The complexity of Ca2+ patterns observed in eukaryotic cells, including plants, has led to the hypothesis that specific patterns of Ca2+ propagation, termed Ca2+ signatures, encode information and relay it to downstream elements (effectors) for translation into appropriate cellular responses. Ca2+-binding proteins (sensors) play a key role in decoding Ca2+ signatures and transducing signals by activating specific targets and pathways. Calmodulin is a Ca2+ sensor known to modulate the activity of many mammalian proteins, whose targets in plants are now being actively characterized. Plants possess an interesting and rapidly growing list of calmodulin targets with a variety of cellular roles. Nevertheless, many targets appear to be unique to plants and remain uncharacterized, calling for a concerted effort to elucidate their functions. Moreover, the extended family of calmodulin-related proteins in plants consists of evolutionarily divergent members, mostly of unknown function, although some have recently been implicated in stress responses. It is hoped that advances in functional genomics, and the research tools it generates, will help to explain themultiplicity of calmodulin genes in plants, and to identify their downstream effectors. This review summarizes current knowledge of the Ca2+–calmodulin messenger system in plants and presents suggestions for future areas of research.
VI.CaM and plant responses to environmental stimuli 52
VII.Conclusions and future studies 58
1. Ca2+ as a signal carrier in living organisms
The concept of Ca2+ as a regulatory molecule in living organisms emerged over a hundred years ago with the finding of its effects on muscle contraction (Brini & Carafoli, 2000) and it has since gained recognition as the most ubiquitous cellular second messenger. It is believed that due to its relatively high abundance (mM range) in the water environments where living cells evolved, and its ability to efficiently co-ordinate with organic molecules and form complexes with phosphate, Ca2+ exclusion from the primordial intracellular environment was imperative for the evolution of cells and likely preceded its role as a signal carrier. Later, mechanisms of stimulus-induced Ca2+ release coupled to biochemical responses evolved. Currently, the roles of Ca2+ as a signal carrier are the subject of intensive investigations in both animals and plants. Ca2+ has been implicated in many aspects of the eukaryotic life cycle, development, and response to environmental cues. Consequently, elucidating the roles of Ca2+ as a regulatory molecule is expected to impact upon medicine and agriculture. Space limitations have excluded the possibility of covering all aspects of Ca2+/CaM signalling in plants and thus the authors apologise to any researchers whose findings we were not able to discuss.
Organisms continually monitor their environment and must respond to changes with adaptive or defensive mechanisms. Such responses are initiated at the molecular and biochemical levels and require the co-ordination of cellular events to ensure that a response is appropriate for a given stimulus. Sophisticated intra- and intercellular communication networks have therefore evolved in living organisms to convey information about a perceived stimulus to the cellular machinery responsible for mediating the responses. Various molecules, termed second messengers (e.g. Ca2+, cyclic nucleotides, phospholipids), are used by organisms to encode information and deliver it downstream to elements (proteins) which decode/interpret signals and initiate cellular responses (e.g. changes in enzyme activity, gene expression, and cytoskeletal rearrangement). A current hypothesis is that spatial and temporal changes in the levels of second messengers are used to encode information about the cellular responses needed for a given stimulus.
One of the important questions, concerning the role of Ca2+ as a second messenger, addresses how cells control response specificity given the multitude of extracellular stimuli that lead to intracellular Ca2+ fluxes. Studies over the past decade have revealed the complex nature of Ca2+ signalling. Different stimuli elicit Ca2+ transients which are distinct in their subcellular localization, amplitude, duration, frequency of oscillation and mode of spatial propagation. These properties are highly co-ordinated and regulated by the spatial distribution of Ca2+-release channels and Ca2+ pumps throughout the cell. Fine analysis of the origin of the Ca2+ signals revealed elementary units of Ca2+ transients termed ‘blips’, resulting from a single Ca2+ channel opening, and ‘puffs’, which originate from synchronized clusters of channels (Lipp et al., 1997; Plattner et al., 1999; Swillens et al., 1999; Fall et al., 2000). The complex nature of Ca2+ patterns is linked to downstream decoding mechanisms that somehow translate these patterns into distinct cellular responses. For example, there is evidence that the frequency of Ca2+ oscillations carries information that is decoded by downstream cellular targets (De Koninck & Schulman, 1998; Dolmetsch et al., 1998). The importance of Ca2+ oscillations as information carrier in plants was also recently reported (Allen et al., 2000).
Additional complexity resides in the subcellular distribution of Ca2+ within the cytosol and organelles. Of particular interest is the relationship between cytosolic and nuclear Ca2+. Ca2+ signals in the cytosol and the nucleus appear to have distinct functions in both animals (Badminton et al., 1998; Hardingham & Bading, 1998; Berridge et al., 2000) and plants (Van der Luit et al., 1999). However, the origin of nuclear Ca2+ has been a matter of debate. Recent investigations, in both animals (Badminton et al., 1998) and plants (Pauly et al., 2000), suggest that although nuclear Ca2+ signals in many cases reflect the patterns of cytosolic Ca2+, the nucleus is also capable of generating stimulus-induced signals independently.
The importance of Ca2+ in mediating plant responses to external stimuli of biotic and abiotic origin is now well established (Knight et al., 1996; Levine et al., 1996; Epstein, 1998; Liu & Zhu, 1998; McAinsh & Hetherington, 1998; Sanders et al., 1999; Allen et al., 2000; Chatterjee et al., 2000; Kiegle et al., 2000). Plant researchers face the same basic questions being addressed by researchers of animal cells regarding the mechanisms that generate and translate Ca2+ signals into appropriate cellular processes. The technologies developed in recent years to measure intracellular Ca2+ distribution in plants are helping to advance our knowledge of Ca2+ signalling (Knight et al., 1991; Read et al., 1993; McAinsh & Hetherington, 1998; McAinsh & Staxen, 1999; Brownlee, 2000). Recently developed techniques for measuring Ca2+ in plant cells include the so called ‘cameleons’, Ca2+-binding proteins based on two variants of the green fluorescent proteins flanking CaM fused to a CaM binding domain (Allen et al., 1999). In addition, recent genetic manipulations allow measurements of Ca2+ in specific cell types transformed with the aequorin gene fused to a heterologous yeast GAL4 promoter. Aequorin expression is driven by the yeast GAL4 transcription factor expressed in specific cell types of enhancer-trap transgenic lines (Kiegle et al., 2000). As further progress is made in understanding Ca2+ signals, parallel studies on the downstream cellular targets of Ca2+ are helping to provide a more complete picture of Ca2+ signal transduction mechanisms. Undoubtedly, the biochemical properties, expression dynamics and spatial distribution of the Ca2+ sensors contribute to the specificity in the cellular processes evoked in response to distinct Ca2+ signals.
2. Ca2+-modulated proteins
Ca2+-dependent modulation of cellular processes occurs via intracellular Ca2+-binding proteins, also known as Ca2+ sensors, of which CaM is one of the best characterized. CaM has no catalytic activity of its own but, upon binding Ca2+, it activates numerous target proteins involved in a variety of cellular processes. Recent reviews on CaM summarize the roles of CaM as a Ca2+ signal transducer in plants (Snedden & Fromm, 1998; Zielinski, 1998) and animals (Van Eldik & Watterson, 1998; Chin & Means, 2000). In addition, a number of recent reviews are available on the subject of other families of important Ca2+ sensors present in plants, such as the Ca2+-dependent (CaM-independent) protein kinases (CDPKs) (Harmon et al., 2000) and annexins, whose association with membranes through binding of anionic phospholipids is Ca2+ dependent (Lim et al., 1998; Hofmann et al., 2000). A brief account of the important features of CaM as a Ca2+ transducer is presented below.
Most proteins that function as intracellular transducers of Ca2+ signals contain a common structural motif, the ‘EF hand’ (Natalie et al., 1989), which is a helix-loop-helix structure that binds a single Ca2+ ion. These motifs typically occur in closely linked pairs, interacting through antiparallel β-sheets (Natalie et al., 1989). This arrangement is the basis for cooperativity in Ca2+ binding. The superfamily of EF-hand proteins is divided into several classes based on differences in number and organization of EF-hand pairs, amino acid sequences within or outside the motifs, affinity to Ca2+ and/or selectivity and affinity to target proteins (Natalie et al., 1989; Crivici & Ikura, 1995). CaM is an acidic EF-hand protein present in all eukaryotes. The CaM prototype is composed of 148 amino acids arranged in two globular domains connected with a long flexible helix. Each globular domain contains a pair of intimately linked EF hands (Fig. 1a).
One of the intriguing properties of CaM is its ability to activate numerous target proteins that share very little amino acid sequence similarity in their binding sites. Thus, the positioning of CaMBDs within CaM targets must be empirically determined. Nevertheless, the majority of known target sites for CaM are composed of a stretch of 12–30 contiguous amino acids with positively charged amphiphilic characteristics and a propensity to form an α-helix upon binding to CaM (Fig. 1b). This affords a tremendous potential for variability in primary sequence (i.e. target diversity). In addition, CaMBDs, or closely juxtaposed regions, often also function as autoinhibitory or pseudosubstrate domains, maintaining the target in an inactive state in the absence of a Ca2+ signal (James et al., 1995). X-ray diffraction and NMR studies of CaM have provided a model for the structural basis of CaM – target interactions (Crivici & Ikura, 1995). The binding of Ca2+ to CaM (Kd in the range of 10−7–10−6 M) exposes two hydrophobic surfaces surrounded by negative charges, one in each globular domain. Ca2+/CaM may then bind to its targets (affinity in the nanomolar range), mainly by hydrophobic interactions with long hydrophobic side chains in the target sites. Electrostatic interactions contribute to the stability of the CaM-target complex. It is noteworthy that the affinity of CaM for Ca2+ may be increased in the presence of targets (Chin & Means, 2000) thereby providing a more sensitive mechanism of signal response. In solution, the two globular domains wrap around the target, forming an almost globular structure (Fig. 1b). However, functions of CaM in the Ca2+-free state have also been described (Chin & Means, 2000) and raise the possibility that Ca2+-free functions of CaM occur also in plants.
II. CaM isoforms and CaM-like proteins
In addition to the evolutionarily conserved form of CaM, plants possess an extended family of CaM isoforms and CaM-like proteins (Snedden & Fromm, 1998; Zielinski, 1998). This is, undoubtedly, one of the most interesting differences between Ca2+ signalling in plants and animals. Although some of these proteins have been implicated in stress responses, most remain uncharacterized and their roles unknown. Furthermore, as genome sequencing projects progress, the list of potential CaM-like proteins in higher plants is growing rapidly. Although mammals possess a few CaM-like proteins (e.g. calmodulin-like-protein (CLP), calmodulin-like skin protein (CLSP); Koller & Strehler, 1988; Mehul et al., 2000), these are mostly forms closely related to the conserved CaM. At present, designation of proteins (and predicted proteins based on genomic sequence data) as isoforms of CaM or as CaM-like proteins remains arbitrary. Rather than attempt to define a nomenclature for CaMs and their relatives at this point, we hope that, given the recent completion of the Arabidopsis thaliana genome sequence, researchers in Ca2+ signalling will take up the challenge of identifying, annotating and phylogentically comparing the various predicted CaMs and CaM-like proteins as a step toward characterizing the Ca2+ receptors in plants. At present, CaMs in Arabidopsis range from about 90% identical (CaMs 2, 3, 5) to human CaM, to < 30% identity for proteins predicted for various expressed sequence tags (ESTs), genomic DNA, or cDNAs contributed to the GenBank and annotated as CaM-like (W. A. Snedden, unpublished). For the purpose of the present discussion, proteins having the arbitrary value of at least 40% identity with Arabidopsis CaM2 (a conserved form) will be designated as CaM-like. Most of the putative proteins described as CaM-like in the Genebank have been annotated as such based upon sequence-homology searches that identified one or more EF hands within their predicted sequence. Outside of this conserved Ca2+-binding motif, the level of sequence homology varies considerably. Nevertheless, the presence of many CaM-like genes in plants (the maximal recorded is at least 15 CaM genes in hexaploid wheat; Yang et al., 1996), raises some interesting questions. For example, do all plant CaM-like proteins serve as transducers of Ca2+ signals, and if so, are there specific downstream effectors of these CaM-like proteins? It is expected that answers to these questions will be forthcoming as research into Ca2+ signalling progresses. Presently however, there is enough evidence to suspect that at least some of the CaM isoforms and CaM-like proteins identified to date have roles in signal transduction. This review does not include sequence comparisons of CaMs. Therefore, for plant CaM sequence information the readers are advised to refer to recent reviews (Snedden & Fromm, 1998; Zielinski, 1998), or specific research articles discussed in this review and elsewhere.
1. Structural comparisons
Our understanding of the structural characteristics of most plant CaMs and CaM-like proteins is rather limited. However, a comparison of the predicted tertiary structure of TCH2 to Paramecium CaM (a conserved form) has revealed some interesting features (Khan et al., 1997). TCH2 encodes a predicted protein of 161 residues containing four EF hand motifs, and shows about 40% sequence identity to conserved CaMs. Interestingly, its name derives from the fact that it belongs to a family of CaM-like genes that are touch-inducible (Braam & Davis, 1990) and also inducible by a variety of other stimuli (Braam et al., 1997). Modelling of TCH2 (in the Ca2+-bound form) suggests that it shares structural features in common with other CaMs yet also possesses a number of unique features. Like most CaMs, TCH2 is composed of two distinct globular domains separated by a central linker helix (see Fig. 1a). Interestingly however, the linker region in TCH2 contains five glycines within a six residue span and is thus predicted to possess increased flexibility in solution when compared with conserved CaM. As targets for TCH2 have not been identified, it is unclear as to how this increased flexibility might affect TCH2–effector interactions but it is possible that it could influence target binding and/or specificity. Furthermore, TCH2 is relatively enriched in positively charged residues suggesting that its targets are more negatively charged than those of the conserved CaM. Another notable feature of TCH2 is the possible existence of a disulphide bridge near the C-terminus. This is particularly interesting given that cysteine residues are often rare in, or absent from, CaMs. Future studies should focus upon identifying TCH2 effectors and providing comparative analyses of TCH2–target interactions with those established for conserved CaM. As more CaM-like proteins are identified, additional modelling studies will be an important part of the characterization needed to uncover their roles. For example, many of the CaM-like proteins in plants possess C-terminal or N-terminal extensions relative to conserved CaM. These extensions vary in length from a few to dozens of residues and may contain additional EF hand motifs such as TCH3 (Braam & Davis, 1990; Snedden & Fromm, 1998), or motifs affecting subcellular targeting (e.g. a CaaX-box prenylation site; Rodriguez-Concepcion et al., 1999). Modelling analyses need to be performed in conjunction with molecular, biochemical and genetic studies in order to elucidate the functions of these unusual extensions. Domain swapping experiments, where regions within a single CaM isoform are repositioned within the molecule (Cobb et al., 1999), or regions from related CaMs are exchanged to create chimeric isoforms (Lee et al., 1997), have complemented structural studies to help provide a picture of how CaMs interact with their targets. Extending such studies to various CaM-like proteins should help reveal their functions.
2. Biochemical studies
One important question arizing from the multiplicity of CaMs in plants is whether specific CaMs might differentially regulate (i.e. activate or inactivate to varying degrees) common targets. It is conceivable that different CaMs could compete for targets in vivo, and thus relative abundance, subcellular distribution, and effector-binding affinities of the various isoforms may play an important role in their ability to regulate a particular target. Additionally, different binding affinities of various CaMs for Ca2+ could theoretically modulate their ability to activate targets and hence create subsets of CaMs conditioned to interpret particular Ca2+ signatures. A number of studies have attempted to address such questions by performing comparative analyses of CaM-dependent enzyme activities using different CaMs. SCaM-4 is a soybean CaM isoform showing about 80% identity to conserved CaM and was among the first isoforms to be studied biochemically (Lee et al., 1995a). Interestingly, although SCaM-4 activates mammalian cyclic nucleotide phosphodiesterase to a level comparable to that observed using conserved CaM, it is unable to activate the CaM-dependent plant enzyme NAD kinase (Lee et al., 1995a). Reciprocal domain swapping experiments using modified recombinant proteins, where domains from SCaM-1 (the conserved form) and SCaM-4 were exchanged to create chimeras, revealed that domain I plays a significant role in NAD kinase activation (Lee et al., 1997). Moreover, through a series of point-mutations, critical residues were identified within domain I responsible for the differential activation. It is noteworthy that, in a related study by Liao et al. (1996) using CaM mutants, the C-terminal region of CaM was also implicated in activating NAD kinase, implying that both C- and N-terminal regions of CaM contribute to enzyme activation. Theoretically, one means of regulating signal transduction pathways could involve the reciprocal activation and inhibition of two different enzymes by a single modulator such as CaM. Such a model has been proposed based upon enzyme-response studies carried out using SCaM-1 and SCaM-4, the conserved and divergent isoforms of soybean CaM, respectively, and the enzymes calcineurin and nitric oxide synthase (Cho et al., 1998). Although this study used a heterologous system (mammalian enzymes and plant CaM isoforms), it nonetheless demonstrated the potential of CaM isoforms to simultaneously activate one enzyme and competitively inhibit another. These principles need to be tested in vivo but raise exciting possibilities for the modulation and co-ordination of signalling pathways. Remarkably, the inhibition of calcineurin by SCaM-1 is the consequence of a single point mutation involving a methionine to valine alteration (Kondo et al., 1999). These findings reflect the sensitivity of CaM targets to the primary sequence of CaM and underscore the potential for a sophisticated regulatory network given the diversity of CaMs in plants. An attempt to extend these findings was performed by Lee et al. (2000), again using SCaM-1 and SCaM-4, but with a variety of CaM-regulated enzymes from both plants and animals. The results suggest a complexity of interactions of CaMs with their targets: some enzymes (e.g. plant Ca2+-ATPase) were maximally activated by either CaM isoform, other enzymes (e.g. plant glutamate decarboxylase) showed a modest isoform-sensitivity, and, for several enzymes, a complete lack of activation by SCaM-4 was observed (Lee et al., 2000). Moreover, differences between the two isoforms were observed in the level of Ca2+ required for half-maximal activation of targets. It has been suggested that the multiplicity of CaM isoforms, with potentially different affinities toward Ca2+, may afford added flexibility to cells when decoding intracellular Ca2+ signals (Snedden & Fromm, 1998; Zielinski, 1998). CaM–target interaction also influences CaM’s responsiveness to Ca2+ (Olwin et al., 1984) thus adding yet another layer of regulatory control. The potential of different CaMs to compete for targets in vivo is supported by an additional in vitro study by Lee et al. (1999). SCaM-1 and SCaM-4 exhibit comparable expression patterns and are able to bind (competitively) to similar proteins from a range of soybean tissues and subcellular fractions. Taken in conjunction with studies showing differential activation/inhibition of potential targets, these data collectively suggest that CaMs could compete for targets in vivo thereby regulating pathways through a combination of CaM : target affinity, stoichiometry, and responsiveness to Ca2+.
Another interesting example of differential effects of CaM isoforms involves the interaction between nuclear proteins and DNA (Szymanski et al., 1996). Arabidopsis CaM-3, though a relatively conserved CaM, when compared to other isoforms, preferentially enhanced the association of DNA-binding proteins with a region of the CaM-3 promoter. Additionally, CaM-3 can bind the recombinant transcription factor TGA3 and enhance its binding to a region within the CaM-3 promoter. Although evidence for CaM-3/TGA3 interaction in vivo is not yet available, these findings raise the intriguing possibility of an autoregulatory mechanism for controlling transcription of CaM-3. It has been proposed that, although seemingly paradoxical, increasing the level of Ca2+ receptors (i.e. CaMs), in addition to providing stoichiometric levels of CaMs for target activation, may provide a mechanism to prepare for long-term or subsequent stimulation (Zielinski, 1998). Differential effects of two potato CaM isoforms, PCM1 and PCM6, have been reported on the autophosphorylation and kinetic properties of tobacco CCaMK but not lily CCaMK (Liu et al., 1998). In addition, three Arabidopsis CaM isoforms (CaMs-2, -4, and -6) showed a differential effect on activation of mammalian cyclic nucleotide phosphodiesterase, with CaM-2 being the least efficient (Reddy et al., 1999). This same study compared the binding affinities of these CaMs to Arabidopsis kinesin-like CaM-binding protein (KCBP) using various techniques and found that CaMs-4 and -6 exhibit a twofold higher affinity than CaM-2 toward KCBP. Importantly, this comparison utilized CaMs and a CaM-target from the same plant thereby lending further support to the hypothesis that CaMs may compete for common targets in vivo.
It is worth noting also that the post-translational status of CaMs may alter their ability to interact with and activate targets. For example, trimethylation of CaM commonly occurs in eukaryotes and significantly affects its ability to activate plant NAD kinase (Roberts et al., 1986). A conserved lysine (residue 115) in CaM is the target site for a CaM-specific methyltransferase (Roberts et al., 1986). Although some discrepancy exists, this lysine residue may also be a target for ubiquitination (Gregori et al., 1987; Ziegenhagen & Jennissen, 1990) thereby contributing to CaM turnover. Using a series of domain exchange and site-directed mutagenesis studies, it has been demonstrated that both sequence and structural components contribute to the methylation of CaM (Cobb et al., 1999; Cobb & Roberts, 2000). Consequently, as various CaM isoforms have not been tested for their susceptibility to methylation, caution should be exercised when making assumptions based solely upon sequence data (i.e. absence of methylation motifs). Moreover, although plant NAD kinase is the only target known to be sensitive to the methylation state of CaM, as new CaM targets in plants are identified, their sensitivity should be evaluated in order to ascertain the broader significance of CaM methylation in plants. Evidence for the in vivo role of CaM methylation in plants was recently reported. A CaM-like protein from petunia (designated CaM53), containing a 35 amino acid extension with a prenylation site at the C-terminus, is associated with the plasma membrane when the protein is prenylated (Rodriguez-Concepcion et al., 1999). The prenylated C-terminus cysteine residue becomes a substrate for a prenyl-cysteine methyltransferase that carboxymethylates the prenylated CaM53. This methylation is required for the association of CaM53 with the plasma membrane (Rodriguez-Concepcion et al., 2000). Interestingly, when prenylation of CaM53 is inhibited, it is localized predominantly in the nucleus (Rodriguez-Concepcion et al., 1999).
CaM in animals is subjected to phosphorylation (Sacks & McDonald, 1988; Sacks et al., 1989; Leclerc et al., 1999). It is still unknown whether plant CaM isoforms are phosphorylated. If so, what role might such modifications have in Ca2+-mediated signal transduction? One interesting difference between most plant CaM isoforms and animal CaM is the fact that whereas mammalian CaM possesses two tyrosine residues at positions 99 and 138, all plant CaMs have Tyr-138 but only a few have Tyr-99 (e.g. petunia CaM72 and soybean SCaM-4; Snedden & Fromm, 1998). Thus it is certainly possible that plant CaM isoforms differ in their phosphorylation patterns. Further, as phosphorylation of CaM affects interaction with its targets (Leclerc et al., 1999), it would be of interest to assess whether plant CaMs, particularly CaM72 and SCaM-4, are subjected to similar post-translational modifications. In general, there remains much to be learned about the extent of post-translational modifications of CaM in plants. Certainly, a survey of the various types of modifications occurring needs to be conducted and would help direct future research in this area.
3. Expression and genetic studies
Insight into the roles of CaMs and CaM-like proteins in plants has also come from expression analyses, particularly during stress responses. The proteins encoded by the TCH genes (touch-induced) were among the first stimuli-inducible CaM-like proteins (with the exception of TCH4 which encodes a xyloglucan endotransglycoslyase) identified in plants (Braam & Davis, 1990). TCH1 encodes a conserved CaM whereas TCH2 and TCH3 predict proteins with about 40% and 60% amino acid identity to CaM, respectively. However, both TCH2 and TCH3 predict proteins that are larger than CaM, particularly TCH3, which possesses several repeat domains that endow it with two additional EF hands near the N-terminus (Sistrunk et al., 1994). Whether this alters it’s Ca2+ sensitivity or target specificity is not yet known. Although a wide variety of stimuli lead to the induction of different TCH genes, including mechanical manipulation, heat shock, cold shock, phytohormone exposure, and increases in external Ca2+, the magnitude and temporal characteristics of induction differ among the TCH family members (Braam, 1992a, 1992b; Antosiewicz et al., 1995; Polisensky & Braam, 1996; Braam et al., 1997). A role in stress-response for the TCH CaM-like proteins seems likely and it will be interesting to discover what their downstream effectors might be. Further insight into the role of TCH3 has been provided by tissue distribution studies (Sistrunk et al., 1994; Antosiewicz et al., 1995). Although present in a variety of tissues, a common theme is expression in regions believed to experience mechanical strain such as root/shoot junctions, branch points, and vascular tissues. Similarly, cell-specific expression analysis of CaM genes in hexaploid wheat revealed the association of a specific CaM subfamily (SF-2) with fibre development (Yang et al., 1998). Based upon collective data, it is likely that certain CaM-like proteins in various, if not all, plants participate in Ca2+-dependent cellular processes associated with cell reinforcement of expansion.
Though not as well characterized as the TCH family, a number of other CaMs and CaM-like genes show interesting patterns of expression. MBCaM-1 and MBCaM-2 are Vigna radiata genes that encode proteins essentially identical to conserved CaM but which show differential expression in response to stimuli (Botella & Arteca, 1994). Similar findings have been reported for Nicotiana plumbaginifolia NpCaM-1 and NpCaM-2 which encode identical CaMs but respond to different stimuli (van der Luit et al., 1999). Furthermore, differences in either tissue and cell-type expression, stimuli response, or expression during development have been reported for multiple CaMs in potato (Takezawa et al., 1995), wheat (Yang et al., 1998), Arabidopsis (Gawienowski et al., 1993; Verma & Upadhyaya, 1998), and tomato (Depege N et al., 1999). In most studies, only a few members of the CaM family were assessed, nonetheless, collectively, these findings support the suggestion that a multigene family could be used for fine-tuning responses to Ca2+ signals based upon selective expression of Ca2+ receptors (Toutenhoofd & Strehler, 2000). In addition, several studies have implicated roles during stress responses for some of the more divergent CaM-like proteins. Hra32 encodes a novel CaM-like protein whose transcripts accumulate rapidly in bean during the hypersensitive-response (HR) induced by Pseudomonas syringae (Jakobek et al., 1999). Other stimuli also lead to induction of Hra32 but with a more transient kinetic profile. Hra32 predicts a c. 17 kDa protein whose similarity to CaM is restricted predominantly to the EF hand regions. It is noteworthy that roles for Ca2+ and CaM during the HR are well established (Levine et al., 1996; Grant & Mansfield, 1999; Heo et al., 1999; Blume et al., 2000). An Arabidopsis gene, AtCP1, predicts a protein with similarity to Hra32, and is strongly induced by NaCl but not ABA treatments (Jang et al., 1998). The latter reflects specificity of response given that both ABA and NaCl mediate pathways known to involve Ca2+ (Knight et al., 1997; Wu et al., 1997; Bressan et al., 1998; Pei et al., 2000). Although the data are correlative, it is tempting to speculate that Hra32 and AtCP1 proteins may function as intracellular Ca2+ interpretors of HR- and NaCl-mediated Ca2+ fluxes, respectively, but conclusions must await further study. Additional support for the involvement of CaM isoforms during pathogen-response comes from expression and genetic studies of soybean SCaM-4 and SCaM-5 (Heo et al., 1999). Further details on the involvement of these and other CaMs in pathogenesis responses are described in chapter VI.
Functional and genetic studies have implicated specific CaMs in nuclear signalling (discussed in chapter IV of this review). As more CaMs and CaM-like proteins are identified, it is important to move beyond expression analyses into functional studies employing the collective tools of genetics, biochemistry, and molecular biology. Notably, for most CaMs and CaM-like proteins, it remains to be demonstrated that their changes in expression during development or stress represent adaptive responses. Analyses of gene knock-out lines may help resolve this issue but the multiplicity of CaMs, and hence the potential ability of other family members to complement mutations, make any single experimental approach difficult. Nevertheless, the fact that plants express such an interesting diversity of Ca2+-binding proteins (CaMs, CaM-like proteins, CDPKs, various EF-hand containing proteins) is testimony to the fact that Ca2+ plays a central role in plant signal transduction. Elucidating the roles of Ca2+ receptors is a vital step toward unravelling the complexities of Ca2+ signals. The various mechanisms that may determine the specific functions of distinct CaM-like proteins are depicted in Fig. 2.
4. Extracellular CaM
Although CaM is considered primarily an intracellular molecule, recent studies have demonstrated the presence of extracellular CaM. In animals, CaM released at sites of tissue injury, or possibly by specific mechanisms in the endothelium, can bind to receptors, modulating the activities of inflamatory cells (Houston et al., 1997). Consistent with these findings, a number of extracellular CaM-binding proteins were found in body fluids of animals (WenQiang et al., 1997). In plants, extracellular CaM has also been observed and the occurrence of extracellular CaM-binding proteins reported (Jun et al., 1996). Does CaM function as an extracellular signalling molecule? The answer to this question is not yet clear but is certainly worth pursuing. Indeed, the role of extracellular CaM has been investigated in a few plant species. Extracellular CaM accelerated the proliferation of suspension-cultured cells, cell wall regeneration in protoplasts, and gene expression (Sun et al., 1995; Zhu et al., 1998). An interesting role for extracellular CaM has been proposed with respect to pollen germination and tube growth (Ma & Sun, 1997; Ma et al., 1999). In this study, extracellular CaM was suggested to function upstream of heteromeric G-proteins in promoting pollen tube growth. Biochemical evidence demonstrated that extracellular CaM stimulated GTPase activity in pollen plasma membrane vesicles (Ma et al., 1999). CaM-regulated enzymes may indeed function in the extracellular domain. It has been shown that overexpression in transgenic Arabidopsis of the CaM-binding isoform of apyrase (Atap1) results in increased apyrase activity in the extracellular domain. It is presently unknown which members of the CaM family might be involved in extracellular functions but it would be particularly interesting to determine if specific isoforms are transported to the extracellular domain, or if CaMs are subject to post-translational modifications that enable them to reach the extracellular domain and to function in an environment that is very different from the cytosol regarding Ca2+ concentration and pH. It is also intriguing to speculate on the possible occurrence of CaM receptors in plant plasma membranes, and on CaM acting as a peptide messenger.
III. CaM-target proteins
1. The repertoire of CaM-target proteins
More than a decade after discovering the occurrence of CaM in plants (Charbonneau & Cormier, 1979) only a handful of proteins had been identified as being regulated by CaM, and no molecular studies on CaM targets in plant cells had been conducted. Of the few CaM targets that were identified by the early 1990s, NAD kinase was the best studied (Jarrett et al., 1982; Roberts et al., 1986; Roberts & Harmon, 1992). Curiously, the gene encoding NAD kinase has yet to be cloned. The advent of molecular procedures initiated in the early 1990s to isolate plant CaM-target proteins not only facilitated the isolation and identification of the cellular targets of plant CaM, but also enabled an in-depth investigation of their function in vitro as purified recombinant proteins, as well as in planta investigations using reverse genetic approaches. Molecular approaches for screening cDNA expression libraries using labelled recombinant CaM as a probe (Fromm & Chua, 1992; Lu & Harrington, 1994; Liao & Zielinski, 1995) proved a powerful tool, and screening of libraries derived from various plant species, organs and cell types, and from plants exposed to a variety of stimuli and stress situations, revealed a large number of unique clones encoding CaM binding proteins. In general, the proteins which have been isolated to date from several laboratories can be divided into the following groups: proteins of unknown function, with no resemblance to sequences deposited in the gene banks. By definition, these are presently considered unique to plants, and include, for example, a maize root tip protein (Reddy et al., 1993), TCB60, a heat shock repressed tobacco protein (Dash et al., 1997), and a pollen-specific protein (Safadi et al., 2000); proteins resembling CaM-regulated proteins of known functions from other organisms such as Ca2+-ATPases (SCA1, Chung et al., 2000), protein kinases (e.g. CCaMK, Patil et al., 1995), and CNGCs (Arazi et al., 1999, 2000a; Kohler & Neuhaus, 2000); proteins of known function in other organisms but regulated by CaM only in plants, such as glutamate decarboxylase (Snedden et al., 1995, 1996) and apyrase (Hsieh et al., 2000); and previously described plant proteins, which initially had not been identified as CaM-binding proteins (e.g. SAUR1; Yang & Poovaiah, 2000a).
Table 1 summarizes the repertoire of plant proteins for which there is some evidence to suggest their regulation by (or association with) CaM. The amount of supportive evidence varies considerably, and the list also includes proteins whose genes have not been cloned. In the future, researchers should consider the following criteria for identifying a protein as being regulated by CaM: does the native (and recombinant, where applicable) target protein bind CaM? Is the affinity of interaction within the expected physiological concentration range of the target? Does the target associate with CaM in vivo? Can an effect of CaM on target activity, or other properties, be demonstrated? Affirmative answers to all these questions strongly suggest a physiological role for CaM in target regulation in vivo. On the other hand, negative answers to some of these questions do not exclude such a possibility. In general, Table 1 should be read with caution, as definitive evidence of a regulatory role for CaM in vivo is lacking for most of the proteins listed. Rather, this table serves to point out areas of future research that might prove fruitful given some of the findings that have been reported in the literature.
Table 1. Plant proteins reported to be associated with CaM regulation
Glutamate decarboxylase (GAD)
– catalyses conversion glutamate to γ-aminobutyric acid (GABA)
A review of the list of plant CaM-regulated proteins (Table 1) reveals several interesting points, the first being that the number of different types of confirmed and putative CaM-regulated targets has already reached 30. This number will likely increase and future screening with other CaM-related proteins as bait may reveal the identity of additional cellular targets. Furthermore, the application of advanced bioinformatic approaches, such as the development of algorithms to predict CaMBDs among Genebank sequences could prove very useful to identify candidates as potential CaM targets. Although bioinformatic predictions of CaMBDs need to be treated cautiously and confirmed through empirical testing, algorithms for the prediction of CaM-binding sites in proteins are available. Interested readers are encouraged to visit the comprehensive CaM-related web site maintained by the Ikura laboratory at the University of Toronto (http://calcium.oci.utoronto.ca/ctdb/noflash.htm).
Another aspect emerging from Table 1 is the classification of the CaM-regulated proteins into numerous functional groups involved in regulating metabolism, motility, phytohormone responses, ion transport, protein folding, DNA binding, protein phosphorylation, phospholipid metabolism, and unidentified functions. Indeed, the diversity of roles for CaM in plants is reminiscent of that in other eukaryotes. However, many of the CaM targets listed in Table 1 appear to be unique to plants.
Not surprisingly, many of the proteins listed in Table 1 are directly associated with responses to stress. These include GAD, NAD kinase, apyrase, SOD, glyoxalase, NtER1, Ca2+ ATPases, chaperonins, MDR and TCB60. This is consistent with the important role of Ca2+ in mediating plant responses to many stresses. Interestingly, homologues of some of the best characterized CaM targets in animals such as adenylyl cylcase, phosphodiesetarse, nitric oxide synthase, CaM kinases and phosphatases (calcineurin) have either not yet been identified in plants, or do not appear to be CaM-regulated. It may be that the extended family of CDPKs (Harmon et al., 2000) and the chimeric CCaMK (Patil et al., 1995) present in plants have evolved to perform the cellular tasks ascribed to CaMKs in mammals.
2. Regulatory mechanisms of CaM – target interactions
In chapter II the interaction of CaM with target proteins is discussed with respect to the CaM protein involved. However, the mounting data obtained from the analysis of CaM-binding domains of plant proteins revealed interesting features underlying regulatory mechanisms controlling CaM–target interactions. Of particular interest is the finding that some plant gene families encode isoforms that bind CaM as well as isoforms that are incapable of binding CaM. For example, Atap1 (Arabidopsis apyrase 1; Table 1) is a CaM-binding protein, whereas Atap2 is not. Comparison of the presumed CaMBD of apyrase 1 with the homologous region of apyrase 2 shows that apyrase 2 has additional glutamate residues, which likely abolish CaM binding, at least in vitro (Steinebrunner et al., 2000). Another interesting case is the tomato gene encoding diacylglycerol kinase (Table 1), which has two forms derived from alternative splicing, only one of which includes the C-terminal exon containing the CaMBD (Snedden & Blumwald, 2000). In addition to the occurrence of CaM binding and nonbinding isoforms, the CaMBDs within a protein family may vary in other properties that could contribute to different modes of regulation. Within the GAD family in Arabidopsis, some isoforms (e.g. GAD1) contain conserved serine and threonine residues within the CaMBD, in the same positions as in GAD from other plants. By contrast, GAD2 in Arabidopsis is lacking serine and threonine residues (Zik et al., 1998, Table 1). It is tempting to speculate that within the GAD family, different isoforms have distinct modes of regulation of interaction with CaM, which are regulated by the phosphorylation state of the CaMBD. Interestingly, the GAD2 isoform that lacks the potential phosphorylation sites is also the one that is ubiquitously expressed in plant tissues, whereas GAD1 expression is confined to roots (Zik et al., 1998). Although phosphorylation may play a role in inhibiting CaM binding to its targets, the possible effect of CaM – target interactions on target phosphorylation may also be considered. In this regard, pea apyrase was found to be phosphorylated by casein kinase II in vitro. This phosphorylation does not affect CaM binding. However, CaM does inhibit the phosphorylation of apyrase (Hsieh et al., 2000). These observations suggest that future functional analysis of CaM isoforms, should take into consideration the possible post-translational modifications of the CaMBDs, as well as other parameters discussed in chapter II (e.g. expression profiles of the proteins analysed).
The following chapters discuss cellular processes regulated by CaM and CaM-like proteins in plants, with reference to some of the CaM proteins discussed in chapter II, and CaM-target proteins presented in Table 1.
IV. Ca2+/CaM and nuclear functions
1. Ca2+ signalling, CaM and CaM-regulated proteins in the nucleus
In the past decade considerable progress has been made in our understanding of Ca2+ signalling in the nucleus. Previously, it was largely assumed that Ca2+ moved passively through nuclear pores and that Ca2+ concentration in the nucleus was a mere reflection of cytosolic Ca2+ concentration. Recently, however, innovative imaging technologies have helped unveil the complexity of nuclear Ca2+ signals and the information emerging is exciting though still somewhat controversial. There is consensus that Ca2+ signals occur in the nucleus, which modulate a number of important processes (Lipp et al., 1997; Hardingham & Bading, 1998; Berridge et al., 2000; Bootman et al., 2000). However, it remains a debated topic as to whether nuclei generate cytosolic-independent Ca2+ signals and whether the nuclear envelope is able to regulate the propagation of cytosolic Ca2+ signals (see Bootman et al., 2000).
In plants, CaM has been shown to be present in the nucleus of a few species (Vantard et al., 1985; Dauwalder et al., 1986; Schuurink et al., 1996). An example of the subcellular distribution of CaM in tobacco cotyledon cells using highly specific monoclonal antibodies against a petunia CaM is shown in Fig. 3. This example reveals the localization of CaM in the cytoplasm and in the nucleus, and the absence of CaM in the vacuole. Higher resolution immunofluorescent localization studies of CaM in endosperm cells (Vantard et al., 1985) revealed that CaM is associated with the mitotic apparatus, particularly with the asterlike centres and kinetochore microtubules. Furthermore, colocalization of Ca2+ and CaM was also reported and a role for CaM in kinetochore microtubule dynamics was suggested (Vantard et al., 1985). Evidence for the effects of CaM on cortical microtubules was reported later (Fisher et al., 1996). Although the mechanism regulating microtubule organization by Ca2+/CaM is not fully understood, an important advance in our understanding of this process emerged from the discovery of a kinesin-like CaM-binding motor protein (designated KCBP; Narasimhulu et al., 1997). The subcellular localization of KCBP, as determined by immunohistochemistry, revealed its association with the preprophase band, the mitotic spindle and the phragmoplast (Bowser & Reddy, 1997). The association of KCBP with microtubule arrays in dividing cells suggests that this motor protein is involved in their formation and likely participates in the functions associated with these structures. Moreover the interaction of KCBP with tubulin subunits is modulated by Ca2+ and this enhances microtubule bundling (Kao et al., 2000). Therefore, it is likely that KCBP is involved in establishing microtubule arrays during different stages of cell division and that Ca2+/CaM regulates the formation of these microtubules. This is consistent with recent evidence for the role of KCBP in cell division (Vos et al., 2000). Interestingly, disruption of a KCBP gene in Arabidopsis by T-DNA insertion prevents trichome development (Oppenheimer et al., 1997). In vitro studies of the interaction of KCBP with assorted Arabidopsis CaM isoforms revealed that isoforms bound to KCBP with different affinities (Reddy et al., 1999). It is currently unknown which of theses isoforms might be regulating KCBP in vivo.
Recently, convincing evidence for the role of specific plant CaM-related proteins in cytosolic – nuclear communication has been reported. A CaM-like protein from petunia (designated CaM53), containing a 35 amino acid extension with a prenylation site at the C-terminus, is associated with the plasma membrane when the protein is prenylated. If, however, prenylation is inhibited, either due to specific mutations, or the nutritional state of the plant, CaM53 is found predominantly in the nucleus (Rodriguez-Concepcion et al., 1999, 2000). Translocation of CaM53 to the nucleus is facilitated by carbon starvation (e.g. in dark), whereas association with the plasma membrane is enhanced in the light, or in the dark if a metabolized carbon source (e.g. sucrose but not mannitol) is provided (Rodriguez-Concepcion et al., 1999). Thus, CaM53 may play a role in conveying signals between the plasma membrane and the nucleus, associated with changes in the carbon status of the plant. A similar mechanism probably operates in other plants, as a rice protein (designated CaM61) with high similarity to the petunia CaM53 was reported (Xiao et al., 1999). These CaM-like proteins represent excellent candidates for a cytosolic – nuclear signalling network, and identifying their cellular targets remains an area for further studies.
2. CaM regulated transcription
In mammalian cells, transcription of several genes is regulated by protein kinases that are located in the nucleus and regulated by Ca2+/CaM (e.g. protein kinase II and IV; Heist & Schulman, 1998). Recent elegant experiments using intracellular Ca2+ chelators to block specifically either cytosolic or nuclear Ca2+ signalling, revealed that Ca2+ signalling in the nucleus itself is required for transcriptional activation of some genes, an example of which is gene activation by the cyclic AMP responsive element binding protein (CREB) transcription factor upon its phosphorylation by the nuclear localized CaM-kinase IV (Hardingham & Bading, 1998). This transcriptional activation could be further modulated by translocation of CaM into the nucleus (Deisseroth et al., 1998; Hardingham & Bading, 1998). As movement of CaM into the nucleus is controlled, at least in part, by cytosolic Ca2+ oscillations (Craske et al., 1999; Teruel & Meyer, 2000; Teruel et al., 2000), it appears that transcriptional regulation by nuclear-localized Ca2+/CaM is co-ordinated by Ca2+ signalling in both the nucleus and the cytosol.
Ca2+/CaM may also modulate nuclear transcription indirectly by regulating post-translational modifications of cytosolic transcription factors. A well studied example of this type of control is calcineurin, the Ca2+/CaM-regulated protein phosphatase which dephosphorylates transcription factors such as NAFT (Rao et al., 1997; Masuda et al., 1998). In addition to the Ca2+/CaM-regulated kinase/phosphatase modulation of transcription factors in the cytosol and in the nucleus, other transcription factors may respond to Ca2+ signals by the direct binding of Ca2+/CaM. This type of interaction may modulate either the binding of transcription factors to DNA, or the ability of transcription factors to interact with other proteins that have an effect on gene transcription without necessarily having a direct effect on DNA-binding properties. A group of mammalian helix-loop-helix (bHLH) transcription factors, whose ability to bind DNA is diminished by direct binding of CaM to their DNA-binding domains was reported (Corneliussen et al., 1994). Apparently, this interaction is of a novel type. Namely, unlike most CaM–target interaction, which are composed of both strong hydrophobic and polar elements, bHLH–CaM interactions are highly polar in nature (Hermann et al., 1998; Onions et al., 2000). It should, however, be noted that a physiological role for CaM interactions with the bHLH transcription factors remains to be demonstrated. An interesting example of a transcription factor that is modulated by direct binding of Ca2+ through four EF-hand motifs, and not via CaM, is the human downstream regulatory element antagonist modulator (DREAM) transcription repressor (Carrion et al., 1999). Upon cell stimulation, Ca2+ binding to DREAM prevents its repressor function.
In spite of the limited research of CaM in the plant nucleus, recent investigations of Ca2+ signalling in the nucleosol and cytosol of plant cells are consistent with studies of mammalian cells. Van der Luit et al. (1999) used tobacco seedlings transformed with a construct that encodes a fusion protein between nucleoplasmin (a nuclear protein) and aequorin (a fluorescent Ca2+ reporter protein), which was targeted to the nucleus. Studies of Ca2+ signalling in these plants, and a comparison to transgenic plants expressing cytosolic aequorin, revealed that wind and cold shock evoked separate Ca2+ signalling pathways. Wind induced the expression of the NpCaM-1 CaM gene by a Ca2+ signalling pathway operational in the nucleus, whereas expression of the same gene in response to cold shock is regulated by a pathway operational in the cytosol. Therefore, distinct Ca2+ signalling pathways in the nucleus and in the cytosol occur both in plants and in animals. Nevertheless, debate continues concerning the ability of animal nuclei to generate cytosolic-independent changes in nuclear Ca2+ concentration (Bootman et al., 2000). However, a recent report suggests that plant nuclei are indeed capable of generating independent Ca2+ signals. Using tobacco, Pauly et al. (2000) were able to demonstrate that nuclei from lysed protoplasts, expressing nuclear-targeted aequorin, were able to generate a cytosolic-independent Ca2+ signal when challenged with mastoporan, a stimulator of G-protein coupled receptors. Moreover, the Ca2+ concentration in these nuclei was not sensitive to high levels of exogenous Ca2+ indicating that some Ca2+-exclusion mechanism must be present in the nuclear membrane. This study also raises questions about the possibility of G-protein coupled receptors in nuclear membranes but the potential for mastoporan to permeabilize membranes to Ca2+ (Matsuzaki et al., 1996) should be kept in mind. Consequently, though these findings need to be confirmed using intact cells and physiologically relevant stimuli, they represent an exciting advance in our understanding of nuclear Ca2+ signalling in plants, and are consistent with recent evidence for ATP-dependent Ca2+ uptake by isolated plant cell nuclei (Bunney et al., 2000), and earlier reports on the occurrence of Ca2+ ATPases in the plant nuclear envelope (Evans & Williams, 1998).
The regulation of transcription by Ca2+/CaM in plants was suggested in the case of phytochrome-mediated photo-transduction (Neuhaus et al., 1993; Bowler & Chua, 1994; Bowler et al., 1994). However, the mechanism involved in this transcriptional regulation remains unclear. Recent evidence for the localization of phytochrome itself in the nucleus, at least in some cells (Gil et al., 2000), is interesting in this regard. DNA-binding proteins are among the possible targets of CaM in plant cell nuclei. For example, the Arabidopsis TGA3 bZIP transcription factor was reported to bind CaM in vitro (Szymanski et al., 1996). More recently, a novel Arabidopsis DNA-binding protein was reported to bind CaM with high affinity (in the low nM range), to be localized in the nucleus, and to possess a domain capable of activating transcription in yeast cells (Table 1 and Bouché, 2000). Therefore, it is likely that some components of the transcriptional machinery in plants are regulated by CaM. This regulation may occur, at least in part, by direct interaction of Ca2+/CaM with transcription factors. On the other hand, as opposed to mammalian cells, there is as yet no evidence for a plant CaM-regulated kinase like the mammalian CaM-kinase II, or CaM-kinase IV, which are important downstream nuclear targets of CaM involved in transcription regulation. However, plants do possess a class of protein kinases regulated by Ca2+ binding to a visinin-like domain, and by Ca2+/CaM binding to a different domain. These are designated chimeric Ca2+/CaM-dependent kinases (CCaMKs; Patil et al., 1995). In theory, these kinases could serve as Ca2+ sensors/transducers during signal propagation. In addition, plants possess a family of Ca2+-dependent CaM-independent protein kinases (CDPKs) that might play a role in controlling nuclear gene expression (Harmon et al., 2000). Identifying their in vivo targets, and localizing subcellularly those targets, are the next logical steps toward dissecting cytosolic and nuclear transduction pathways.
The overall picture emerging from studies of animal and plant cells is of an important role for CaM in nuclear functions. From a mechanistic point of view it should be noted that CaM may modulate nuclear functions either by interacting with target proteins in the cytosol, which consequently transduce a signal to the nucleus, or by responding to Ca2+ signals in the nucleus (or by both possibilities). A schematic model with examples of CaM responding to Ca2+ signals in the cytosol or in the nucleosol is shown in Fig. 4. An example of transcriptional regulation by CaM in response to cytosolic Ca2+ is the activation of the cytosolic Ca2+/CaM-regulated calcineurin phosphatase, which upon dephosphorylation of certain transcription factors, such as nuclear factor of activated T-cells (NFAT) in mammalian cells, induces their translocation to the nucleus (Fig. 4a). Molecular and physiological evidence for plant signalling pathways involving calcineurin-like proteins have been reported recently (Kudla et al., 1999) but a CaM-dependent phosphatase in plants has yet to be cloned. These signalling pathways appear to play a role in responses to salinity stress both in plants (Pardo et al., 1998; Kudla et al., 1999) and in yeast (Lippuner et al., 1996).
CaM-regulated proteins have also been identified in association with chromatin in both animals (Wu & Means, 2000) and plants (Hsieh et al., 1996). This suggests that CaM may play different roles associated with chromatin function in addition to transcriptional regulation.
V. Regulation of ion transport
1. CaM regulated Ca2+ pumps
Among the best studied functions of CaM is regulation of ion flux, particularly the regulation of ion transporters controlling Ca2+ homeostasis. The maintenance of a low Ca2+ concentration (approx. 10−7 M) is essential for the proper functioning of the cell, as sustained increases in cytosolic Ca2+ would lead to cell death. Hence, in eukaryotes a network of Ca2+ pumps and Ca2+ exchangers operate to maintain this low cytosolic Ca2+ level against a 1 000–10 000-fold concentration gradient across the plasma membrane, or endomembranes. At the same time, the evolution of Ca2+ as a cellular second messenger in eukaryotes is associated with the occurrence of an intricate network of Ca2+ release channels in the plasma membrane and in endomembranes, regulated by various physical and chemical signals. CaM plays an important role in regulating Ca2+ transporters. The function of CaM in these processes may be either direct, by binding to transporters, or indirect, by modifying the activity of regulatory proteins such as kinases and phosphatases.
An important group of transporters in eukaryotes are the Ca2+-ATPases and Ca2+ exchangers that actively drive, against an electrochemical gradient, Ca2+ out of the cytosol and into either the extracellular matrix or intracellular stores. In mammals, the major machinery responsible for Ca2+ efflux are Ca2+ exchangers, particularly Na+/Ca2+ exchangers, and Ca2+-ATPases (Guerini, 1998). By contrast, in plants, Ca2+ pumps (Ca2+-ATPases) seem to have the major role in Ca2+ efflux, and less is known about Ca2+ exchangers. In mammals, two classes of Ca2+-ATPases have been described; those targeted to endomembranes (type IIA), and those targeted to plasma membranes (type IIB). Only type IIB Ca2+-ATPases are regulated by Ca2+/CaM. The functions and diversity of mammalian Ca2+-ATPases have been recently reviewed (Guerini, 1998). Recent reviews on plant Ca2+-ATPases (Evans & Williams, 1998; Geisler et al., 2000; Sze et al., 2000) should prove valuable for plant researchers.
The major physiological role of Ca2+-ATPases is to restore and maintain Ca2+ homeostasis by pumping Ca2+ out of the cytosol, but evidence for their involvement in the temporal and spatial propagation of complex Ca2+ waves is emerging (Plieth, 1999). The activation of type IIB pumps by Ca2+/CaM therefore provides a feedback mechanism to respond quickly to Ca2+ fluxes. Although it is likely that mammalian and plant Ca2+-ATPases perform similar functions, there are significant differences between the two systems. For example, there are important distinctions between plants and animals regarding the subcellular distribution of type IIB Ca2+-ATPases. In mammalian cells, type IIB Ca2+-ATPases are found exclusively in the plasma membrane whereas in plants, CaM-regulated ATPases are present in endomembranes including the endoplasmic reticulum (Harper et al., 1998; Hong et al., 1999), and tonoplast (Malmstrom et al., 1997, 2000) as well as in the plasma membrane (Askerlund, 1997; Bonza et al., 2000; Chung et al., 2000). To date, about eight type IIB Ca2+-ATPases have been found in Arabidopsis with the evidence for their identification as such varying from unequivocal (genetic, biochemical, molecular studies) to putative (descriptions of ESTs having sequence similarity to known type IIB Ca2+-ATPases) (Geisler et al., 2000; Sze et al., 2000). It is noteworthy that evidence for the subcellular localization of CaM to various membranes in plants (Collinge & Trewavas, 1989; Rodriguez-Concepcion et al., 1999) is consistent with reports on the regulation and distribution of Ca2+ pumps. Although there are differences in intracellular Ca2+ sequestration between plants and animals (the former utilizing primarily the vacuole, the latter predominately the ER), it remains unclear as to why endomembrane type IIB Ca2+-ATPases are found only in plants. Their generalized structure consists of 10 membrane-spanning domains, a large central cytoplasmic loop, and an overall mass of c. 116 and c. 130 kDa for endomembrane and plasma membrane type IIB Ca2+-ATPases, respectively (Geisler et al., 2000).
Interestingly, the CaMBD of plant type IIB Ca2+-ATPases is present near the N-terminus, in contrast to the type IIB Ca2+-ATPases in mammals where the CaMBD is invariably in the C-terminal (Carafoli & Brini, 2000). The N-terminal CaMBDs of plant type IIB Ca2+-ATPases are not particularly well-conserved but do show a predicted secondary structure common among CaMBDs (propensity to form an amphiphilic alpha-helix) and bind CaM in the nM range (Table 1). Although the positioning of the CaMBDs differ between plant and animal type IIB Ca2+-ATPases, a similar mechanism of activation seems to be employed; relief of autoinhibition by the binding of Ca2+/CaM (Askerlund, 1997; Harper et al., 1998; Carafoli & Brini, 2000). Based upon models developed in mammalian systems, it is likely that, in the absence of Ca2+, the autoinhibitory region of plant type IIB Ca2+-ATPases interacts with a region(s) within the pump to maintain it in an inactive state. The binding of Ca2+/CaM relieves this autoinhibition thereby activating the pump. It is noteworthy that the regions comprising autoinhibitory and CaMBDs within plant and animal type IIB Ca2+-ATPases are typically juxtaposed and sometimes overlapping (Geisler et al., 2000; Carafoli & Brini, 2000). In addition to regulation by CaM, reversible phosphorylation seems to be a mechanism of post-translational regulation for both plant and animal Ca2+ pumps. ACA2 is an Arabidopsis Ca2+-ATPase located in the endoplasmic reticulum whose activity can be stimulated by Ca2+/CaM or inhibited by the phosphorylating activity of CPK1 (a CDPK) (Hwang et al., 2000). Moreover, prior binding of Ca2+/CaM to ACA2 prevents inhibitory phosphorylation and thus ACA2 may represent a point of crosstalk between different Ca2+ signalling pathways based upon the initial kinetics of Ca2+-mediated phosphorylation and Ca2+-mediated CaM binding. Given the multiplicity of CDPK and CaM isoforms in plants, it seems likely that similar examples involving other plant Ca2+ pumps will be uncovered in the near future.
Another type of CaM-stimulated Ca2+-pump, CAP1, has recently been identified in maize (Subbaiah & Sachs, 2000). This pump shows sequence similarity to mammalian SERCA-type pumps. However, unlike other CaM-binding Ca2+-ATPases found in plants which possess an N-terminal CaM-binding domain, CAP1 has a C-terminal CaM-binding domain, a feature of mammalian PM-type pumps. CAP1 is therefore a hybrid pump in some respects. Although CAP1 is predicted (based upon sequence homologies) to be localized to the ER, the subcellular distribution, and physiological function, of CAP1 remains unclear (Subbaiah & Sachs, 2000). Nevertheless, the presence of multiple CaM-regulated Ca2+ pumps likely affords the cells an additional layer of control when fine-tuning intracellular Ca2+ levels. Despite the recent advances in identification of type IIB Ca2+-ATPases in plants, most remain poorly characterized and their individual cellular roles are unclear.
2. CaM regulation of ion channels
Ion channels are subject to various noncovalent or covalent structural modifications mediated by a variety of cytoskeletal and signalling proteins (Sheng & Pak, 2000). These modifications enhance their functional flexibility and can alter voltage- or ligand-sensitivity, probability of opening, rate of desensitization or cause inactivation. There is accumulating evidence that several ion channels are directly modulated by CaM binding (Van Eldik & Watterson, 1998). Activation of sodium and potassium channels in Paramecium is believed to involve direct CaM binding. Ca2+/CaM inhibits the opening of sarcoplasmic reticulum Ca2+ release channels (ryanodine receptors), and several CaM-binding sites have been identified. Trp and trpl are Drosophila ion channels that were cloned on the basis of their CaM binding properties, and may be important in phototransduction. Mammalian NMDA (Krupp et al., 1999) and olfactory cyclic-nucleotide gated ion channels (Grunwald et al., 1999) are both inhibited by Ca2+/CaM through identified CaM-binding sites, and this inhibition has important implication for synaptic physiology and olfaction, respectively. The small conductance Ca2+-activated K+ channel has been shown to be activated directly by Ca2+/CaM, with identified regions mediating CaM interaction (Xia et al., 1998), and it was shown that CaM plays a pivotal role in both Ca2+-dependent inactivation and facilitation of L-type Ca2+ channels (Zuhlke et al., 1999) via an IQ-like motif found on the carboxy tail of the α1c pore forming subunit (Peterson et al., 1999; Zuhlke et al., 1999). CaM was also shown to bind to analogous IQ regions of N-, P/Q-, and R-type Ca2+ channels, suggesting that CaM-mediated regulation may be widespread in the Ca2+ channel family (Peterson et al., 1999). Another example of CaM-modulated ion channel activity is the inositol 1,4,5-trisphosphate (IP3) receptor in animal cells. These endomembrane receptors control Ca2+ release from internal stores in response to the secondary messenger IP3 and thus play an important role in the integration of phosopholipid and Ca2+ signalling. CaM binding inhibits IP3 binding to these receptors thereby inhibiting Ca2+ release. Interestingly, CaM binds to IP3 receptors in both a Ca2+-dependent and Ca2+-independent fashion (Patel et al., 1997). However, the extent of inhibition and the requirement for Ca2+ during CaM-mediated inhibition appears to differ among IP3 receptor isoforms (Vanlingen et al., 2000).
By comparison, much less is known about CaM interactions with ion channels in plants. It was shown by patch-clamp techniques that the activity of slow vacuolar (SV) channels from barley aleurone cells is modulated by Ca2+/CaM. Increasing external free Ca2+ stimulated their activity, whereas CaM antagonists were inhibitory (Bethke & Jones, 1994). Kurosaki et al. (1994) suggested the presence of plant plasma membrane ion channels that are activated by cAMP, permeable to Ca2+, and are negatively regulated by CaM, and further discussed the cross-talk of cAMP and Ca2+ cascades in plants (Kurosaki, 1997). Yet, the molecular identity of these putative CaM-regulated channels remains unknown. Nevertheless, recently, a novel family of channel proteins from barley (HvCBT1; Schuurink et al., 1998), Arabidopsis (AtCNGC1-AtCNGC6; Kohler et al., 1999; Leng et al., 1999; Sunkar et al., 2000), and tobacco (NtCBP4; Arazi et al., 1999, 2000a) was reported. Structurally, these proteins all contain six putative transmembrane domains, a presumed pore region located between the fifth and sixth transmembrane domains, and a putative cNMP-binding domain coinciding with a high-affinity CaM-binding site (Arazi et al., 2000a). These proteins are most similar in overall structure to mammalian cyclic nucleotide-gated nonselective cation channels which have been implicated in Ca2+ signal transduction (Grunwald et al., 1999). Interestingly, in Arabidopsis, different members of this protein family show distinct interactions with Arabidopsis CaM isoforms in vitro (Kohler & Neuhaus, 2000). A schematic description of CaM-regulated ion transporters is presented in Fig. 5.
3. Crosstalk of cyclic nucleotide and Ca2+ signalling: the channel connection
Proteins that bind cNMPs (cAMP or cGMP) share a structural domain of c. 100–120 residues. The best studied is the prokaryotic catabolite gene activator. Other proteins known to contain these domains are cAMP- and cGMP-regulated protein kinases, vertebrate CNG-gated ion channels (Shab & Corbin, 1992) and several eag related K+ channels (Warmke & Ganetzki, 1994). In plants, the inward rectifying K+ channels that have been identified in Arabidopsis (KAT1 and AKT1) (Anderson et al., 1992; Sentenac et al., 1992) and potato (KST1; Mueller-Roeber et al., 1995), contain a region with similarity to the cyclic nucleotide-binding domains of animal proteins and thus represent putative targets for cNMPs. For KAT1 and AKT1, it could be shown that cGMP can modulate their activity in heterologous systems (Hoshi, 1995; Gaymard et al., 1996). Similarly, Leng et al. (1999) reported on the activation of the Arabidopsis CNGC2 by cNMPs in mammalian cells.
Cyclic nucleotide-binding domains are not sequence but structure conserved (Shab & Corbin, 1992). X-ray crystallography of CAP showed that this domain is composed of three α-helices and a distinctive eight-stranded, antiparallel β-barrel structure. The cyclic nucleotide binds within a pocket formed by the αC-helix and β-barrel. Cross-talk between cNMPs and Ca2+/CaM is known to occur in animal rod and olfactory CNG channels (Grunwald et al., 1999). These channels serve as final targets for cNMPs and a functionally significant feature is their permeability for Ca2+. Ca2+/CaM decreases the apparent affinity constant of the channels for cNMPs, resulting in a decreased cation influx through the channels (Grunwald et al., 1999). A similar mechanism may exist with plant CNGCs. However, in contrast to the mammalian channels, those of plants have overlapping Ca2+/CaM and cNMP-binding domains (Fig. 6). These channels provide another example of similarities in CaM-regulated proteins in animals and plants but with different arrangements of their CaM-binding domains, as in the case of Ca2+-ATPases (see above). Interestingly, in phytochrome-mediated light signal transduction a Ca2+/CaM-dependent pathway mediating chloroplast development, and a cGMP-mediated pathway mediating chalcone synthase expression operate in a reciprocally repressive manner; microinjection of Ca2+/CaM inhibits anthocyanin production, whereas cGMP microinjection inhibits CAB expression (Bowler & Chua, 1994). The mechanism underlining this regulation is unknown. However it is not unlikely that the molecular basis for this crosstalk involves a protein (perhaps a channel) with overlapping Ca2+/CaM- and cGMP-binding domains.
It should also be noted that whereas CaM can directly bind and modulate the activity of certain ion channels, these may also be modulated indirectly by Ca2+/CaM- and cyclic nucleotide dependent protein kinases (Guerini, 1998), or by other kinases such as protein kinase C (Guerini, 1998). In plants, CDPKs were reported to modulate the phosphorylation and/or the activity of a putative soybean ion channel (nodulin 26; Lee et al., 1995b), an anion (Cl−) channel (Pei et al., 1996), H+-ATPase (Lino et al., 1998), potassium channel (KAT1; Li et al., 1998), and Ca2+-ATPases (Hwang et al., 2000). In general, regulation of ion flux is likely an important point of interaction between signalling pathways utilizing Ca2+ and CaM, cNMPs, and other secondary messengers, and likely plays an important role in co-ordinating this cross-talk.
VI. CaM and plant responses to environmental stimuli
In view of the role Ca2+ plays in mediating plant responses to biotic (Levine et al., 1996) and abiotic (Knight, 2000) stimuli, it is not surprising that CaM, as an important cellular Ca2+ receptor, is involved in mediating these responses. Table 2 summarizes the evidence for the involvement of CaM in plant responses to various biotic and abiotic signals, some of which are discussed in greater detail in this chapter.
Table 2. Involvement of calmodulin in plant responses to environmental stimuli and stresses
Stimulus or stress treatment
CaM, CaM-like, and CaM-binding protein/gene implicated in the response*
*The calmodulin (CaM) like proteins and CaM-binding proteins implicated in the responses are indicated. ‘CaM’ appears where no specific CaM isoform or CaM-target protein has been reported. §The type of evidence supporting the involvement of CaM is indicated as follows: MC, molecular characterization of a gene/cDNA; EX, gene expression; BC, biochemical analysis of protein function in vitro, protein–protein interactions (yeast 2-hybris system, coimmunoprecipitation), and ligand–protein interactions; PS, physiological studies (Ca2+ measurements, or measurement of metabolic activity); PR, pharmacological studies using chemical agonists and antagonists of CaM, or microinjection of signalling components into plant cells; RG, functional studies by reverse genetics (transgenic plants and/or gene knockout).
One of the best studied stress responses leading to rapid modulation of metabolic enzyme activity by Ca2+/CaM in plants is that of glutamate decarboxylase, the enzyme catalysing the conversion of L-glutamate to γ-aminobutyric acid (GABA). The variety of physical and chemical stimuli shown to induce GABA production in plants was summarized in recent reviews (Shelp et al., 1999; Snedden & Fromm, 1999). These stimuli include temperature shock, mechanical manipulation, hypoxia, water stress, and other treatments. The synthesis of GABA in response to external stimuli is often quite rapid. Certain stimuli, such as cold shock and mechanical perturbations, appear to elicit a very rapid (within seconds to minutes) increase in GABA synthesis, reaching 100-fold from the basal level (Ramputh & Bown, 1996). Consequently, it is likely that this phenomenon is a result of changes in the activity of GAD rather than an effect on transcription or translation. The temporal changes in cytosolic Ca2+ levels differ according to the nature of stimulation applied. For example, touch or low temperature elicit very rapid and transient increases in cytosolic Ca2+ levels in tobacco epidermal cells (Knight et al., 1991; McAinsh & Hetherington, 1998). These findings correlate well with previous descriptions of rapid GABA synthesis as a consequence of cold shock or mechanical manipulation (Snedden & Fromm, 1999) and with evidence that GAD activation in response to cold requires both Ca2+ and CaM (Cholewa et al., 1997). Similarly, the slower rise in cytosolic Ca2+ concentration observed during hypoxia in plant cells (Bush, 1995) is reminiscent of studies showing hypoxia-induced accumulation of GABA during several hours of treatment.
Evidence for the regulation of GAD by Ca2+/CaM is consistent with analysis of the purified recombinant enzyme. Using purified recombinant GAD from petunia, it was possible to demonstrate (Snedden et al., 1996) that Ca2+ exerts influence upon GAD activity via CaM, and that GAD is virtually inactive in the absence of Ca2+ and CaM at physiological pH (7.0–7.5). The levels of Ca2+ (K0.5 c. 0.8 µM) and CaM (K0.5c. 15 nM) required for half-maximal activation of the recombinant GAD are in accord with physiological estimations of these signal molecules (Roberts & Harmon, 1992; Bush, 1995) and are similar to values reported for other purified CaM-binding proteins in plants (Table 1). The proper regulation of GAD and other GABA shunt genes is important for normal plant development and amino acid metabolism (Baum et al., 1996; H. Fromm, unpublished). However, the physiological role of enhanced GABA synthesis in stress situations is unclear. Recent studies have demonstrated that plant amino acid transporters which facilitate the movement of compatible solutes such as proline and glycine-betaine, are also capable of selectively transporting GABA (Breitkreuz et al., 1999; Schwacke et al., 1999). This raises the interesting possibility that GABA may function as a compatible solute during stress response and merits further investigation. GAD and GABA are associated with nitrogen and carbon metabolism via the GABA shunt, a pathway which feeds, ultimately, into the Krebs cycle (Busch & Fromm, 1999; Shelp et al., 1999). In addition, other recent studies revealed a form of CaM-binding GAD in yeast. Overexpression of this GAD in yeast cells confers tolerance to oxidative stress whereas disruption of the GAD gene increases the sensitivity to oxidative stress (Coleman et al., 2001). Although the role of CaM in regulating yeast GAD has not been investigated, these recent findings may provide clues as to the role of GAD in stress tolerance in plants.
CaM appears to play a signalling role in other metabolic pathways. Plants are unique among eukaryotes in producing their own carbohydrates from inorganic carbon, the synthesis of which is controlled by developmental as well as environmental signals. Studies have implicated Ca2+/CaM in phototransduction pathways that control chloroplast development (Bowler & Chua, 1994). This developmentally regulated process establishes the machinery for the photosynthetic activity of plants whereby carbon is fixed into sugars. In addition, CaM was also suggested to participate in sugar sensing and/or sugar signal transduction (Smeekens & Rook, 1997). More recently, evidence for the involvement of CaM in transducing signals that signify changes in the carbon status of the plant was reported. The petunia CaM-related protein designated CaM53, which contains an extension of 35 amino acids with a prenylation site at the C-terminus, was shown to be translocated to the nucleus in response to carbon starvation (e.g. dark), but associated with the plasma membrane in the light or in the dark in the presence of an exogenous carbon source (Rodriguez-Concepcion et al., 1999). This suggests that CaM53 may have distinct cellular targets, dependent upon the carbon status of the cell, in the plasma membrane and in the nucleus.
An interesting but controversial role for CaM within chloroplasts was suggested almost two decades ago (Jarrett et al., 1982; Muto, 1982). The presence of CaM-binding proteins in chloroplasts (Roberts et al., 1983), and effects of CaM antagonists on inhibiting chloroplast functions (Barr & Crane, 1982) were also reported. However, these early reports should be considered with caution either because high quality antibodies against plant CaM were unavailable at the time, or due to limitations concerning the cell fractionation procedures used. Nevertheless, the possible presence of CaM in chloroplasts is intriguing from an evolutionary perspective. Whereas CaM is considered a typical eukaryotic protein, others suggest that prokaryotes do possess CaM-like proteins (Pettersson & Bergman, 1989; Onek & Smith, 1992; Onek et al., 1994; Smith, 1996). In particular, cyanobacteria, which are believed to be related to the prokaryotic ancestors of higher plant chloroplasts, possess a 22-kDa protein reported to cross react with antibodies against spinach CaM, and to activate known CaM-target proteins such as NAD kinase in a Ca2+-dependent manner (Onek et al., 1994). Others have reported the detection of Ca2+/CaM-regulated adenylate cyclase in cyanobacteria (Bianchini et al., 1990). Thus, it is possible that CaM-like proteins were involved in photosynthesis-related activities before the appearance of eukaryotic photosynthetic organisms. The issue is intriguing also in view of evidence for Ca2+ diurnal fluctuations in chloroplasts and in response to light/dark transitions (Johnson et al., 1995). However, in spite of the evidence for an involvement of CaM in various aspects of carbon metabolism, its presence in chloroplasts remains controversial. The recent finding of the chloroplast chaperonine 10 as a CaM-binding protein (Yang & Poovaiah, 2000c) does not clarify the issue since this chaperonine is nuclear encoded and may be regulated by CaM before being transported into the chloroplast. It would be of interest to investigate the potential modulation of chaperonine activity by CaM within the chloroplast environment, to identify other potential CaM targets in the chloroplast and, most important in this regard, to obtain conclusive evidence proving or refuting the occurrence of CaM in chloroplasts.
2. Mechanical, light, and gravity stimuli responses mediated by CaM
There is ample evidence for a role of Ca2+ in regulating the establishment of plant form in response to external and internal mechanical stimuli (reviewed by Trewavas & Knight, 1994; Knight et al., 1995; Trewavas & Malho, 1997). Mechanical signals have dramatic effects on plant morphology such as stem length and thickening, cell wall composition (e.g. lignification), direction of pollen tube growth, and wrapping of tendrils around supportive surfaces. The role of distinct CaM-related genes in mediating plant responses to mechanical stimuli was reinforced by the characterization of a group of touch-inducible genes (TCH genes) in Arabidopsis (Braam & Davis, 1990). TCH genes are rapidly induced by various mechanical stimuli (touch, rain, wounding), in some cases reaching an increase of 100-fold in mRNA levels within minutes. Moreover, the effect of the mechanical stimulation on expression appeared to be proportional to the strength of the stimulus; stronger mechanical stimuli resulted in a higher increase in their mRNA levels. These studies are reminiscent of earlier reports, based on pharmacological studies, implicating CaM in thigmomorphogenesis (Jones & Mitchell, 1989).
Ca2+ was also suggested to function as a mediator of phytochrome responses (Roux, 1983, 1992; Shacklock et al., 1992), and based upon the use of CaM antagonists, CaM was suggested to be one of the downstream effectors (Datta et al., 1985). More recent studies provided further evidence for these earlier observations. Microinjection of CaM into a phytochrome-deficient mutant of tomato allowed the identification of Ca2+/CaM-dependent and -independent pathways for light-mediated gene activation and chloroplast development (Neuhaus et al., 1993). Ca2+/CaM microinjection leads to the induction of genes associated with the development of the photosynthetic complexes (e.g. CAB genes), whereas genes that are involved in anthocyanin biosynthesis (e.g. the gene encoding chalcone synthase) can be induced by microinjection of cGMP. Moreover, these signalling molecules operate in a reciprocally repressive manner; Ca2+/CaM microinjection inhibits anthocyanin production, whereas cGMP microinjection inhibits CAB expression (Bowler & Chua, 1994). Recent studies demonstrated that phytochrome may control Ca2+ permeability of the plasma membrane and induces transient changes in cytosolic Ca2+ (Volotovski, 1998). Evidence for the translocation of phytochromes into the nucleus in a light-dependent manner suggests that CaM might be involved in light-dependent processes in the nucleus, such as transcription (Nagy & Schaefer, 2000). This is particularly interesting in view of the light/dark-regulated translocation of specific CaM isoforms into the nucleus (Rodriguez-Concepcion et al., 1999). In addition, UV-B, but not UV-A light, appears to modulate chalcone synthase gene expression through a CaM-mediated pathway (Christie & Jenkins, 1996). Similarly, microinjection of CaM into the zygotes of fucoid algae enhanced photopolarization of the zygote axis (Love et al., 1996). In general, however, the targets of CaM that mediate these responses remain to be determined.
The response of plants to gravity, as an asymmetric growth of cells on opposite sides of an organ, involves asymmetric distribution of Ca2+ (e.g. in the fern Ceratopteris richardii;Chatterjee et al., 2000), although an earlier report suggested that the gravitropic response in Arabidopsis roots is not associated with detectable changes in Ca2+ (Legue et al., 1997). Various lines of evidence suggest that CaM is involved in the response to gravity. Changes in CaM activity and localization are associated with maize root cap and plumule, which have gravisensing cells (Stinemetz & Evans, 1985; Roux & Dauwalder, 1985). Low concentrations of CaM antagonists inhibit gravitropic curvature without affecting growth (Sinclair et al., 1996). More direct evidence has emerged from analysis of the Arabidopsis agravirtropic mutant, agr-3. This mutant, when compared to WT, shows reduced levels of gravity-induced CaM expression (Sinclair et al., 1996). In addition, a protein kinase that binds CaM has been detected in maize root cap (Lu et al., 1996). As the gravisensing response involves changes in cell polarity, cytoskeleton and organelle distribution, it is possible that CaM is a regulator of these cellular processes, possibly via a kinesin-like motor protein. Auxin redistribution is an important step in the asymetrical growth required for gravity stimulated curvature. The recent identification of an auxin responsive protein (SAUR1) with CaM-binding properties (Yang & Poovaiah, 2000a), may suggest that hormonal regulation in the gravitropic response is also regulated by CaM. Moreover, in some plants, the gravity response is a light-dependent process mediated by phytochrome. Since CaM participates in the response to both environmental signals, it is possible that cross talk between them involves a CaM-regulated pathway.
The perception of temperature change by plants and the subsequent initiation of adaptive biochemical responses constitutes a process known as thermotolerance. Heat stress results in massive changes in plant metabolism, gene expression, protein folding machinery, and, eventually, growth and development. Various studies suggest that adaptive responses to heat shock are, at least in part, mediated by Ca2+ ions (Gong et al., 1998). Pre-incubation of plants with Ca2+ improves their thermotolerance, whereas increased thermosensitivity occurs when plants are treated with Ca2+ chelators (Gong et al., 1998). In addition, the magnitude of the cytosolic Ca2+ signals that are measured in response to heat shock are proportional to the severity of the stress treatment (Gong et al., 1998).
Little is known about the downstream CaM targets of heat-induced Ca2+ signals. One candidate involved in heat shock changes in plant metabolism is the CaM-regulated GAD. Heat shock induces GABA accumulation in roots of Arabidopsis, and this heat-induced GABA accumulation is abolished if plants are pretreated with CaM antagonists or Ca2+ channel blockers (Locy et al., 2000). The measured changes in GABA levels are not associated with changes in GAD levels, hence activation of GAD via Ca2+/CaM most likely accounts for GABA accumulation.
In addition to the metabolic changes in response to heat stress, other cellular functions involve CaM-target proteins. The wheat FK506-binding proteins FKBP73 and FKBP77 (Table 1) bind CaM and function in protein folding associated with their peptidyl prolyl cis-trans-isomerase activity. FKBP77 expression is induced by heat shock (Kurek et al., 1999), and both FKBPs form complexes with HSP90, another heat shock protein (Reddy et al., 1998). In addition, the cytosolic HSP70 protein from maize was also recently found to interact with CaM, a phenomenon reported to inhibit the intrinsic ATPase activity of HSP70 (Sun et al., 2000). Other CaM-binding proteins of unknown function, whose expression is induced or suppressed by heat shock, have also been reported (Lu et al., 1995). These findings imply that CaM-related proteins are likely involved in modulating biochemical aspects of plant adaptation to heat stress and thermotolerance.
The involvement of Ca2+ in responses to cold stress has been described in various organisms including bacteria (Torrecilla et al., 2000), mosses (Russell et al., 1996), higher plants (Knight et al., 1991), protozoa (Kuriu et al., 1997) and vertebrates (Job et al., 1983; Sweet & Welsh, 1988). There is also accumulating evidence that CaM is involved in mediating at least some of these responses in plants (Braam & Davis, 1990), and other organisms (Kuriu et al., 1997). CaM-related touch-induced genes were shown to be up-regulated by cold shock (Braam & Davis, 1990; Polisensky & Braam, 1996). Moreover, the increased expression of some of the TCH genes requires an intracellular Ca2+ increase. Refinement of the subcellular pattern of Ca2+ transients in tobacco by using aequorin targeted either to the nucleus or cytosol revealed that expression of a specific CaM gene was induced by wind and cold shock (Van der Luit et al., 1999). Interestingly, wind-induced CaM expression was found to be regulated by Ca2+ signalling operational in the nucleus, whereas expression of the same gene in response to cold is regulated by a pathway operational in the cytoplasm. In Paramecium, studies of CaM gene mutants have implicated CaM and CNGCs in cold stress responses (Kuriu et al., 1997). The presence of a large family of CaM-binding CNGCs in plants (Table 1) should prompt further studies into their possible roles during cold stress response.
4. Responses to anoxia and salinity
A role for Ca2+ during responses to hypoxia has been documented in plants (Subbaiah et al., 1994a,b, 1998) and animals (e.g. Beitner-Johnson et al., 1998). In plants responding to anoxia, CaM was suggested to regulate GAD activity (Aurisano et al., 1995). In addition, CaM is involved in regulating Ca2+ homeostasis by activating plasma- and endomembrane Ca2+-ATPases. Recently, a SERCA-type Ca2+-ATPase with a CaMBD was reported (designated CAP1) whose expression is enhanced in response to anoxia as well as other stresses (Subbaiah & Sachs, 2000). The expression of a recently characterized plasma membrane soybean Ca2+-ATPase (SCA1) is highly and rapidly induced in response to salt stress (Chung et al., 2000).
The role of Ca2+/CaM in response to salinity stress appears to have similarities in organisms from yeast (Lippuner et al., 1996), through plants (Pardo et al., 1998) to mammals (Tremblay et al., 1991). Of particular interest are emerging similarities between yeast and plants with regard to the involvement of calcineurin-like proteins in salt tolerance (Lippuner et al., 1996; Liu & Zhu, 1998; Pardo et al., 1998). Plants transformed with a regulatory subunit of yeast calcineurin and a truncated form of the catalytic subunit (to yield a constitutively activated phosphatase) exhibited substantial salt tolerance in comparison to WT plants (Pardo et al., 1998). These findings suggest that plants, like yeast, may possess a CaM-dependent calcineurin, which functions within transduction pathways required for salt stress adaptation.
5. Response to heavy metals
CaM is an important cellular target for various heavy metals. Lead (Pb2+) can occupy all four Ca2+-binding sites of CaM simultaneously (Ouyang & Vogel, 1998) and can activate CaM to over 90% of its potential activity with Ca2+. Other metals that can bind at the known Ca2+-binding sites of CaM are Sr2+ and Cd2+. Metals like Mg2+, Zn2+, K+ and Na+, may either bind with low affinity to some of the Ca2+-binding sites, or interact with CaM through auxiliary sites (Ouyang & Vogel, 1998) and exert different effects on CaM functions. Consequently, it is not surprising that certain heavy metals have been found to mimic the effect of Ca2+ in signal transduction. For example, Pb2+ can act as a Ca2+ substitute in second messenger mediated metabolic control (Goldstein, 1993). The similarity in the protein binding sites for Pb2+ and Ca2+ is also consistent with evidence that Pb2+ entry into animal (Simons & Pocock, 1987; Tomsig & Suszkiw, 1991) and plant (Huang & Cunningham, 1996) cells occurs, at least in part, through Ca2+-permeable channels. The recently discovered plasma membrane CaM-binding CNGCs (Table 1) may serve as entry pathways for certain heavy metals (Arazi et al., 1999, 2000b; Sunkar et al., 2000). Indeed, a reverse genetic approach revealed that over-expressing a member of this tobacco protein family conferred hypersensitivity to Pb2+ and enhanced Pb2+ accumulation in transgenic plants (Arazi et al., 1999). By contrast, transgenic plants expressing a truncated non-functional form of the same protein, from which regulatory domains were removed, displayed attenuated uptake of Pb2+ and improved tolerance to this toxic metal (Sunkar et al., 2000). A similar phenotype was exhibited by an Arabidopsis mutant with a T-DNA insertion in the CNGC1 gene (Sunkar et al., 2000).
Aluminium, although not a heavy metal by definition (density 2.7 g per cm−3), is an abundant metal comprising nearly 10 percent of the crust of the earth, and exerts negative effects on plants leading to major problems in world agriculture. Recent studies suggest that extracellular CaM may be involved in the effects of aluminium on pollen germination and tube elongation (Ma et al., 2000). Aluminium was also reported to have an effect on cytosolic calcium homeostasis in root hairs (Jones et al., 1998). However, the mechanisms involved in aluminium toxicity are still not clear.
6. Involvement of CaM in plant defence against pathogens
Several lines of evidence implicate CaM in plant responses to pathogens. Plants expressing a mutated CaM that is incapable of being methylated at lysine 115 exhibited a more rapid and stronger active oxygen burst than that observed in control plants challenged with the same stimuli (Harding et al., 1997). It was suggested that the CaM-regulated NAD kinase may be a downstream target which, by altering NAD(H)/NADP(H) homeostasis, potentiates active oxygen species production by NADPH oxidase during defence response (Harding et al., 1997). Further evidence for the involvement of specific CaM isoforms in plant responses to pathogens emerged from expression in transgenic tobacco of two soybean CaM isoforms, SCaM-4 and SCaM-5 (Heo et al., 1999). Both genes are rapidly induced by pathogens or fungal elicitors. This induction is abolished by Ca2+ chelators and is mimicked in the presence of Ca2+ ionophores implying a role for Ca2+ in gene induction. Interestingly, other stimuli, such as exposure to ABA, known to involve Ca2+ signalling, did not induce the expression of these genes suggesting a specificity of response. Transgenic tobacco overexpressing SCaM-4 or SCaM-5 showed increased levels of systemic-acquired-resistance (SAR) genes, spontaneous HR-associated lesions (in the absence of pathogens), increased pathogen resistance, but not elevated levels of salicylic acid (SA). Remarkably, transgenic plants showed enhanced SAR gene expression even in the presence of the bacterial nahG gene (which degrades endogenous SA) indicating an SA-independent pathway of gene induction correlated with SCaM expression. Taken together, these data make a compelling case for the involve-ment of SCaM-4 and -5 in pathogen response. Another CaM-like protein associated with the hypersensitive response to pathogens is Hra32, which is induced by Pseudomonas syringae (Jakobek et al., 1999).
It is not yet clear which cellular targets are modulated by these CaM isoforms during pathogen response. However, recent genetic studies of plant – pathogen interactions identified a gene designated DND1 which is essential for the hypersensitive response to pathogens, although it is not required for activation of other defence responses. Cloning of DND1 revealed its identity as a CaM-binding CNGC (designated CNGC2;Clough et al., 2000). This gene is similar to other CNGC-like genes isolated from other plant species (Table 1). The current working model is that these channels are activated by cyclic nucleotides (cGMP or cAMP) in response to an extracellular stimulus (e.g. elicitor), and closed (inactivated) as a result of the increase in cytosolic Ca2+ via CaM, as the cNMP- and CaM-binding domains coincide (Arazi et al., 2000a).
VII. Conclusions and future studies
Plants possess tremendous flexibility in modulating their metabolic processes and growth patterns in response to environmental stimuli. An extreme example of this flexibility would be the inhibition of seed germination or bud emergence in the absence of appropriate environmental signals. In other examples, the pattern of growth, plant morphology, and timing of different life cycle stages are often dictated by environmental conditions. Whereas plants are able to respond to many environmental stimuli with a certain degree of specificity, it is the integration of signals that is required for optimization of plant growth. The evidence for roles of Ca2+ and Ca2+ sensors as key components in the messenger systems evoked during signalling and stimuli response is mounting. Moreover, the complexity and multiplicity of the Ca2+/CaM messenger system in plants vs that found in animals, is probably a key factor in the morphogenic and developmental plasticity of plants. Indeed, the evidence for a continuously changing Ca2+ messenger system in plants in response to environmental stimuli was suggested to be comparable to learning in animals, a process associated with increased neural connections within a neural network (Trewavas, 1999). There remains a great deal to learn about the various components involved in the signal transduction networks within plant cells. The CaMs and their relatives represent an excellent system for addressing questions about the decoding and interpretation mechanisms underlying Ca2+ signals. Plants possess multiple CaM genes encoding virtually identical proteins, a family of closely related CaMs, and an extended family of evolutionarily divergent CaM-like proteins. In order to understand the roles of these proteins, and to ultimately delineate the signalling networks in plant cells, a concerted effort and multidisciplinary approach will be required. In particular, further research is needed in the following areas if we are to help achieve that goal. A detailed phylogenetic comparison of all (sequence-confirmed) members of the extended CaM family should be conducted following the completion of the Arabidopsis thaliana genome sequence. This will provide a frame of reference for subsequent functional analyses and may shed light upon subgroupings within the CaM family. Comprehensive analyses are needed on the mRNA and protein expression patterns, at the tissue, cellular, and subcellular levels, for the various CaM family members. Subcellular localization studies may prove to be of particular interest if microdomains of Ca2+ sensors, or specialized, multiprotein signalling complexes exist within cells. In addition, it is imperative to examine the developmental and stress-responsive expression profiles for each family member as this may provide insight into their cellular functions. It is expected that advances in high-throughput (transcript) expression analysis using DNA microarrays will help facilitate these studies. Target identification studies need to be performed for the various members of the CaM family. In vitro analyses have revealed interesting aspects of competition and target specificity among the CaM family. However, this work must be extended to include all members of the family, and more importantly, to examine in vivo interactions between CaMs and their targets. Advances in imaging (such as in vivo FRET studies) and mass-spectrometry (MALDI-TOF, SELDI-TOF) technologies to look at protein–protein interactions, in addition to conventional methods (coimmunoprecipitations, colocalization, interactive cloning) should help provide clues as to the effector identities and specificity of the various CaM members. Similarly, assessing the Ca2+-binding affinities of the CaM family, in the presence of their targets, will help to determine which CaMs are responsive to particular amplitudes of Ca2+ signals in vivo. Coupled with expression and subcellular localization information, a holistic picture of Ca2+ sensing and interpretation is expected to emerge. (iv) Genetic and reverse genetic analyses of Ca2+/CaM signalling pathways (e.g. Krysan et al., 1996; Bouché, 2000) must be intensified to provide the plant material necessary to ultimately link molecules to cellular function. It is anticipated that as more knock-out transgenic lines become available, an assessment of the involvement in development and stress-response for each family member (individually or in combinations) could be systematically conducted. Again, initial expression studies may help provide the clues as to the growth conditions needed to observe physiological effects of particular gene knock-outs. Lastly, it is expected that various genetic approaches will continue to yield information serendipitously on the roles of different CaMs (e.g. broad-spectrum analyses of stress-induced genes, mutant screens, protein-interaction screens).
In parallel, efforts to refine our understanding of the spatio-temporal complexities of Ca2+ signals, and the mechanisms that generate and propagate them, must be continued. New technologies to measure Ca2+ signals with high resolution within the cell should be accompanied by biochemical and molecular studies to associate these Ca2+ signals with the spatial distribution of Ca2+ transporters, Ca2+-responsive proteins and their downstream effectors. Dynamic cellular processes should be monitored at the single molecule level in real time wherever possible. Findings from the various subdisciplines of plant science (development, metabolism, etc.) need to be incorporated into a comprehensive model of how plants use second messengers to organize their responses to external stimuli. The past decade has seen exciting advances in our knowledge of Ca2+ signalling, yet there remains a tremendous amount to learn before we will be able to provide an encompassing picture of the signalling networks operating within plant cells.
We wish to thank Professor S.J. Roux for providing data prior to publication, and for critical reading of the manuscript.