Identification of a novel divergent calmodulin isoform from soybean which has differential ability to activate calmodulin-dependent enzymes.

Calmodulin plays pivotal roles in the transduction of various Ca-mediated signals and is one of the most highly conserved proteins in eukaryotic cells. In plants, multiple calmodulin isoforms with minor amino acid sequence differences were identified but their functional significances are unknown. To investigate the biological function of calmodulins in the regulation of calmodulin-dependent enzymes, we cloned cDNAs encoding calmodulins in soybean. Among the five cDNAs isolated from soybean, designated as SCaM-1 to −5, SCaM-4 and −5 encoded very divergent calmodulin isoforms which have 32 amino acid substitutions from the highly conserved calmodulin, SCaM-1 encoded by SCaM-1 and SCaM-3. SCaM-4 protein produced in Escherichia coli showed typical characteristics of calmodulin such as Ca-dependent electrophoretic mobility shift and the ability to activate phosphodiesterase. However, the extent of mobility shift and antigenicity of SCaM-4 were different from those of SCaM-1. Moreover, SCaM-4 did not activate NAD kinase at all in contrast to SCaM-1. Also there were differences in the expression pattern of SCaM-1 and SCaM-4. Expression levels of SCaM-4 were approximately 5-fold lower than those of SCaM-1 in apical and elongating regions of hypocotyls. In addition, SCaM-4 transcripts were barely detectable in root whereas SCaM-1 transcripts were as abundant as in apical and elongating regions of hypocotyls. In conclusion, the different biochemical properties together with differential expression of SCaM-4 suggest that this novel calmodulin may have different functions in plant cells.

Calmodulin, a highly conserved and ubiquitous protein in eukaryotes, mediates Ca 2ϩ signals to various target proteins (1). A variety of regulatory enzymes and proteins such as protein kinases, ion channels, Ca 2ϩ pumps, nitric oxide synthetase, inositol trisphosphate kinase, cyclic nucleotide phospho-diesterase, and NAD kinase are known to be regulated by Ca 2ϩ and calmodulin (2)(3)(4). While a great deal of information has been known for biological roles of calmodulin in animal cells, very little is known about the roles of calmodulin in plant cells. This is mainly due to the absence of purified calmodulindependent enzymes and/or their genes in plants. As an effort to investigate the biological role(s) of calmodulin in plants, calmodulin genes in various plant species have been cloned and characterized recently (4). Interestingly, in Arabidopsis, cDNAs encoding multiple calmodulin isoforms have been isolated although the degree of sequence divergence is minor, only 6 amino acid differences between the two most divergent isoforms among Arabidopsis calmodulins (5)(6)(7). This is very notable because, in animal cells, only a single form of calmodulin is produced by a calmodulin multigene family (8,9). However, it has not been determined whether plant calmodulin isoforms have the same biochemical properties such as calcium-binding abilities and activation of calmodulin-dependent enzymes. Also the biological role of multiple calmodulin isoforms in vivo is completely unknown.
To understand Ca 2ϩ /calmodulin-mediated signal transduction mechanisms in plants, first we cloned calmodulin cDNAs from soybean. From these cDNA clones, we have identified two novel divergent calmodulin isoforms. One of the calmodulin isoforms showed several distinct characteristics to other highly conserved plant calmodulin isoforms. Here we describe the structural and functional differences between the novel divergent calmodulin isoform SCaM-4 and the highly conserved calmodulin SCaM-1.

MATERIALS AND METHODS
Isolation of Calmodulin cDNAs from Soybean-A cDNA library was constructed in ZAPII (Stratagene) from half-apical and half-elongating regions of hypocotyls (0.3-1.3 cm section below cotyledon tissues) of 4-day-old etiolated soybean (Glycine max L. cv. Williams) seedlings. The library was screened by using the ECL gene detection system (Amersham) with a rice calmodulin genomic clone, cam-2, (10) as a probe. Nucleotide sequences were determined from both strands of cDNAs using a Taq dye primer cycling sequencing kit on a 373A automatic DNA sequencer (Applied Biosystems Inc.). Nucleotide and deduced amino acid sequences were analyzed using the GCG sequence analysis program (11). For phylogenetic analysis of calmodulin, amino acid sequences of calmodulins were aligned with CLUSTAL V (12) and by eye with LINEUP of the GCG sequence analysis program. Those positions not occupied in all OTUs were excluded from the alignment (available upon request) leaving 130 sites for phylogenetic inference. Distance between sequences was measured as number of amino acid substitutions/site using a ␥ distribution for the variability of substitution rate across positions (13). For this, a neighbor-joining tree (14) was constructed using the proportion of amino acid differences between sequences, from which the ␥ parameters were estimated. The value thus determined (2.162) was used to estimate numbers of substitutions/site * This work was supported by a grant from the Korea Science and Engineering Foundation to the Plant Molecular Biology and Biotechnology Research Center. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  between sequences using the ␥ correction. The resulting distance matrix yielded the final neighbor joining tree. The reliability of branches was estimated by bootstrapping using the same ␥ parameter.
Production of Calmodulin Proteins in E. coli-A T7 expression vector, pET-3d, was used for production of SCaM-1 and SCaM-4 protein in Escherichia coli BL21(pLysS)DE3 (15). The start codon of SCaM-1 cDNA was modified to introduce an NcoI site by PCR 1 with a mutant oligomer (5Ј-CAATCACCATGGCAGAT-3Ј) and the T3 oligomer (5Ј-ATTAACCCTCACTAAAG-3Ј) as primers. The PCR product digested with NcoI and BamHI was subcloned into the NcoI and BamHI site of the pET-3d vector. Integrity of the SCaM-1/pET-3d construct was verified by nucleotide sequencing. SCaM-4 cDNA had an NcoI site at the start codon and was processed as SCaM-1 without a modification. Calmodulin proteins expressed in E. coli were purified to homogeneity by Ca 2ϩ -dependent phenyl-Sepharose (Pharmacia) column chromatography as described (16,17). Protein concentration was determined using a Protein assay kit (Bio-Rad) with bovine serum albumin and bovine brain calmodulin (Sigma) as standards.
Preparation of Antibodies and Immunoblotting-Polyclonal antibodies against two SCaM isoforms, SCaM-1 and SCaM-4, were prepared by immunizing goats subcutaneously with 10 mg of each purified SCaM protein in the Freund's complete adjuvant. Subsequent boosting injections were done at 3-week intervals with 1 mg of protein in the Freund's incomplete adjuvant. To determine cross-reactivities among SCaM-1, -4, and bovine brain calmodulin, unpurified antisera were used in immunoblot experiments. Anti-bovine brain calmodulin polyclonal antibody was purchased from ICN-Biomedicals. Purified SCaM proteins and bovine brain calmodulin (50 ng) were electrophoresed on 13.5% SDS-polyacrylamide gels either in the presence of 5 mM EGTA or 5 mM CaCl 2 in the SDS sample buffer (18). Proteins were transferred onto a PVDF membrane (Millipore) and incubated with either anti-SCaM-1 or anti-SCaM-4 antibody. Protein bands were detected using the ECL system after incubating with horseradish peroxidase-conjugated antigoat IgG (ICN Biomedicals). Polyclonal antiserum to each isoform was purified to remove any remaining minor cross-reactivities by two successive antigen affinity chromatographies (19) using SCaM-1-Sepharose or SCaM-4-Sepharose columns prepared by conjugating purified SCaM-1 or SCaM-4 to CNBr-activated Sepharose 4B according to the procedure recommended by the manufacturer (Pharmacia LKB Biotechnology Inc.). Briefly, anti-SCaM-4 antiserum was first loaded onto a SCaM-4-Sepharose column, and bound fractions were passed through a SCaM-1-Sepharose column to remove any cross-reactive antibody to SCaM-1. Anti-SCaM-1-specific antibody was prepared similarly. Plant extracts were prepared from leaves of mature chinese cabbage, rice, tobacco, tomato, and from seedling tissues of soybean and Arabidopsis as described (20).
Assays of Calmodulin-dependent Enzymes-Phosphodiesterase activity was assayed using commercially available bovine heart calmodulin-deficient phosphodiesterase (Boehringer Mannheim) as described (21). NAD kinase was partially purified from pea seedlings by successive protamine sulfate precipitation, polyethylene glycol precipitation, and DEAE-Sephacel column chromatography procedure as described (22). Effluents from the DEAE-Sephacel column were used for NAD kinase assays without further purification. NAD kinase assay was done as described (23) with varying amount of activator calmodulins. As controls, NAD kinase activations were examined with reaction mixtures either in the presence of 5 mM EGTA or absence of exogenous activator calmodulin to verify that our NAD kinase preparation was free of endogenous calmodulin contamination and that the activation of NAD kinase was a calcium-dependent process. To determine activation parameters, curve fittings were done by the aid of the GraFit software (24) through non-linear regression analysis with a modified Hill equation (25) where v is the observed rate, V max is the maximal activity, [CaM] is the concentration of added calmodulin, K act is the concentration of calmodulin required for half-maximal activity, and n is the Hill coefficient.
Southern and Northern Blot Analyses-Genomic DNA was prepared from soybean leaves as described by Dellaporta et al. (26) and further purified by a cesium chloride/ethidium bromide density gradient ultra-centrifugation (27). Ten g of purified DNA was digested with appropriate restriction endonucleases, size fractionated on a 0.8% agarose gel, and transferred onto a Hybond-Nϩ membrane (Amersham). Hybridization was carried out in the buffer of 6 ϫ SSPE, 5 ϫ Denhardt's reagent, 1% SDS, 100 g/ml denatured calf thymus DNA, 1 ϫ 10 6 counts/min/ml probe at 65°C. 32 P-Labeled probe was made from a 678-bp EcoRI fragment of SCaM-1 which contains the entire calmodulin coding region by the random primer labeling method (28). The hybridized blot was finally washed in 0.1 ϫ SSC, 0.1% SDS at 65°C. For Northern blot analyses, tissues and organs from various parts of 4-day etiolated soybean seedlings and 5-week-old mature soybean plants were divided into sections as described previously (29). Total RNA was extracted from these sections according to the acid guanidium thiocyanate-phenol-chloroform extraction method (30) and further purified by ultracentrifugation (27). Twenty g of total RNAs were electrophoresed on 2% formaldehyde/agarose gels and transferred onto GeneScreen Plus membranes (New England Nuclear). Hybridizations were carried out as described above except for hybridization temperature of 60°C and final washes in 0.5 ϫ SSC, 0.1% SDS at 55°C. Gene-specific probes were made from the 3Ј-untranslated region of each cDNA: a 276-bp HaeIII-XhoI fragment of SCaM-1, a 251-bp EcoRI fragment of SCaM-2, a 368-bp HaeIII-XhoI fragment of SCaM-3, a 347-bp EcoRI-XhoI fragment of SCaM-4, and a 313-bp EcoRI-XhoI fragment of SCaM-5.
RT-PCR Analyses-Three g of total RNA was reverse-transcribed in a 20-l reaction volume containing 50 mM Tris-Cl, pH 8.3, 75 mM KCl, 1.5 mM MgCl 2 , 10 mM dithiothreitol, 500 M each dNTP, 10 ng of oligo(dT) 17 , 40 units of ribonuclease inhibitor (Promega), 200 units of SUPERSCRIPT II (Life Technologies, Inc.) at 42°C for 30 min. The reaction mixture of first strand cDNA was diluted with 30 l of water and used for PCR amplification. Primers for the PCR were: 5Ј-TCA-CAACAAAGGAGCTT-3Ј and 5Ј-TCTCATCAACCTCTTCATC-3Ј for specific co-amplification of SCaM-1 and SCaM-3, and 5Ј-ATACACCAT-GGCAGATA-3Ј and 5Ј-TCTGCTCCACCTCTTCATC-3Ј for SCaM-4. Internal standards were generated by making small internal deletions in original cDNA clones as described (29). The sizes of predicted amplified products were 290 bp for SCaM-1 cDNA, 119 bp for SCaM-1 internal standard, 380 bp for SCaM-4 cDNA, and 206 bp for SCaM-4 internal standard. PCR was carried out in a 50-l reaction volume containing 3 l of the diluted first strand mix, 10 mM Tris-Cl, pH 9.0, 50 mM KCl, 1.5 mM MgCl 2 , 0.1% Triton X-100, 200 M each dNTP, 30 ng of each primer, 100 pg of internal standard, 5 units of Taq polymerase using a Gene-Amp PCR system 9600 (Perkin-Elmer). Amplification was carried out at the condition of a 1-min denaturation at 94°C, followed by 25 cycles of 94°C for 30 s, 42°C for 10 s, and 72°C for 1 min. After PCR, 20 l was removed and electrophoresed in a 2% agarose gel. The amplified products were transferred onto a GeneScreen plus membrane, hybridized with a cDNA probe, and visualized by the ECL gene detection system.

RESULTS
Isolation of Calmodulin cDNAs from Soybean-Five cDNA clones encoding calmodulin isoforms were isolated by screening a soybean cDNA library with a rice calmodulin genomic clone as a probe (10). Among 86 positives out of 5 ϫ 10 4 plaques screened, 84 cDNA clones were grouped into three different calmodulin cDNAs which were designated as SCaM-1, -2, and -3. Two additional cDNA clones, SCaM-4 and -5, were isolated from a single weekly hybridizing plaque in both cases. To verify that these five cDNAs were derived from the soybean genome, genomic Southern blot analysis was performed. Probe generated from the calmodulin coding region of SCaM-1 recognized multiple hybridizing bands of more than five in each enzyme digestion as shown in Fig. 1. In addition, genomic Southern blot analyses using gene-specific probes prepared from the 3Ј-untranslated region of each cDNA showed specific hybridizing bands to each probe (data not shown), indicating that the five cDNAs were derived from genuine transcripts of a soybean calmodulin gene family.
Comparison of the Deduced Amino Acid Sequences of SCaMs with Other Calmodulins-SCaM-1, -2, and -3, encoded calmodulins which were very similar to other plant calmodulins such as alfalfa, barley, and Arabidopsis calmodulins (4). As shown in Fig. 2, both SCaM-1 and SCaM-3 encoded the same calmodulin isoform, SCaM-1, and have an identical amino acid 1 The abbreviations used are: PCR, polymerase chain reaction; CNBr, cyanogen bromide; PDE, 3Ј,5Ј-cyclic nucleotide phosphodiesterase; PVDF, polyvinyledene difluoride; RT-PCR, reverse transcription-coupled polymerase chain reaction; SCaMs, soybean calmodulins; bp, base pair(s). sequence to that of alfalfa calmodulin (32). SCaM-2 differed by only two amino acid residues from SCaM-1 and was identical to barley calmodulin (33). However, SCaM-4 and -5 encoded novel divergent calmodulins, and the two isoforms were different from SCaM-1 by 32 amino acid residues out of 149 amino acid residues in both cases. Among them, at least 10 substitutions were non-conservative exchanges which had not been found in any of plant and animal calmodulins. The extent of amino acid substitutions found in SCaM-4 and -5 was very surprising since potato calmodulin (34), the most divergent calmodulin isoform isolated in plants, has 10 amino acid substitutions from SCaM-1. Thus these results indicate that SCaM-4 and -5 are the most divergent calmodulins identified so far in both animals and plants. To see the primary sequence relationship of the two divergent SCaM isoforms with other known calmodulins, a phylogenetic analysis was performed using amino acid sequences of 33 known calmodulins by the neighbor-joining method (see "Materials and Methods"). The neighbor-joining tree, Fig. 3, showed a close relationship of SCaM-1 and SCaM-2 to most of plant calmodulins as expected. Potato, tomato, and one petunia calmodulin isoform constitute a separate group, which reflects primary structural diversities of them. However, two novel SCaM-4 and -5 did not belong to either of the two plant calmodulin groups and instead constituted a new independent group. Furthermore, the long branch length for SCaM-4 and -5 suggests a release of functional constraint and attainment of new functions for the two soybean calmodulin isoforms.
Characteristics of SCaM-1 and SCaM-4 Proteins Expressed in E. coli-To investigate significance of multiple amino acid substitutions at the protein level, two representative calmodulin isoforms, SCaM-1 and SCaM-4 proteins, were produced in E. coli (see "Materials and Methods") and purified to homogeneity by Ca 2ϩ -dependent hydrophobic interaction chromatog-raphy (17). During the purification, the two proteins behaved similarly with respect to heat stability and elution profiles on a phenyl-Sepharose column. First, we investigated electrophoretic mobility shift of the two calmodulin isoforms upon Ca 2ϩ binding which is a typical characteristic of calmodulins (35,36). Both SCaM-1 and SCaM-4 proteins showed Ca 2ϩ -dependent electrophoretic mobility shifts. However, the extent of shift was different among calmodulin isoforms (Fig. 4). SCaM-4 showed the greatest degree of mobility shift and the extent of shift was approximately 2-fold greater than that of SCaM-1. This is very interesting because SCaM-1 and -4 have nearly identical calculated molecular weights of 16,862 and 16,819 and isoelectric points of 4.26 and 4.36, respectively. These results suggest that SCaM-4 may become a more compact structure than SCaM-1 and bovine calmodulin upon Ca 2ϩ binding.
Also, we investigated antigenic differences among the two isoforms and bovine calmodulin. Unpurified antisera raised against SCaM-1 and SCaM-4 in goats were examined for crossreactivity by immunoblot analysis. As shown in Fig. 5, anti-SCaM-4 antiserum recognized SCaM-4 as expected but not SCaM-1 and bovine calmodulin. In contrast, anti-SCaM-1 antisera recognized SCaM-1 and bovine calmodulin but not SCaM-4, suggesting that SCaM-1 has a more closely related structure to bovine calmodulin than SCaM-4. However, after overexposure of immunoblots, faint cross-reacting bands were observed in both blots, indicating that the two isoforms may have common but very weak antigenic epitopes (data not shown). Thus the two calmodulin isoforms have different major antigenic determinants despite 70% of amino acid identity. Control experiments with anti-bovine brain calmodulin antibody showed a poor cross-reaction with SCaM-4 in contrast to a strong cross-reaction with SCaM-1 as effective as bovine brain calmodulin (data not shown). These data strongly suggest that the multiple amino acid substitutions may confer significant differences on the protein structure of SCaM-4.
Differential Activation of Phosphodiesterase and NAD Kinase by SCaM-1 and SCaM-4 -To assess the ability of the divergent calmodulin, SCaM-4, to activate calmodulin-dependent enzymes, phosphodiesterase and NAD kinase assays were performed. For these assays, bovine heart calmodulin-deficient cyclic nucleotide phosphodiesterase was purchased from a commercial source, and NAD kinase was partially purified from pea seedlings as described (see "Materials and Methods"). As shown in Fig. 6A, both of the two calmodulin isoforms activated phosphodiesterase equally well and the half-maximal activation values of the two isoforms were 7.63 and 6.17 nM for SCaM-1 and SCaM-4, respectively. In addition, maximal activation values of the two SCaM isoforms for PDE were not significantly different from each other. These results clearly indicate that SCaM-4 is a bona fide functional calmodulin isoform despite of its primary structural diversity. However, surprisingly, when examined for the activation of pea seedling NAD kinase, the divergent SCaM-4 did not activate NAD kinase at all even at 500-fold higher concentration than that of SCaM-1 for a maximal activation of NAD kinase (Fig. 6B). SCaM-1 activated NAD kinase approximately 5-fold higher than bovine brain calmodulin (data not shown). This is most likely due to the absence of post-translational modification of SCaM-1 protein prepared in E. coli because trimethylation of calmodulin has been shown to decrease NAD kinase activation by 4-fold (37). In the presence of EGTA, both enzymes did not show any activity regardless of the presence of activator calmodulins. Thus, the activation of both enzymes by calmodulin is a calcium-dependent process.
Differential Expression of SCaM-1 and SCaM-4 in Various Tissues-To examine the expression pattern of the five calmodulin genes, we carried out Northern blot analyses using total RNA. RNA gel blots prepared from total RNA of various tissues and organs were hybridized with each gene-specific probe prepared from the 3Ј-untranslated region of each cDNA (see "Materials and Methods"). We first examined total calmodulin gene expression with a probe made from the coding region of SCaM-1 since this probe cross-hybridized with all of other calmodulin cDNAs under hybridization conditions used. As shown in Fig. 7, abundant expression of calmodulin genes was observed in the apical and elongating regions of seedling hypocotyls and stem and root of mature plants. SCaM-1 and -2 had similar expression patterns to that of total calmodulin genes in all tissues we examined. The expression pattern of SCaM-3 was similar to that of SCaM-1, but expression levels were lower in apical region of hypocotyls, pods, and nodules. Thus, the similar expression pattern of SCaM-1, -2, and -3 in all tissues and organs examined suggests that expression of these genes may be regulated by a similar transcriptional control. The expression of SCaM-4 was barely detected with total RNA gel blots probably due to a low expression level of the gene. However, when poly(A) RNA was used for Northern blot analysis, the expression of SCaM-4 was clearly visible in apical and elongating regions of hypocotyls of soybean seedlings (Fig. 7,  panel C). To quantitate the relative expression level of SCaM-1 and SCaM-4, semi-quantitative RT-PCR analysis was performed. The primer set used for the amplification of SCaM-1 mRNA was designed to co-amplify SCaM-1 and SCaM-3 transcripts because both genes encode the same protein, SCaM-1. The specificity of these primers for SCaM-1, -2, and -3 were tested using SCaM cDNA clones (data not shown). As shown in Fig. 7D, the result of RT-PCR assay for SCaM-1 and SCaM-4 was consistent with that of Northern analysis. The two isoform genes were actively expressed in apical and elongating regions of hypocotyls although the level of SCaM-4 mRNA was at least 5-fold lower than that of SCaM-1. In mature hypocotyls, both genes were expressed at very low but similar levels so that bands were detected only after prolonged exposure of the blot (data not shown). Interestingly, in roots SCaM-4 mRNA was almost undetectable whereas expression level of SCaM-1 was as abundant as for that of apical and elongating regions of hypocotyls. The differential expression patterns of SCaM-1 and -4 suggest that these genes may be subject to completely dif-ferent transcriptional regulations.
Presence of SCaM-4 Homologs in Other Plants-The finding of the functionally different calmodulin isoform in soybean prompted us to investigate the presence of SCaM-4 homologs in various plants at the protein level. A Western blot analysis with SCaM-4-specific antibody (see "Materials and Methods" for preparation) was employed to study this possibility using protein extracts from six different plant species including five dicot and one monocot plants. As shown in Fig. 8A, soybean extract and all other plant extracts contained protein species recognized by the anti-SCaM-4-specific antibody, suggesting that SCaM-4 homologs or closely related proteins may be expressed in other plants. The multiple bands migrated between 14.4 and 21.5 kDa may represent either a group of proteins closely related to SCaM-4 which have different electrophoretic mobilities, such as SCaM-5, or post-translationally modified forms of SCaM-4 homologs (38). When the same blot was reprobed with anti-SCaM-1-specific antibody to examine crossreactivity, as shown in Fig. 8B, a different immunostaining pattern was obtained, which further ensured no detectable cross-reactivity of antibodies used in this experiment. Interestingly, the protein level of SCaM-4 is relatively lower than that of SCaM-1. These results indicate that other plants may also have both forms of calmodulin isoforms, SCaM-1 and SCaM-4 homologs.
FIG. 6. Activation of Ca 2؉ /calmodulin-dependent enzymes by SCaM isoforms. Dose-response curves of PDE and NAD kinase are shown. Data points represent means of three (for PDE) or five (for NAD kinase) independent assay results, and error bars represent standard deviations. Fitted curves are drawn from the Hill equation as described under "Materials and Methods." Panel A, activation of PDE by SCaM-1 and SCaM-4. PDE assay was done using calmodulin-deficient bovine heart PDE and cAMP as substrate. Activity of PDE was monitored with varying amounts of calmodulins and expressed as a relative activity to that of PDE in the presence of 100 nM activator SCaM-1. K act values of SCaM-1 and SCaM-4 for PDE were 7.63 and 6.17 nM, and the Hill coefficients were 1.87 and 1.67, respectively. Panel B, NAD kinase assay. Pea NAD kinase was partially purified and assayed as described under "Materials and Methods." The activity of NAD kinase is expressed as a relative activity to that of NAD kinase in the presence of 80 nM SCaM-1. K act value of SCaM-1 for NAD kinase was 8.07 nM and the Hill coefficient was 1.86.  17 primer and subsequently PCR amplified using isoform-specific oligonucleotide primers. Amplified products were separated on a 2% agarose gel, hybridized with each cDNA probe, and visualized by the ECL gene detection system. Closed arrow heads (ñ) indicate 290-bp amplified cDNA copies of mRNAs encoding SCaM-1 (left panel) and 380-bp amplified cDNA copies of mRNAs encoding SCaM-4 (right panel), respectively. Internal standards co-amplified for semiquantitive analysis are indicated by open arrow heads (î). DISCUSSION Calmodulin is one of the most highly conserved proteins in higher eukaryotes. The essential role of calmodulin in a variety of cellular processes may be the reason for strict conservation of the primary structure of calmodulin during evolution. However, recent studies in a plant system revealed the presence of multiple calmodulin isoforms in a single organism (4 -7). This is very interesting because there exists only a single form of calmodulin in animal systems although calmodulin is encoded by multiple genes (8,9). This implicates that the plant system may have a unique feature although the overall Ca 2ϩ /calmodulin-mediated signal transduction mechanism is similar to that of animal systems. Plant calmodulin isoforms found so far have minor amino acid sequence divergency such that the most divergent plant calmodulin isoform, potato calmodulin, has the difference of 10 amino acid residues from SCaM-1 (34). Most of the other plant calmodulins have only one to six amino acid difference and showed more than 90% identities at the amino acid sequence level. In contrast to this, newly identified SCaM-4 and -5 have only 78% identities to SCaM-1 and appear to be the most divergent calmodulin isoforms found yet. Therefore, the study presented here not only increased the number of calmodulin isoforms in plant systems but also, more importantly, greatly extended the primary structural diversities of functional calmodulins.
SCaM-4 had distinct characteristics compared to the highly conserved SCaM-1 such as differences in the extent of Ca 2ϩdependent electrophoretic shift, antigenicity, and, most importantly, the ability to activate calmodulin-dependent enzymes. To our surprise, SCaM-4 did not activate NAD kinase at all although its ability to activate phosphodiesterase was as good as SCaM-1. In addition, SCaM-4 did not inhibit the activation of NAD kinase by SCaM-1 even in the presence of 100-fold molar excess of SCaM-4 to that of SCaM-1, suggesting that SCaM-4 may not bind NAD kinase at the assay condition used (data not shown). The reason for this difference may be due to structural difference between SCaM-1 and SCaM-4. Previous studies on the relationship between structure and function of calmodulin by site-directed mutageneses or isolation of mutants showed that only single critical amino acid substitution could change the ability of calmodulin to activate target enzymes (39 -41). Thus the failure of SCaM-4 to activate NAD kinase is most likely due to the divergent primary structure of SCaM-4. However, we cannot exclude the possibility that plant phosphodiesterase is differently regulated by these isoforms.
It is very intriguing to have such divergent calmodulin isoforms in plants compared to animal systems. Although, currently there is no definitive answers to the role(s) of the divergent isoforms, our studies with SCaM-1 and SCaM-4 provide important clues. The differential activation of calmodulin-dependent enzymes by different calmodulin isoforms suggests that each isoform may have its own target enzymes although some of them may be shared. This implicates that, in plant, there may exist calmodulin isoform-specific processes. However, this notion needs further studies on the identification of more calmodulin-dependent enzymes. Currently a very limited number of calmodulin-dependent enzymes have been identified or cloned in plants, which makes these studies difficult. Another possible function of plant calmodulin isoforms is signal-, tissue-, and/or developmental stage-specific roles. Consistent with this proposal, differential expression of two mungbean calmodulin isoforms in response to various stimuli have been reported recently (42). Interestingly, the two mungbean calmodulin isoforms MBCaM-1 and MBCaM-2 have identical amino acid sequences to SCaM-1 and SCaM-2, respectively. Thus it is plausible that SCaM-1 and -2 may respond similarly to these various signals although SCaM-1, -2, and -3 showed the same expression patterns in various tissues we examined. Also the different expression level and patterns of SCaM-4 compared to other calmodulin genes support the notion of the tissue-and/or organ-specific functions of calmodulin isoforms. Further studies on the response of each calmodulin gene to various stimuli will clarify these possibilities.
Currently we are in the process of generating transgenic tobacco plants with sense and antisense constructs of SCaM-1 and -4 to investigate the biological function of each isoform in vivo. However, until now we were not very successful in obtaining transgenic plants. One of reasons may be due to the essential role(s) of calmodulin in various processes in cells, thus plants with perturbation of calmodulin expression level may not be able to survive.