Calcium in Plants

Abstract

Calcium is an essential plant nutrient. It is required for various structural roles in the cell wall and membranes, it is a counter‐cation for inorganic and organic anions in the vacuole, and the cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt) is an obligate intracellular messenger coordinating responses to numerous developmental cues and environmental challenges. This article provides an overview of the nutritional requirements of different plants for Ca, and how this impacts on natural flora and the Ca content of crops. It also reviews recent work on (a) the mechanisms of Ca²⁺ transport across cellular membranes, (b) understanding the origins and specificity of [Ca²⁺]_cyt signals and (c) characterizing the cellular [Ca²⁺]_cyt‐sensors (such as calmodulin, calcineurin B‐like proteins and calcium‐dependent protein kinases) that allow plant cells to respond appropriately to [Ca²⁺]_cyt signals.

INTRODUCTION

Calcium is an essential plant nutrient. As the divalent cation (Ca²⁺), it is required for structural roles in the cell wall and membranes, as a counter‐cation for inorganic and organic anions in the vacuole, and as an intracellular messenger in the cytosol (Marschner, 1995). Calcium deficiency is rare in nature, but excessive Ca restricts plant communities on calcareous soils. Calcium is taken up by roots from the soil solution and delivered to the shoot via the xylem. It may traverse the root either through the cytoplasm of cells linked by plasmodesmata (the symplast) or through the spaces between cells (the apoplast). The relative contributions of the apoplastic and symplastic pathways to the delivery of Ca to the xylem are unknown (White, 2001). However, the movement of Ca through these pathways must be finely balanced to allow root cells to signal using cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt), control the rate of Ca delivery to the xylem, and prevent the accumulation of toxic cations in the shoot.

Calcium enters plant cells through Ca²⁺‐permeable ion channels in their plasma membranes (White, 2000). Since a high [Ca²⁺]_cyt is cytotoxic, a submicromolar [Ca²⁺]_cyt is maintained in unstimulated cells by Ca²⁺‐ATPases and H⁺/Ca²⁺‐antiporters (Sze et al., 2000; Hirschi, 2001). These enzymes remove cytosolic Ca²⁺ to either the apoplast or the lumen of intracellular organelles, such as the vacuole or endoplasmic reticulum (ER). The rapid influx of Ca²⁺ through cation channels in the plasma membrane, tonoplast and/or ER generates [Ca²⁺]_cyt perturbations that initiate cellular responses to a diverse range of developmental cues and environmental challenges (White, 2000; Sanders et al., 2002). Proteins that change conformation or catalytic activity upon binding Ca²⁺, such as calmodulin (CaM), calcineurin B‐like proteins (CBLs) and Ca²⁺‐dependent protein kinases (CDPKs), allow the cellular perception and transduction of the [Ca²⁺]_cyt signal. These proteins are termed ‘[Ca²⁺]_cyt sensors’. It is speculated that cellular responses to specific biotic and abiotic stimuli are encoded by distinct [Ca²⁺]_cyt perturbations and are transduced by particular [Ca²⁺]_cyt sensors. Much current work on Ca in plants is dedicated to understanding the nature and specificity of [Ca²⁺]_cyt signalling and response networks.

This article provides an overview of recent work on Ca in plants. First, it discusses the Ca requirements of different plant species, the mechanisms of Ca uptake and delivery to the xylem, and the impact of these on natural flora and the Ca content of crops. It then highlights the insights made possible by recent advances in electrophysiology, microscopy and plant molecular biology that have enabled researchers to characterize Ca²⁺ transporters in cellular membranes, begin to unravel the origins and specificity of [Ca²⁺]_cyt signals, and identify the [Ca²⁺]_cyt‐sensors that allow plant cells to respond to [Ca²⁺]_cyt perturbations.

NUTRITION

The calcium requirements of plants

Plants growing with adequate Ca in their natural habitats have shoot Ca concentrations between 0·1 and 5 % d. wt (Marschner, 1995). These values reflect both Ca availability in the environment and the contrasting Ca requirements of different plant species. Calcium deficiency is rare in nature, but may occur on soils with low base saturation and/or high levels of acidic deposition (McLaughlin and Wimmer, 1999). By contrast, several costly Ca‐deficiency disorders occur in horticulture (Fig. 1; Shear, 1975). These generally arise when sufficient Ca is momentarily unavailable to developing tissues. Deficiency symptoms are observed (a) in young expanding leaves, such as in ‘tipburn’ of leafy vegetables, (b) in enclosed tissues, such as in ‘brown heart’ of leafy vegetables or ‘black heart’ of celery, or (c) in tissues fed principally by the phloem rather than the xylem, such as in ‘blossom end rot’ of watermelon, pepper and tomato fruit, ‘bitter pit’ of apples and ‘empty pod’ in peanut. They occur because Ca cannot be mobilized from older tissues and redistributed via the phloem. This forces the developing tissues to rely on the immediate supply of Ca in the xylem, which is dependent on transpiration. Transpiration is low in young leaves, in enclosed tissues and in fruit. Other physiological disorders, such as ‘cracking’ in tomato, cherry and apple fruit, occur in tissues lacking sufficient Ca upon hypo‐osmotic shock (following increased humidity or rainfall), presumably as a result of structural weaknesses in cell walls. When excessive Ca is present in the rhizosphere solution, plants may suffer Ca toxicity. This may prevent the germination of seeds and reduce plant growth rates (Fig. 2). In cultivated tomato, one symptom of excess calcium is the development of tiny yellowish flecks or ‘gold spot’ in the cell walls around the calyx and shoulders of the fruit (Fig. 1). These flecks are crystals of calcium oxalate and their abundance is increased by high humidity and high Ca fertilization (Bekreij et al., 1992).

Fig. 1. Calcium disorders in horticultural crops: (a) cracking in tomato fruit; (b) tipburn in lettuce; (c) calcium deficiency in celery; (d) blossom end rot in immature tomato fruit; (e) bitter pit in apples; (f) gold spot in tomato fruit with calcium oxalate crystals (inset). Photographs (A–E) are from the HRI collection and (F) is courtesy of Lim Ho (HRI‐Wellesbourne).

Fig. 2. The relationships between Ca concentration in the soil solution ([Ca²⁺]_ext) and (A) shoot dry weight and (B) shoot Ca content of two calcifuge species Juncus squarrosus (filled circles) and Nardus stricta (filled triangles), the mineral‐tolerant Siegelingia decumbens (open circles), and the calcicole species Origanum vulgare (filled squares). Plants were grown for approx. 4 weeks after germination in a quartz sand to which was added a complete mineral solution containing various concentrations of CaCl₂. Data are taken from Jefferies and Willis (1964), who observed (a) that Juncus squarrosus did not establish at [Ca²⁺]_ext ≥ 0·8 mm and (b) that Origanum vulgare showed symptoms of Ca deficiency and was unable to survive at [Ca²⁺]_ext less than about 0·8 mm.

Ecologists have classified plant species into calcifuges, which occur on acid soils with low Ca, and calcicoles, which occur on calcareous soils. The Ca concentrations in calcifuge and calcicole plants growing in their natural habitats differ markedly. However, it is the ability to tolerate excessive Al, Mn and Fe that largely determines the flora of acid soils, and an insensitivity to Fe‐ and P‐deficiencies that determines the flora of calcareous soils (Lee, 1999). Nevertheless, calcifuges generally grow well at low Ca²⁺ concentrations in the rhizosphere ([Ca²⁺]_ext) and respond little to increased [Ca²⁺]_ext, which may even inhibit growth (Fig. 2). Conversely, the mechanisms that enable calcicole plants to maintain low [Ca²⁺]_cyt in their natural habitat are believed to restrict their growth at low [Ca²⁺]_ext by inducing Ca‐deficiency (Fig. 2; Lee, 1999). This is consistent with the phenotype of plants overexpressing Ca²⁺‐transporters that remove Ca²⁺ from the cytoplasm to the vacuole which show Ca‐deficiency symptoms at low [Ca²⁺]_ext (Hirschi, 2001). Hence, the optimal [Ca²⁺]_ext for a plant in hydroponics often approximates the [Ca²⁺]_ext of its natural habitat.

Distinct relationships between shoot Ca concentration ([Ca]_shoot) and [Ca²⁺]_ext have been attributed to contrasting Ca ‘physiotypes’ (Kinzel and Lechner, 1992). The ecology of different Ca physiotypes is thought to reflect their ability to utilize Ca as an osmoticum in primarily xerophytic, calcareous environments. One characteristic of calcicole plants, such as the Crassulaceae, Brassicaceae and Fabaceae, is a high soluble Ca concentration. In these plants, which are also termed ‘calciotrophs’ (Kinzel, 1982), Ca accumulation is stimulated greatly by increasing [Ca²⁺]_ext. By contrast, calcifuges generally have a low soluble Ca concentration. They include the ‘potassium plants’, such as the Apiales and Asterales, which are characterized by high shoot K/Ca quotients, and the ‘oxalate plants’, which have high tissue oxalate concentrations. Plants that accumulate oxalate can be subdivided into species that contain soluble oxalate and those in which Ca‐oxalate is precipitated. Interestingly, the uptake of Ca does not appear to increase with increasing [Ca²⁺]_ext in plants containing soluble oxalate, such as the Oxalidaceae (Kinzel and Lechner, 1992). In plants that precipitate Ca‐oxalate such as the Caryophyllales families Caryophyllaceae, Chenopodiaceae and Polygonaceae, there is a proportional increase in both Ca and oxalate concentration with increasing [Ca²⁺]_ext (Libert and Franceschi, 1987; Kinzel and Lechner, 1992).

Calcium uptake and movement to the shoot

Calcium is acquired from the soil solution by the root system and translocated to the shoot via the xylem. The Ca flux to the xylem is high, and a rate of 40 nmol Ca h^–1 g^–1 f. wt root is not unreasonable in an actively growing plant (White, 1998). The delivery of Ca to the xylem is restricted to the extreme root tip and to regions in which lateral roots are being initiated (Clarkson, 1993; White, 2001). In these regions a contiguous, Casparian band between endodermal cells is absent or disrupted, and/or the endodermal cells surrounding the stele are unsuberized. The Casparian band restricts the apoplastic movement of solutes (Clarkson, 1984, 1993; White, 2001) and suberization prevents Ca²⁺ influx to endodermal cells (Moore et al., 2002). These observations suggest that Ca might reach the xylem solely via the apoplast in regions where the Casparian band is absent or disrupted, or circumvent the Casparian band by entering the cytoplasm of unsuberized endodermal cells when the Casparian band is present (Clarkson, 1984, 1993; White 2001). These are referred to as the apoplastic and symplastic pathways, respectively.

Each pathway of Ca movement across the root confers distinct advantages and disadvantages. The apoplastic pathway allows Ca to be delivered to the xylem without impacting on the use of [Ca²⁺]_cyt for intracellular signalling (White, 1998). Intracellular signalling requires [Ca²⁺]_cyt to be maintained at submicromolar levels in the resting cell and to increase rapidly in response to developmental cues or environmental challenges. Since the Ca²⁺ fluxes required for [Ca²⁺]_cyt signalling are minute compared with those required for adequate nutrition, both these requirements for [Ca²⁺]_cyt signalling might be compromised by high nutritional Ca²⁺ fluxes through root cells (White, 2001). However, the Ca flux to the xylem through the apoplastic pathway is influenced markedly by transpiration, which could lead to vagaries in the amount of Ca supplied to the shoot and the development of Ca disorders (Marschner, 1995; McLaughlin and Wimmer, 1999). Furthermore, the apoplastic pathway is relatively non‐selective between divalent cations (White, 2001; White et al., 2002b), and its presence could result in the accumulation of toxic solutes in the shoot. By contrast, the symplastic pathway allows the plant to control the rate and selectivity of Ca transport to the shoot (Clarkson, 1993; White, 2001). It is thought that Ca²⁺ enters the cytoplasm of endodermal cells through Ca²⁺‐permeable channels on the cortical side of the Casparian band, and that Ca²⁺ is pumped from the symplast by the plasma membrane Ca²⁺‐ATPases or Ca²⁺/H⁺‐antiporters of cells within the stele. By regulating the expression and activity of these transporters, Ca could be delivered selectively to the xylem at a rate consistent with the requirements of the shoot.

Several lines of evidence suggest that both apoplastic and symplastic pathways contribute to Ca delivery to the xylem. First, since Ca is delivered to the xylem in regions of the root where the Casparian band is fully developed and apoplastic Ca transport is restricted, some Ca might bypass the Casparian band through the cytoplasm of endodermal cells (Clarkson, 1984, 1993; White, 2001). Secondly, although Ca²⁺ channels and Ca²⁺‐ATPases are present and thermodynamically capable of catalysing Ca²⁺ influx and efflux across the plasma membrane of root endodermal cells, it has been calculated that there is insufficient ATP to power (Flowers and Yeo, 1992) and insufficient proteins to catalyse (White, 1998, 2001) the observed Ca²⁺ fluxes solely through the symplast. Thirdly, if Ca reached the xylem solely by a symplastic pathway, its accumulation in the shoot would be expected to show the hallmarks of protein‐catalysed transport, which it does not. For example, both Ca²⁺ channels in the plasma membrane of root cells and Ca²⁺‐ATPases discriminate between divalent cations, but there seems to be no discrimination between Ca²⁺, Ba²⁺ and Sr²⁺ in their transport to the shoot (White, 2001). Furthermore, there seems to be no competition, or interactions, between Ca²⁺, Ba²⁺ and Sr²⁺ during their transport to the shoot, and the accumulation of divalent cations in the shoot is often linearly related to their concentrations in the rhizosphere solution (White, 2001). Although the relative contributions of the apoplastic and symplastic pathways to the delivery of Ca to the xylem are unknown, it is likely that a functional separation of apoplastic Ca²⁺ fluxes (for transfer to the shoot) and symplastic Ca²⁺ fluxes (for cell signalling) would enable the root to fulfil the demand of the shoot for Ca without compromising intracellular [Ca²⁺]_cyt signals.

The Ca concentration in xylem sap ([Ca]_xylem) is influenced greatly by [Ca²⁺]_ext, and [Ca]_xylem between 300 µm and 16·5 mm has been reported (White et al., 1992; De Silva et al., 1998). The relative proportion of Ca²⁺ to total Ca also varies, with organic acids, such as malate and citrate, chelating Ca²⁺ in the xylem sap (Marschner, 1995). When abundant Ca is present in the xylem sap, there is a close relationship between Ca distribution to the shoot and transpiration. Within the leaf, Ca follows the apoplastic route of the transpiration stream and accumulates in either the mesophyll cells, trichomes or epidermal cells adjacent to guard cells, depending on the plant species (Karley et al., 2000). Both the [Ca²⁺]_cyt in guard cells and the closing of stomata in detached epidermal strips are sensitive to apoplastic Ca²⁺ concentrations within the range of [Ca²⁺]_xylem (McAinsh et al., 1995) and the ability of some calcicole species, such as Leontodon hispidus and Centaurea scabiosa, to tolerate high [Ca²⁺]_ext may be related to their ability to accumulate Ca in their trichomes (De Silva et al., 1996, 1998).

The phylogeny of shoot calcium concentration

When grown under identical conditions, the [Ca]_shoot of different plant species differs markedly (Figs 2 and 3 and Table 1; Broadley et al., 2003a). A large proportion of this variation can be attributed to the phylogenetic division between eudicots and monocots (Thompson et al., 1997; Broadley et al., 2003a). Eudicots generally have a higher [Ca]_shoot than monocots. Within the eudicots, orders within both the rosid (Cucurbitales, Rosales, Malvales and Brassicales) and asterid (Apiales, Asterales, Lamiales and Solanales) clades have the highest [Ca]_shoot. Within the monocots, [Ca]_shoot is higher in the non‐commelinoid orders (e.g. Asparagales) than in the commelinoid orders (e.g. Poales). Phylogenetic differences in [Ca]_shoot have not been resolved at lower taxonomic levels, but it is noteworthy that distinct Ca physiotypes (‘calciotrophes’, ‘potassium plants’ and ‘oxalate plants’) occur in particular plant families (Kinzel, 1982) and that traits such as the prevalence, shape and tissue distributions of Ca oxalate crystals, are used as taxonomic characters (e.g. Franceschi and Horner, 1980; Kuo‐Huang et al., 1994; Wu and Kuo‐Huang, 1997; Prychid and Rudall, 1999; Caddick et al., 2002).

Fig. 3. The relationships between root CEC, derived from a literature survey, and shoot calcium concentration, derived from both literature and experimental data, for angiosperm orders (Table 1). Circles represent shoot Ca content data from a literature survey. Filled circles and linear regression represent orders with n ≥ 3 species sampled. Triangles represent shoot Ca content estimated in a phylogenetically balanced experiment.

Table 1.

The mean relative shoot calcium concentration of angiosperm orders derived from both a literature survey and a phylogenetically balanced experiment, and their mean relative root cation exchange capacity (CEC)

	Literature		Literature		Experiment
Order	Root CEC (meq 100 g–1 d. wt root)	No. spp.	Relative shoot [Ca]	No. spp.	Relative shoot [Ca]	No. spp.
Apiales	40·67	3	1·90	5	1·21	4
Aquifoliales	27·28	1	–	–	–	–
Asparagales	27·08	6	1·22	4	1·10	13
Asterales	40·34	12	1·78	11	1·19	15
Brassicales	36·39	7	2·19	15	2·96	3
Caryophyllales	32·12	12	1·57	17	1·07	7
Cucurbitales	43·35	4	3·19	2	3·21	1
Dipsacales	35·01	1	–	–	1·04	2
Ericales	22·54	2	–0·01	1	1·04	1
Fabales	33·03	23	1·53	78	1·66	9
Gentianales	32·21	1	–	–	1·36	4
Geraniales	28·41	4	1·53	2	–	–
Lamiales	25·96	2	1·72	5	1·84	7
Laurales	–2·07	1	–	–	–	–
Liliales	24·06	2	–	–	–	–
Malpighiales	23·18	1	1·12	2	1·07	5
Malvales	34·91	1	3·04	2	2·50	3
Myrtales	26·21	1	0·70	1	1·13	6
Poales	14·94	50	0·37	48	0·43	18
Ranunculales	55·42	4	1·69	2	–	–
Rosales	22·54	4	2·73	1	1·25	3
Sapidales	17·98	6	–	–	1·29	4
Solanales	31·22	4	1·73	8	1·60	4
Vitaceae	15·67	2	–	–	–	–

Data for shoot Ca concentration were taken from Broadley et al. (2003a) and for root CEC from the comparative studies cited by Asher and Ozanne (1961), Heintz (1961), Crooke and Knight (1962), Fréjat et al. (1967), Snaydon and Bradshaw (1969) or Wacquant (1977).

All literature data sets were subjected to a residual maximum likelihood (REML) analysis, using a procedure described previously (Broadley et al., 2001). This procedure adjusted for differences in between‐study means to generate mean relative shoot Ca concentration (206 plant species in 18 orders from 244 studies) and mean relative root CEC (154 plant species from 93 studies).

Data are expressed as order means from n species sampled.

Although the Ca physiotype of a plant determines its ability to accumulate Ca in the shoot (Kinzel and Lechner, 1992; Broadley et al., 2003a), genotypic differences in the activities of Ca²⁺ transporters in root cell membranes (Hirschi, 2001; White et al., 2002a) and/or in the relative contributions of symplastic and apoplastic pathways to the delivery of Ca to the xylem (White, 2001) will contribute much to phylogenetic variation in [Ca]_shoot. Historically, [Ca]_shoot has been correlated with the cation exchange capacity (CEC) of plant roots (Fig. 3). The CEC is located in the root apoplast, and is attributed to the free carboxyl groups of galacturonic acids of cell wall pectins in the middle lamella (Haynes, 1980; Sattelmacher, 2001). If the pectin contents of shoot and root cell walls are similar, it is unsurprising that the phylogenetic variation in CEC in monocot roots parallels that in pectin content of shoot cell walls (Jarvis et al., 1988). The pectin contents of shoot cell walls are low in the Cyperaceae, Poaceae, Juncaceae and Restionaceae families (all assigned to the Poales) and intermediate in the commelinoid monocot families Commelinaceae (Commelinales), Bromeliaceae (unassigned to order) and Typhaceae (Poales) and in the non‐commelinoid Pandanaceae (Pandanales) family. The pectin content of shoot cell walls is similar to that of dicots in the commelinoid monocot families Musaceae (Zingiberales) and Sparganiaceae (Poales) and in the non‐commelinoid monocot families Velloziaceae (Pandanales), Alliaceae Amaryllidaceae, Asphodelaceae, Iridaceae, Orchidaceae (all Asparagales), Alismataceae, Araceae, Juncaginaceae, Potamogetonaceae, Zosteraceae (all Alismatales) and Liliaceae (Liliales).

In regions of the root where the Ca flux to the xylem is apoplastic, CEC may exert a direct effect on the transport of Ca²⁺ to the shoot. However, several reviews have asserted that there is no direct connection between the capacity of roots to bind cations in the free space and their active (energy‐dependent) accumulation (e.g. Haynes, 1980). Nevertheless, fixed negative charges, and charge screening, can influence both the absolute and relative concentrations of cations in the apoplast, especially at low ionic activities. Thus, root CEC could determine shoot cation content indirectly by affecting the rate and selectivity of cation uptake into the symplast in addition to cation transport through the apoplast (Asher and Ozanne, 1961; Wacquant, 1977). The effects of CEC on cation transport may have ecological implications. It has been suggested that, in solutions of low ionic strength, plants with higher root CEC compete for divalent cations more effectively, whereas plants with lower root CEC compete for monovalent cations more effectively (Smith and Wallace, 1956; Asher and Ozanne, 1961).

Quantifying the phylogenetic impact on [Ca]_shoot has several uses. First, since [Ca]_shoot correlates with ecological traits (Kinzel, 1982; Thompson et al., 1997; Grime, 2001), predictions of the responses of plant communities to environmental change can be improved using this information. Secondly, knowledge of phylogenetic variation in [Ca]_shoot can be used to predict the movement of Ca (and chemically similar elements such as Sr, Mg and Ba; White, 2001; Broadley et al., 2003b) from soils to the shoots of plant species whose transfer coefficients are unknown and, thereby, improve nutrient and contaminant cycling models (Broadley et al., 2001). Thirdly, appreciation of phylogenetic influences on [Ca]_shoot can improve the delivery of Ca to the human diet. Unsurprisingly, Ca‐deficiency disorders have been observed in populations whose dietary habits have changed from bean‐rich to rice‐rich sources of food (Graham et al., 2001). Phylogenetic information can help identify crops that are either predisposed to higher [Ca]_shoot or offer genetic potential for breeding.

CALCIUM TRANSPORTERS IN CELLULAR MEMBRANES

Ca²⁺ efflux from the cytosol: Ca²⁺‐ATPases and H⁺/Ca²⁺‐antiporters

The removal of Ca²⁺ from the cytosol against its electrochemical gradient to either the apoplast or to intracellular organelles requires energized, ‘active’ transport. This is catalysed by Ca²⁺‐ATPases and H⁺/Ca²⁺‐antiporters (Fig. 4). By removing Ca²⁺ from the cytosol these enzymes perform several important functions (Sze et al., 2000; Hirschi, 2001): (1) they maintain a low [Ca²⁺]_cyt in the resting (unstimulated) cell appropriate for cytoplasmic metabolism; (2) they restore [Ca²⁺]_cyt to resting levels following a [Ca²⁺]_cyt perturbation, thereby influencing the magnitude, kinetics and subcellular location of [Ca²⁺]_cyt signals; (3) they replenish intracellular and extracellular Ca²⁺ stores for subsequent [Ca²⁺]_cyt signals and permit the generation of local [Ca²⁺]_cyt oscillations through their interplay with Ca²⁺ channels (Klüsener et al., 1995; Harper, 2001); (4) they provide Ca²⁺ in the ER for the secretory system to function (Blatt, 2000b; Ritchie et al., 2002); (5) they remove divalent cations, such as Mg²⁺, Mn²⁺, Ni²⁺ or Zn²⁺, from the cytosol, to support the specialized biochemistry of particular organelles and to prevent mineral toxicities (Hirschi, 2001; Wu et al., 2002). The relative importance of Ca²⁺‐ATPases and H⁺/Ca²⁺‐antiporters in each of these functions is unknown. Hirschi (2001) suggested that the Ca²⁺‐ATPases, which have high affinity (K_m = 1–10 µm; Evans and Williams, 1998) but low capacity for Ca²⁺ transport, are responsible for maintaining [Ca²⁺]_cyt homeostasis in the resting cell, whereas the H⁺/Ca²⁺‐antiporters, which have lower affinities (K_m = 10–15 µm) but high capacities for Ca²⁺ transport, are likely to remove Ca²⁺ from the cytosol during [Ca²⁺]_cyt signals and thereby modulate [Ca²⁺]_cyt perturbations. Consistent with this hypothesis is the observation that an H⁺/Ca²⁺‐antiporter, but not the vacuolar Ca²⁺‐ATPase, resets [Ca²⁺]_cyt in yeast following hypertonic shock (Denis and Cyert, 2002). However, the Arabidopsis de‐etiolated 3 (det3) mutant, which has reduced tonoplast H⁺‐ATPase and (presumably) H⁺/Ca²⁺‐antiporter activity, has a constitutively high [Ca²⁺]_cyt (Allen et al., 2000), suggesting that H⁺/Ca²⁺‐antiporters contribute to [Ca²⁺]_cyt homeostasis, and plants poisoned by the V‐type H⁺‐ATPase inhibitor bafilomycin show greater [Ca²⁺]_cyt elevations in response to hypo‐osmotic shock (Takahashi et al., 1997).

Fig. 4. Cartoon illustrating the subcellular location of Ca²⁺ transporters in Arabidopsis thaliana based on Sze et al. (2000), Sanders et al. (2002) and White et al. (2002). In the plasma membrane there are hyperpolarization‐activated Ca²⁺ channels (HACC, possibly encoded by annexin genes), depolarization‐activated Ca²⁺ channels (DACC, one of which may be encoded by TPC1), Ca²⁺‐permeable outward rectifying K⁺ channels (KORC, encoded by SKOR and GORK), voltage‐insensitive cation channels (VICC, probably encoded by the CNGC and GLR genes) and Ca²⁺‐ATPases (one of which is ACA8). Biochemical and electrophysiological evidence indicates that IP₃‐receptors, IP₆‐receptors, cADPR‐activated (ryanodine)‐receptors and two types of voltage‐gated Ca²⁺ channels (the depolarization‐activated SV channels and the hyperpolarization‐activated Ca²⁺ channels) are present in the tonoplast together with Ca²⁺‐ATPases (including ACA4) and H⁺/Ca²⁺‐antiporters encoded by the CAX genes. There is also biochemical and electrophysiological evidence for the presence of NAADP‐receptors, IP₃‐receptors, cADPR‐activated (ryanodine)‐receptors and depolarization‐activated Ca²⁺ channels in the endoplasmic reticulum (ER), together with Ca²⁺‐ATPases (including ECA1 and ACA2). The Ca²⁺‐ATPase ACA1 is located in the plastid inner membrane.

Plant Ca²⁺ ATPases belong to one of two major families (Evans and Williams, 1998; Geisler et al., 2000; Sze et al., 2000; Axelsen and Palmgren, 2001; Garciadeblas et al., 2001). The first family (the P‐type ATPase IIA family) lacks an N‐terminal autoregulatory domain. Four members of this family have been identified in the arabidopsis genome (termed AtECAs 1 to 4 by Axelsen and Palmgren, 2001). They are likely to be present in the plasma membrane, tonoplast and the ER/Golgi apparatus. In tomato, two transcripts of a type IIA Ca²⁺‐ATPase (LeLCA1) were observed in phosphate‐starved roots that correlated with two distinct protein isoforms (120 and 116 kDa) located in the plasma membrane and tonoplast of root cells, respectively (Navarro‐Avino et al., 1999). The second family of plant Ca²⁺ ATPases (the P‐type ATPase IIB family) is characterized by an autoinhibitory N‐terminal domain that contains a binding site for Ca‐CaM plus a serine‐residue phosphorylation site. Their catalytic activity can be modulated by [Ca²⁺]_cyt either through activation upon binding CaM or by inhibition following phosphorylation by Ca²⁺‐dependent protein kinases (CDPK; Hwang et al., 2000). Since CaM binding‐sites are generally quite diverse, each type‐IIB Ca²⁺‐ATPase may have a different affinity for CaM or may bind a different CaM isoform. Ten members of the type‐IIB Ca²⁺‐ATPase family have been identified in the arabidopsis genome (termed AtACAs 1, 2 and 4 and AtACAs 7 to 13 by Axelsen and Palmgren, 2001). These reside on various cellular membranes including the plasma membrane (AtACA8), the tonoplast (AtACA4), ER (AtACA2) and the plastid inner membrane (AtACA1). Several explanations for the abundance of Ca²⁺‐ATPase isoforms, and also the presence of several isoforms on the same cellular membrane (such as AtECA1 and AtACA2 in the root ER) have been proposed (Geisler et al., 2000; Sze et al., 2000; Axelsen and Palmgren, 2001). Such explanations typically suggest that individual isoforms are functionally distinct and specialized to specific cellular processes requiring distinct spatial or temporal expression. They also imply a requirement for CaM‐independent and CaM‐dependent regulation of Ca²⁺‐ATPase activities in the modulation of [Ca²⁺]_cyt perturbations during cell signalling. Multiple Ca²⁺‐ATPase genes are required to effect this because alternative splicing events are rare in plants. Intriguingly, the expression of many Ca²⁺‐ATPases is increased upon exposure to high salinity or high [Ca²⁺]_ext, and some Ca²⁺‐ATPase genes are expressed only under stress conditions (Geisler et al., 2000; Garciadeblas et al., 2001). This may reflect a role in maintaining [Ca²⁺]_cyt homeostasis or in reducing Na⁺ influx to the cytosol in saline environments.

The H⁺/Ca²⁺‐antiporters present in the plasma membrane and tonoplast have been characterized biochemically (Evans and Williams, 1998; Sanders et al., 2002). These have a lower affinity for Ca²⁺ than the Ca²⁺‐ATPases, and may also transport Mg²⁺. The stoichiometry of the dominant H⁺/Ca²⁺‐antiporter in the tonoplast is apparently 3H⁺/1Ca²⁺ (Blackford et al., 1990). Eleven genes encoding putative H⁺/Ca²⁺‐antiporters (AtCAX) have been identified in the genome of Arabidopsis thaliana (Hirschi, 2001; Mäser et al., 2001). The transporters AtCAX1, AtCAX2 and AtCAX4 are located at the tonoplast (Hirschi, 2001; N. Cheng et al., 2002, 2003). The AtCAX1 antiporter exhibits both a high affinity and high specificity for Ca²⁺. By contrast, the AtCAX2 transporter is a high‐affinity, high capacity H⁺/heavy metal cation antiporter. The cation specificity of other AtCAXs is unknown, but may reside in a specific stretch of nine amino acids termed the ‘Ca²⁺ domain’ (Shigaki et al., 2002). The increased transport capacity of truncated versions of AtCAX1, AtCAX3 and AtCAX4, and their relative inhibition by synthetic peptides corresponding to these deletions, suggest that they are subject to autoinhibition by their N‐termini in an isoform‐specific manner (N. Cheng et al., 2002; Pittman et al., 2002a, b). How, and whether, this regulatory mechanism operates in vivo has yet to be revealed.

The AtCAXs have homologues in other plant species and their physiological roles have been investigated using transgenic plants (Hirschi, 2001). Transgenic tobacco overexpressing AtCAX1 exhibits Ca‐deficiency disorders, such as tipburn, metal‐hypersensitivity and susceptibility to chilling, that can be reversed by increasing Ca supply. Since these plants have increased [Ca]_shoot, Hirschi (2001) speculated that such phenotypes resulted from depleted [Ca²⁺]_cyt, suggesting that the principal role of AtCAX1 was to maintain [Ca²⁺]_cyt homeostasis by removing excess cytosolic Ca²⁺ to the vacuole, and noted that the expression of AtCAX1 and AtCAX3 (but not AtCAX2 or AtCAX4) was increased by raising Ca supply (Shigaki and Hirschi, 2000; Hirschi, 2001; N. Cheng et al., 2002).

Ca²⁺ influx to the cytosol: calcium channels

Calcium‐permeable channels have been found in all plant membranes (Fig. 4). They have been classified on the basis of their voltage‐dependence into depolarization‐activated (DACC), hyperpolarization‐activated (HACC) and voltage‐independent (VICC) cation channels (White, 2000; Miedema et al., 2001; Sanders et al., 2002). The presence of diverse classes of Ca²⁺‐permeable channels in a particular membrane is thought to enable physiological flexibility.

The principal roles of Ca²⁺‐permeable channels in the plasma membrane appear to be in cell signalling, but they may also contribute to nutritional Ca²⁺ fluxes in particular cell types (White, 1998, 2000; Miedema et al., 2001). Several types of DACCs have been observed in the plasma membrane of plant cells (White, 1998, 2000). Although each has distinct pharmacological and electrophysiological properties, all are permeable to both monovalent and divalent cations. They may therefore contribute to the uptake of essential or toxic cations in addition to Ca²⁺. Most DACCs activate significantly at voltages more positive than about –150 to –100 mV under physiological conditions (White, 1998). The dominant DACCs in protoplasts from arabidopsis tissues and carrot suspension cells appear to be controlled by cytoskeletal interactions and stabilized by the disruption of microtubules (Thion et al., 1996, 1998). It is argued that DACCs transduce general stress‐related signals since plasma membrane depolarization is common to many stimuli, occurs by many diverse mechanisms, and is likely to increase [Ca²⁺]_cyt throughout the cell periphery (White, 1998, 2000). However, specific roles for DACCs acting in tandem with cytoskeletal rearrangements have been proposed in the acclimation of chilling‐resistant plants to low temperatures (Mazars et al., 1997; White, 1998; Xiong et al., 2002) and in the interactions of plants with microbes (White et al., 2002a). The outward‐rectifying K⁺ channels (KORCs) found in the plasma membrane of plant cells are also Ca²⁺‐permeable DACCs (White et al., 2002a). These channels activate significantly at voltages more positive than about –50 mV under most physiological conditions and catalyse a large K⁺ efflux simultaneously with a small Ca²⁺ influx (White, 1997; Gaymard et al., 1998; De Boer, 1999; Roberts and Snowman, 2000). The Ca²⁺ influx through KORCs might increase [Ca²⁺]_cyt to coordinate ion transport, metabolism and gene expression. An elaborate model of how negative feedback through [Ca²⁺]_cyt might control the loading of K⁺ into the root xylem by KORCs has been proposed by De Boer (1999).

Patch‐clamp electrophysiological techniques have identified HACCs in root cells (Kiegle et al., 2000a; Véry and Davies, 2000; Foreman et al., 2003), onion epidermal cells (Pickard and Ding, 1993), suspension‐cultured tomato cells (Blumwald et al., 1998), leaf mesophyll cells (Stoelzle et al., 2003) and stomatal guard cells (Hamilton et al., 2000, 2001; Pei et al., 2000; Murata et al., 2001; Perfus‐Barbeoch et al., 2002). These channels are permeable to many divalent cations including Ba²⁺, Ca²⁺, Mg²⁺, Mn²⁺, Cd²⁺ and Zn²⁺. They activate at voltages more negative than about –100 to –150 mV at physiological [Ca²⁺]_cyt, but increasing [Ca²⁺]_cyt shifts their activation potential to more positive voltages in root hairs (Véry and Davies, 2000) or more negative voltages in guard cells (Hamilton et al., 2000). At the apex of root hairs and pollen tubes, and in cells of the root elongation zone, the regulation of HACCs by [Ca²⁺]_cyt may provide a positive feedback system to generate and maintain the elevated [Ca²⁺]_cyt required for cell expansion (White, 1998, 2000; Véry and Davies, 2000; Miedema et al., 2001; Demidchik et al., 2002a). Mechanosensitive HACCs may orchestrate the changes in morphology induced by gravity, touch or flexure in a similar manner (Pickard and Ding, 1993; Bibikova et al., 1999). Reactive oxygen species also increase the activity of HACCs in root hairs (Foreman et al., 2003), guard cells (Pei et al., 2000; Murata et al., 2001; Klüsener et al., 2002; Köhler et al., 2003) and suspension‐cultured cells (Lecourieux et al., 2002) through the activity of NADPH oxidases. Oxidative bursts have been correlated with elevated [Ca²⁺]_cyt, the initiation of [Ca²⁺]_cyt waves, elongation growth, altered gene expression and the induction of antimicrobial activities. Elicitor‐activated HACCs, such as those activated by a heterotrimeric G‐protein/protein kinase cascade in suspension‐cultured tomato cells, are thought to raise [Ca²⁺]_cyt to initiate cellular responses to pathogens (Blumwald et al., 1998). In guard cells, HACCs have a central role in closing stomata during water stress. Increasing ABA concentration shifts the activation potential of HACCs to more positive voltages, thereby promoting their opening, and the subsequent entry of Ca²⁺ not only depolarizes the plasma membrane but also initiates the [Ca²⁺]_cyt‐dependent events, including [Ca²⁺]_cyt‐dependent Ca²⁺ release (CICR) from intracellular stores, that lead to stomatal closure (Blatt, 2000a, b; White, 2000; Murata et al., 2001; Schroeder et al., 2001; Köhler and Blatt, 2002).

Many distinct VICCs are present in the plasma membrane of plant cells, which differ in cation selectivity, voltage‐dependence and pharmacology (White et al., 2002a; Demidchik et al., 2002b). Nevertheless, all VICCs appear to be permeable to both monovalent and divalent cations (Davenport and Tester, 2000; Demidchik and Tester, 2002; Demidchik et al., 2002a, b) and are likely to provide a weakly voltage‐dependent Ca²⁺ influx to cells under physiological ionic conditions (White and Davenport, 2002; Demidchik et al., 2002a). It has been suggested that Ca²⁺ influx through VICCs, which are open at physiological voltages and are generally insensitive to cytoplasmic modulators, is required to balance the perpetual Ca²⁺ efflux through Ca²⁺‐ATPases and H⁺/Ca²⁺‐antiporters to maintain [Ca²⁺]_cyt homeostasis in an unstimulated plant cell (White and Davenport, 2002). This hypothesis is supported both by the pharmacology of Ca²⁺ influx to plant root cells and the linear dependence of their [Ca²⁺]_cyt on cell membrane potential (Demidchik et al., 2002a). Furthermore, VICCs appear to be the only Ca²⁺‐permeable channels open at the resting potential of most plant cells.

There has been some recent speculation on the identity of genes encoding Ca²⁺‐permeable channels in the plasma membranes of plant cells (Demidchik et al., 2002b; Véry and Sentenac, 2002; White et al., 2002a). It has been suggested (a) that homologues of the arabidopsis AtTPC1 gene (Furuichi et al., 2001) encode DACCs regulated by [Ca²⁺]_cyt, (b) that homologues of the arabidopsis AtSKOR (Gaymard et al., 1998) and AtGORK (Ache et al., 2000) genes encode KORCs, (c) that the annexin genes encode HACCs and (d) that genes from the cyclic‐nucleotide gated channel (CNGC) and glutamate receptor (GLR) families encode VICCs. Since CNGCs have a binding site for CaM within a C‐terminus binding site for cyclic nucleotide monophosphates, CNGC could impose a ‘coincidence hierarchy’ for [Ca²⁺]_cyt signalling by integrating the response to two different signalling cascades acting simultaneously. The low‐affinity cation transporter in the plasma membrane of wheat root cells (TaLCT1) could also facilitate Ca²⁺ uptake into plant cells (Clemens et al., 1998), but it is not known whether this protein forms an ion channel.

Several Ca²⁺‐permeable channels co‐reside in the tonoplast (Allen and Sanders, 1997; White, 2000; Sanders et al., 2002). At least three types of pharmacologically distinct depolarization‐activated Ca²⁺‐permeable channels have been observed, of which the SV type is the most common. The SV channels are permeable to monovalent and divalent cations and catalyse a small Ca²⁺ influx to the cytoplasm under physiological conditions (Pottosin et al., 1997, 2001; White, 2000). They are regulated by many effectors, suggesting that they could be pivotal in effecting the coincidence control of [Ca²⁺]_cyt signalling (Sanders et al., 1999, 2002; White, 2000). Increasing [Ca²⁺]_cyt or decreasing vacuolar Ca²⁺ concentration ([Ca²⁺]_vac) promotes their opening at physiological trans‐tonoplast voltages. Their response to [Ca²⁺]_cyt is sensitized both by CaM and by cytosolic Mg²⁺. Cytoplasmic alkalinization or vacuolar acidification also promotes their opening under physiological conditions. The SV channels are regulated by [Ca²⁺]_cyt‐dependent phosphorylation at two sites. Phos phorylation at one site is inhibitory, whereas phosphorylation at the other site activates the channel. It has been proposed that [Ca²⁺]_cyt‐dependent regulation of SV channels might prevent an excessive rise in [Ca²⁺]_cyt or modulate the kinetics of changes in [Ca²⁺]_cyt (Sanders et al., 1999). The activity of SV channels is also reduced by 14‐3‐3 proteins (van den Wijngaard et al., 2001). The genes(s) encoding SV channels are unknown, but KCO1 has been implicated in their formation, since SV‐channel currents in mesophyll protoplasts from the arabidopsis kco1 mutant are smaller than those from wild‐type plants (Schonknecht et al., 2002). At least two types of pharmacologically distinct hyperpolarization‐activated, Ca²⁺‐permeable channels (HACC) have been reported in plant vacuoles (Allen and Sanders, 1997; White, 2000). These open at trans‐tonoplast voltages within the physiological range (–20 to –70 mV) and catalyse Ca²⁺ influx to the cytoplasm. Both tonoplast HACCs show spontaneous changes in their kinetic behaviour, which has been interpreted as resulting from interactions with (hypothetical) regulatory ligands. One type of tonoplast HACC is inhibited by a [Ca²⁺]_cyt above 1 µm, whereas the other type is insensitive to [Ca²⁺]_cyt. The opening of the second type of HACC is promoted by increasing [Ca²⁺]_vac and/or vacuolar alkalinization.

Several highly selective Ca²⁺ channels activated by cytosolic second messengers (IP₃, IP₆ or cADPR) are also present in the tonoplast (Allen and Sanders, 1997; White, 2000; Sanders et al., 2002). All mediate Ca²⁺ influx to the cytoplasm. The IP₃‐dependent Ca²⁺ channels are activated half‐maximally by IP₃ concentrations as low as 200 nm. They open at physiological trans‐tonoplast voltages, and their opening is promoted by tonoplast hyperpolarization. They may have a role in turgor regulation in response to salt and hyper‐osmotic stresses (Allen and Sanders, 1997; Drøbak and Watkins, 2000; DeWald et al., 2001; Xiong et al., 2002), in nastic movements (Kim et al., 1996), in gravitropic movements of roots (Fasano et al., 2002) and pulvini (Perera et al., 1999), in stomatal closure (Staxén et al., 1999; Blatt, 2000a, b; Ng et al., 2001b; Schroeder et al., 2001; Klüsener et al., 2002) and in pollen tube elongation (Malhó et al., 1998; Rudd and Franklin‐Tong, 2001). They may also have roles in pollen tube self‐incompatibility (Malhó et al., 1998; Rudd and Franklin‐Tong, 2001) and in plant defence responses (Mithöfer et al., 1999; Sanders et al., 1999; Blume et al., 2000). The cADPR‐dependent Ca²⁺ channels also open at physiological trans‐tonoplast voltages, but are inhibited by [Ca²⁺]_cyt greater than 600 nm (Leckie et al., 1998). The pharmacology of plant cADPR‐dependent Ca²⁺ channels resembles that of the cADPR‐activated, ryanodine‐receptor channels found in the endomembranes of animal cells (Allen and Sanders, 1997). They are activated half‐maximally by 20–40 nm cADPR, 20 nm ryanodine or millimolar concentrations of caffeine. They are inhibited by micromolar concentrations of ruthenium red and procaine. The cADPR‐dependent Ca²⁺ channels have been implicated in ABA‐signalling pathways leading to cold acclimation, desiccation tolerance (Wu et al., 1997) and stomatal closure (Leckie et al., 1998; Klüsener et al., 2002), in circadian [Ca²⁺]_cyt rhythms and in the activation of plant defence responses (Durner et al., 1998; Klessig et al., 2000). Evidence for Ca²⁺ channels activated by sub‐micromolar concentrations of IP₆ and having a role in stomatal closure is circumstantial but persuasive (Lemtiri‐Chlieh et al., 2000).

A variety of voltage‐dependent and ligand‐activated Ca²⁺ channels have also been revealed in the ER using biochemical or electrophysiological techniques (White, 2000). Depolarization‐activated Ca²⁺ channels were observed when ER vesicles from the touch‐sensitive tendrils of Bryonia dioica (BCC1) (Klüsener et al., 1995) or the root tips of Lepidium sativum (LCC1) (Klüsener and Weiler, 1999) were incorporated into planar lipid bilayers. Their activities were modulated by the Ca²⁺ gradient across the ER membrane and enhanced by cytoplasmic acidification. An interesting model for the generation of oscillations in [Ca²⁺]_cyt through BCC1 has been proposed (Klüsener et al., 1995). At the resting membrane potential of the ER, which is assumed to be close to zero, and with submicromolar Ca²⁺ activities in the lumen of the ER, BCC1 is likely to be closed. However, if the lumenal Ca²⁺ activity is increased, for example by the activity of an ER Ca²⁺‐ATPase, BCC1 will open to release Ca²⁺ into the cytoplasm. This reduces the lumenal Ca²⁺ and the channel recloses. This cycle will generate transient local elevations of [Ca²⁺]_cyt. The frequency of these elevations could be determined by modulation of BCC1, the Ca²⁺‐ATPase or the Ca²⁺ gradient across the ER. In addition to these voltage‐dependent Ca²⁺ channels, biochemical studies indicate the presence of Ca²⁺ channels activated by cADPR (Navazio et al., 2001), NAADP (nicotinic acid adenine dinucleotide phosphate) (Navazio et al., 2000) and (possibly) IP₃ (Muir and Sanders, 1997). To date, no genes encoding ER Ca²⁺ channels have been identified in plants, and there appear to be no plant genes homologous to the IP₃‐receptors, cADPR‐receptors, NAADP‐receptors or endomembrane voltage‐gated channels of animals.

A hyperpolarization‐activated Ca²⁺‐permeable channel has been recorded in excised patches from the nuclear envelope of red beet cells (Grygorczyk and Grygorczyk, 1998). This channel was unaffected by changes in [Ca²⁺]_cyt, but when the [Ca²⁺] of the perinuclear space (the lumen of the nuclear envelope) exceeded about 1 µm the channel opened at physiological voltages. Based on this observation, a role for this channel in regulating Ca²⁺‐dependent nuclear processes was proposed (Grygorczyk and Grygorczyk, 1998). Interestingly, annexin‐like proteins with the potential to form Ca²⁺‐permeable cation channels have been located at the perinuclear membrane and in the nucleolus (Clark et al., 1998; Kovács et al., 1998; de Carvalho‐Niebel et al., 2002).

CYTOSOLIC CALCIUM SIGNALS

The evolution of [Ca²⁺]_cyt signalling and the [Ca²⁺]_cyt ‘signature’

The [Ca²⁺]_cyt of plant cells increases in response to many developmental cues and environmental challenges (Table 2). This is considered essential for producing a physiological response. It is thought that elevating [Ca²⁺]_cyt is a primitive, and universal, response to stress. Sanders et al. (1999) observed that the low solubility product of Ca²⁺ and phosphate would have necessitated a [Ca²⁺]_cyt lower than the [Ca²⁺] of seawater to maintain energy metabolism. They presumed that this required the early evolution of mechanisms to remove Ca²⁺ from the cytoplasm, and noted that a homeostatically maintained submicromolar [Ca²⁺]_cyt would have been ideal for the subsequent evolution of [Ca²⁺]_cyt signalling systems, since it would confer sensitivity and speed to any signal. Sanders et al. (1999) also noted that the chemistry of Ca²⁺, which can coordinate six to eight uncharged oxygen atoms, had fortuitously made possible the evolution of proteins that change conformation upon binding Ca²⁺, allowing the cellular perception and transduction of a [Ca²⁺]_cyt signal.

Table 2.

Examples of the developmental processes and responses to abiotic and biotic challenges initiated by a perturbation in cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt)

Developmental process or environmental challenge	Characteristic [Ca2+]cyt perturbation	Stores releasing Ca2+ to cytosol	Selected references
Pollen tube elongation	Oscillation of high apical [Ca2+]cyt	Apoplast and internal	Malhó and Trewavas, 1996; Holdaway‐Clarke et al., 1997; Malhó et al., 1998; 2000; Messerli et al., 2000; Rudd and Franklin‐Tong, 2001
Pollen tube self‐incompatibility response	Intracellular [Ca2+]cyt wave in shank	ApoplastInternal (IP3‐dependent)	Rudd and Franklin‐Tong, 2001; Straatman et al., 2001; Franklin‐Tong et al., 2002
Cell polarity after fertilization	Intracellular [Ca2+]cyt wave from sperm fusion site leading to sustained [Ca2+]cyt elevation	Apoplast	Antoine et al., 2001
Cell division	Elevated [Ca2+]cyt		Bush, 1995
Seed germination (giberellins)	Slow rise in [Ca2+]cyt		Bush, 1995; Anil and Sankara Rao, 2001
Apoptosis	Slow, sustained [Ca2+]cyt elevation		Levine et al., 1996
Red light	Elevated [Ca2+]cyt	Apoplast	Shacklock et al. 1992; Malhó et al., 1998
Blue light	Brief spike in [Ca2+]cyt (seconds)	Apoplast	Malhó et al., 1998; Baum et al., 1999
Circadian rhythms	Circadian [Ca2+]cyt oscillation		Johnson et al., 1995; Wood et al., 2001
Stomatal closure (ABA, sphingosine‐1‐phosphate)	(1) Elevated [Ca2+]cyt at cell periphery(2) Elevated [Ca2+]cyt around vacuole(3) Oscillations in [Ca2+]cyt	(1) Apoplast(2) Vacuole(3) Apoplast and internal	McAinsh et al., 1992; Allen et al., 1999, 2000; Blatt, 2000a, b; White, 2000; Anil and Sankara Rao, 2001; Evans et al.. 2001; Ng et al., 2001a, b; Schroeder et al., 2001; Klüsener et al., 2002
CO2	Elevated [Ca2+]cyt in guard cells	Apoplast	Webb et al., 1996
Increasing apoplastic Ca2+	Oscillations in [Ca2+]cyt of guard cells	Apoplast	McAinsh et al., 1995; Allen et al., 1999, 2000
Auxin responses	(1) Slow, prolonged [Ca2+]cyt increase		Felle, 1988; Malhó et al., 1998; Ng et al., 2001b; Plieth, 2001; Plieth and Trewavas, 2002
	(2) Oscillations in [Ca2+]cyt
Xylem K+ loading	Elevated [Ca2+]cyt		De Boer, 1999
Exocytosis	Elevated [Ca2+]cyt		Battey et al., 1999; Camacho and Malhó, 2003
Root cell elongation	Sustained [Ca2+]cyt elevation	Apoplast	Cramer and Jones, 1996; Demidchik et al., 2002a
Root hair elongation	Sustained high apical [Ca2+]cyt	Apoplast	Wymer et al., 1997; White, 1998; Bibikova et al., 1999
Inhibition of cyclosis	Elevated [Ca2+]cyt		Ayling and Clarkson, 1996
Nodulation (nod factors)	Initial [Ca2+]cyt rise then oscillations in [Ca2+]cyt	Apoplast	Cárdenas et al., 2000; Wais et al., 2000; Walker et al., 2000; Lhuissier et al., 2001; Shaw and Long, 2003
Senescence	Sustained [Ca2+]cyt elevation		Huang et al., 1997
UV‐B	Slow [Ca2+]cyt rise, elevated [Ca2+]cyt sustained for several minutes	Apoplast	Frohnmeyer et al., 1999
Heat‐shock	Elevated [Ca2+]cyt sustained for 15–30 min	Apoplast and internal (IP3‐dependent)	Gong et al., 1998; Malhó et al., 1998
Cold‐shock	(1) Single brief [Ca2+]cyt spike (seconds)(2) Oscillations in [Ca2+]cyt	(1) Apoplast	Knight et al., 1991; Malhó et al., 1998; White, 1998; Plieth et al., 1999; van der Luit, 1999; Allen et al., 2000; Kiegle et al., 2000b; Knight, 2000; Cessna et al., 2001; Plieth, 2001
Slow cooling	Biphasic(1) Brief [Ca2+]cyt spike (seconds)(2) Slow [Ca2+]cyt elevation (minutes)	(1) Apoplast(2) Apoplast and internal (IP3‐dependent)	Knight et al., 1996; Plieth et al., 1999; Knight, 2000; Knight and Knight, 2000; Moore et al., 2002
Oxidative stress (paraquat, superoxide, H2O2, ozone)	(1) Brief [Ca2+]cyt spike (2) Sustained [Ca2+]cyt elevation(3) Oscillations in [Ca2+]cyt	(1) Apoplast(2) Apoplast and internal (IP3‐dependent)	Price et al., 1994; Levine et al., 1996; McAinsh et al., 1996; Knight et al., 1998; Malhó et al., 1998; Clayton et al., 1999; Allen et al., 2000; Kawano and Muto, 2000; Knight, 2000; Klüsener et al., 2002; Lecourieux et al., 2002
Anoxia	Biphasic(1) Slow spike (duration of minutes)(2) Sustained [Ca2+]cyt elevation (hours)	(1) Apoplast(2) Internal including mitochondria	Subbaiah et al., 1994, 1998; Sedbrook et al., 1996; Malhó et al., 1998; Plieth, 2001
Drought/hyper‐osmotic stress (mannitol)	Biphasic(1) Slow spike (duration of minutes)(2) Sustained [Ca2+]cyt elevation (hours)	Apoplast and vacuole	Knight et al., 1997, 1998; Malhó et al., 1998; Kiegle et al., 2000b; Cessna et al., 2001; Pauly et al., 2001; Plieth, 2001
Salinity (NaCl)	BiphasicTissue [Ca2+]cyt wave(1) Slow spike (duration of minutes)(2) Sustained [Ca2+]cyt elevation (hours)(3) Reduced [Ca2+]cyt (days)	Apoplast and vacuole (IP3‐dependent)	Knight et al., 1997; Kiegle et al., 2000b; Knight, 2000; DeWald et al., 2001; Pauly et al., 2001; Moore et al., 2002; Halperin et al., 2003
Hypo‐osmotic stress	Biphasic(1) Small [Ca2+]cyt elevation(2) Large [Ca2+]cyt elevation	(1) Apoplast(2) Apoplast and internal (IP3‐dependent)	Takahashi et al., 1997; Malhó et al., 1998; Knight, 2000; Cessna and Low, 2001; Cessna et al., 2001; Pauly et al., 2001; Plieth, 2001
Mechanical stimulation (motion, touch, wind)	Single brief [Ca2+]cyt spike (seconds)Tissue [Ca2+]cyt wave	Internal	Knight et al., 1991, 1992; Haley et al., 1995; Legue et al., 1997; Malhó et al., 1998; van der Luit, 1999; Plieth, 2001; Fasano et al., 2002
Aluminium stress	Elevated [Ca2+]cyt		Zhang and Rengel, 1999
Pathogens (elicitors)	Biphasic(1) Slow spike (duration of minutes)(2) Sustained [Ca2+]cyt elevation (hours)(3) Oscillations in [Ca2+]cyt	(1) Apoplast(2) Apoplast and internal (IP3‐dependent)	Knight et al., 1991; Malhó et al., 1998; Mithöfer et al., 1999; Blume et al., 2000; Fellbrich et al., 2000; Grant et al., 2000; Cessna and Low 2001; Cessna et al. 2001; Rudd and Franklin‐Tong, 2001; Klüsener et al., 2002; Lecourieux et al., 2002
	(the relative magnitude of different phases varies with elicitor identity)

The kinetics of the [Ca²⁺]_cyt perturbations and the location of stores releasing Ca²⁺ into the cytosol associated with each developmental process or environmental challenge are indicated.

To effect an appropriate physiological response to a particular stimulus, it is thought that the [Ca²⁺]_cyt perturbation (termed the [Ca²⁺]_cyt ‘signature’) elicited by each developmental cue or environmental challenge is unique. This uniqueness is manifest in the sub‐cellular location and/or the kinetics or magnitude of the [Ca²⁺]_cyt perturbation (McAinsh and Hetherington, 1998; Trewavas, 1999; Rudd and Franklin‐Tong, 2001). An increase in [Ca²⁺]_cyt is effected by Ca²⁺ influx to the cytosol either from the apoplast, across the plasma membrane, or from intracellular stores such as the ER or vacuole. This Ca²⁺ influx is mediated by Ca²⁺‐permeable ion channels, and their type, cellular location and abundance will influence the spatial characteristics of [Ca²⁺]_cyt perturbations. Since the diffusion of Ca²⁺ within the cytoplasm is low (Clapham, 1995), and the buffering of Ca²⁺ in the cytoplasm is high (0·1 to 1 mm; Malhó et al., 1998; Trewavas, 1999), the opening of a Ca²⁺ channel produces a local increase in [Ca²⁺]_cyt that dissipates rapidly after the channel has closed. The subcellular localization of Ca²⁺ channels is therefore critical for the targeting of different cellular processes. In some circumstances, the proteins responding to the changes in [Ca²⁺]_cyt must either be associated with the Ca²⁺ channel itself or tethered closely to the membrane. Thus, the opening of Ca²⁺ channels influences local biochemical processes.

To coordinate cellular responses, [Ca²⁺]_cyt ‘waves’ are produced within the cytoplasm by the successive recruitment of receptive Ca²⁺ channels. Various authors have speculated how these waves might be initiated and propagated. Trewavas (1999) suggested that a local elevation of [Ca²⁺]_cyt might generate soluble second messengers, such as IP₃ or cADPR, that diffuse through the cytoplasm to activate a relay of spatially separated Ca²⁺‐channels. This might occur during the [Ca²⁺]_cyt waves observed in the shank of poppy pollen tubes during the self‐incompatibility response (Straatman et al., 2001; Rudd and Franklin‐Tong, 2001) or in plant cells responding to salt stress (Drøbak and Watkins, 2000). Antoine et al. (2001) suggested that the [Ca²⁺]_cyt wave that crosses the egg following fertilization was the result of the successive activation of (mechanosensitive) Ca²⁺ channels in the plasma membrane radiating from the site of sperm entry, and Franklin‐Tong et al. (2002) suggested that repetitive Ca²⁺ influx across the plasma membrane contributed to the [Ca²⁺]_cyt waves that occurred in the shank of poppy pollen tubes during the self‐incompatibility response. The spatial changes in elevated [Ca²⁺]_cyt that occur following the application of ABA to guard cells, first close to the plasma membrane and subsequently adjacent to the vacuole (McAinsh et al., 1992; Allen et al., 1999), are thought to reflect the sequential opening of hyperpolarization‐activated Ca²⁺ channels at the plasma membrane and then second‐messenger activated Ca²⁺ channels in the tonoplast (Grabov and Blatt, 1998; White, 2000; Schroeder et al., 2001). In addition to these subcellular [Ca²⁺]_cyt waves, ‘waves’ of cells with high [Ca²⁺]_cyt may also propagate through plant tissues. These can be induced in root tissues by mechanical stimulation (Legué et al., 1997; Fasano et al., 2002) or saline shock (Moore et al., 2002), in cotyledons by cold shock (Knight et al., 1993) and in leaves by chilling plant roots briefly (Campbell et al., 1996). Electrical action potentials, osmotic perturbations or chemical signals may trigger these waves.

Although an elevated [Ca²⁺]_cyt is necessary for signal transduction, a prolonged increase in [Ca²⁺]_cyt is lethal. Indeed, sustained high [Ca²⁺]_cyt is implicated in apoptosis, both during normal development (e.g. in tissue patterning and xylogenesis) and in hypersensitive responses to pathogens (Levine et al., 1996). To effect other responses, [Ca²⁺]_cyt perturbations must be either of low amplitude or transient. Transient increases in [Ca²⁺]_cyt can be single (spike), double (biphasic) or multiple (oscillations). Unique [Ca²⁺]_cyt spikes can be generated by delaying the [Ca²⁺]_cyt rise, or by altering the rate of change of [Ca²⁺]_cyt, the maximal [Ca²⁺]_cyt reached or the duration [Ca²⁺]_cyt is above a certain threshold. Oscillations can differ in their [Ca²⁺]_cyt amplitudes, periodicity or duration (Evans et al., 2001). The production of [Ca²⁺]_cyt waves allows specific spatiotemporal [Ca²⁺]_cyt perturbations to be generated that may differ in their cellular location, rate and extent of propagation and/or their [Ca²⁺]_cyt amplitude during propagation (Malhó et al., 1998). However, despite the seductive logic of a [Ca²⁺]_cyt signature for each developmental cue or environmental challenge, plant tissues are comprised of populations of heterogeneous cells that may have contrasting abilities to generate [Ca²⁺]_cyt signatures. For example, (a) within the root, the [Ca²⁺]_cyt perturbations induced by mechanical perturbation (Legué et al., 1997), salinity, osmotic stress, cold shock or slow cooling (Kiegle et al., 2000b; Moore et al., 2002) differ markedly between cell types, and (b) shoot cells exhibit a biphasic [Ca²⁺]_cyt perturbation during anoxia, whereas only a slow increase in [Ca²⁺]_cyt is observed in root cells (Sedbrook et al., 1996; Plieth, 2001). Indeed, even cells of the same type, such as the two guard cells of a stomate, seldom generate an identical [Ca²⁺]_cyt in response to a defined stimulus (Allen et al., 1999). This heterogeneity may arise from differences in cellular cytology, the distribution of cytoplasmic Ca²⁺ buffers and/or the activities of Ca²⁺ transporters.

The [Ca²⁺]_cyt signature: temporal and spatial aspects

Several abiotic challenges result in an immediate, transient increase in [Ca²⁺]_cyt that is restored to basal levels within minutes (Table 2). Such challenges include mechanical perturbation (Knight et al., 1991, 1992; Haley et al., 1995; Legué et al., 1997; Malhó et al., 1998; van der Luit, 1999; Plieth, 2001; Fasano et al., 2002) and rapid cooling for brief periods, termed ‘cold‐shock’ (Knight et al., 1991; Plieth et al., 1999; van der Luit, 1999; Kiegle et al., 2000b; Plieth, 2001). By contrast, sustained cooling below a threshold temperature results in a biphasic response in [Ca²⁺]_cyt, in which an initial transient increase in [Ca²⁺]_cyt is followed by a more prolonged, second transient elevation of [Ca²⁺]_cyt (Plieth et al., 1999; Knight, 2000; Knight and Knight, 2000; Moore et al., 2002). Heat shock, acute salt (NaCl) stress, hyper‐osmotic (mannitol) stress, annoxia and exposure to oxidative stress also elicit an immediate, transient increase in [Ca²⁺]_cyt in plant cells, which is followed by a more prolonged elevation of [Ca²⁺]_cyt lasting many minutes or hours (Table 2; McAinsh et al., 1996; Sedbrook et al., 1996; Knight et al., 1997, 1998; Gong et al., 1998; Clayton et al., 1999; Kiegle et al., 2000b; Knight, 2000; DeWald et al., 2001; Plieth, 2001; Lecourieux et al., 2002; Moore et al., 2002). Biphasic [Ca²⁺]_cyt perturbations have been observed in plant cells in response to diverse pathogens and elicitors, but the kinetics of these responses varies greatly (Lecourieux et al., 2002). Rudd and Franklin‐Tong (2001) speculated that the slow generation of a sustained [Ca²⁺]_cyt elevation was a common feature of these responses, and the sustained increase in [Ca²⁺]_cyt alone was subsequently correlated with the induction of defence responses (Cessna and Low, 2001; Lecourieux et al., 2002).

Several developmental cues and environmental challenges produce [Ca²⁺]_cyt oscillations (Table 2). The duration, periodicity and amplitude of [Ca²⁺]_cyt oscillations vary considerably, and their form is often dependent on the strength and combination of specific stimuli (Felle, 1988; McAinsh et al., 1995; Allen et al., 1999; Staxén et al., 1999; Evans et al., 2001). Periodicities may range from the circadian oscillations observed in various plant cells (Johnson et al., 1995; Wood et al., 2001), through the slow oscillations observed in coleoptile cells following the addition of auxin (periodicity ≈30 min; Felle, 1988) and in guard cells either spontaneously or in response to stimuli such as ABA, [Ca²⁺]_ext, ozone, cold‐shock and elicitors (periodicity >3 min; Grabov and Blatt, 1998; Allen et al., 1999, 2000; Staxén et al., 1999; Evans et al., 2001; Ng et al., 2001a; Klüsener et al., 2002), to the rapid oscillations in [Ca²⁺]_cyt observed at the tip of growing pollen tubes (periodicity 40–75 s; Holdaway‐Clarke et al., 1997; Malhó et al., 1998, 2000; Messerli et al., 2000; Rudd and Franklin‐Tong, 2001) and in root hairs during nodulation (periodicity 1–2 min; Wais et al., 2000; Walker et al., 2000; Shaw and Long, 2003). It is noteworthy that oscillations can be induced in guard cells by releasing caged Ca²⁺ into the cytosol (McAinsh et al., 1995) and oscillatory [Ca²⁺]_cyt waves can be induced in pollen tubes by releasing caged Ca²⁺ or IP₃ into the cytosol (Malhó and Trewavas, 1996; Malhó, 1998). Several hypotheses have been proposed for the generation of [Ca²⁺]_cyt oscillations. These include models based on the successive emptying and refilling of finite Ca²⁺ stores (Klüsener et al., 1995; Harper, 2001), on the successive activation and deactivation of target Ca²⁺ channels through co‐incident signalling cascades (Sanders et al., 1999; White, 2000) or by the stretching and relaxation of a membrane (Holdaway‐Clarke et al., 1997).

Different stimuli generate contrasting spatial [Ca²⁺]_cyt perturbations by mobilizing Ca²⁺ from different cellular stores and/or by activating Ca²⁺ channels in a restricted location (Table 2; Sanders et al., 1999). For example, Ca²⁺ influx from the apoplast is implicated in elevating [Ca²⁺]_cyt during brief cold‐shock (White, 1998, 2000; Plieth et al., 1999; Kiegle et al., 2000b; Knight, 2000; Plieth, 2001) and oxidative stress (Knight et al., 1997; Clayton et al., 1999), but release from internal stores contributes significantly to the [Ca²⁺]_cyt elevations in response to prolonged cooling, salinity, osmotic stresses or pathogens (Knight et al., 1996, 1997; Blume et al., 2000; Knight, 2000; Pauly et al., 2001; Lecourieux et al., 2002; Moore et al., 2002) and dominates the [Ca²⁺]_cyt elevations in response to mechanical perturbations (Haley et al., 1995; Legué et al., 1997) and anoxia (Subbaiah et al., 1994, 1998). Van der Luit et al. (1999) demonstrated that separate nuclear [Ca²⁺] and [Ca²⁺]_cyt perturbations were initiated in tobacco seedlings in response to wind and cold shock, respectively, and that each led to the expression of a different CaM isoform. Pauly et al. (2001) similarly proposed that different cytoplasmic and nuclear signals were involved in discriminating between hypo‐ and hyper‐osmotic shock.

Graded [Ca²⁺]_cyt responses to stimuli

Perturbations in [Ca²⁺]_cyt often show a graded response that is proportional to the strength of the stimulus. For example, the frequency and amplitude of oscillations in the [Ca²⁺]_cyt of guard cells varies directly with the apoplastic ABA (Staxén et al., 1999) or sphingosine‐1‐phosphate (Ng et al., 2001a) concentration, with corresponding effects on the kinetics of stomatal closure. The form of the [Ca²⁺]_cyt perturbations that occur during cooling are a complex function of the rate of cooling, the duration of cooling, and the magnitude of the temperature drop (Plieth et al., 1999). Plant cells respond to a brief cold‐shock with an immediate, transient increase in [Ca²⁺]_cyt that is proportional to the rate of cooling, whereas they respond to sustained cooling below a threshold temperature with a prolonged biphasic [Ca²⁺]_cyt perturbation (Table 2). The magnitude of the [Ca²⁺]_cyt elevation induced in roots following their reorientation in a gravitational field is proportional to the angle of displacement up to a maximum of 135°, which corresponds to the optimal angle for amyloplasts to slide down the cell wall (Plieth and Trewavas, 2002), and the [Ca²⁺]_cyt perturbation induced by mechanical stimulation is increased in proportion to the stress and/or the time an organ is in motion (Knight et al., 1992; Haley et al., 1995). Similarly, the magnitude of the initial [Ca²⁺]_cyt elevation induced in tobacco suspension cells by hypo‐ or hyper‐osmotic shock depends upon the change in the osmolarity of the bathing medium (Takahashi et al., 1997; Pauly et al., 2001), the magnitude of the [Ca²⁺]_cyt perturbation in response to elicitors increases proportionally to the elicitor concentration applied (Mithöfer et al., 1999; Blume et al., 2000; Lecourieux et al., 2002) and the magnitude of the [Ca²⁺]_cyt elevation induced by oxidative stress increases in proportion to the concentrations of ozone, H₂O₂ or methyl viologen applied (Price et al., 1994; Levine et al., 1996; McAinsh et al., 1996; Clayton et al., 1999; Lecourieux et al., 2002).

Explicit [Ca²⁺]_cyt perturbations produce defined physiological responses

Explicit [Ca²⁺]_cyt signatures are responsible for the activation of specific genes in animal cells (Dolmetsch et al., 1997, 1998; Li et al., 1998). This is consistent with the hypothesis that explicit [Ca²⁺]_cyt signatures produce defined physiological responses to specific developmental cues or environmental challenges. However, there have been few attempts to generate explicit [Ca²⁺]_cyt perturbations artificially in plants. Although several experiments have demonstrated that elevating [Ca²⁺]_cyt using ionophores induces the expression of Ca²⁺‐dependent genes and physiological responses to chilling and ABA (Monroy and Dhinsa, 1995; Sheen, 1996), the [Ca²⁺]_cyt perturbations produced by ionophores are unlikely to have the same form as the [Ca²⁺]_cyt perturbations elicited by these effectors. This suggested to Plieth (2001) that the form of the [Ca²⁺]_cyt signature was inconsequential for these responses. Indeed, Plieth (2001) further argued that the notion of an explicit [Ca²⁺]_cyt signature entirely responsible for a defined physiological response was fundamentally doubtful. He cited these observations: (a) both osmotic stress and salt stress induced similar [Ca²⁺]_cyt perturbations, but result in different levels of expression of the p5cs gene (Knight et al., 1997), and (b) almost any [Ca²⁺]_cyt perturbation could be elicited in plant cells with an appropriate manipulation of temperature (Plieth, 2001), to suggest that factors other than, or in addition to, [Ca²⁺]_cyt were involved in producing an appropriate response to a particular challenge. Consistent with this suggestion, [Ca²⁺]_cyt transients have recently been found to be necessary, but not sufficient, for the expression of ABA‐inducible genes in plant cells (Webb et al., 2001). Nevertheless, several biochemical and genetic consequences have been attributed to specific [Ca²⁺]_cyt signatures.

An insight into the [Ca²⁺]_cyt signatures required for specific cellular responses has been obtained by correlating biochemical or genetic responses with particular characteristics of [Ca²⁺]_cyt perturbations common to diverse stimuli or by dissecting [Ca²⁺]_cyt perturbations using pharmaceuticals. Thus, Cessna and Low (2001) used pharmaceuticals to alter the [Ca²⁺]_cyt perturbations in response to hypo‐osmotic shock to demonstrate that only the Ca²⁺ flux from internal stores during the second transient increase in [Ca²⁺]_cyt initiated an oxidative burst. They speculated that this was a consequence of either a local elevation of [Ca²⁺]_cyt and/or a local distribution of the enzymes generating H₂O₂. Similarly, Lecourieux et al. (2002) demonstrated that only the second, sustained increase in [Ca²⁺]_cyt elicited by cryptogein was required for a hypersensitive response and cell death. Blume et al. (2000) demonstrated that only a sustained [Ca²⁺]_cyt elevation induced phytoallexin synthesis in response to elicitors, and Clayton et al. (1999) showed that the induction of glutathione‐S‐transferase (GST) gene expression by ozone relied only on the second, transient [Ca²⁺]_cyt elevation. Finally, the changes in [Ca²⁺]_cyt perturbations following combinations of oxidative stress and hyper‐osmotic stress correlated well with the expression of Ca²⁺‐regulated osmotic stress induced genes (p5cs and rab18), and the acquisition of osmotic stress tolerance (Knight et al., 1998).

A direct demonstration that an explicit [Ca²⁺]_cyt perturbation was required to produce a defined physiological response was performed recently on arabidopsis guard cells (Allen et al., 2001). Artificially elevating [Ca²⁺]_cytin guard cells through hyperpolarization (Grabov and Blatt, 1998; Allen et al., 2000, 2001) or H₂O₂ (Allen et al., 2000; Pei et al., 2000) can induce stomatal closure. However, although a single elevation of [Ca²⁺]_cyt was sufficient for immediate stomatal closure, prolonged stomatal closure could only be induced if the initial increase in [Ca²⁺]_cyt was followed by [Ca²⁺]_cyt oscillations with a specific periodicity (Allen et al., 2001). Interestingly, appropriate [Ca²⁺]_cyt oscillations are induced in guard cells by ABA and stomates of the ABA‐insensitive mutant gca2, which shows aberrant [Ca²⁺]_cyt oscillations in response to ABA, can be induced to close when appropriate [Ca²⁺]_cyt are produced artificially (Allen et al., 2001).

Habituation, ‘learning’ and ‘memory’

Trewavas (1999) has likened the [Ca²⁺]_cyt signalling network to a basic cellular ‘memory’. He noted that, in common with a neural network, [Ca²⁺]_cyt signals have the following properties: they may be (a) spatially structured, through the occurrence and location of cellular components; (b) subject to coincidence control, since certain elements are the targets of multiple [Ca²⁺]_cyt cascades and can process or block specific signals arriving coincidentally; (c) synchronized, for example by membrane depolarization or a sudden increase in the concentration of a diffusable second messenger; (d) modified by exposure to specific challenges, through changes in the abundance of cellular components involved in [Ca²⁺]_cyt signals; and (e) predisposed by previous challenges to generate an appropriate [Ca²⁺]_cyt signal in response to a current challenge. He suggested that these prerequisites for cellular ‘learning’ allow plant cells to respond ‘intelligently’ to the challenges they experience through an innate phenotypic and physiological plasticity.

Previous sections have discussed the spatial distribution of Ca²⁺‐transporters within a plant cell and their modulation by components of diverse intracellular signalling cascades such as [Ca²⁺]_cyt, cytoplasmic pH, reactive oxygen species, kinases, phosphatases, cNMPs, IP₃ and cADPR. Such Ca²⁺ transporters could be a point of convergence, and integration, of many developmental and environmental signals. This phenomenon is termed ‘cross‐talk’ and might be a mechanism whereby plants develop cross‐tolerance to various biotic and abiotic stresses (Bowler and Fluhr, 2000; Knight, 2000). Similarly many targets of [Ca²⁺]_cyt signals are also regulated by components of diverse signalling cascades (see below).

There is considerable evidence that [Ca²⁺]_cyt signatures are modified by previous experience. A diminished [Ca²⁺]_cyt elevation upon repetitive stimulation and/or a refractory period during which an increase in [Ca²⁺]_cyt cannot be elicited by the same developmental cue or environmental challenge is commonly observed. The magnitude of the [Ca²⁺]_cyt perturbation elicited by reorientation in a gravitational field (Plieth and Trewavas, 2002), touch (Legué et al., 1997) or wind‐induced motion (Knight et al., 1992) becomes progressively smaller upon repeated stimulation and a refractory period of several minutes is required before a full response is observed again. Following one brief pulse of phototropically active blue light, arabidopsis seedlings will not respond maximally again for several hours (Baum et al., 1999). A second exposure to an elicitor does not influence [Ca²⁺]_cyt for several hours after its initial application (Blume et al., 2000), and plant cells challenged with H₂O₂ fail to respond to H₂O₂ again for several hours (Price et al., 1994). An anoxic treatment reduces the magnitude of the initial [Ca²⁺]_cyt peak and delays the sustained elevation of [Ca²⁺]_cyt in response to a second anoxic treatment (Sedbrook et al., 1996), and the magnitude of [Ca²⁺]_cyt perturbations is diminished by continued chilling and rewarming (Plieth et al., 1999). These observations may indicate a desensitization of signalling cascades or a depletion of stores releasing Ca²⁺ to the cytosol. By contrast, the magnitude of the second (vacuolar) [Ca²⁺]_cyt elevation observed during slow cooling was increased following a period of cold acclimation (Knight and Knight, 2000), and a pretreatment with mannitol increased the magnitude of the [Ca²⁺]_cyt perturbations subsequently induced by hyper‐osmotic shock (Knight et al., 1998).

There is also evidence that the [Ca²⁺]_cyt signatures elicited by one environmental challenge can be modified by prior exposure to a contrasting one. For example, the magnitude of the [Ca²⁺]_cyt perturbations in response to oxidative stress was reduced by a prior exposure to hyper‐osmotic stress (Knight et al., 1997; Knight, 2000), and the magnitude of the [Ca²⁺]_cyt perturbations in response to hyper‐osmotic stress was reduced by a prior exposure to oxidative stress (Knight et al., 1998; Knight, 2000). These observations may imply cross‐talk between the signalling cascades and/or Ca²⁺ channels and stores responding to these challenges. However, prior oxidative stress did not affect the [Ca²⁺]_cyt transient observed upon cold shock or touch (Price et al., 1994; Knight et al., 1998), suggesting that contrasting signalling cascades and/or Ca²⁺ stores and channels are recruited by these challenges. Similarly, during the refractory period following heat shock, during which additional heat shocks fail to elevate [Ca²⁺]_cyt, [Ca²⁺]_cyt can be elevated by cold shock or mechanical perturbation (Gong et al., 1998), and during the refractory period following wind‐induced motion, cells can still raise [Ca²⁺]_cyt in response to cold shock (Knight et al., 1992). The cross‐talk (and lack of cross‐talk) between diverse signalling cascades responding to each facet of the environment is thought to optimize a plant’s response to its surroundings.

RESPONDING TO CYTOSOLIC CALCIUM SIGNALS

To respond appropriately to a specific [Ca²⁺]_cyt perturbation, a cell must activate a unique combination of Ca²⁺‐binding proteins. These [Ca²⁺]_cyt sensors include CaMs, CaM‐like proteins, calcineurin B‐like (CBL) proteins and Ca²⁺‐dependent protein kinases (CDPKs). Many of these proteins bind Ca²⁺ using a helix–loop–helix structure termed the ‘EF hand’, which binds a single Ca²⁺ molecule with high affinity (Strynadka and James, 1989). Frequently, pairs of EF hands interact through antiparallel β‐sheets, which allows cooperativity in Ca²⁺ binding. When Ca²⁺ binds to [Ca²⁺]_cyt sensors their structural and/or enzymatic properties change and their subsequent interactions with target proteins can alter solute transport and enzymatic activities, cytoskeletal orientation, protein phosphorylation cascades and gene expression. It is believed that these changes result in stress tolerance and/or a developmental switch. The form of the physiological response is determined not only by the [Ca²⁺]_cyt perturbation itself, but also by the expression of the [Ca²⁺]_cyt sensors, their affinities for both Ca²⁺ and target proteins, and the abundance and activity of the target proteins. Since different cell types, and probably even individuals of the same cell type, have contrasting transcript, protein and enzyme profiles, a similar [Ca²⁺]_cyt perturbation will result in individual responses, which may contribute to phenotypic plasticity (Gilroy and Trewavas, 2001).

Calcium‐binding proteins: CaM and CaM‐related proteins

CaM is found in the apoplast and in the cytosol, ER and nucleus of plant cells. Within the cytosol, the estimated CaM concentration is 5 to 40 µm (Zielinski, 1998). Calmodulin has been implicated in Ca²⁺‐dependent responses to light, gravity, mechanical stress, phytohormones, pathogens, osmotic stress, salinity, heavy metals, xenobiotics, annoxia, oxidative stress, heat shock and chilling (Zielinski, 1998; Snedden and Fromm, 2001; Reddy, 2001; Rudd and Franklin‐Tong, 2001; Fasano et al., 2002). Calmodulin is a small (17 kDa), highly conserved, acidic protein with two globular domains each containing two EF hands connected by a flexible α‐helical spacer (Zielinski, 1998; Reddy, 2001; Snedden and Fromm, 2001; Luan et al., 2002). A Ca²⁺/CaM complex generally interacts with target proteins, although there are exceptions, such as the interaction of myosin‐like proteins with CaM alone (Reddy, 2001). The binding of Ca²⁺ to CaM, which has a K_d between 10^–7 and 10^–6 m, exposes a hydrophobic surface on each globular domain. This enables the Ca²⁺/CaM complex to wrap around its target protein and bind to its CaM‐binding site with an affinity in the nanomolar range through non‐specific van der Waals interactions. The co‐location of two EF hands in each globular domain allows Ca²⁺ to bind cooperatively, which ensures that CaM activation occurs over a narrow range of Ca²⁺ concentrations. The affinity of CaM for Ca²⁺ may be profoundly influenced by the presence of particular target proteins (Zielinski, 1998). Calmodulins bind to many different proteins implicated in diverse physiological processes including cation transport (including [Ca²⁺]_cyt homeostasis), cytoskeletal rearrangements and cell division, phytohormone and phospholipid signalling, disease resistance (including the oxidative burst) and stress tolerance (Zielinski, 1998; Reddy, 2001; Snedden and Fromm, 2001; Luan et al., 2002; Reddy et al., 2002). Calmodulins can also regulate gene expression by binding to specific transcription factors (Szymanski et al., 1996; Reddy et al., 2000; Yang and Poovaiah, 2000; Bouché et al., 2002).

Small gene families encode CaM isoforms in plants. Many species possess several CaM genes encoding identical proteins as well as other genes encoding divergent isoforms (Zielinski, 1998; Snedden and Fromm, 2001). It is thought that the presence of genes encoding identical CaMs may reflect a need for diverse tissue‐specific, developmental or stress‐induced expression, and the presence of genes encoding different CaM isoforms may reflect a requirement for specific interactions with target proteins (Reddy, 2001; Snedden and Fromm, 2001). Indeed, it has been suggested that particular CaM isoforms transduce specific environmental or developmental signals. Consistent with this hypothesis, it has been shown that environmental challenges rapidly up‐regulate the expression of different CaM isoforms. For example, a subset of CaM genes is up‐regulated by touch in various plants (Braam et al. 1997; Verma and Upadhyaya, 1998; Zielinski, 1998), cold shock and wind stimuli lead to the expression of different CaM isoforms (van der Luit et al., 1999), and specific CaM genes are up‐regulated in response to wounding or pathogens (Heo et al., 1999; Yamakawa et al., 2001), auxin or salinity (Botella and Arteca, 1994). Many CaM isoforms show tissue‐specific and/or developmentally regulated expression (Gawienowski et al. 1993; Takezawa et al., 1995; Yang et al., 1998; Yamakawa et al., 2001; Duval et al., 2002), and CaM isoforms can differentially regulate target enzymes or have contrasting affinities for Ca²⁺ and CaM‐binding peptides (Liao et al., 1996; Liu et al., 1998; Reddy et al., 1999; Köhler and Neuhaus, 2000; Lee et al., 2000; Duval et al., 2002; Zielinski, 2002). Calmodulins can also be post‐translationally trimethylated, which influences both their stability and physiological activities (Zielinski, 1998). All these properties are consistent with [Ca²⁺]_cyt signatures producing unique biochemical and physiological consequences in cells expressing specific CaM isoforms and target proteins. Interestingly, target proteins themselves may have CaM‐binding or non‐binding isoforms (Snedden and Fromm, 2001), which presumably allows for [Ca²⁺]_cyt‐dependent and [Ca²⁺]_cyt‐independent regulatory cascades.

Plants also possess ‘CAM‐like’ proteins. These have between one and six EF hands and a limited homology to CaM (defined arbitrarily as <75% identity with canonical CaM isoforms; Reddy, 2001; Snedden and Fromm, 2001; Luan et al., 2002; Zielinski, 2002). In arabidopsis, they include: CaBP‐22 (Ling and Zielinski, 1993), TCH2 and TCH3 (Braam et al., 1997), AtCP1 (Jang et al., 1998), centrins (Cordeiro et al., 1998), NADPH oxidases (AtrbohA–F; Torres et al., 1998), homologues of the rice ABA‐inducible EFA27 protein (Frandsen et al., 1996) and Ca²⁺‐binding protein phosphatases such as ABI1 and ABI2 (Leung et al., 1997). These proteins have been implicated in cellular responses to diverse environmental, developmental and pathological challenges.

Calcium‐binding proteins: calcineurin B‐like proteins

Calcineurin B‐like proteins (CBL) possess three EF hands (Luan et al., 2002). In arabidopsis there are at least ten AtCBL genes, including AtSOS3 (AtCBL4), which encodes a Ca²⁺‐sensor protein involved in salt tolerance (Liu and Zhu, 1998; Luan et al., 2002; Xiong et al., 2002). Many CBLs have a conserved myristoylation site in their N‐termini, which allows membrane association. It has been proposed that myristoylation of CBLs alters their cellular location and intracellular interactions (Luan et al., 2002). It is thought that particular CBLs transduce specific environmental or developmental signals. Consistent with this hypothesis is the induction of AtCBL1 expression by drought, salinity, cold and wounding (Kudla et al., 1999; Piao et al., 2001) and the accumulation of AtCBL1 and AtCBL2 transcripts in leaves upon illumination (Nozawa et al., 2001). A family of SNF1‐like serine/threonine protein kinases (AtCIPK1 through AtCIPK25) has been identified as Ca²⁺‐dependent targets for AtCBLs (Shi et al., 1999; Halfter et al., 2000; Albrecht et al., 2001; Y. Guo et al., 2001; Luan et al., 2002). The CIPKs interact with CBLs through a unique 24‐amino‐acid domain termed the ‘NAF domain’ (Albrecht et al., 2001) and require divalent‐cation cofactors for their activity (Luan et al., 2002). Both AtCIPK1 and AtCIPK2 require Mn²⁺ as a cofactor and AtSOS2 (AtCIPK24) requires Mg²⁺. Each CBL protein may interact with several CIPKs. For example, AtCBL1 interacts with numerous AtCIPKs (Shi et al., 1999) and AtSOS3 interacts with AtSOS2 (AtCIPK24) and at least seven other AtCIPKs (Halfter et al., 2000; Y. Guo et al., 2001). Conversely, certain AtCIPKs can interact with several AtCBLs (Shi et al., 1999; Kim et al., 2000; Albrecht et al., 2001; Y. Guo et al., 2001). Common interactions between AtCBLs and AtCIPKs may allow cross‐talk between signalling cascades, whereas preferential associations between CBLs and CIPKs are thought to underpin specific signalling cascades. In this context it is noteworthy that the expression of CIPKs (in addition to CBLs) can be tissue specific and/or regulated by environmental stresses such as cold, drought, salinity, wounding or nutrient starvation (Albrecht et al., 2001; Y. Guo et al., 2001; Kim et al., 2003).

Calcium‐binding proteins: calcium‐dependent protein kinases

The activity of many protein kinases can respond to [Ca²⁺]_cyt signals directly. These can be placed in one of four classes: Ca²⁺‐dependent protein kinases (CDPKs), CDPK‐related proteins (CRKs), CaM‐dependent protein kinases (CaMKs) and chimeric Ca²⁺‐ and CaM‐dependent protein kinases (CCaMKs). McAinsh and Hetherington (1998) proposed an interesting model for generating specific physiological responses to [Ca²⁺]_cyt perturbations through the action of Ca²⁺‐dependent and Ca²⁺‐independent protein kinases and phosphatases. In this model, specific [Ca²⁺]_cyt perturbations influence the degree of phosphorylation of a target protein leading to contrasting or graded responses.

The CDPKs are ubiquitous in plants. There are at least 34 genes encoding CDPKs in the arabidopsis genome (Harmon et al., 2001; S. H. Cheng et al., 2002) and similar numbers in other plant species. They generally have four EF hands at their C‐terminus that bind Ca²⁺ to activate their serine/threonine kinase activity. They act as monomers and many may be autoinhibited by autophosphorylation of a pseudosubstrate domain (S. H. Cheng et al., 2002). Different CDPKs have contrasting affinities for Ca²⁺ (Lee et al., 1998) and the binding of Ca²⁺ to some CDPKs is modulated by lipids, interactions with 14‐3‐3 proteins or phosphorylation (Reddy, 2001; S. H. Cheng et al., 2002; Sanders et al., 2002). No CDPKs appear to be integral membrane proteins, but many are associated with the cytoskeleton, nucleus, plasma membrane and ER (Reddy, 2001; Sanders et al., 2002). At least 24 arabidopsis CDPKs can potentially undergo myristoylation and palmitoylation at their N‐termini, which may facilitate their association with membranes (S. H. Cheng et al., 2002; Xiong et al., 2002).

The CDPKs are capable of converting [Ca²⁺]_cyt signals into biochemical and genetic consequences through the phosphorylation of diverse target proteins including membrane solute transporters (including the Ca²⁺‐ATPase, AtACA2), ion and water channels, NADPH oxidases, enzymes involved in carbon and nitrogen metabolism, cytoskeletal proteins, proteases and DNA‐binding proteins (Reddy, 2001; Rudd and Franklin‐Tong, 2001; S. H. Cheng et al., 2002; Sanders et al., 2002). They are implicated in pollen development, control of the cell cycle, phytohormone signal transduction, light‐regulated gene expression, gravitropism, thigmotropism, nodulation, cold acclimation, salinity tolerance, drought tolerance and responses to pathogens (Sheen, 1996; Saijo et al., 2000; Anil and Sankara Rao, 2001; Reddy, 2001; Romeis et al., 2001; S. H. Cheng et al., 2002; Xiong et al., 2002; Lee et al., 2003). It is thought that the possession of many CDPKs with contrasting Ca²⁺ affinities and target proteins allows plant cells to respond appropriately to specific [Ca²⁺]_cyt perturbations (Anil and Sankara Rao, 2001; S. H. Cheng et al., 2002; Sanders et al., 2002). In various plant species specific CDPKs are induced by cold, drought, salinity, annoxia, mechanical stress, wounding and pathogen elicitors (Urao et al., 1994; Breviario et al., 1995; Monroy and Dhinsa, 1995; Botella et al., 1996; Tähtiharju et al., 1997; Yoon et al., 1999; Saijo et al., 2000; Anil and Sankara Rao, 2001; Romeis et al., 2001; Chico et al., 2002; Lee et al., 2003). In addition, the activities of CDPKs are affected by post‐translational modifications during development (Anil et al., 2000) or in response to environmental challenges such as wounding or pathogens (Romeis et al., 2001).

Other protein kinases responding to [Ca²⁺]_cyt are less well characterized than the CDPKs. There are at least seven CRKs in the arabidopsis genome that are structurally similar to CDPKs, but have degenerate or truncated EF hands that may not be able to bind Ca²⁺ (Reddy, 2001). Orthologues of these are present in many plant species. Several CaMKs have also been cloned from arabidopsis and other plants (Zhang and Lu, 2003). Their kinase activity is activated by CaM‐dependent autophosphorylation and their catalytic activity is also modulated by CaM. They are expressed highly in rapidly growing cells and tissues of the root and flower (Zhang and Lu, 2003). Chimeric CCaMKs are expressed in the anthers of several plant species, including lily and tobacco (Liu et al., 1998), but none have been identified in arabidopsis (Zhang and Lu, 2003). The CCaMKs possess a CaM‐binding domain and three EF hands. They require only Ca²⁺ for autophosphorylation, but Ca²⁺ and CaM for substrate phosphorylation. Liu et al. (1998) demonstrated that different CaM isoforms have contrasting effects on substrate phosphorylation by lily and tobacco CCaMKs, suggesting that these kinases could respond to specific environmental or developmental challenges.

Calcium‐binding proteins: proteins without EF hands

Several proteins lacking EF hands are also capable of binding Ca²⁺ (Reddy, 2001; Anil and Sankara Rao, 2001). For example, the activity of phospholipase D (PLD), which cleaves membrane phospholipids into a soluble head group and phosphatidic acid, is regulated by [Ca²⁺]_cyt through a Ca²⁺/phospholipid binding‐site termed the ‘C2 domain’ (Wang, 2001). Phospholipase D activity is implicated in cellular responses to ethylene and ABA, α‐amylase synthesis in aleurone cells, stomatal closure, pathogen responses, leaf senescence and drought tolerance (Ritchie et al., 2002). Plants possess several PLD isoforms that differ in their affinity for Ca²⁺ and their modulation by phosphoinositides, free fatty acids and lysolipids (Wang, 2001). These biochemical modulators of PLD activity are the substrates or products of phospholipase C, which generates IP₃ and diacylglycerol, phospholipase A₂ and diacylglycerol kinase, both of which are regulated by CaM. It has therefore been suggested that [Ca²⁺]_cyt signalling cascades might coordinate the activities of these diverse enzymes to effect specific responses to contrasting environmental or developmental stimuli (Wang, 2001).

In annexins, Ca²⁺ is bound by the ‘endonexin fold’, which contains a characteristic GXGT‐{38}‐(D/E) motif (Delmer and Potikha, 1997). This often appears only in the first quarter of plant annexins, although animal annexins generally possess four such motifs. The binding of Ca²⁺ to annexins enables them to associate with membranes to form cation‐channels (Hofmann et al., 2000; White et al., 2002a). In plants, annexins are encoded by small gene families. Seven annexin genes (AnnAt1‐7) have been identified in arabidopsis (Clark et al., 2001; White et al., 2002a). Most annexin genes are expressed throughout the plant, but annexins are especially abundant in highly secretory cell types, where they are located at the cell periphery and may have a role in membrane fusion, membrane trafficking and/or secretion (Clark et al., 2001). The increased expression of certain annexins upon specific developmental or environmental challenges has linked them with Ca²⁺ signals during root nodulation, pathogen attack, ABA responses, fruit ripening and cold acclimation (White et al., 2002a).

Other Ca²⁺‐binding proteins include calreticulin, calsequestrin, calnexin and BiP, which sequester Ca²⁺ in the ER. These proteins are implicated in Ca²⁺ homeostasis, protein folding and post‐translational modifications (Crofts and Denecke, 1998; Michalak et al., 1998). The expression of calreticulin is increased in reproductive tissues. Calreticulin possesses two Ca²⁺‐binding domains. The first is a proline‐rich, low‐capacity/high‐affinity Ca²⁺‐binding domain (K_d =1 µm; 1 mol Ca²⁺ mol protein^–1) that contains a highly conserved stretch of amino acids (KPEDWD). The second is a high‐capacity/low‐affinity Ca²⁺‐binding domain (K_d = 2 mm; 25 mol Ca²⁺ mol protein^–1) comprising acidic amino acid residues (generally glutamate and aspartate). Both calreticulin and BiP, and a Ca²⁺‐binding protein (PCP) expressed in Brassica pistils and implicated in pollen–pistil interactions (Furuyama and Dzelzkalns, 1999) also possess high‐capacity/low‐affinity Ca²⁺‐binding domains. Calnexin possesses a proline‐rich low‐capacity/high‐affinity Ca²⁺‐binding domain.

Calcium and CaM‐binding transcription factors

Changes in [Ca²⁺]_cyt may influence nuclear Ca²⁺ activities, but cell nuclei are also capable of initiating autonomous Ca²⁺ signals (van der Luit et al., 1999; Pauly et al., 2001). Perturbations in nuclear Ca²⁺ activities are thought to regulate DNA transcription, the cell cycle during mitosis and meiosis, DNA replication and DNA fragmentation during programmed cell death (Snedden and Fromm, 2001). Signal transduction through Ca²⁺/CaM has been inferred in several instances (Yang and Poovaiah, 2000; Rudd and Franklin‐Tong, 2001; Snedden and Fromm, 2001; Townley and Knight, 2002). It is possible for Ca²⁺/CaM to interact with target proteins in the cytosol, which subsequently transduce a signal to the nucleus, or with target proteins in the nucleus itself (Snedden and Fromm, 2001). Several CaM isoforms have been located in the nuclei of plant cells. In animals, [Ca²⁺]_cyt signals appear to direct the translocation of CaM into the nucleus and environmental signals have been shown to alter the distribution of petunia CaM53 (Rodríguez‐Concepción et al., 1999) and rice OsCaM61 (Dong et al., 2002) between the plasma membrane and the nucleus by influencing their prenylation. Transcription factors are among the target proteins for Ca²⁺/CaM in plant cell nuclei (Szymanski et al., 1996; Reddy et al., 2000; Yang and Poovaiah, 2000; Bouché et al., 2002; Reddy et al., 2002). Such interactions have been suggested in the transduction of phytohormone signals, phytochrome‐ responses and acclimation to low temperatures. Interestingly, the expression of some CaM isoforms might be autoregulated through interactions between Ca²⁺/CaM and transcription factors (Szymanski et al., 1996). It has been suggested that this could supply stoichiometric levels of CaM isoforms for the activation of target proteins and, thereby, provide a mechanism for their long‐term or subsequent activation. A CaM‐like protein has also been implicated in the suppression of post‐transcriptional gene silencing in plants (Anandalakshmi et al., 2000). In addition to Ca²⁺/CaM signal transduction cascades, CaM‐independent CDPKs have also been implicated in the regulation of nuclear gene expression (Patharkar and Cushman, 2000) and a nuclear transcription factor with two EF hands (AtSUB1) has been implicated in transcriptional responses to blue and far‐red light (H. Guo et al., 2001). The arabidopsis genome contains two homologues of AtSUB1: AtSUL1 and AtSUL2.

PERSPECTIVES

It has long been known that Ca is an essential element for plants, and that plant species differ in both the amounts of Ca they require and their tolerance of Ca in the rhizosphere. These differences between plant species not only influence the natural flora of calcareous soils but also have consequences for breeding crops to improve the delivery of Ca to the human diet. Recent research in this area has concentrated on elucidating the physiology and biochemistry underpinning contrasting Ca physiotypes and in developing a phylogenetic framework to describe the variation in [Ca]_shoot within the angiosperms. This work will be used to inform the modelling of responses of plant communities to environmental change, of nutrient and contaminant partitioning within an ecosystem, and of the delivery of Ca to the food chain.

More recently, research on Ca in plants has been focused at the cellular level. This is presumably a consequence of the realization, some two decades ago, that [Ca²⁺]_cyt was an obligate intracellular messenger coordinating the cellular responses to numerous developmental cues and environmental challenges. Since then, biochemists have characterized the Ca²⁺‐ATPases and H⁺/Ca²⁺‐antiporters required to maintain submicromolar [Ca²⁺]_cyt in a resting cell, and electrophysiologists have characterized the Ca²⁺‐permeable channels that raise [Ca²⁺]_cyt to initiate [Ca²⁺]_cyt signals. Genes encoding these transport proteins have been identified, and their spatial and temporal expression patterns, subcellular locations, functional properties and physiological roles are being studied using a variety of biochemical, physiological, molecular biological and genetic techniques. In parallel, the origin and specificity of [Ca²⁺]_cyt signals are being addressed. In particular, the hypothesis that each developmental cue or environmental challenge produces a unique [Ca²⁺]_cyt signature to elicit specific cellular responses is being tested using cell‐specific targeting of transgenic aequorin (Knight et al., 1991) and cameleon (Allen et al., 1999) Ca²⁺‐reporters in combination with single‐cell transcript profiling (Karrer et al., 1995; Brandt et al., 2002). Researchers are also identifying the cellular [Ca²⁺]_cyt‐sensors that allow plant cells to respond appropriately to [Ca²⁺]_cyt signals, and elucidating their biochemical properties and interactions with target proteins. The involvement of specific [Ca²⁺]_cyt‐sensors in particular physiological responses is indicated by their diverse tissue expression patterns and by their contrasting changes in expression in response to developmental or environmental stimuli. This hypothesis will be tested through the analysis of transgenic plants mis‐expressing [Ca²⁺]_cyt‐sensors.

Undoubtedly, the application of functional genomics will provide further insights to the biology of Ca in plants. For example, oligonucleotide microarrays might be used to identify the genetic consequences of acclimation to low or high [Ca²⁺]_ext. This will provide a unique insight into the physiology of Ca accumulation. Genetic responses to changes in [Ca²⁺]_ext might include the differential expression of transport proteins, of enzymes in the biochemical pathways containing compounds that chelate Ca, or of transcriptional cascades that lead to the modification of plant anatomy and/or morphology. These responses could not be predicted a priori. In parallel, genes that impact on Ca accumulation in the shoot might be identified through the resolution of quantitative trait loci (QTL) detected using mapping populations that show variation in their ability to accumulate Ca in the shoot. These are available for several plant species. Information from both microarray and QTL analyses could be used to formulate hypotheses on the impact of specific genes on [Ca]_shoot, which might be tested by investigating the phenotype of transgenic plants mis‐expressing candidate genes. Such approaches could ultimately lead to the development of crops that accumulate less of contaminants, such as radiostrontium, or deliver more Ca to the food chain. The challenge of the immediate future will be to integrate the massive data sets arising from plant ‘omic’ technologies into a cohesive conceptual framework for Ca in plants that embraces both its roles in biochemistry and cell signalling.

ACKNOWLEDGEMENTS

We thank Graham Seymour (HRI) for comments on cell wall composition, Lim Ho (HRI) for his photographs of horticultural disorders, Marc Knight (Oxford) for the overview of [Ca²⁺]_cyt signals he presented at ‘Calcium 2002’ at HRI‐Wellesbourne, and Tim Flowers (Sussex) for his editorial encouragement. The research in our laboratories is supported by the Biotechnology and Biological Sciences Research Council (UK) and by the Department for Environment, Food and Rural Affairs (UK).

Received: 7 January 2003; Returned for revision: 26 March 2003; Accepted: 6 June 2003    Published electronically: 21 August 2003