Calcium in plant defence-signalling pathways

Authors

  • David Lecourieux,

    1. UMR INRA 1088/CNRS 5184/Université de Bourgogne, Plante-Microbe-Environnement, 17 rue de Sully, BP 86510, 21065 Dijon cedex, France
    Search for more papers by this author
  • Raoul Ranjeva,

    1. UMR CNRS 5546 Université Paul Sabatier, Signaux et Messages Cellulaires chez les Végétaux, Pôle de Biotechnologies Végétales, 24 chemin de Borde Rouge, BP 17, Auzeville, 31326 Castanet-Tolosan, France;
    Search for more papers by this author
  • Alain Pugin

    1. UMR INRA 1088/CNRS 5184/Université de Bourgogne, Plante-Microbe-Environnement, 17 rue de Sully, BP 86510, 21065 Dijon cedex, France
    Search for more papers by this author

Author for correspondence: Alain Pugin Tel: +33 03 80 69 31 77 Fax: +33 03 80 69 32 26 Email: pugin@dijon.inra.fr

Abstract

Contents

  • I. Introduction 250
  • II. Increases in free cellular [Ca2+]: an early and key event in plant defence signalling 251
  • III. Relationships between [Ca2+]cyt increases and defence-signalling pathways 256
  • IV. Calcium targets involved in plant defence reactions 258
  • V. Concluding remarks 262
  • Acknowledgements 263

  • References 263

Summary

In plant cells, the calcium ion is a ubiquitous intracellular second messenger involved in numerous signalling pathways. Variations in the cytosolic concentration of Ca2+ ([Ca2+]cyt) couple a large array of signals and responses. Here we concentrate on calcium signalling in plant defence responses, particularly on the generation of the calcium signal and downstream calcium-dependent events participating in the establishment of defence responses with special reference to calcium-binding proteins.

Abbreviations
[Ca2+]cyt

cytosolic free calcium concentration

CaM

calmodulin

CBL

calcineurin B-like proteins

CIPK

CBL-interacting protein kinase

CDPK

Ca2+-dependent protein kinases

OG

oligogalacturonides

ER

endoplasmic reticulum

HR

hypersensitive response

ABA

abscisic acid

[Ca2+]nuc

nuclear free calcium concentration

IP3

1,4,5-trisphosphate

ROS

reactive oxygen species

NO

nitric oxide

cADPR

cyclic ADP ribose

LEAC

large conductance elicitor-activated ion channel

MAPK

mitogen-activated protein kinase

GFP

green fluorescent protein

SV channel

slow vacuolar channel

H2O2

hydrogen peroxide

CNGC

cyclic nucleotide-gated ion channel

SA

salicylic acid

SIPK

SA-induced protein kinase

WIPK

wound-induced protein kinase

PICBP

pathogen-induced CaM-binding protein

Rboh

respiratory burst oxidase homologue

NOS

nitric oxide synthase

PAL

phenylalanine ammonia lyase

TMV

Tobacco mosaic virus

PLD

phospholipase D

BcPG1

endopolygalacturonase 1

VICC

voltage-independent calcium channel.

I. Introduction

The calcium ion is now firmly established as a second messenger in numerous plant signalling pathways, conveying a wide range of environmental and developmental stimuli to appropriate physiological responses. Under resting conditions, the cytosolic free calcium concentration ([Ca2+]cyt) is maintained between 100 and 200 nm (Bush, 1995), 104 times less than that in the apoplastic fluid and 104 to 105 less than that in cellular organelles, providing the potential for the ready import of Ca2+ into the cytosol where it acts as a second messenger. [Ca2+]cyt is rigorously regulated by the coordination of passive fluxes (Ca2+ channels) and active transport (Ca2+-ATPases and Ca2+-antiporters) across the plasma membrane and/or endomembranes, and the buffering capacity of the cytosol (Bush, 1995; Sanders et al., 1999, 2002).

Changes in [Ca2+]cyt have been reported in response to various signals (Table 1), including hormones, light, abiotic stress and microbial elicitors (Sanders et al., 1999; Reddy, 2001; Rudd & Franklin-Tong, 2001; White & Broadley, 2003). How cells encode/decode the Ca2+ signals produced by a plethora of biotic and abiotic stimuli is one of the most fascinating questions in cell biology. During the past two decades, studies on both animal and plant cells provide evidence that the spatial and temporal changes in [Ca2+]cyt (referred to as ‘calcium signature’) caused by a given signal contribute to the specificity of the biological outcome (Dolmetsch et al., 1998; McAinsh & Hetherington, 1998; Trewavas & Malho, 1998; Knight, 2000; Allen & Schroeder, 2001; Ng & McAinsh, 2003; Hetherington & Brownlee, 2004). These changes may proceed as single calcium transients, oscillations, or repeated spikes with specific subcellular location, lag time, amplitude and frequency. Moreover, the same signal induces different calcium signatures depending on the organ, the tissue, or the cell type in a tissue (McAinsh & Hetherington, 1998; Reddy, 2001; Moore et al., 2002; Hetherington & Brownlee, 2004). [Ca2+]cyt elevation may be caused by an uptake of Ca2+ from the extracellular medium, or by Ca2+ mobilization from organelles, and/or by both. The origin of calcium signals may be important in the physiological response (Knight et al., 1996; van der Luit et al., 1999). The information encoded in transient Ca2+ changes is decoded by an array of Ca2+-binding proteins giving rise to a cascade of downstream effects, including altered protein phosphorylation and gene expression patterns (Rudd & Franklin-Tong, 2001; Sanders et al., 2002). In plants the many Ca2+-binding proteins fall into two main classes, referred to as sensor relays and sensor responders, respectively (Luan et al., 2002; Sanders et al., 2002; Reddy & Reddy, 2004). Sensor relays such as calmodulin (CaM), CaM-related proteins and calcineurin B-like proteins (CBL) function through bimolecular interactions. They undergo a conformational change induced by Ca2+ before interacting with and changing the activity or structure of the target proteins. Sensor responders such as the Ca2+-dependent protein kinases (CDPK) function at first through intramolecular interactions and undergo a Ca2+-induced conformational change that alters the protein's own activity or structure (Harmon et al., 2000; Zhang & Lu, 2003; Harper et al., 2004). These two modes of decoding calcium signals are used extensively in plants to provide a series of regulatory modules allowing flexibility and diversity of biological responses triggered by calcium. Moreover, these calcium targets should be close to calcium channels because of the low diffusion of calcium. This observation suggests the presence of precisely localized and complex molecular networks able to decode specific [Ca2+]cyt increases. Collectively, the available data indicate that specificity in the Ca2+-signalling system results from a multifactorial decision process, ranging from a specific Ca2+ signature to the availability of a specific set of calcium sensors and their target proteins that are coupled to downstream components in a precise place.

Table 1.  Examples of [Ca2+]cyt perturbations reported in plant cells in responses to abiotic or biotic stresses (modified from White & Broadley, 2003)
Stimulus[Ca2+]cyt responseReferences
  1. Additional examples of [Ca2+]cyt perturbations and the location of stores releasing Ca2+ into the cytosol associated with various developmental processes and environmental challenges are extensively reported by White & Broadley (2003).

PathogensBiphasicKnight et al. (1991); Mithöfer et al. (1999)
(1) Slow spike (min)Blume et al. (2000); Grant et al. (2000); Fellbrich et al. (2000); Klüsener et al. (2002)
(2) Sustained elevation (h)Lecourieux et al. (2002); Poinssot et al. (2003)
(3) Oscillations in [Ca2+]cyt 
Oxidative stress(1) Brief [Ca2+]cyt spikePrice et al. (1994); Levine et al. (1996); Clayton et al. (1999); Allen et al. (2000); Kawano & Muto (2000); Klüsener et al. (2002); Lecourieux et al. (2002)
(2) Sustained [Ca2+]cyt elevation 
(3) Oscillations in [Ca2+]cyt 
Heat shockElevated [Ca2+]cyt sustained for 15–30 minGong et al. (1998); Mahlóet al. (1998)
Cold shock(1) Single brief [Ca2+]cyt spike (s)Knight et al. (1991); Mahlóet al. (1998); Knight (2000)
(2) Oscillations in [Ca2+]cyt 
Red lightElevated [Ca2+]cytShacklock et al. (1992); Mahlóet al. (1998)
Blue lightBrief spike in [Ca2+]cyt (s)Mahlóet al. (1998); Baum et al. (1999)
UV-BSlow [Ca2+]cyt rise, elevated [Ca2+]cyt sustained for several minutesFrohnmeyer et al. (1999)
AnoxiaBiphasicSubbaiah et al. (1994); Sedbrook et al. (1996)
(1) Slow [Ca2+]cyt spike (min) 
(2) Sustained [Ca2+]cyt elevation (h) 
Hypo-osmotic stressBiphasicTakahashi et al. (1997); Mahlóet al. (1998); Knight (2000); Pauly et al. (2001)
(1) Small [Ca2+]cyt elevation 
(2) Large [Ca2+]cyt elevation 
Drought-hyper-osmoticBiphasic stressKnight et al. (1997); Kiegle et al. (2000)
(1) Slow [Ca2+]cyt spike (min)Mahlóet al. (1998); Cessna & Low (2001); Pauly et al. (2001)
(2) Sustained [Ca2+]cyt elevations (h) 
Mechanical stimulationBrief [Ca2+]cyt spike (s)Knight et al. (1991); Haley et al. (1995); Legue et al. (1997); Plieth (2001)
Tissue [Ca2+]cyt wave 
Abscisic acidOscillation in [Ca2+]cytMcAinsh et al. (1992); Allen et al. (2000); Schroeder et al. (2001)
Elevated [Ca2+]cyt at cell periphery 
Elevated [Ca2+]cyt around vacuole 

Transient changes in permeability of the plasma membrane to Ca2+ are a common early event in plant defence signalling (Atkinson et al., 1996; Jabs et al., 1997; Wendehenne et al., 2002), and Ca2+ plays a pivotal role in activating the plant's surveillance system against attempted microbial invasion (Nürnberger & Scheel, 2001). Thus elicitor-responsive Ca2+-permeable ion channels located at the plasma membrane of plant cells mediate elicitor-induced Ca2+ entry, resulting in transient changes in [Ca2+]cyt. Recent descriptions of Ca2+ signatures induced in plant cells in response to various elicitors provide precious information about the role of Ca2+ in mediating biotic signals (Table 1). A better understanding of the physiological significance of the [Ca2+]cyt changes in a number of plant–pathogen interactions will involve the identification and functional analysis of downstream targets of cytosolic Ca2+. However, most intracellular target proteins that sense and relay Ca2+ signatures toward the appropriate defence responses remain to be identified. This review summarizes the state of knowledge regarding generation of the calcium signal and the downstream calcium-dependent events involved in the establishment of defence responses. Data highlighting the key role of calcium in response to Nod factors during symbiotic interactions have been covered by recent reviews (Oldroyd & Downie, 2004; Bothwell & Ng, 2005) and are not discussed here.

II. Increases in free cellular [Ca2+]: an early and key event in plant defence signalling

1. Ca2+ signatures in different plant cell–elicitor systems

Early pharmacological and 45Ca2+-based approaches led to the conclusion that activation of defence responses depends on Ca2+ influxes from the apoplast into the cytosol of plant cells (Cramer et al., 1985; Kurosaki et al., 1987; Stäb & Ebel, 1987; Atkinson et al., 1990; Conrath et al., 1991; Mathieu et al., 1991; Nürnberger et al., 1994; Tavernier et al., 1995). However, our knowledge about the intracellular Ca2+ homeostasis in plant cells, and particularly its modulation during plant defence responses, remained limited until the development of optical methods for in vivo[Ca2+] monitoring and imaging (Plieth, 2001). All the methods described for monitoring cell calcium variations have advantages and drawbacks, as mentioned in various reviews (Rudd & Franklin-Tong, 1999; Plieth, 2001; Mithöfer & Mazars, 2002). Ca2+-sensitive dyes have been used to analyse [Ca2+]cyt fluctuations occurring during plant–pathogen interactions (Messiaen et al., 1993; Xu & Heath, 1998). However, many side-effects encountered with the fluorescent probes, such as their general resistance to entry into plant cells, their buffering capacity or toxicity, greatly limited their application. Nowadays, most calcium signalling studies on plant cells are performed using the aequorin technology based on bioluminescence. Aequorin is a Ca2+-binding photoprotein found in jellyfish (Fig. 1a; Knight et al., 1991; Mithöfer & Mazars, 2002), and aequorin-transformed plant cell suspensions have been exploited to monitor and compare specific [Ca2+]cyt fingerprints in different plant cell–elicitor systems. However, in these experimental conditions the [Ca2+] changes are the average of many cells and do not necessarily reflect the [Ca2+] fluctuations in individual cells. The fluorescence resonance energy transfer (FRET)-based calcium indicators, cameleons, initially used to monitor calcium oscillations in guard cells (Allen et al., 2000), should improve the resolution of the elicitor-induced [Ca2+] variations in single cells (Plieth, 2001).

Figure 1.

Measuring intracellular free calcium concentrations in plant cells using the aequorin technology. (a) Mechanism of light emission by aequorin upon Ca2+ binding. Aequorin is a Ca2+-binding photoprotein composed of an apoprotein (apoaequorin, 22 kDa) and a prosthetic group consisting of a luciferin molecule named coelenterazine. After spontaneous reconstitution with the hydrophobic luminophor coelenterazine in the presence of molecular oxygen, the functional holoprotein aequorin behaves as a calcium-sensitive luminescent protein. The binding of three Ca2+ ions per aequorin molecule triggers a conformational change of the holoprotein, which therefore acts as an oxygenase. This causes coelenterazine to be oxidized to coelenteramide, accompanied by the release of CO2 and concomitant emission of blue light detected with a luminometer. (b) Targeting aequorin into plant cells. Although aequorin can be introduced into plant cells using direct microinjection, it is generally introduced into plants by stable transformation techniques. One advantage of the transgene aequorin approach is that the probe can be stably expressed in the cytoplasm, but also targeted to different organelles (dotted arrows 1–5), subcellular compartments or tissues by appropriate targeting signal sequences or fusion constructs (for technical details see Plieth, 2001; Mithöfer & Mazars, 2002).

Using Nicotiana plumbaginifolia cells expressing aequorin in the cytosol, Lecourieux et al. (2002) monitored changes in [Ca2+]cyt after treatments with cryptogein (a proteinaceous elicitor secreted by the oomycete Phytophthora cryptogea) or oligosaccharidic elicitors (Fig. 2). The [Ca2+]cyt signatures observed with each elicitor differ in term of intensity, kinetics and duration. In N. plumbaginifolia cells challenged with 1 µm cryptogein (Fig. 2a), a lag phase preceded a 6-min transient and rapid [Ca2+]cyt increase, which peaked at 2.4 µm after 5 min, then decreased to 0.35 µm. The first peak was followed immediately by a second [Ca2+]cyt increase, which reached 0.75 µm at 20 min after the beginning of the treatment, then decreased slowly but did not return to the background level even after 2.5 h treatment. In comparison, after a lag phase of 20 s, oligogalacturonides (OG, a mixture of oligomers with degrees of polymerization ranging from 7 to 20, used at 100 µg ml−1) treatment, induced a first transient increase in [Ca2+]cyt that peaked at 1.3 µm, within 60–90 s before decreasing (Fig. 2a). A second transient increase in [Ca2+]cyt occurred 4 min after the beginning of treatment, with a maximum value of 0.9 µm. Then [Ca2+]cyt returned to the resting value within 15–20 min. All three other oligosaccharidic elicitors – laminarin (1 mg ml−1), a β-1,3-glucan from the brown algae Laminaria digitata (Klarzynski et al., 2000); chitooligosaccharides (50 µg ml−1); and lipopolysaccharides (100 µg ml−1) from Pseudomonas (Boller, 1995; Müller et al., 2000) induced a signature resembling the OG response (Fig. 2c,d). Interestingly, Ca2+ uptake measured by 45Ca2+ accumulation increased linearly within cells challenged with cryptogein, reaching 50 nmol Ca2+ per 0.1 g FW after a 2.5 h treatment (Fig. 2b). Only 0.9% of the calcium entering into the cells was present as free calcium in cytosol after a 30-min treatment. This ratio fell to 0.05% after 2.5 h, illustrating the ability of cells to buffer and store calcium in organelles through particularly active vacuolar and endoplasmic reticulum (ER) Ca2+-ATPases and Ca2+/H+ antiporters. Efflux systems also involve plasma membrane Ca2+-ATPases that catalyse the removal of Ca2+ from the cytosol to the apoplast (White & Broadley, 2003).

Figure 2.

Effects of different elicitors on [Ca2+]cyt and [Ca2+]nuc changes in aequorin-transformed tobacco cells (from Lecourieux et al., 2002, 2005). (a–d) [Ca2+]cyt elevations in elicitor-treated tobacco cells; (e) [Ca2+]nuc elevations in elicitor-treated tobacco cells. (a) Treatment with H2O, 1 µm cryptogein (Cry) or 100 µg ml−1 OGs. (b) 45Ca2+ influx and [Ca2+]cyt changes in response to 1 µm cryptogein during a 3-h treatment. (c) Treatment with 1 µm cryptogein (Cry) and 0.1 or 1 mg ml−1 laminarin (LAM). (d) Treatment with 100 µg ml−1 OGs, 100 µg ml−1 lipopolysaccharides (LPS) or 50 µg ml−1 chitopentaoses (CP). (e) Tobacco cells were treated with 100 nm cryptogein, 100 µg ml−1 oligogalacturonides (OGs) or 1 mg ml−1 laminarin.

After a lag phase of approx. 60 s, two transient peaks were also observed in aequorin-transformed soybean cells treated with a Phytophthora sojae-derived β-glucan elicitor (200 µg ml−1) or chitin fragments (200 ng ml−1) (Mithöfer et al., 1999). Both elicitors enhanced the cytosolic Ca2+ concentration from resting concentrations as low as 0.1 µm to highest levels of approx. 2 µm. Another elicitor-induced Ca2+ response was described using the parsley cell suspensions/Pep-13 system (Blume et al., 2000). Pep-13 corresponds to a peptide of 13 amino acids within a cell-wall transglutaminase from the phytopathogenic oomycete P. sojae (Nürnberger et al., 1994). After a lag phase of 30–40 s, treatment with the Pep-13 (100 nm) of parsley cells, stably expressing aequorin, revealed a rapid increase in [Ca2+]cyt, which peaked at 1 µm after 2 min and subsequently decreased to a slowly declining plateau of 300 nm during the next 10–40 min (Blume et al., 2000). Similarly, Grant et al. (2000) used aequorin-expressing Arabidopsis plants to investigate the role of [Ca2+]cyt as an important early signal during the gene-for-gene interaction between the bacterial pathogen Pseudomonas syringae pv. tomato harbouring either the avrRpm1 or avrB avirulence gene, and Arabidopsis thaliana carrying the matching resistance gene RPM1. Infiltration of Arabidopsis leaves with concentrated bacterial inoculum (5 × 108 cfu ml−1) carrying either avrRpm1 or avrB produced an early transient increase in [Ca2+]cyt, lasting for approx. 10 min. Then a second and sustained [Ca2+]cyt increase appeared, reaching a maximum after 2 h and prior to the onset of cell collapse characteristic of the hypersensitive response (HR). Hu et al. (2004) observed that oligogalacturonic acid (50 µg ml−1) stimulated a rapid, substantial and transient [Ca2+]cyt elevation in A. thaliana seedlings, whereas Gerber et al. (2004) described a weak and transient increase in [Ca2+]cyt in Nicotiana tabacum cell suspension treated with lipopolysaccharides (100 µg ml−1) isolated from the outer cell wall of Burkholderia cepacia. The endopolygalacturonase 1 (BcPG1) from Botrytis cinerea was characterized as a potent elicitor of defence response in grapevine (Poinssot et al., 2003). When applied to grapevine cells transformed with the gene encoding aequorin, BcPG1 (5 µg g−1 FW cell) induced a biphasic and sustained elevation of [Ca2+]cyt (Vandelle et al., 2006).

In plants, [Ca2+]cyt signalling can take place through complex patterns such as waves and oscillations (Trewavas & Malho, 1998; Rudd & Franklin-Tong, 1999; Allen & Schroeder, 2001). Oscillations in [Ca2+]cyt were observed, especially in plant cells treated with auxin, abscisic acid (ABA), or blue light during circadian rhythms and under anoxia (Felle, 1988; McAinsh et al., 1995; Sedbrook et al., 1996; Wood et al., 2001). Using the stomatal guard cells as a model system, Allen et al. (2000) clearly demonstrated that the hydrogen peroxide (H2O2)-induced [Ca2+]cyt oscillations observed in these cells truly encode information leading to subsequent stomatal closure. These results constitute the first and unique demonstration in plants that the calcium signature contains encrypted information that could be decoded into a specific biological end response (Scrase-Field & Knight, 2003). If [Ca2+]cyt oscillations were reported in Arabidopsis guard cells after treatment with chitosan or yeast elicitors (Klüsener et al., 2002), most studies indicate that plant cells use preferentially transient elevations in [Ca2+]cyt to transduce responses to elicitor treatment. Moreover, these data show that: (1) increase in [Ca2+]cyt is an early and rapid event in the elicitor-sensing mechanism of plant cells; and (2) each elicitor tested triggers specific [Ca2+]cyt increases, giving typical calcium signatures that differ in both kinetics (lag time, peak time, duration) and peak intensities (Blume et al., 2000; Grant et al., 2000; Lecourieux et al., 2002). A sustained [Ca2+]cyt increase is a common characteristic of the Ca2+ signatures induced by cryptogein (1 µm) and Pep-13 (100 nm) in tobacco and parsley cells, respectively, whereas this sustained elevation in [Ca2+]cyt is not observed in response to different oligosaccharidic elicitors. Similarly, a sustained [Ca2+]cyt elevation is observed following the avrRpm1/RPM1 interaction in Arabidopsis. Long-lasting [Ca2+]cyt elevations have been reported to be involved in hypersensitive cell death (Binet et al., 2001; Lecourieux et al., 2002).

Several authors took advantage of the possibility to target the aequorin probe to specific organelles (Fig. 1b; Kendall et al., 1992; Rizzuto et al., 1992; Johnson et al., 1995; Pauly et al., 2001; Sai & Johnson, 2002; Logan & Knight, 2003). This strategy highlighted the physiological significance of changes in free calcium in organelles such as chloroplasts, mitochondria and the nucleus (Johnson et al., 1995; Baum et al., 1999; Goddard et al., 2000; Logan & Knight, 2003). For instance, increases in nuclear free calcium concentration ([Ca2+]nuc) have been reported to occur in plants in response to physical stimuli such as wind, stretch, temperature (van der Luit et al., 1999; Xiong et al., 2004), and osmotic constraints (Pauly et al., 2001). The interplay between nuclear and cytosolic calcium in elaborating a global calcium signature and the elicitation of biological responses has been established (van der Luit et al., 1999; Pauly et al., 2001). The importance of nuclear calcium in signalling processes is underlined by the existence of calcium effectors in the plant nucleus, including CaM, CaM-binding protein, CDPK and Ca2+-CaM-regulated protein phosphatase (Li et al., 1991; Andreeva & Kutuzov, 2001; Dammann et al., 2003; Lee et al., 2003; Bouchéet al., 2005; Levy et al., 2005). It was also shown that Ca2+-ATPase and some components of the phosphoinositide signalling pathway localized to the plant nucleus (Downie et al., 1998; Dröbak & Heras, 2002). Additionally, it has been demonstrated that the plant nucleus is able to generate its own calcium increases, even in an acellular environment (Pauly et al., 2000; Xiong et al., 2004). Using tobacco cell suspensions, we compared the effects of two types of elicitor: proteins leading to necrosis, including harpin and four elicitins; and non-necrotic elicitors, including flagellin (flg22) and two oligosaccharidic elicitors, namely OGs and laminarin (Lecourieux et al., 2005). The proteinaceous elicitors induced a pronounced and sustainable [Ca2+]nuc elevation, relative to the small effects of oligosaccharidic elicitors (Fig. 2). This [Ca2+]nuc elevation, which seems insufficient to induce cell death, is unlikely to result directly from the diffusion of calcium from the cytosol. The [Ca2+]nuc rise depends on free cytosolic calcium, 1,4,5-trisphosphate (IP3), and reactive oxygen species (ROS), but is independent of nitric oxide (NO). The physiological significance of these elicitor-induced [Ca2+]nuc rises remains to be determined. One would expect that nuclear calcium should be responsible for the activation of Ca2+-dependent proteins in the nucleus, and involved in the regulation of nuclear activities such as gene expression (White & Broadley, 2003). Recently, Levy et al. (2005) identified an Arabidopsis gene, IQD1 (IQ-DOMAIN 1), which encodes a novel calmodulin-binding nuclear protein. Interestingly, IQD1 is proposed to integrate intracellular Ca2+ signals towards stimulation of plant defences, including accumulation of the secondary metabolites glucosinolates.

2. Sources of Ca2+ in elicitor-induced [Ca2+] signatures

In eukaryotic cells, various stimuli mobilize different pools of calcium to trigger characteristic changes in [Ca2+]cyt, which are transduced through various Ca2+ channels on different cell membranes (Sanders et al., 2002). Ca2+-permeable channels have been characterized in plant cells, but their physiological roles and probable cooperation in response to various stimuli are poorly understood. Ca2+ channels have been detected in the plasma membrane, vacuolar membrane, ER, chloroplast and nuclear membranes of plant cells (White, 2000). These channels are classified after their voltage dependence, and may be regulated by stretch, interactions with the cytoskeleton, or binding of ligands [IP3, cyclic ADP ribose (cADPR)] or covalent modifications. The electrophysiological properties of all the known Ca2+ channels have been reviewed (White, 2000).

Role of extracellular calcium and different classes of plasma membrane Ca2+ channels in elicitor-induced [Ca2+]cyt signatures  All the studies performed with the aequorin technology indicate that elicitor-induced [Ca2+]cyt elevations predominantly result from a continuous Ca2+ influx through the plasma membrane (Mithöfer et al., 1999; Blume et al., 2000; Lecourieux et al., 2002; Hu et al., 2004; Vandelle et al., 2006). Many Ca2+-permeable channels have been found in plant plasma membranes and have been shown to be implicated in cell signalling (White & Broadley, 2003). Most of these channels are not strictly selective for Ca2+ ions, and facilitate the transport of others cations (Demidchik et al., 2002a; Very & Sentenac, 2002). So far, the elicitor-stimulated Ca2+-permeable channels mediating these Ca2+ entries are still poorly characterized.

Patch-clamping of parsley protoplasts revealed a Ca2+-permeable ion channel of large conductance in the plasma membrane, a large conductance elicitor-activated ion channel (LEAC). The LEAC is specifically activated in a rapid and reversible way on addition of Pep-13 without significant changes in membrane potential (Zimmermann et al., 1997). As similar macroscopic ion fluxes have been detected in other plants on elicitor treatment, the existence of functional homologues of LEAC in these species can be anticipated (Zimmermann et al., 1997). Gelli et al. (1997) reported the activation of a hyperpolarization-dependent Ca2+-permeable current in the plasma membrane of tomato protoplasts in response to race-specific fungal elicitors. Pharmacological studies suggested that the activity of this elicitor-activated Ca2+-permeable channel is modulated by a heterotrimeric G protein-dependent phosphorylation of the channel protein. Different voltage-dependent Ca2+-permeable channels activated by membrane depolarization were identified on the plant plasma membrane (Thuleau et al., 1994; Pineros & Tester, 1997; White, 2000; Davenport & Tester, 2001). A Ca2+-permeable channel that belongs to the two-pore channel family (TPC1) was reported to be involved in cryptogein-induced Ca2+ elevations and defence responses in tobacco BY-2 cells (Kadota et al., 2004). Recently, the Japanese group has demonstrated the involvement of a two-pore channel (OsTPC1) in elicitor-induced defence responses in rice cells. OsTCP1 controls sensitivity to the elicitor and plays a key role in regulating activation of a mitogen-activated protein kinase (MAPK) cascade and hypersensitive cell death (Kurusu et al., 2005). Bombardment of onion epidermal cells with the green fluorescent protein (GFP)–OsTCP1 fusion protein suggested that OsTCP1 functions at the plasma membrane (Kurusu et al., 2005). However, Peiter et al. (2005) showed that the TCP1 gene of Arabidopsis encodes a class of Ca2+-dependent Ca2+-release channel that is known as the slow vacuolar (SV) channel. Transient expression of AtTCP1–GFP fusion protein in Arabidopsis protoplasts revealed a vacuolar membrane localization. This was further confirmed by Western blot using microsomal membranes and intact vacuoles, probed with antibodies against TCP1 (Peiter et al., 2005).

Several studies reported a rapid elicitor-induced plasma membrane depolarization resulting, at least in part, in the activation of anion channels (Ward et al., 1995; Pugin et al., 1997). However, the relationship between Ca2+ influx and anion efflux is still unclear. In few cases, the elicitor-induced Ca2+ influx is inhibited by different anion-channel blockers, suggesting that the anion flux precedes and controls Ca2+ entries (Ward et al., 1995; Ebel & Mithöfer, 1998). A possible function of these anion channels might be to initiate or amplify plasma membrane depolarization, which in turn may activate Ca2+ voltage-dependent channels. By contrast, Ca2+ influx was shown to be a prerequisite for the activation of plasma membrane anion channels in several systems (Ward et al., 1995; Jabs et al., 1997; Wendehenne et al., 2002). In cryptogein-treated tobacco cells, the major calcium influx did not result from plasma membrane depolarization. Instead, the Ca2+ influx occurred upstream and triggered anion efflux and plasma membrane depolarization, which in turn may mobilize some Ca2+ voltage-dependent channels (Pugin et al., 1997; Wendehenne et al., 2002).

Recently, it has been shown that H2O2 produced by tobacco cells in response to elicitors (Lecourieux-Ouaked et al., 2000) participates in [Ca2+]cyt increase, probably through the activation of H2O2-sensitive Ca2+ channels located to the plasma membrane (Lecourieux al., 2002). Consistently, H2O2 was previously shown to trigger calcium influx in tobacco (Price et al., 1994; Takahashi et al., 1998; Kawano & Muto, 2000), and [Ca2+]cyt increase was reported to be involved in ROS-mediated cell death (Levine et al., 1996). Moreover, recent studies have reported that ABA activates a hyperpolarization-dependent Ca2+-permeable channel in the plasma membrane of Arabidopsis guard cells, leading to Ca2+ influx and to an increase in [Ca2+]cyt (Hamilton et al., 2000; Pei et al., 2000). It has been demonstrated that ABA increases the level of ROS in an NAD(P)H-dependent manner, and that ROS stimulate a hyperpolarization-activated Ca2+ influx current in the plasma membrane termed ICa (Pei et al., 2000). In good agreement, Foreman et al. (2003) nicely showed that NADPH oxidases control development by making ROS that regulate plant cell expansion through the activation of hyperpolarization-activated Ca2+ channels. Interestingly, Klüsener et al. (2002) found that yeast elicitor and chitosan, both eliciting plant defence responses, also activate this ICa current in Arabidopsis guard cells, and that this activation requires cytosolic NAD(P)H.

Finally, recent data illustrate the involvement of cyclic nucleotide-gated ion channels (CNGC) in plant defence (Talke et al., 2003; Yoshioka et al., 2006). Interestingly, the absence of functional AtCNGC2, a plasma membrane CNGC permeable to Ca2+ (Clough et al., 2000), characterizes the Arabidopsis dnd1 mutant (DND for defence no death). The dnd1 mutant fails to produce the HR in response to avirulent P. syringae but expresses systemic acquired resistance constitutively (Yu et al., 1998). The Arabidopsis HLM1/DND2, another member of the CNGC family (AtCNGC4), is also an essential signalling component in the HR (Balague et al., 2003; Jurkowski et al., 2004). Klessig's laboratory previously identified the A. thaliana mutant constitutive expresser of PR gene22 (cpr22), which displays constitutive activation of multiple defence responses (Yoshioka et al., 2001). Recently, the same group showed that the phenotype conferred by cpr22 is attributable to a deletion that generates a chimeric CNGC-encoding gene designated AtCNGC11/12 (Yoshioka et al., 2006). The analysis of the Arabidopsis genome sequence revealed the presence of 20 genes encoding putative CNGCs (Talke et al., 2003), suggesting the possible involvement of additional CNGCs in the Ca2+-dependent signalling pathways leading to defence responses.

Contribution of organelles as internal pools of Ca2+ in elicitor-induced [Ca2+]cyt signatures  Although external Ca2+ appears to be the main provider for elicitor-induced [Ca2+]cyt changes, these data do not rule out the implication of internal Ca2+ released in the cytosol from organelles. Calcium channels located on vacuolar or ER membranes, and ligands gated by IP3 or cADPR, are involved in many physiological processes in plants (Sanders et al., 1999), especially in ABA signalling (Wu et al., 1997; Leckie et al., 1998). In tobacco cells challenged with cryptogein, and in parsley cells treated with Pep-13, preincubation with neomycin (a phospholipase C antagonist that inhibits IP3-mediated Ca2+ release) significantly reduced the first transient [Ca2+]cyt elevation, whereas the sustained [Ca2+]cyt increase is not affected, suggesting the contribution of IP3-dependent internal Ca2+ release to the transient [Ca2+]cyt peak (Blume et al., 2000; Lecourieux et al., 2002). In soybean cells treated with β-glucans and in tobacco cells challenged with OGs, neomycin drastically inhibits the second transient [Ca2+]cyt peak without affecting the first one (Mithöfer et al., 1999; Lecourieux et al., 2002). Using grapevine cells, Vandelle et al. (2006) showed the involvement of Ca2+ from intracellular stores in BcPG1-induced [Ca2+]cyt elevations. These results suggest that at least two different Ca2+ stores might be involved in the elicitor-induced [Ca2+]cyt elevations. In addition, the participation of cADPR-dependent Ca2+ channels is suggested in various defence responses (Durner et al., 1998; Mithöfer et al., 1999; Blume et al., 2000; Klessig et al., 2000; Lamotte et al., 2004).

III. Relationships between [Ca2+]cyt increases and defence-signalling pathways

Recognition and perception of elicitors by plant cells through specific receptors lead to modulation of the defence-signalling cascades. These signalling pathways include the regulation of protein kinases, protein phosphatases, phospholipases, G-proteins, NADPH oxidases, ion channels, and the production of various metabolites such as reactive oxygen intermediates, nitric oxide, salicylic acid (SA), ethylene and jasmonic acid, in an orderly and timely fashion, to activate defence gene expression (Nürnberger & Scheel, 2001).

1. Receptor-mediated activation of the [Ca2+]cyt responses in elicitor-treated plant cells

Recent data showed a receptor-mediated activation of the [Ca2+]cyt signatures in elicitor-treated plant cells. The sequence of events triggered by cryptogein includes the high-affinity binding of the elicitor on plasma membrane glycoprotein(s) of tobacco cells (Wendehenne et al., 1995; Bourque et al., 1998, 1999). Taking advantage of four different elicitins that bind with comparable affinities to the same binding sites, and that trigger the same effects but with different magnitudes (Bourque et al., 1999), Lecourieux et al. (2002) monitored [Ca2+]cyt increases during competition assays in vivo using the most efficient (cryptogein) and the less efficient (cinnamomin) elicitins. Increasing concentrations of cinnamomin revealed a shift of the cryptogein-induced Ca2+ signature towards the cinnamomin-induced Ca2+ signature. These results indicate that the typical [Ca2+]cyt increases induced by cryptogein depend on specific interactions with the high-affinity binding sites characterized previously.

In parsley cells, activation of elicitor-induced reactions is mediated through binding of Pep-13 to a 100-kDa plasma membrane receptor protein (Nürnberger et al., 1994). Using a series of structural derivatives of Pep-13, Blume et al. (2000) nicely demonstrated that the Pep-13-stimulated increase in [Ca2+]cyt is a receptor-mediated process.

2. Protein phosphorylation and MAPK activation

Post-translational modification of proteins by reversible phosphorylation is a key process regulating many functions in plants, including defence responses induced by elicitors (Dietrich et al., 1990; Felix et al., 1991). Modifications of the phosphorylation status of proteins have often been reported during elicitor treatment (Lecourieux-Ouaked et al., 2000; Peck et al., 2001). In several plant cell–elicitor systems, phosphorylation events were described both upstream and downstream of the elicitor-induced Ca2+ influx. For example, defence-related responses including the early and large Ca2+ influx are prevented by the general serine/threonine protein kinase inhibitor staurosporine in cryptogein-treated tobacco cells. In contrast, calyculin A, a serine/threonine protein phosphatase type 1 and 2A inhibitor, partially mimicked elicitation in the absence of elicitor (Lecourieux-Ouaked et al., 2000). These data suggest the involvement of protein phosphorylation/dephosphorylation event(s) between the cryptogein recognition (possibly the receptor itself) and the first Ca2+-permeable channel activated in response to the elicitor. Accordingly, Lecourieux et al. (2002) showed that the cryptogein-induced [Ca2+]cyt increase is fully inhibited by staurosporine. Moreover, [Ca2+]cyt signature might be involved in the Ca2+ influx-dependent phosphorylation of 12 polypeptides detected in cryptogein-treated tobacco cells (Lecourieux-Ouaked et al., 2000). The Ca2+-dependent phosphorylation of these 12 polypeptides could involve directly (e.g. CDPK or CBL-interacting protein kinase, CIPK) or indirectly (e.g. MAPK) Ca2+-regulated protein kinases.

Nühse et al. (2003) identified synthaxin among the phosphorylated proteins isolated from plasma membrane in Arabidopsis cells elicited with the flagellin peptide (flg22). Synthaxin is phosphorylated in a Ca2+-dependent manner and may be a primary target of kinases on calcium flux induced by flg22. Because synthaxins are known to be involved in membrane fusion and exocytosis, the authors suggested that the Ca2+ signal might stimulate exocytosis of defence-related proteins and compounds.

MAPK cascades are major components downstream of receptors or sensors that transduce extracellular stimuli into intracellular responses in eukaryotic cells. Several MAPK cascades were shown to be associated with the induction of plant defence responses (Zhang & Klessig, 2001; Jonak et al., 2002). Pharmacological approaches indicated the involvement of calcium signalling upstream of MAPK pathways in plant defence responses (Lebrun-Garcia et al., 1998; Romeis et al., 1999; Link et al., 2002). In aequorin-transformed tobacco cells, both [Ca2+]cyt elevations and MAPK (SA-induced protein kinase, SIPK and wound-induced protein kinase, WIPK) activation induced with OGs did not exceed 15 min. By contrast, cryptogein treatment led to a fast and sustained MAPK activation for at least 2 h, which superimposed to the cryptogein-induced [Ca2+]cyt increase. Suppression of the sustained [Ca2+]cyt increase in cryptogein-treated cells abolished the activation of both MAPKs. Taking into account that both SIPK and WIPK in vitro activities did not require calcium, these data led to the conclusion that the sustained activation of SIPK/WIPK was caused by Ca2+-dependent events upstream of the MAPK module, and/or by a Ca2+-dependent inhibition of negative regulators of MAPK (Lecourieux et al., 2002). These data are in accordance with recent work by Kurusu et al. (2005) showing that overexpression of a putative voltage-gated Ca2+ channel in rice resulted in enhanced activation of a MAPK cascade on elicitor treatment. In contrast, Lee et al. (2001) reported that harpin, produced by the bean halo-blight pathogen P. syringae pv. phaseolicola, induces the activation of a MAPK in N. tabacum cv. Samsung without extracellular Ca2+ requirement. Similarly, in BcPG1-treated grapevine cells, MAPK activation does not act downstream of [Ca2+]cyt variations (Vandelle et al., 2006).

3. ROS and NO production

The rapid production of ROS, termed oxidative burst, is an early inducible plant reaction during pathogen invasion or on treatment with elicitors (Wojtaszek, 1997). Whereas Ca2+ fluxes and protein phosphorylation are required for the controlled generation of H2O2 (Neill et al., 2002), [Ca2+]cyt-independent H2O2 accumulation has also been reported (Chandra et al., 1997; Mithöfer et al., 2001).

In fact, most defence events depend on Ca2+ influx, but downstream events may amplify Ca2+ signalling by increasing Ca2+ influx and/or Ca2+ release from internal stores. For instance, H2O2 production from the Ca2+-dependent activation of a NADPH oxidase (Pugin et al., 1997; Keller et al., 1998; Torres et al., 1998) could trigger a Ca2+ influx (Price et al., 1994; Levine et al., 1996; Takahashi et al., 1998; Kawano & Muto, 2000; Pei et al., 2000). The relationship between the NADPH oxidase-dependent H2O2 production and a subsequent [Ca2+]cyt elevation was reported in the cryptogein/tobacco system (Lecourieux et al., 2002), and clearly demonstrated through stimulation of an NADPH-dependent, hyperpolarization-activated Ca2+ influx in Arabidopsis guard cells treated with yeast elicitors or chitosan (Klüsener et al., 2002). Interestingly, an emerging sum of evidence suggests that the ROS-mediated plasma membrane Ca2+ channel activation may be of more general importance in plant signal transduction and development. Indeed, the AtRbohD and AtRbohF (for Arabidopsis respiratory burst oxidase homologue) NADPH oxidases have a dual function in contributing to pathogen-associated ROS production and in mediating ABA signal transduction (Torres et al., 1998; Kwak et al., 2003).

Generation of NO was also found to be one of the early events involved in the activation of plant defence after pathogen attacks (Klessig et al., 2000; Wendehenne et al., 2001). To date, the relationships between both messenger Ca2+ ions and NO molecules become more apparent. When added to tobacco cells, cryptogein induced a very fast Ca2+-dependent NO production (Foissner et al., 2000; Lamotte et al., 2004). Characterization of a plant NOS (AtNOS1, for A. thaliana nitric oxide synthase) has revealed that the enzyme contains CaM-binding motifs and that the full activation of NOS requires both Ca2+ and CaM (Guo et al., 2003). Moreover, elicitor-induced NO production can be involved in [Ca2+]cyt increases (Durner et al., 1998). Lamotte et al. (2004) showed that cryptogein-induced NO production participates in [Ca2+]cyt changes observed in tobacco cells on cryptogein treatment. Their data suggest that NO promotes [Ca2+]cyt elevations through the mobilization of intracellular pools of Ca2+, potentially via cyclic guanilate monophosphate- or cADPR-gated Ca2+-permeable channels (Wendehenne et al., 2004). Similarly, Vandelle et al. (2006) reported the key role of NO production in grapevine cells treated with BcPG1. In this model, NO accumulation depends on [Ca2+]cyt, and NO is involved in [Ca2+]cyt homeostasis by regulating Ca2+ fluxes across the plasma membrane and activating Ca2+ release from internal stores. Further investigation is needed to identify Ca2+-permeable channels modulated directly or indirectly by NO production, and to understand the complexity of the relationships between Ca2+ and NO in the signalling cascade leading to plant defence activation.

4. Gene expression, phytoalexin production and cell death

There are only a few examples of direct links between calcium signatures and biological outcomes in plants, most using the stomatal guard cells as a model system (Scrase-Field & Knight, 2003; Hetherington & Brownlee, 2004). Concerning defence-associated responses, the data suggest the involvement of [Ca2+]cyt changes in defence gene expression, phytoalexin accumulation and HR-related cell death.

Changes in gene expression have been reported in numerous plant–pathogen interaction systems (Rushton & Somssich, 1998). These ‘defence genes’ encode a broad range of proteins with different functions, such as the pathogenesis-related PR proteins with antimicrobial activities (Stintzi et al., 1993), and different enzymes involved in the phenylpropanoid biosynthetic pathways leading to the production of phytoalexins and lignins. Uptake of Ca2+ from the extracellular medium is usually required for elicitor-induced phytoalexin production (Kurosaki et al., 1987; Stäb & Ebel, 1987; Ebel, 1995; Tavernier et al., 1995). Using aequorin-expressing plant cells, Mithöfer et al. (1999) and Blume et al. (2000) demonstrated that the elicitor-induced increase in [Ca2+]cyt is causally involved in accumulation of phytoalexins. A prolonged [Ca2+]cyt elevation appeared to be correlated with phytoalexin accumulation in elicitor-stimulated soybean or parsley cells (Mithöfer et al., 1999; Blume et al., 2000). Accordingly, Lecourieux et al. (2002) showed that suppression of the sustained [Ca2+]cyt increase in cryptogein-treated tobacco cells suppressed the accumulation of transcripts corresponding to phenylalanine ammonia lyase (PAL), the gene encoding the enzyme committed in the first specific biosynthetic step leading to phenylpropanoid derivatives. Moreover, the fact that PAL mRNA levels did not change in OG-treated cells reinforced the possibility that the second sustained [Ca2+]cyt increase, triggered specifically by cryptogein but not by oligosaccharidic elicitors, is a determinant for defence gene expression.

In animals, Ca2+ is a key regulator of cell survival, but Ca2+ can also induce cell death under various pathological conditions (Hajnoczky et al., 2003). The dual role of Ca2+– survival factor or ruthless killer – is under intensive investigation. It is becoming clear that cellular Ca2+ overload or perturbation of intracellular Ca2+ compartmentalization and cytosolic homeostasis can cause cytotoxicity and trigger either apoptotic or necrotic cell death (Orrenius et al., 2003). The HR in plants is a programmed cell death often associated with disease resistance (Dangl et al., 1996). Using the cowpea–cowpea rust fungus pathosystem, Xu & Heath (1998) found that a slow and prolonged elevation in [Ca2+]cyt is involved in establishment of the HR. Similarly, Grant et al. (2000) reported sustained [Ca2+]cyt elevation in Arabidopsis cells during the avrRpm1/RPM1 gene-for-gene interaction, this prolonged increase in [Ca2+]cyt being necessary for induction of the hypersensitive cell death. In tobacco cells challenged with cryptogein, hypersensitive cell death was related to the intensity of calcium influx and to microtubule depolymerization (Binet et al., 2001). Moreover, the expression of hsr203J, a gene associated with hypersensitive cell death (Marco et al., 1990; Pontier et al., 1994; Tronchet et al., 2001) and the manifestation of cell death, are both strictly dependent on the sustained [Ca2+]cyt elevation in cryptogein-treated cells (Lecourieux et al., 2002). In soybean cells, Levine et al. (1996) described a signal function for Ca2+, downstream of the oxidative burst in the activation of a physiological cell death program that appears similar to apoptosis in animals. The induction of the hypersensitive cell death of tobacco cells that followed treatments with the elicitin INF1 required Ca2+ (Sasabe et al., 2000). Overexpression of a Ca2+-permeable channel in rice led to enhanced sensitivity to elicitor, resulting in hypersensitive cell death (Kurusu et al., 2005). Finally, the analysis of Arabidopsis mutants that failed to produce HR cell death in response to avirulent pathogen infection, or that showed aberrant regulation of cell death, led to the identification of cyclic nucleotide-gated channels (AtCNGC2 and AtCNGC4, respectively) that mediate the entry of Ca2+ (AtCNGC2) and other cations (Clough et al., 2000; Balague et al., 2003; Jurkowski et al., 2004). Collectively, these data suggest that the control of ion permeation, including Ca2+, through CNG channels may be a critical component of HR and defence responses (Talke et al., 2003). Recently, it has been speculated that CNG channels belong to the class of voltage-independent calcium channels (VICCs), which are open at all physiological membrane potentials (Demidchik et al., 2002a, 2002b; White & Davenport, 2002; White, 2004). It has been suggested that VICCs/CNGCs maintain the basal Ca2+ influx necessary for [Ca2+]cyt homeostasis, and thereby contribute to Ca2+-signalling processes indirectly rather than directly (White, 2004). In the future, an important task will be to investigate the possible relationships between the CNGC and Ca2+ signalling in order to understand how this ion channel family is involved in hypersensitive cell death and mediates defence responses.

IV. Calcium targets involved in plant defence reactions

A large number of Ca2+ sensors have been characterized in plants, including calmodulin, CBL and Ca2+-regulated protein kinases (Reddy, 2001). These Ca2+-binding proteins contain helix–loop–helix motif(s) called EF-hand domains that bind Ca2+ with high affinity (Zielinski, 1998). In addition, several other plant proteins bind Ca2+ but do not contain EF-hand motifs (Reddy, 2001). Whereas accumulating data report that changes in free Ca2+ concentration play a key second messenger role in plant defence-signalling pathways, only few of the Ca2+-binding proteins involved in the corresponding cascades have been identified. In this section we provide information on the Ca2+-binding proteins implicated in plant defence responses (Fig. 3).

Figure 3.

Schematic representation of Ca2+ sensors and their target proteins known to be part of defence-signalling pathways. On pathogen perception, increases in [Ca2+]cyt can be decoded by various Ca2+ sensors including calmodulins and Ca2+-dependent protein kinases. Once activated by Ca2+ ions, these Ca2+ sensors interact with downstream effectors that modulate numerous biochemical and cellular functions involved in defence responses (section IV).

1. Calmodulin and calmodulin-binding proteins

Calmodulin is a highly conserved and broadly distributed Ca2+-binding protein, which acts as a multifunctional intermediary connecting Ca2+ signals to the activation of other cellular components (Zielinski, 1998; Chin & Means, 2000). CaM isoforms are small acidic proteins that undergo conformational changes on binding of four Ca2+ ions to their four EF-hand domains. On binding Ca2+, hydrophobic surfaces are exposed and allow the Ca2+-CaM complex to interact with, and to regulate the activity of, an array of target proteins. In animal cells, changes in [Ca2+]cyt regulate CaM by both targeting its subcellular distribution and inducing a variety of conformational states in CaM that result in specific target activation (Chin & Means, 2000).

A prominent feature of CaM in higher plants is the expression of multiple isoforms (Reddy, 2001; Luan et al., 2002). One mechanism by which plant cells may transduce Ca2+ signals to build specific physiological responses involves the differential expression of these CaM isoforms. Heo et al. (1999) demonstrated that specific CaM isoforms are involved in plant defence against fungal pathogens and Tobacco mosaic virus (TMV). Expressed at a low level in control soybean cells, two divergent CaM isoforms (SCaM-4 and -5) are highly induced either by a fungal elicitor or by pathogen attack, whereas three other SCaM genes encoding conserved CaMs are not inducible. Transgenic tobacco plants overexpressing SCaM-4 and -5 showed spontaneous lesions and constitutive expression of systemic acquired resistance-associated genes independently of SA production. Moreover, these transgenic plants exhibit enhanced resistance to a wide spectrum of pathogens, suggesting that specific CaM isoforms are part of an SA-independent signalling cascade leading to disease resistance. Similarly, Yamakawa et al. (2001) studied the expression profile of 13 tobacco CaM genes in response to pathogen infection and wounding. They showed a predominant accumulation of NtCaM-1, -2 and -13 transcripts and NtCaM-13 protein in tobacco leaves infected with TMV. In addition, the authors showed a specific degradation of NtCaM proteins by proteasome in response to wounding. Similar results provided by Bergey & Ryan (1999) revealed an accumulation of CaM mRNA and CaM protein in tomato plants after wounding or systemin treatment. Collectively, these data indicate that the level of individual CaM proteins is differentially regulated both transcriptionally and post-transcriptionally in plants on exposure to pathogen infection and/or wounding. Moreover, Ca2+-CaM interactions with specific target proteins depend on CaM isoforms, explaining differential transduction pathways of the Ca2+ signatures towards specific Ca2+-dependent defence responses (Reddy, 2001).

CaMs transmit Ca2+ signals by interacting with a number of target proteins, and several CaM-binding proteins have been identified in plants (Reddy, 2001; Luan et al., 2002; Bouchéet al., 2005). However, the roles of CaM-binding proteins in the induction of defence responses are still poorly known (Bouchéet al., 2005). Harding et al. (1997) identified NAD kinase as a CaM-binding protein involved in the generation of the elicitor-induced oxidative burst. These authors suggested that Ca2+/CaM-dependent NAD kinase activation enhances ROS production by increasing the NAD(H)/NADP(H) ratio. Reddy et al. (2003) isolated a new protein, PICBP (for pathogen-induced CaM-binding protein) that binds CaM in a Ca2+-dependent manner. The expression of PICBP in Arabidopsis is induced in response to several avirulent pathogen species, and in response to inducers of plant defence such as SA and H2O2. Furthermore, PICBP is constitutively expressed in the accelerated-cell-death22 Arabidopsis mutant (acd22). These data suggest a role for PICBP in Ca2+-mediated defence signalling and cell death.

The CNG channels may be regulated by Ca2+-CaM, as suggested by the existence of a high-affinity CaM-binding site overlapping the cyclic nucleotide-binding domain (Arazi et al., 2000). Talke et al. (2003) proposed a model for the integration of CNG channels in plant signalling networks, in which CNGCs constitute a link between cyclic nucleotide and Ca2+ signals. In this model, Ca2+-activated CaM is assumed to bind to and to inhibit CNGCs, allowing tight regulation of the Ca2+ signal.

Absence of mildew resistance locus o (MLO), a transmembrane CaM-binding protein in barley, leads to disease resistance against the pathogenic powdery mildew and to a deregulated leaf cell death, indicating that MLO has a negative regulatory function in plant defence and cell death (Buschges et al., 1997; Kim et al., 2002a, 2002b). Interestingly, loss of CaM binding lowers the ability of MLO negatively to regulate defence against powdery mildew in vivo, suggesting that an increase in [Ca2+]cyt might promote resistance suppression. A rapid [Ca2+]cyt elevation was observed in barley epidermal cells on mildew infection, reinforcing the idea that the MLO-dependent defence suppression might be boosted soon after pathogen challenge via a Ca2+/CaM-dependent process. Additional data suggest that the CaM-dependent MLO protein has ambivalent functions, carrying both resistance-mediating and resistance-suppressing activities. Indeed, the powdery mildew-resistant mlo plants exhibit enhanced susceptibility to the hemibiotrophic rice blast fungus Magnaporthe grisea and the necrotrophic fungus Bipolaris sorokiniana (Jarosch et al., 1999; Kumar et al., 2001). A possible explanation is that mlo mutations favoured these pathogens in their necrotrophic lifestyles by triggering mesophyll cell death on inoculation, as suggested by Levine et al. (1996). Another possibility is that MLO may modulate different resistance mechanisms effective against biotrophic or necrotrophic fungi. Thus MLO may ensure a balance between mutually inhibitory responses to different pathogen species (Panstruga & Schulze-Lefert, 2003).

AtNOS1, a protein displaying NOS activity, has been identified recently in Arabidopsis, and mutant plants lacking AtNOS1 are more susceptible to pathogenic bacteria. AtNOS1 contains CaM-binding motifs, and full activation of the enzyme needs both Ca2+ and CaM (Guo et al., 2003; Zeidler et al., 2004). In cryptogein-treated cells, NO synthesis is tightly regulated by a signalling cascade involving Ca2+ influx (Lamotte et al., 2004). Therefore one hypothesis is that NO production observed in defence responses is stimulated by CaM and changes in [Ca2+]cyt.

The tobacco early ethylene-responsive gene NtER1 encodes a CaM-binding protein. NtER1 is developmentally regulated and acts as a trigger for senescence and death, indicating clearly the involvement of Ca2+/CaM-mediated signalling in ethylene action (Yang & Poovaiah, 2000). NtER1 and five related genes in Arabidopsis (AtSR for A. thaliana signal-responsive gene) are rapidly and differentially induced by defence-related molecules such as methyl jasmonate, H2O2, and SA (Yang & Poovaiah, 2002), suggesting the involvement of CaM-binding AtSR in defence-signalling pathways. Moreover, the AtSR proteins are located in the nucleus, show specific DNA binding activity, and interact with the CGCG cis-elements. These observations suggest that AtSR proteins may regulate the expression of defence genes through their binding to CGCG boxes. Another calmodulin-binding nuclear protein, IQD1 (IQ-DOMAIN 1), was identified recently and proposed to integrate intracellular Ca2+ signals towards stimulation of plant defences (Levy et al., 2005).

2. Ca2+-dependent protein kinases

The CDPKs, encoded by 34 different genes in Arabidopsis, constitute one of the largest families of potential Ca2+ sensors in plants (Harmon et al., 2000; Cheng et al., 2002; Hrabak et al., 2003). CDPKs are activated in response to various stimuli known to trigger [Ca2+]cyt changes, including osmotic stresses, drought, low temperature stresses and pathogen-derived elicitors (Lee & Rudd, 2002; Cheng et al., 2002). These observations suggested that Ca2+-stimulated CDPKs decode and translate calcium signals in different biological contexts.

Evidence for a potential role of CDPKs in plant defence came first from expression analysis of genes encoding CDPK during plant–pathogen interactions. Several CDPK transcripts were reported to accumulate in tissues from tobacco, maize, tomato or pepper in response to fungal infection and treatment with pathogen elicitors (Yoon et al., 1999; Murillo et al., 2001; Chico et al., 2002; Chung et al., 2004). Romeis et al. (2000, 2001) provided strong evidence for the involvement of CDPKs in controlling tobacco responses to pathogen attack. Using tobacco cell cultures stably transformed with the tomato resistance gene Cf-9, which provides resistance to Cladosporium fulvum with the corresponding Avr9 avirulence gene, Romeis et al. (2000) observed an Avr9/Cf-9-dependent activation of a membrane-associated CDPK. This particular form of CDPK acts independently or upstream of a signalling pathway for AVR9-induced ROS production. In tobacco, expression of both NtCDPK2 and NtCDPK3 is induced in response to elicitation, and virus-induced gene silencing of the NtCDPK2/3 gene family resulted in a reduced HR after race-specific elicitation, indicating a role of the corresponding proteins in a defence-related signalling cascade (Romeis et al., 2001).

A better understanding of the functions of CDPKs in the signalling defence cascades needs the identification of substrates phosphorylated by pathogen-induced CDPKs. CDPK substrates have been identified, and their identities suggest potential regulatory roles in gene expression, metabolism, signalling pathways, transport of ions and water across membranes, and cytoskeleton dynamics (Harmon et al., 2000). However, only two proteins were reported as potential substrates for CDPK in the defence response context. PAL, a key enzyme in defence against pathogen attack, may be phosphorylated in French bean cells challenged with elicitor (Allwood et al., 1999, 2002). Recently, Cheng et al. (2001) have shown that a CDPK, namely the Arabidopsis AtCPK1, is able to phosphorylate PAL in vitro. The involvement of AtCPK1 in the defence responses remains to be demonstrated, as well as the biological significance of PAL phosphorylation. Plasma membrane-associated NADPH oxidase is another putative substrate for CDPK. Indeed, Xing et al. (2001) observed that ectopic expression of a heterologous CDPK could enhance NADPH oxidase activity and stimulate an oxidative burst in tomato protoplasts. However, it is still unclear whether the CDPK phosphorylates directly and thereby activates NADPH oxidase, or phosphorylates an upstream component involved in the regulation of the enzyme activity.

3. Other EF-hand motif-containing Ca2+-binding proteins

In addition to CaM, a number of proteins containing EF-hand motifs have been characterized in plants (Reddy, 2001), but only few have been shown to be implicated in pathogen-mediated defence responses.

Plant Rboh proteins are plasma membrane-bound enzymes responsible for the pathogen-induced oxidative burst in plants, and homologous to the catalytic subunit gp91phox of the neutrophil NADPH oxidase (Torres et al., 2002; Yoshioka et al., 2003). Keller et al. (1998) showed that the oxidative burst caused by Rboh is regulated through Ca2+-dependent mechanisms. Indeed, plant Rboh enzymes contain two EF-hand motifs that are not present in mammalian NADPH oxidase gp91phox. Recently, Sagi & Fluhr (2001) demonstrated that Ca2+ concentrations shown to induce plant Rboh activity in vitro are consistent with low-affinity EF-hands, and may reflect the concentrations of Ca2+ released and sequestered around membrane-bound plant NADPH oxidases during plant defence reaction. In plant disease resistance, direct activation of Rboh by [Ca2+]cyt elevations may be important for rapid production of ROS in challenged cells. Moreover, because the cytosolic diffusion of free Ca2+ is highly restricted (Sanders et al., 2002), Ca2+-stimulated Rboh enzymes may be positioned close to Ca2+ channels localized on the plasma membrane. Thus Ca2+ binding to Rboh enzymes may provide spatial control for localized generation of reactive oxygen intermediates immediately adjacent to sites of attempted pathogen infection.

Originally identified in the unicellular green alga Chlamydomonas reinhardtii, centrins were also found in plant cells (Schiebel & Bornens, 1995; Reddy, 2001). Centrins are CaM-like proteins with four Ca2+-binding EF-hand motifs. In animal and algal cells, centrins have been found to be associated with the cytoskeleton, and a role in microtubule severing and cytoskeleton reorganization has been shown (Salisbury, 1995). A centrin gene was shown as an early induced gene when Arabidopsis plants are inoculated with avirulent strains of bacteria (Cordeiro et al., 1998). Infection of parsley cells with Phytophthora infestans is associated with a rapid translocation of cytoplasm and nucleus to the fungal penetration site, a response that is mediated by depolymerization of the microtubular network (Gross et al., 1993). Based on these observations, Cordeiro et al. (1998) suggest that centrin could be involved in the intracellular reorganization during early infection. Along the same lines, the arrangement of microtubules and microfilaments was found to play an important role in the expression of nonhost resistance in barley (Kobayashi et al., 1997), whereas Binet et al. (2001) reported that a rapid and Ca2+-dependent disruption of microtubular cytoskeleton is associated with cell death in cryptogein-treated tobacco cells. Takezawa (2000) showed a rapid accumulation of CCD-1 mRNA in elicitor-treated wheat cells. CCD-1 encodes a 14-kDa Ca2+-binding protein that shares homology with the C-terminal half domain of centrin. In addition, PvHra32, a gene that is highly expressed during hypersensitive reaction in bean tissue challenged with P. syringae, has been shown to encode a small Ca2+-binding protein (17 kDa) with four EF-hand motifs (Jakobek et al., 1999). Taken together, these data suggest a role for centrins in defence responses, but clear evidence is still missing.

A new family of Ca2+ sensors in plants, referred to as calcineurin B-like (CBL) proteins, consists of proteins similar to both the regulatory B-subunit of calcineurin and the neuronal Ca2+ sensor in animals (Liu & Zhu, 1998; Luan et al., 2002). To transmit Ca2+ signals, CBLs interact with a family of protein kinases in a Ca2+-dependent manner. These kinases, referred to as CIPKs (CBL-interacting protein kinases), are serine–threonine protein kinases related to the sucrose nonfermenting protein kinase from yeast. To date, only few CBL–CIPK pairs have been shown to exert a physiological function. Several lines of evidence indicate that different CBL–CIPK interactions are involved in the signalling cascades induced in response to a variety of abiotic stresses such as wounding, cold, drought and high salt (Luan et al., 2002). With at least 10 CBLs and 25 CIPKs encoded in the Arabidopsis genome, many functional CBL–CIPK pairs can be formed that potentially function in a large array of signalling processes involving Ca2+ signalling. Although expected, the involvement of CBL–CIPK pairs in the induction of defence responses remains to be demonstrated.

4. Ca2+-binding proteins without EF-hands

Many Ca2+-binding proteins do not contain EF-hand structural motifs, but other domains such as the C2 domain (Reddy, 2001). The C2 domain is a Ca2+–phospholipid-binding site, and Ca2+ binding is coordinated by four to five amino acid residues provided by bipartite loops (Rizo & Südhof, 1998). In mammals, C2 domains have been found in more than 100 proteins, most of which are involved in lipid metabolism, signal transduction or membrane trafficking (Rizo & Südhof, 1998). These domains often mediate Ca2+-dependent phospholipid binding, and thus play an important role in associating proteins having a C2 domain with substrates and membranes. In plants, copine and phospholipase D, two proteins that contain a C2 domain, have been shown to be involved in defence responses (Jambunathan et al., 2001; Laxalt & Munnik, 2002).

Copines belong to a highly conserved class of proteins encoded by a multigenic family in plants, animals and protozoa (Tomsig & Creutz, 2002). They bind membrane phospholipids due to the presence of two C2 domains in the N-terminal portion that are activated by calcium. The C-terminal half of the copine molecule, called the A domain, may be involved in targeted protein–protein interactions. In animals, it was recently observed that copines mediate the Ca2+-dependent association of target proteins with phospholipids, a phenomenon that could influence intracellular localization and the activities of the target proteins (Tomsig & Creutz, 2002). Additional data from Tomsig et al. (2003) suggest that copines may represent a universal transduction pathway for Ca2+. In plants, copines were described as playing a role in membrane trafficking and as being transcriptionally regulated by the environmental conditions (Hua et al., 2001). Jambunathan et al. (2001) identified an Arabidopsis copine mutant, cpn1-1 (for copine1-1), which displayed aberrant regulation of cell death including a lesion mimic phenotype and an accelerated HR. The improved disease resistance and accelerated HR in cpn1-1 plants support a role for CPN1 as a negative regulator of plant defence responses, including the HR. The authors suggested that a pathogen-triggered [Ca2+]cyt elevation may activate the repressive function of CPN1, possibly by triggering the localization of CPN1 to a membrane where it could exert a repressive influence on cell death signal transduction. These data are in accordance with genetic evidence indicating that the HR is under negative as well as positive control to prevent uncontrolled cell death and lesion expansion after HR initiation (Dietrich et al., 1994). Recently, Jambunathan & McNellis (2003) reported a strong, rapid and specific accumulation of CPN1 gene transcripts in response to pathogen inoculation. CPN1 transcript accumulation is responsive to gene-for-gene-mediated signalling, as well as to treatment with pathogenic bacteria. It is not unusual for pathogen-induced genes to be induced by both virulent and avirulent pathogens, although induction by avirulent pathogens is generally much stronger and more rapid than that by virulent pathogens, as observed with CPN1. Together, these results indicate that CPN1 plays a role in plant disease-resistance responses, possibly as a suppressor of defence responses including hypersensitive cell death (Jambunathan & McNellis, 2003). Moreover, the authors suggest that CPN1, as a Ca2+-dependent membrane-associated protein, might be involved in determining the specificity of Ca2+ signalling and preventing inappropriate defence responses.

Phospholipase D (PLD) is an enzyme widely distributed in bacteria, fungi, plants and animals, which breaks the phosphodiester bond of phospholipids producing alcohol and phosphatidic acid (Exton, 2002). In animal cells, activation of PLD generates signalling messengers and is involved in a wide range of cellular processes, including hormone action, meiosis, defence response and vesicular trafficking. PLD is a major family of phospholipases in plants, and PLD activation has been associated with abscisic acid treatment, osmotic stress wounding and plant defences. Several reports implicate PLD in plant–pathogen interactions, with data indicating the induction of PLD gene expression and showing an elicitor-induced PLD activation (Laxalt & Munnik, 2002; Wang, 2002). In contrast to animal or yeast PLDs, the C2 domain is present in all characterized plant PLDs, making Ca2+ an important regulator (Wang, 2002). A positive correlation between increased [Ca2+]cyt levels and increased PLD activity was reported in plant tissues (De Vrije & Munnik, 1997). The functional link between elicitor-induced [Ca2+]cyt elevation, PLD activation and defence-response expression has to be demonstrated. Potentially, Ca2+ may associate directly with PLD through the C2 domain, and such binding could induce conformational changes in the enzyme to facilitate its binding to a membrane surface and/or its activation for catalysis.

V. Concluding remarks

Calcium is an essential second messenger in many pathways, especially in plant responses to pathogen attacks. To understand how Ca2+ is involved in generating plant defence responses, specific spatiotemporal changes in [Ca2+]cyt, described as ‘calcium signatures’, have been reported in several plant–elicitor systems and are described in this review. However, the calcium-signature hypothesis, used as a concept to explain specificity in signalling pathways involving Ca2+, should be used cautiously, as mentioned recently by Scrase-Field & Knight (2003); Plieth (2001, 2005). To date, few data clearly demonstrate that Ca2+ signatures encode specific information leading to a specific end response (Hetherington & Brownlee, 2004). Another possibility is that Ca2+ operates as a simple on–off binary switch to activate Ca2+-dependent components. In this case, the Ca2+ signal builds an appropriate end response only if associated with other signalling components, the signal specificity being the responsibility of some signalling players other than calcium (Scrase-Field & Knight, 2003). Many studies have demonstrated that defence-signalling pathways require not only calcium, but also its combination with other second-messenger molecules generating a spatially and temporally complex array of simultaneously operating signal transmitters. For example, calcium can be positioned as a key player in the signalling cascade induced by cryptogein in tobacco cells, because this ion is absolutely required to trigger most of the downstream cellular events listed so far (Fig. 4). However, the signal amplification involves other signalling molecules, including H2O2, NO, cADPRc and IP3, and the specificity of the response probably relies on the distribution of these second messengers and targeted proteins in time and space. Moreover, we should bear in mind that changes in free calcium concentration with physiological significance are also apparent in other organelles, such as chloroplasts, mitochondria and nucleus. For instance, the data provided by Lecourieux et al. (2005) open up the possibility of a role for nuclear Ca2+ in plant defence responses. One might expect that nuclear Ca2+ should be responsible for activation of Ca2+-dependent proteins shown to be located in the nucleus and involved in the regulation of nuclear activities. A possibility is that [Ca2+]nuc elevation modulates the assembly and composition of transcriptional complexes that, in turn, control the expression of defence genes. In chloroplasts, a specific [Ca2+] rise might modulate the biosynthesis of precise secondary metabolites that can act as antimicrobial molecules. In mammals, [Ca2+] elevations in mitochondria result in mitochondrial dysfunction that, in turn, leads to apoptotic cell death (Kim et al., 2005). The involvement of mitochondrial [Ca2+] in the establishment of the HR typified by a cell death program remains an open question.

Figure 4.

Ca2+ fluxes and Ca2+-dependent events in cryptogein-treated tobacco cells. Calcium fluxes and the resulting elevations in free-Ca2+ concentrations are major elements of the cryptogein-induced signalling pathway and regulate numerous downstream responses in tobacco cells (the cryptogein-induced signal transduction was recently reviewed by Garcia-Brugger et al. 2006). The binding of cryptogein to its putative plasma membrane receptor rapidly triggers a large and sustained Ca2+ influx that leads to a biphasic [Ca2+]cyt elevation. This Ca2+ signature involves both extracellular and intracellular Ca2+ pools. The initial, large entry of Ca2+ into the cell is mediated by an unknown Ca2+ channel and requires previous phosphorylation events. This first Ca2+ entry induced various events including anion effluxes, production of reactive oxygen species, synthesis of nitric oxide, and plasma membrane depolarization that can, in turn, stimulate Ca2+ fluxes. The subsequent Ca2+ influxes from the apoplast can take place through H2O2-stimuled Ca2+ channels, and possibly through voltage-gated Ca2+ channels. Nitric oxide production can lead to cADPR accumulation, this second messenger being able to stimulate specific Ca2+ channels. Additionally, the Ca2+-dependent microtubule depolymerization could result in stimulation of plasma membrane-localized Ca2+ channels that interact with the cytoskeleton. The increase in [Ca2+]cyt is also caused by release from internal Ca2+ pools through IP3- and cADPR-regulated Ca2+ channels. This cryptogein-induced [Ca2+]cyt rise is necessary for the [Ca2+]nuc elevation, which is caused by H2O2-stimuled or IP3-gated calcium channels rather than by a diffusion of Ca2+ from the cytosol. Whereas it is clearly established that [Ca2+]cyt elevation is required for various cryptogein-induced cellular responses (such as gene expression, phytoalexin production and cell death), the physiological significance of the [Ca2+]nuc rise remains to be elucidated. R, cryptogein receptor; PP, protein phosphatase; PK, protein kinase; SOD, superoxide dismutase.

Although it is now clearly established that Ca2+ is a key element of signalling pathways mobilized during plant–pathogen interactions, a major challenge remains to understand how alterations in [Ca2+] are integrated into defence signalling networks. So far, little is known about the sensor proteins that recognize various elicitor-induced calcium signals and relay these signals into downstream specific defence responses (Fig. 3). Future work associating genetic and proteomic approaches will help identify these Ca2+ sensors among hundreds of candidates (Reddy & Reddy, 2004). Functional analyses of these Ca2+-signalling components will yield important insights that will allow us to decipher the role of Ca2+ in the establishment of plant defences.

Acknowledgements

The authors acknowledge funding support from the Centre National de la Recherche Scientifique, the Institut National de la Recherche Agronomique, the Conseil Régional de Bourgogne, and the Ministère de l’Education Nationale de l’Enseignement Supérieur et de la Recherche. We thank David Wendehenne and Fatma Lecourieux-Ouaked for their comments on the manuscript.

Ancillary