of Arabidopsis thaliana seedlings was shown to regulate the expression of 484 genes involved in signalling, defence, metabolism (especially flavonoid metabolism) and detoxification (Downie et al., 2004). Taken together, these data effectively indicate that methanol is not only a by-product of PME activity, but could also act as an interplant alarm signal. However, the mechanism by which methanol activates signalling pathways that translate the methanol signal into an output defence response has yet to be determined. The cytosolic free Ca²⁺ concentration ([Ca²⁺]_cyt) in plant cells frequently changes rapidly and dynamically in response to pathogen attacks (Lecourieux et al., 2006; Ma et al., 2011; Zhang et al., 2014). Such [Ca²⁺]_cyt variations could be promoted by various volatiles with different kinetic patterns and intensities, depending on the nature of the volatile (Asai et al., 2009; Zebelo et al., 2012). Elevations in [Ca²⁺]_cyt are sensed by Ca²⁺-binding proteins (calmodulin, calcium-dependent protein kinases or calcineurin B-like proteins) which relay or decode the encoded Ca²⁺ signals into specific cellular and physiological responses in order to survive pathogen attacks (Zhang et al., 2014). We thus tested whether methanol could induce [Ca²⁺]_cyt variations in cultured cells and seedlings and then looked to see whether these variations could be related to some classical plant defence responses, namely reacive oxygen species (ROS) generation, ion flux regulation and ethylene synthesis (Wu et al., 2014).

MATERIALS AND METHODS

Suspension cell cultures

Arabidopsis thaliana L. (Col-0) suspension cells and Nicotiana tabacum (BY-2) cell suspensions containing recombinant apoaequorin targeted to the cytosol were grown at pH 5.8 in Gamborg medium and Murashige and Skoog (MS) medium, respectively. They were maintained at 24 ± 2 °C, under continuous darkness and continuous shaking (on a gyratory shaker) at 120 rpm. Suspension cells were sub-cultured weekly using 1:10 and 1:40 dilutions for A. thaliana and tobacco, respectively. All experiments were performed at 22 ± 2 °C using log-phase cells (4 and 6–7 d after sub-culture for A. thaliana cells and tobacco cells, respectively).

Seedling culture

Arabidopsis thaliana L. ecotype Wassilewskija (Ws) seeds containing recombinant cytosolic apoaequorin were sterilized in 1 % (w/v) sodium hypochlorite, and allowed to germinate on sterilized MS agar plates containing vitamin B5 but lacking 2,4-dichlorophenoxy acetic acid (2,4-D). The seedlings were grown on the agar plates under a 12/12 h under dark/light regime (45 μmol m^–2 s^–1) for 20 d at 22 ± 2 °C.

Cytosolic calcium measurements

Cell suspensions of A. thaliana and tobacco, or A. thaliana seedlings expressing apoaequorin in the cytosol were used to record cytoplasmic Ca²⁺ variations (Knight et al., 1991). Aequorin was reconstituted by overnight incubation of the cell suspensions or seedlings in Gamborg or MS medium containing 12.5 µm native coelenterazine. For cell suspensions, 500 µL aliquots of cell suspension were directly transferred carefully to a luminometer glass tube. Treatments with non-lethal amounts of methanol (Supplementary Data Fig. S1) were performed by gentle pipette injection in the luminometer tubes. To record the effect of volatile methanol on seedlings, after aequorin was reconstituted, one seedling was carefully transferred on the upper part of a luminometer glass tube (their wet status allowing the seedling to adhere to the tube wall). Then various amounts of methanol were able to become volatile when added with the pipette on a small cotton puff placed in the bottom of the luminometer tube, which was then tightly sealed with parafilm. The luminescence counts were recorded continuously at 0.2 s intervals with a FB12-Berthold luminometer. At the end of each experiment, the residual aequorin was discharged by addition of 1 mL of a 1 m CaCl₂ solution dissolved in 100 % methanol. The resulting luminescence was used to estimate the total amount of aequorin for each condition. Calibration of calcium levels was performed using the equation: pCa = 0.332588(–logk) + 5.5593, where k is a rate constant equal to luminescence counts per second divided by total remaining counts (Knight et al., 1996). Data are expressed as micromolar and are means ± s.e.

Electrophysiology

Individual A. thaliana cultured cells were impaled in the culture medium with borosilicate capillary glass (Clark GC 150F) micropipettes (resistance: 50 MΩ when filled with 600 mm KCl). The main ion concentrations in the Gamborg medium after 4 d were 9 mm K⁺ and 11 mm NO₃^– (Reboutier et al., 2002). Individual cells were voltage-clamped using an Axoclamp 2B amplifier (Axon Instruments, Foster City, CA, USA) for discontinuous single electrode voltage clamp (dSEVC) experiments as previously described (Reboutier et al., 2002; Errakhi et al., 2008a, b; Tran et al., 2013). Voltage and current were digitized using a computer fitted with a Digidata 1320A acquisition board (Axon Instruments). The electrometer was driven by pClamp software (pCLAMP8, Axon Instruments). Data are expressed as milliVolts or percentages, and are means ± s.e.

Monitoring of ROS production

The production of ROS was monitored by the chemiluminescence of the Cypridina luciferin analogue (CLA) as previously described (Kadono et al., 2010; Tran et al., 2013). CLA is known to react mainly with O₂·^– and ¹O₂ with light emission (Nakano et al., 1986). Chemiluminescence from CLA was monitored using a FB12-Berthold luminometer with a signal integrating time of 0.2 s. The ROS scavengers 1,2-dihydroxybenzene-3,5-disulphonic acid disodium salt (Tiron) and 1,4-diazabicyclo[2.2.2]octane (DABCO) were added 20 min prior to methanol treatment.

Ethylene measurement

A 2.5 mL aliquot of tobacco cell suspension was sub-cultured in 5 mL flasks tightly closed with serum caps, maintained at 22 °C under constant shaking. After 2 h, a 2 mL gas sample was taken from each flask and injected into a gas chromatograph (GC) (Hewlet Packard 5890 series II) equipped with a flame ionization detector (FID) and an activated alumina column (6 mm in internal diameter, 50 cm long, 50–80 mesh) for ethylene determination (Errakhi et al., 2008a). Results are presented as the means of 3–6 measurements ± s.e. and are expressed as picomoles of ethylene produced per 1 g of fresh matter.

Determination of vaporized methanol

Vaporized methanol samples were analysed using a GC (Shimadzu GC-2014, Tokyo, Japan) with FIDs, fitted with a packed column (Porapak-Q 50/80), equipped with a C-R8A Chromatopac Data Processor (Shimadzu, Tokyo, Japan). Temperatures at the injection ports, column and detectors were maintained at 150, 80 and 200 °C, respectively. The flow rate of carrier gas (N₂) was 50 mL min^–1. The retention time for the methanol peak was 5.75 min. For calibration, standard methanol vapour (22.4 %, v/v) was prepared by evaporating the drops (32 µL) of liquid methanol under vacuum at 100°C in the 100 mL gas-sampling bottle. Vaporized methanol samples were mixed with fresh air and injected to a GC. In order to estimate the methanol vapour concentration derived from liquid methanol in the sealed 3.5 mL glass tubes (with plant seedlings hanging on the inner wall), a specific model container allowing both passive evaporation of methanol from the liquid methanol droplet applied onto the cotton puff, and collection of an air sample without altering the inner pressure was designed. Sampled air was used for methanol determination with GC. Evaporation of methanol (10 µL, liquid) was allowed for 0.25, 0.5, 1, 2, 5, 10, 15, 30 and 60 min, and the temporal profile of methanol evaporation was monitored.

Statistics

Significant differences between treatments were determined by the Mann–Whitney test, and P-values <0.05 were considered significant.

RESULTS

Methanol induced a biphasic variation of cytosolic Ca²⁺ in cultured cells

Addition of methanol induced a rapid biphasic variation of [Ca²⁺]_cyt composed of an early peak followed by a second slower transient increase in A. thaliana cells (Fig. 1A) and tobacco cells (Supplementary Data Fig. S3A). The first and the second peaks of [Ca²⁺]_cyt appeared to be dose dependent, both increasing after addition of 0.1–5 % methanol in A. thaliana cells (Fig. 1B). The methanol-induced first short-lived increase in [Ca²⁺]_cyt could be reduced upon a 20 min pre-treatment with a plasma membrane (PM) Ca²⁺ channel blocker, lanthanum (500 µm LaCl₃) (Fig. 1C, D; Supplementary Data Fig. S2C, D) or a Ca²⁺ chelator 1,2-bis(o-aminophenoxy)ethane-N,N,N,N-tetraacetic acid (1 mm BAPTA) (Fig.1C, D). The second peak of [Ca²⁺]_cyt was even more reduced, by about 80 %, after such pre-treatments (Fig. 1C, D; Supplementary Data Fig. S3B). In order to assess the possible involvement of Ca²⁺ intracellular stores in the methanol-induced variations in [Ca²⁺]_cyt, cells were incubated with dantrolene and 8-bromo-cADP-ribose (8-Br-cADPr), ryanodine receptor (RyR) agonists, or caffeine, that allows depletion of the intracellular Ca²⁺ store, 20 min prior to methanol treatment. The level of the second methanol-induced increase in [Ca²⁺]_cyt was inhibited by >70 % with caffeine, dantrolene and 8-Br-cADPr, whereas the first rapid variation was only slightly reduced (Fig. 1C, D). These data indicated that methanol could induce a complex calcium signalling pathway, involving different source of Ca²⁺ in plant cells.

Fig. 1.

Effect of methanol on the cytosolic Ca²⁺ concentration in cultured cells. (A) Typical changes in [Ca²⁺]_cyt measured by using cell suspensions derived from A. thaliana transformed by the apoaequorin gene upon addition of various concentrations of methanol ranging from 0.1 to 2 % (0.5–10 µL in 500 µL of Gamborg medium). (B) Mean values of Δ[Ca²⁺]_cyt for the first and the second peak depending on the volume of methanol added. *Significantly different from the control level (addition of 25 µL of H₂O). (C) Changes in [Ca²⁺]_cyt induced by 5 % (v/v) methanol after pre-treatment with the Ca²⁺ channel inhibitor (500 µm La³⁺), the calcium chelator (1 mm BAPTA) or caffeine (10 mm) allowing depletion of the Ca²⁺ internal stores. (D) Mean values of methanol-induced Δ[Ca²⁺]_cyt on the first and second peak after pre-treatment with 500 µM La³⁺, 1 mm BAPTA, 10 mm caffeine or inhibitors of endomembrane Ca²⁺-permeable channels (100 µm dantrolene and 10 µm 8 bromo-cADPr). Data correspond to the means of at least five independent experiments ± s.d. *Significantly different from the [Ca²⁺]_cyt variation induced by 1 % methanol. (A) and (C) are presented with an extended scale in Supplementary Data Fig. S2 for a better resolution of the firts peaks.

Ca²⁺ influx is an upstream event of methanol-induced depolarization of plant cells

Variations of cell polarization and anion currents are some of the earliest signalling events detectable upon plant–pathogen interactions (Errakhi et al., 2008a; Wu et al., 2014) or volatile application (Zebelo et al., 2012). Thus, methanol’s effects were further tested on cell polarization and anion currents using dSEVC on A. thaliana cultured cells. As reported by previous dSEVC studies (Errakhi et al., 2008a; Kadono et al., 2010; Tran et al., 2013), in control conditions, cultured cells display whole-cell currents being mainly carried by anion currents, which are the main hallmarks of slow anion channels (Reboutier et al., 2002) and were sensitive to unrelated anion channel inhibitors (Reboutier et al., 2002, 2005; Kadono et al., 2010). This explains the value of the resting membrane potential (V_m) of –40 ± 5 mV (n = 25) as previously reported (Reboutier et al., 2002; Haapalainen et al., 2012; Tran et al., 2013). However, in these conditions, addition of methanol induced a depolarization of the cell PM, reaching its maximal amplitude within about 1 min, followed by a repolarization to the control level (Fig. 2A). These transient variations of membrane potential (V_m) were dose dependent, being significant from 1 % methanol (Fig. 2A, B). Methanol also induced a transient increase in anion current of –1.22 ± 0.55 nA (n = 5) for 5 % methanol (Fig. 2C, D) from a control value of –0.95 ± 0.15 nA (n = 15). Upon repolarization of the cells, the anion current level decreased to the control level recorded before methanol addition (Fig. 2C). The transient regulation of these currents certainly explains the observed transient depolarization. Pre-treatment of the cells with La³⁺ or BAPTA abolished the methanol-induced increase in anion currents and depolarization (Fig. 2B, D). This indicates that the cytosolic increase in Ca²⁺ is an upstream event, when compared with methanol-induced anion current regulation and PM depolarization.

Fig. 2.

Methanol-induced transient changes in membrane potential and anion current of cultured cells. (A) Typical recording of the membrane potential (V_m) of an A. thaliana cultured cell after addition of methanol. The bar illustrates the time when methanol was added to the 500 µL of Gamborg medium. (B) Mean values of the maximal depolarization upon addition of various doses of methanol or after pre-treatment with the Ca²⁺ channel inhibitor (500 µm La³⁺) or the calcium chelator (1 mm BAPTA). (C) Typical methanol-induced changes in anion currents (recorded using dSEVC) during depolarization and repolarization, and corresponding current–voltage curves (V_h = holding potential, V_m = membrane potential). (D) Mean values of anion current increases (recorded 1.9 s after the imposition of a –200 mV pulse) at the maximal depolarization induced by 5 % methanol alone or after pre-treatment with 500 µm La³⁺ or 1 mm BAPTA. Results are presented as means of at least five measurements ± s.d. *Significantly different from the control level recorded before methanol addition.

Interaction between methanol-induced ROS generation and Ca²⁺ variations in cultured cells

The ROS generation induced by methanol was characterized using tobacco cells. Addition of methanol (1–5 %) induced a dose-dependent short-lived increase in ROS (Fig. 3A) that could be abolished upon a pre-treatment with DABCO, a scavenger of singlet oxygen (¹O₂) (Fig. 3B). This first peak could be followed by a second slower transient increase (Fig. 3B). Tiron, a scavenger of superoxide anion (O₂·^–), failed to decrease the early short-lived increase in ROS (Fig. 3B). In the set of experiments in which a second increase in ROS was recorded, this increase was reduced by DABCO but also by Tiron. The first early peak of ROS could not be reduced by La³⁺ or BAPTA, but these inhibitors decreased the second increase in ROS (Fig. 3B). These data suggest that methanol could induce a rapid increase in ¹O₂ and a slightly delayed increase in O₂·^–, the latter possibly being dependent on the early induced ¹O₂ generation and Ca²⁺ influx.

Fig. 3.

Methanol-induced reactive oxygen species (ROS) generation in BY2 cultured cells. (A) Mean values of fast ROS generation in tobacco cells upon addition of methanol ranging from 0.1 to 5 % (v/v). Typical kinetics of the fast ROS generation are presented in the inset. (B) Mean biphasic ROS generation induced by methanol (1 %) and effects of pre-treatments with a Ca²⁺ channel inhibitor (500 µm La³⁺), a calcium chelator (1 mm BAPTA), a scavenger of ¹O₂ (2.5 mm DABCO) or a scavenger of O₂·^– (Tiron 5 mm). (C) Quantification of the second ROS peak. (D) Effect of DABCO and Tiron on the changes in [Ca²⁺]_cyt induced by 3 % methanol. Results are presented as means of at least five measurements ± s.d. *Significantly different from the control level (addition of 25 µL of H₂O). (B) is presented with an extended scale in Supplementary Data Fig.S6 for a better resolution of the first peaks.

Therefore, we tested the impact of this ROS pharmacology on the methanol-induced [Ca²⁺]_cyt variations. DABCO, the scavenger of ¹O₂, almost eliminated the first and second peak of Ca²⁺ (Fig. 3C), suggesting that the rapid ¹O₂ generation could be an upstream event even when compared with the methanol-induced Ca²⁺ variations. On the other hand, Tiron, the scavenger of O₂·^–, decreased the methanol-induced increases in [Ca²⁺]_cyt (Fig. 3C), indicating that O₂·^– generation could also participate in the methanol-induced [Ca²⁺]_cyt variations, implying complex interactions between ROS generation and [Ca²⁺]_cyt variations.

Methanol induced a calcium-dependent ethylene production in cultured cells

Methanol was notably shown to induce an increase in gene expression of the ethylene biosynthetic enzyme 1-aminocyclopropane-1-carboxylate (ACC) oxidase (Downie et al., 2004). We thus investigated eventual methanol-induced ethylene synthesis and its dependence on Ca²⁺ influx in cultured cells. Accumulation of ET was observed in the cell culture flask, reaching 60 % with 1 % methanol after 4 h (Fig. 4) and 300 % with 5 % methanol (not shown). Pre-treatment of suspension cells by α-aminoisobutyric acid (AIB), an inhibitor of ACC oxidase, or by aminooxyacetic acid (AOA), an inhibitor of ACC synthase, at 200 µm each, for 4 h inhibited the ethylene synthesis induced by methanol (Fig. 4). The addition of BAPTA or La³⁺ reduced the methanol-induced ethylene synthesis, indicating that the cytosolic Ca²⁺ increase is an upstream event in the pathway leading to ethylene synthesis (Fig. 4).

Fig. 4.

Methanol-induced ethylene synthesis in cultured cells. Methanol-induced synthesis of ethylene in tobacco cells after a 4 h treatment and its inhibition by pre-treatment with 200 µm α-aminoisobutyric acid (AIB), an inhibitor of ACC oxidase, or 200 µm aminooxyacetic acid (AOA), an inhibitor of ACC synthase. Pre-treatment with a Ca²⁺ channel inhibitor (500 µm La³⁺) or a calcium chelator (1 mm BAPTA) also reduced the methanol-induced ethylene production after 4 h. Data are expressed as picomoles of ethylene produced per gram of fresh weight, and are representative of at least four independent experiments, and error bars correspond to the standard errors. *Significantly different from the control level (addition of 5 µL H₂O).

Volatile methanol induced [Ca²⁺]_cyt variations in A. thaliana seedlings

To assess further the impact of methanol on Ca²⁺ signalling in plants, we examined the time-course of the volatile methanol-induced increase in [Ca²⁺]_cyt in A. thaliana seedlings. Different amounts of liquid methanol were dropped on a cotton puff in the bottom of the luminometer tubes, allowing the release of gaseous methanol to seedlings placed on the upper part of the tubes. Volatile methanol promoted increases in [Ca²⁺]_cyt with different kinetic patterns and intensities for methanol volumes >1 µL (Fig. 5A). However, the maximal increases in [Ca²⁺]_cyt seemed to be dose dependent, along with the increased concentration of vaporized methanol determined at the 15 min time point [and not dependent on the added methanol volumes (Fig. 5B)]. Note that we have preliminarily examined the temporal profile of passive methanol vaporization from liquid methanol (10 µL) dropped onto a cotton puff placed on the bottom of a sealed tube (3.5 mL), attaining a saturated level (approx. 14 %, v/v) within an initial 1–2 min (data not shown).

Fig. 5.

Effect of volatile methanol on cytosolic Ca²⁺ concentrations of seedlings. (A) Changes in [Ca²⁺]_cyt were measured by using seedlings from A. thaliana transformed by the apoaequorin gene. Different amounts of liquid methanol (0.5–10 µL) were dropped on a cotton puff in the bottom of the luminometer tubes, allowing the release of gaseous methanol to seedlings. (B) Maximal volatile methanol-induced increases in [Ca²⁺]_cyt and vapour methanol recovered reported as a function of applied liquid methanol. Data correspond to the means of at least three independent experiments ± s.d. *Significantly different from the control level before methanol addition.

In order to discriminate between the impact of volatile methanol on aerial parts and on roots, seedlings were incised and incubated overnight with coelenterazine to adapt the root and leaf parts after mechanical damage due to incision. Although shorter, the luminescence recorded for the aerial parts appeared more important than that observed from roots (Fig. 6), suggesting that the volatile methanol-induced increase in [Ca²⁺]_cyt in seedlings was mainly due to the leaves.

Fig. 6.

Effect of volatile methanol on cytosolic Ca²⁺ concentrations in roots and aerial parts of seedlings. (A) Changes in [Ca²⁺]_cyt were measured by using only roots or aerial parts of excised seedlings from A. thaliana transformed by the apoaequorin gene. A 10 µL aliquot of liquid methanol was dropped on a cotton puff in the bottom of the luminometer tubes, allowing the release of gaseous methanol to roots or aerials parts. (B) Maximal volatile methanol-induced increases in [Ca²⁺]_cyt for roots and aerial parts. Data correspond to the means of at least three independent experiments ± s.d. *Significantly different from the control level before methanol addition.

DISCUSSION

It is well established that Ca²⁺ is a ubiquitous secondary messenger for cellular signalling in various stresses comprising plant immune responses (Lecourieux et al., 2006; Ma et al., 2011; Seybold et al., 2014). Cytosolic Ca²⁺ variations were also reported in volatile-sensing mechanisms (Walter et al., 2007; Chen et al., 2008; Asai et al., 2009; Zebelo et al., 2012; Zhang et al., 2015), but to our knowledge nothing was reported on methanol, although it was shown to upregulate defence genes and active defensive reactions (Dorokhov et al., 2012; Komarova et al., 2014). In this study, we showed that exogenous methanol could induce [Ca²⁺]_cyt increases in plant cells when applied in the liquid phase or as a volatile. The methanol released in the demethylation of pectin splits into the gas and liquid phases according to Henry’s Law, methanol emissions in leaf reflecting a dynamic balance between rates of production, phase partitioning, stomatal conductance and transpiration (Harley et al., 2002). The biphasic kinetics of the [Ca²⁺]_cyt increases recorded when methanol was applied in the liquid phase for the suspension cells (Fig. 1A; Supplementary Data Fig. S3A) or seedlings (Supplementary Data Fig. S4A) were more reproducible when compared with volatile methanol applications which displayed more chaotic kinetics (Fig. 5). Root cells seemed less sensitive than leaf cells (Fig. 6), perhaps due to the lack of an effective pathway for processing methanol in roots (Ramirez et al., 2006). However, these [Ca²⁺]_cyt increases could be minimized upon pre-treatment with the PM Ca²⁺ inhibitor La³⁺ when methanol was added in the liquid phase for suspensions cells and seedlings (Figs 1C, D and 3D; Supplementary Data Fig. S3) or as a volatile (Supplementary Data Fig. S5), indicating a rapid influx of Ca²⁺ through PM Ca²⁺ channels upon methanol challenge. The first and second peak delayed Ca²⁺ increases observed upon addition of liquid methanol could also be reduced by the Ca²⁺ chelator BAPTA, reinforcing the apoplast as a source of Ca²⁺ in this process. It is noteworthy that Ca²⁺ was shown to be mobilized from an apoplastic source to contribute to the increase in [Ca²⁺]_cyt level during PME activity (Wu and Jinn, 2010). Thus activation of PME upon pathogen attacks that leads to methanol production (de Gouw et al., 2000; Peñuelas et al., 2005; von Dahl et al., 2007; Körner et al., 2009; Dorokhov et al., 2012) and remobilization of apoplastic Ca²⁺ could lead to [Ca²⁺]_cyt increases. However, the [Ca²⁺]_cyt increases induced by methanol (mainly the second) were also reduced by dantrolene and 8 bromo-cADPr, inhibitors of endomembrane Ca²⁺-permeable channels (RyR-like channels) and by caffeine allowing depletion of the Ca²⁺ internal stores, suggesting that methanol also promotes Ca²⁺ release from intracellular stores. The inhibition of both fast and delayed Ca²⁺ increases by La³⁺ and BAPTA and the inhibition of the delayed Ca²⁺ increase by an inhibitor of internal Ca²⁺ store release suggests a possible Ca²⁺-induced Ca²⁺ release involving PM Ca²⁺ channels that induced a Ca²⁺ release from internal stores (Bewell et al., 1999), as already reported in response to various stimuli in plants (Meimoun et al., 2009a; Liu et al., 2012). It is further noteworthy that [Ca²⁺]_cyt increases were higher in A. thaliana cultured cells than in BY2 tobacco cells, suggesting a stronger sensitivity of A. thaliana cells to methanol, as already reported with thaxtomin A, a toxin inducing Ca²⁺ influx (Meimoun et al. 2009b).

Generation of ROS is postulated to be an integral part of the defence responses of the plant (Lehmann et al., 2015; Camejo et al., 2016) that frequently interacts with Ca²⁺ signalling (Kadono et al., 2010; Frederickson Matika and Loake, 2014). In response to methanol, we could effectively record a very fast dose-dependent ¹O₂ generation in suspension cells (Fig. 3) and seedlings (Supplementary Data Fig. S6). The inhibition of this fast ¹O₂ production allowed elimination of all the [Ca²⁺]_cyt variations, this ¹O₂ production thus appearing as the earliest methanol-induced response we could record. As expected from an ¹O₂-dependent [Ca²⁺]_cyt elevation, BAPTA and La³⁺ were inefficient in reducing the ¹O₂ production which is an upstream event. Such fast ¹O₂ production responsible for activation of Ca²⁺ influx through PM channels was already reported in response to salt stress in various plant cells (Monetti et al., 2014; Ben Hamed et al., 2016). The origin of this fast ¹O₂ production remains unclear, and we cannot even exclude the production of O₂·^– in this fast ROS generation, since Tiron, a scavenger of O₂·^–, could also reduce the fast [Ca²⁺]_cyt variations, although less efficiently. NADPH-oxidases such as AtRbohD and AtRbohB (for respiratory burst oxidase homologue), known to participate in plant–pathogen interactions (Pogány et al., 2009; Nagano et al., 2016), are not the source of this fast early ROS generation since it was still recorded in response to methanol in A. thaliana RBOH mutants AtrbohD and AtrbohB, and the double mutant AtrbohB-D (Supplementary Data Fig. S7A). The mechanism at the origin of this early ROS generation is still to be determined. Several cell wall-located enzymes such as polyamine oxidases (Pottosin and Shabala, 2014) or peroxidases (Kimura and Kawano, 2015) that could be responsible for extracellular ¹O₂ generation (Kawano et al., 1998; Kanofsky, 2000; Guo et al., 2009) are known to be responsible for cell wall ROS production susceptible to control of Ca²⁺ transport across the PM.

However, inhibiting this fast ROS generation also allowed inhibition of the delayed O₂·^– generation, as it could be observed with BAPTA and La³⁺. These data suggest that the delayed O₂·^– generation was dependent on the [Ca²⁺]_cyt variations, highlighting a complex interplay between generation of different ROS and [Ca²⁺]_cyt variations (Fig. 7). These data are reminiscent of what was observed upon salicylic acid treatment (Kimura and Kawano, 2015) for which apoplastic peroxidases are likely to be involved in the earlier phase of oxidative burst, and NADPH oxidases are likely to be involved in the late phase of the oxidative burst, the key signalling event connecting the two phases of oxidative burst being calcium channel activation. Although methanol alone was not shown to be active on ROS generation in tomato, only allowing the increase and prolongation of the flg22-induced oxidative burst (Hann et al., 2014), our data reinforced the hypothesis of a role for methanol in inducing a signalling pathway. Effectively, ROS generations (in addition to [Ca²⁺]_cyt variations) belong to early events responsible for most of the ensuing cascades of chemical and molecular reactions leading to plant defence response (Maffei et al., 2007). We cannot exclude that methanol-induced ROS generation may locally modify the cell wall plasticity. It has been shown that ROS, particularly hydroxyl radicals, may modify the cell wall plasticity (Fry et al., 2001; Schopfer, 2001; Tenhaken, 2015). It is also known that demethylesterified homogalacturonan can form ionic bonds between the negatively charged carboxyl groups of several homogalacturonan chains and Ca²⁺ ions, forming a pectate gel that may provide sufficient stiffness to delay the progression of the pathogen.

Fig. 7.

Putative model for early cellular events induced by pectin methylesterase-derived methanol.

We also observed that methanol could regulate anion currents through the PM, another rapid signalling event frequently detected upon plant–pathogen interactions (Errakhi et al., 2008a, b; Wu et al., 2014). Methanol could effectively induce a transient cell depolarization due to the activation of anion channels. These responses were blocked by La³⁺ and BAPTA, highlighting their dependence on [Ca²⁺]_cyt variations, as already reported in response to thaxtomin and ozone (Errakhi et al., 2008b; Kadono et al., 2010). Partly recovered rapid depolarizations have already been recorded in tomato upon challenge with leaf-emitted volatiles (Zebelo et al., 2012), but their role in VOC-induced signalling was not shown. Here we showed that the methanol-induced regulation of anion channels could participate in the pathway leading to ethylene production since it was reduced by glibenclamide and 9AC, two structurally unrelated anion channel inhibitors (Supplementary Data Fig. S8), as was already reported in response to oxalic acid (Errakhi et al., 2008a). This methanol-induced ethylene production could also be reduced by inhibitors of the ACC synthase and the ACC oxidase. These data are in accordance with the methanol-induced increase in activity or gene expression of ethylene biosynthesis enzymes by methanol and other VOCs (Arimura et al., 2002; Downie et al., 2004; Pelloux et al., 2007). In agreement with the dependence of methanol-induced anion channel regulation on [Ca²⁺]_cyt variations, the ethylene production was also sensitive to BAPTA and La³⁺, as already observed in response to oxalic acid (Errakhi et al., 2008a). Moreover this production could also be reduced by dantrolene and caffeine (Supplementary Data Fig. S8), suggesting that the [Ca²⁺]_cyt increase is a key upstream event in the pathway leading to methanol-induced ethylene production. The ethylene production was also recorded after volatile application of methanol to seedlings of various species (Supplementary Data Fig. S9), indicating that methanol could induce the same pathway in different species. Ethylene, like methanol, is recognized as a plant–plant volatile message controlling plant defences (von Dahl et al., 2007; Broekgaarden et al., 2015). In response to pathogen attack, the methanol-induced ethylene generation could thus consist of an amplification of the volatile message favouring information transfer between damaged and intact parts of the same plant and between plants, as was already reported for ethylene-induced release of VOCs thought to function as indirect defences (Horiuchi et al., 2001; Schmelz et al., 2003).

Although the amount of methanol that may accumulate in cell wall microenvironments is still to be determined, the methanol release is greatly increased through the action of pathogens (Peñuelas et al., 2005; Dorokhov et al., 2012) and it was shown to be emitted at much higher rates than all other detected VOCs upon feeding of larvae of Euphydrya saurinia on Succisa pratensis leaves (Peñuelas et al., 2005). Taken together, our data suggest that methanol could induce [Ca²⁺]_cyt variations that could participate in a signalling pathway allowing plant cell response activation upon activation of PMEs. However, complementary experiments such as phenotyping experiments are needed to ascertain the role of the different players proposed in our model of methanol-induced signalling (Fig. 7).

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Figure S1: effect of methanol on Arabidopsis thaliana cultured cell viability. Figure S2: effect of methanol on cytosolic Ca²⁺ concentration in cultured cells. Figure S3: effect of methanol on cytosolic Ca²⁺ concentration in tobacco cultured cells. Figure S4: effect of methanol on cytosolic Ca²⁺ concentration in Arabidopsis thaliana seedlings. Figure S5: inhibition of volatile methanol-induced increase in cytosolic Ca²⁺ concentrations of seedlings by La³⁺. Figure S6: methanol-induced reactive oxygen species generation in BY2 cultured cells. Figure S7: methanol-induced reactive oxygen species generation in Arabidopsis thaliana seedlings. Figure S8: methanol-induced ethylene synthesis in A. thaliana cells. Figure S9: methanol-induced ethylene production in seedlings of A. thaliana, tobacco and sunflower.