Plant cells are equipped with highly efficient mechanisms to perceive, transduce and respond to a wide variety of internal and external signals during their growth and development. Perception of signals via receptors results in generation or synthesis of nonproteinaceous molecules often termed messengers such as calcium which control diverse cellular processes through calcium sensors which is also known as calcium binding proteins (Bowler and Fluhr, 2000; Reddy, 2001; Snedden and Fromm, 2001; White and Broadley, 2003).The intracellular calcium sensor protein calmodulin (CaM) interacts with a large number of proteins to regulate their biological functions in response to calcium stimulus. This molecular recognition process is diverse in its mechanism, but can be grouped into several classes based on structural and sequence information (Yap et al., 2000). CaM is a small (148 residues) multifunctional acidic protein arranged in two globular domains connected with a long flexible helix. Each globular domain contains a pair of intimately linked EF hands. CaM undergoes conformational changes on binding of four Ca²⁺ ions to their four EF-hand domains. On binding Ca²⁺, hydrophobic surfaces are exposed and allow the Ca²⁺-CaM complex to interact with, and to regulate the activity of, an array of target proteins. The activities of these proteins affect physiological responses to the vast array of specific stimuli received by plant cells (Yang and Poovaiah, 2003). It is a key calcium sensor in all eukaryotes, regulates diverse cellular processes by interacting with other proteins (Reddy et al., 2000).

In plants, CaM regulates a variety of targets including kinases, metabolic proteins, cytoskeletal proteins, ion channels and pumps, and transcription factors (Bouche et al., 2005). CaM is found in the apoplast and in the cytosol, ER and nucleus in plant cells. Within the cytosol, the estimated CaM concentration is 5 to 40 µM (Zielinski, 1998). CaM is the paradigm for EF-hand proteins that are involved in signal transduction (Kawasaki et al., 1998). In addition to the evolutionarily conserved form of CaM, plants possess an extended family of CaM isoforms and CaM-like of the most interesting differences between Ca²⁺ signaling in plants proteins (Snedden and Fromm, 1998). Primary structures of CaM are generally conserved throughout evolution among various organisms. However, in plants, one striking difference is that numerous isoforms of CaM may occur within a single plant species. A large family of genes encoding CaM isoforms from several plants has been identified including Arabidopsis (Arabidopsis thaliana) (Zielinski, 2001), potato (Solanum tuberosum) (Takezawa et al., 1995), soybean (Glycine max) (Lee et al., 1995) and petunia (Petunia hybrida) (Rodriguez-Concepcion et al., 1999). The representative structure of EF-hand of sequenced Eleusine coracana CaM clone is shown in Figure 1.

Figure 1: EF-hand structure generated from cloned CaM of Eleusine coracana based on its amino acid sequence.

Cereals and millets constituted the staple food of the world since their domestication~10,000 years ago which were most important group of cultivated plants in terms of food production and acreage covered, providing>60% of the calories and proteins requirement of our daily diet (O'Kennedy et al., 2006; Varshney et al., 2006). The genomes of cereals and millets are very different in terms of size, ploidy level and chromosome number. Despite these significant differences, comparative mapping analysis reveals colinearity at the great level of genetic markers while genes tend to be highly conserved at the DNA sequence level between different cereals as well as millet genomes (Devos and Gale, 2000; Feuillet and Keller, 2002). Several attempts have been made to reveal the existing synteny and colinearity on the basis of comparative genomics (Kellogg, 1998; Paterson et al., 2004; Paterson et al., 2005; Caetano-Anolles, 2005).

Eleusine coracana, commonly called finger millet or ragi is one of the important minor cereals cultivated mainly in East Africa and Southern India where it makes an important contribution to food security due to the high nutritional value and good storage ability of its grain. Its improvement has, however, lagged behind that of other crops (Srinivasachary et al., 2007). It is an excellent source of calcium (376-515mg/100g) which is far above the ground then the other cereals and millets (Barbeau and Hilu, 1993). The high calcium level in finger millet prompted us to investigate the factors or molecular mechanism responsible for it accumulation. Both environmental and genetic factors influence calcium accumulation in bean seed (Quenzer et al., 1978). However, little is known about factors influencing distribution of calcium and variation of calcium levels in different cereals and millets.

Therefore an attempt has been made in the present study to characterize the CaM of finger millet along with cereals and millets to identify the structural similarity of CaM genes with their possible role in calcium signaling and calcium accumulation in cereals. This paper reports in silico analysis of PCR amplified, cloned and sequenced CaM from finger millet varieties viz. brown, golden and white along with other cereals and millets which were amplified with the same set of primer designed from finger millet EST, of five cereals namely rice (Oryza sativa), maize (Zea mays), sorghum (Sorghum bicolor), barley (Hordeum vulgare) and oat (Avena sativa), and five millets viz. finger millet (Eleusine coracana), barnyard millet (Echinochloa frumentacea), proso millet (Panicum milliaceum Linn.), little millet (Panicum antidotale) and kodo millet (Paspalum scrobiculatum Linn.). All nucleotide sequences of cloned CaM were submitted to GenBank database and have received the accession numbers (Table 1).

Table 1: List of assigned accession number from Gen Bank of sequenced CaM genes from cereals and millets.

Calcium estimation

Calcium content of different cereals and millets was estimated by Atomic Absorption Spectrophotometer (AAS) method. The samples were prepared as described by Barbeau and Hilu (1993). Samples were wet ashed to avoid mineral decomposition and destruction (volatilization) that may occur while ashing in a muffle furnace. Ashed samples were analyzed for calcium by atomic absorption spectrophotometry (AAS) using a Perkin-Elmer 2100 AA Spectrophotometer. Each analysis was done in triplicate.

PCR amplification and cloning

The finger millet EST sequence (CX264726) was downloaded from finger millet EST database of NCBI (http://www.ncbi.nlm. nih.gov/nucest) and CaM specific primer set EF03F/ Ca0R1 (FAACTGTCATGCGTTCATTGG, R- GGAGTCGAGCTGATTTGATGA) was designed using Primer-3 online tool (Untergasser et al., 2007). The genomic DNA of different cereals and millets were isolated by standard method (Murray and Thompson, 1980), quantified and analyzed on agarose gel electrophoresis (Maniatis et al., 1989). PCR amplification was performed as per the standard protocol using 50- 100 ng of template DNA with 60°C annealing temperature based on Tm value of the primers. The amplified products were analyzed on agarose gel and the expected size amplicons were gel eluted using QIAquick Gel Extraction Kit (Qiagen, USA) and cloned in pGEM-Teasy vector (Promega, USA) as per the kit instructions. Putative cloned CaM genes were sequenced using M13 universal primer present in pGEM-T easy vector.

In silico analysis

The vector sequence was wiped off by VecScreen software. The sequences were further subjected to bioinformatics softwares GENESCAN (Stormo, 2000) and FGENESH (Solovyev et al., 2006) for fishing out the probable genes and the putative CDS and were translated to protein sequence using translation tool (http:// ca.expasy.org/tools/dna.html). The translated CaM protein sequences were subjected to protein functional analysis using PFAM version 23.0 (Finn et al., 2006), PROSITE version 20.37 (Castro et al., 2006) and INTERPROSCAN version 4.4 (Quevillon et al., 2005) and SMART version 5.1(a Simple Modular Architecture Research Tool) (Schultz et al., 1998). These sequences of CaM proteins from different cereals and millets were aligned using ClustalW (Thompson et al., 1994) and phylogenetic tree was constructed using UPGMA method. A phylogenetic tree was constructed by MEGA 3.1 software (Tamura et al., 2007). The conserved motifs present in these sequences were analyzed using BLOCKS and MEME (Multiple EM for Motif Elicitation) software version 3.5.7 (Bailey et al., 2006). For motif analysis of CaM, the selection of maximum number of motifs was set to 10 with minimum width of 15 amino acids for cloned CaM sequences and 3 with minimum width of 20 amino acids for the comparative analysis with the published CaM sequences which were retrieved from NCBI database while other factors were of default selections. These selections were made in order to minimize the 'Evalue' of the given parameter based on the probability of finding an equally well conserved pattern in set of sequences. Motifs involved with regulatory region of putative CaM genes were analysed using PLACE tool (Higo et al., 1999) and PLANTCARE tool (Lescot et al., 2002).

Three dimensional structural analyses

For constructing the structures of calmodulin, a template for homology modeling was searched with BLAST program on the Protein Data Bank (www.rcsb.org/pdb/; Berman et al., 2000). Template structure was selected with a cut-off sequence identity of < 40%. Three dimensional structures of calmodulin were modeled using MOE software version 2008.10 (Chemical Computing Group's Molecular Operating Environment) and energies were also minimized using parameters such as Force field= MMFF94X, Gradient= 0.05, Cutoff: On=8, off=10, Solvation: Dielectric=1, Exterior=80. Secondary structure components of the calmodulin sequences were analyzed using SWISS-PDB viewer (http://spdbv.vital-it.ch) and Geno3D (http:// geno3d-pbil.ibcp.fr) tools. Binding of Calcium with different EF hands was performed by MOE and HEX Docking software. Model consistency and viability were appraised by PROCHECK software available online (http://www.ebi.ac.uk./Thorton/software.html) for protein structure verification (Laskowski et al., 1993). Analysis of the modeled structures was performed using the RasMol (Raster Display of Molecules) version 2.7.3.1 (Sayle and Milner-White, 1995).

Determination of calcium content of Cereals and millets

Calcium content was estimated using AAS. The calcium content was highest in finger millet as compared to other cereals and millets. The calcium content in PRM-1 (Brown), PRM-701(golden) and PRM- 801 (white) varieties of finger millet was found to be 314.5, 232 and 144 mg/100g respectively, having high calcium as compare to other cereals and millets (Table 2). The results are in general agreement with previous investigators who have reported that finger millet containing much higher calcium than other cereal grains

Table 2: Total calcium content of different cereals and millets.

PCR amplification of CaM from different cereals and millets

The CaM from different cereals and millets were PCR amplified using same CaM specific primer designed from finger millet ESTs, gel eluted, cloned in pGEMTeasy vector and sequenced. The PCR amplicon of expected size i.e. 613 bp was uniformly observed with all the templates except foxtail millet as shown in Figure 2.

Figure 2: PCR amplifi cation profi ling of cereals and millets using EF-03F/ Ca01 R1 primer set L- 100 bp ladder, Lane 1- Brown, 2-Golden, 3-White, 4- Rice, 5-Wheat, 6- Maize, 7- Oat, 8- Barley, 9- Sorghum, 10-Little millet, 11- Kodo millet, 12- Proso millet, 13- Barnyard millet, 14- Foxtail millet, M-λ DNA / EcoR1 / HindIII Double Digest ladder.

Comparative in silico analysis of CaM genes

The CaM sequences were subjected to BLAST search for deducing similarity with sequences available in different databases. CaM sequences, range from 579 to 623bp, were obtained for different cereals and millets and this variation in sequence length might be due to amplification of variable length of genomic sequence by same CaM specific primer. These sequences were subjected to homology search to confirm their identity for CaM genes except in T. aestivum

Homology search confirm the identity as CaM genes in E. coracana, Z. mays, H. vulgare, A. sativa, E. frumentacea, P. antidotale and P. scrobiculatum while as CaM like gene in O. sativa, S. bicolor, P. milliaceum. Further, the nucleotide sequences were translated to respective protein sequences using translation tool and subjected to protein Blast to reveal the similarity at protein level with other existing CaM proteins to further confirm their identity as CaM. Beside this the sequences were also confirmed by subjecting the nucleotide sequences to gene finding software GENESCAN and FGENESH. All the CaM sequences were aligned using Clustal W to find out the extent of similarity amongst the sequences which showed highly conserved (110 aa) sequences associated with three EF-hand except sorghum which is containing only two EF-hand, though the few variations were also observed in sequences of CaM as shown in Figure 3a.

Figure 3: Multiple Sequence alignment showing conserved amino acids of the CaM protein sequences from cereals and millets. Conserved EF – hand shown in rectangles and yellow dotted portion indicates the conserved amino acid residues among different CaM sequences.

The translated CaM protein sequences subjected to phylogenetic tree construction using a neighbor-joining method, based on amino acid sequence similarity, formed two distinct clusters reveals all cloned CaMs grouped in one cluster except sorghum which is separated from the others (Figure 3b).

In all CaMs, motif-1 was most commonly observed which is functionally related to its calcium binding properties while Motif-2 also having similar function as Motif-1 was present in all CaMs except Rice and Sorghum which might be due to the CaM like gene or single set of primer amplify the different region. Beside this Motif-3 also most frequently present in all CaMs which is functionally unknown. Motif-6, which is having Protein kinase C phosphorylation site, observed only in barley and finger millet varieties (Brown, golden and white) while Motif 4 having Kinase binding site is present in Oat, Kodo millet, Little millet, Barnyard millet as shown in Figure 4. The sequences of all motif shown in Table 3.

Figure 4: Phylogenetic tree based on 12 CaM sequences and schematic distribution of respective conserved motifs using MEME software. Multilevel consensus sequences for the MEME defi ned motifs are listed in Table 3.

Table 3: Multilevel consensus sequences for the MEME defined motifs.

On the other hand these cloned CaM sequences were compared with the published CaM sequences of dicots and monocots. CaM sequences retrieved from NCBI and construct Phylogenetic tree using UPGMA method. These sequences were further subjected to MEME program for motif analysis (Figure 5).

Figure 5: Phylogenetic tree generated from cloned CaM sequences and nine available sequences of CaM retrieved through BLAST search. Multilevel consensus sequences for the MEME defi ned motifs are listed in Table 3.

The translated protein sequence of cloned CaM gene formed cluster with existing sequences of CaM genes of O. sativa, Z. mays, A. thaliana, B. napus and B. oleracea. The phylogenetic tree analysis reveals the different CaMs of dicots and monocots grouped in one cluster and may be more similar during evolution or it may be highly conserved. Motif-1 and Motif-2, functionally related to its calcium binding properties, was most frequently observed in all CaM sequences except sorghum.

Further in silico analysis of CaM sequences reveals variation in EF-hand position when these sequences were subjected to various online bioinformatics tool viz. Pfam (Finn et al., 2006) and SMART (Schultz et al., 1998) as shown in Table 4. The results reveal that EF-hand positions were similar in finger millet genotypes, Oat and Barley though distinctly different from the other cereals and other millets. In all CaM sequences three EF-hand copies were present except Sorghum where only two EF-hand copies observed which might be due to the amplification of the different region by single set of primer.

Table 4: In silico comparative analysis of EF-hand domains in CaM from cereals and millets

The CaM nucleotide sequences were further subjected to online bioinformatics tools PLACE and PLANTCARE software for analyzing different motifs present. In all the CaM sequences, seed specific Skn- 1 motif, CAAT-box present while the GCN-4 motif present only in sorghum and proso millet. Apart from this, the TATA-Box observed in all the CaM sequences of cereals and millets except barley while 5UTR Py-rich stretch region present only in proso millet. The G-Box, I-box, GATA-motif observed in all CaM sequences except rice and proso millet. The GAG-motif absent only in maize while present in remaining CaM sequences. The results shows structural and functional similarity amongst CaM sequences in various cereals and millets in terms of presence and absence of various motifs.

Comparative three dimensional structure analysis of cloned CaMs of cereals and millets

All CaM protein sequences from cereals and millets were subjected to the MOE software to generate three dimensional structures (Figure 6). The protein sequences of CaMs of all the cereals except Sorghum bicolor possess three EF-hands for binding three calcium ions. Presence of two EF hands in Sorghum bicolor might be due to the partial CaM sequence obtained by PCR amplification using same set of primers. Structures of all twelve cloned CaMs were modeled using MOE employing the NMR structure of CaM (PDB accession no. 2F3Y.A) as a template and obtained a refined model after energy minimization. For each CaM molecule eleven structures were generated in the database, out of these which possess maximum score, was selected. The initial energy of the protein was calculated (in kcal/mol) by GizMOE using MMFF94 X force field. The energies of the designed structures were minimized using the energy minimization tool of MOE.

Figure 6: 3D structure of CaM from different cereals and millets using the MOE software. The bound calcium ions were shown in green color.

The minimized structures were finally saved as .pdb files which were validated online by PROCHECK (data not shown). MOE generated structures contained 114, 116, 116, 113, 115, 116, 116, 95, 115, 115, 115 and 115 amino acid residues for Brown, Golden, White, Rice, Maize, Oat, Barley, Sorghum, Little millet, Kodo millet, Proso millet and Barnyard millet respectively. In all cereals and millets the calcium binding loop containing same sequence of amino acid residues which binds with calcium ion except Maize and Rice. The distribution of amino acids in EF-hand, which is involved in binding with calcium, at 1,3,5,7, 9, 12 position are (EFh1 Asp, Asp, Asn, Thr, Asp, Glu, EFh2 - Asp, Asp, Asn, Phe, Ser, Glu, EFh3 - Asp, Asp, Asp, Gln, Asn, Glu) . But in Maize, Proline is present at -X (9^th) position instead of Serine in EFh2. In Rice calcium binding loop has 11 residues in which calcium binds with Asp, Asp, Asn, Asn, Phe and Glu at position 1, 3,5,7,9 and 11th respectively in EF hand1. Sorghum having two EF-hands (EFh2 and EFh3) but the calcium binds with same residues as in others CaM. The interaction energy generated due to the calcium binding is almost similar in all cereals and millets. Brown cultivar of Finger millet has minimum interaction energy among the cereals and millets i.e. 34.2 kcal/mol while golden and white cultivar of finger millet have same interaction energy i.e. 32.0 kcal/mol. Other millets like little, kodo, proso and barnyard have -32.0 kcal/mol, -34.1 kcal/mol, -31.3 kcal/mol and -29.9 kcal/mol interaction energy respectively while other cereals like rice, maize, oat, barley and sorghum have -24.1 kcal/mol, -33.8 kcal/mol, -33.2 kcal/mol, -32.8 kcal/mol and -28.0 kcal/ mol respectively. CaMs in cereals and millet were structurally and functionally similar due to the similar pattern of three dimensional structures.

The Arabidopsis genome is predicted to encode 50 CaM-like proteins (CMLs) with sequence identities ranging from about 20% to 75% at the protein level (Mc-Cormack and Braam 2003; Mc-Cormack et al., 2005). Of the 15 currently known CaM genes in Arabidopsis thaliana, five have 100% similarity to CaM2, and CaM3 and CaM5 calmodulins have 100% sequence identity with that of CaM2 (Zielinski, 2004). In the present report the 100% sequence identity was found in CaMs of finger millet varieties viz. PRM-1 (brown), PRM-701(golden) and PRM-801(white).

A maximum 37 CaMs and related potential calcium sensor proteins (five loci were defined as CaM genes and thirty two additional CaM-like (CML) genes) were identified in the rice genome. Multiple sequence alignment of the rice CaM amino acid sequences with those of typical CaMs from other species, indicates there is high degree of sequence conservation. They have reported that OsCaM1 amino acid sequences are identical to CaMs of barley and wheat (Boonburapong and Buaboocha, 2007). This paper reports, multiple sequence alignment of the finger millet CaM with the CaMs of other cereals and millets, indicates high degree of sequence conservation though few variations were also observed which might be due to the CaM sequences range from 579 to 623bp. The CaM of finger millet genotypes (brown, golden and white) were identical to CaM of barley which reflecting the close relationships among finger millet and barley as evident from the phylogenetic tree construction with other cereal. Phylogenetic tree result outlines the development of CaM in finger millet and other cereals and millets which indicate that CaM is strictly conserved and has same evolutionary pattern. The availability of genetic map of E. coracana (Dida et al., 2007) could be used to investigate the diversity of CaM genes and its possible locations to specific chromosomes in near future. In order to understand the molecular mechanisms associated with calcium accumulation in finger millet, there is also need to identify the CaM isoforms, gene copy number, expression analysis and downstream interaction with target proteins to investigate the genetic mechanism of differential calcium accumulation.

The PCR based amplification of CaM of different cereals and millets using same primer set designed from finger millet was investigated in silico after sequencing of respective PCR amplicons. The homology search, multiple sequence alignment, EF-hand copies, phylogenetic tree construction and motif analysis has clearly revealed the identity of these sequences as CaM of respective cereals and millets. Multiple sequence alignment reveals high degree of sequence conservation in CaM of the cereals and millets though there were some alterations which might be due to the partial sequence ranged from 579 to 623 bp of the CaM obtained in the cereals and millets. The phylogenetic analysis reveals the CaM enjoy common phylogeny in different cereals and millets as well as in dicots viz. Arabidopsis thaliana, Brassica spp. The in silico 3D-structural analysis of cloned sequences showed similar structures and reveals high degree of conserved CaM in cereals and millets. The paper reports the evolutionary colinearity of CaM in the cereals and millets though the CaM of Eleusine coracana and Hordeum vulgare having closed evolutionary relationship as compare to others. The CaM gene of E. coracana is first report based on in silico studies though it has already been reported in Oryza sativa, T. aestivum, H. vulgare and Z. mays. Further research is in progress to find out the isoforms of CaM genes in brown, golden and white varieties of E. coracana to elucidate its role in signal transduction transport and physiological function including its interaction with target proteins.

The authors acknowledge the facilities provided by DBT-funded SUB-DIC, Bioinformatics Centre, Pantnagar for carrying out in silico investigation. The support provided by Director, Experiment Station, G.B. Pant University of Agriculture and Technology, Pantnagar is also thankfully acknowledged.