Further Characterization of Calcineurin B-Like Protein and Its Interacting Partner CBL-Interacting Protein Kinase from Pisum sativum

Abstract

The recently discovered calcium sensor calcineurin B-like proteins (CBLs), and their interacting partners CBL-interacting protein kinases (CIPKs), have emerged as a key network in response to stress and calcium signaling in plants. The studies on CBL and CIPK are so far mainly restricted to Arabidopsis and work on dissecting this pathway in higher plants is inadequate. Our recent studies revealed that calcium sensor CBL from pea gets phosphorylated by pea CIPK. Furthermore, the transcript levels of both the CBL and CIPK from pea were coordinately upregulated in response to various stresses including high salinity, cold, wounding, salicyclic acid and calcium but not to abscisic acid and dehydration. Here we report the results on the computational analysis of EF hands of PsCBL protein, which indicate that it contains all the functional domains required for calcium binding activity. We have also focused on homology based computational modeling of PsCBL and PsCIPK proteins using AtCBL2 and Chk1 as templates respectively, which suggested the high degree of conservation between AtCBL2, PsCBL and OsCBL and between Chk1, PsCIPK and AtCIPK18. Furthermore, the direct interaction of PsCBL and PsCIPK mutant proteins was studied by the yeast 2-hybrid system, which confirmed that FISL domain is an important interaction module required for PsCBL and PsCIPK interaction.

Introduction

Calcium ions (Ca²⁺) function as major secondary-messenger signaling molecule, which plays a fundamental role in plant growth and development under normal as well as stress conditions.^1,2,3 Almost all extracellular signals including plant hormones, light, stress factors, and pathogenic or symbiotic elicitors elicit changes in cellular Ca²⁺ concentration.^3–6 Recently, it is reported that internal concentration of Ca²⁺ in plant cells is also constantly being actively revised.⁷ Relatively, recently discovered calcium sensor calcineurin B-like proteins (CBLs), and their interacting partners CBL-interacting protein kinases (CIPKs), have emerged as key network that plays an important role in plants in response to calcium and stress signaling.^2,3,6,8–10 CIPK initiates a phosphorylation cascade and can regulate the expression of major genes. Though CBL-CIPK network is also widely distributed among higher plants but except Arabidopsis, the complexity and characterization of this pathway remains largely unrevealed. In our previous study¹¹ we have isolated and characterized two genes PsCBL and PsCIPK from Pisum sativum and shown that these two genes are coordinately upregulated in response to various stresses including high salinity, cold, wounding, salicyclic acid and calcium but not to abscisic acid and dehydration. Interestingly, PsCIPK was found to interact and phosphorylates PsCBL. Here we show some of the important aspect of PsCBL and PsCIPK protein, which includes the computational analysis of EF hands of PsCBL protein for helix and coil analysis. We also focused on homology based modeling of PsCBL and OsCBL using AtCBL2 structure as a template as well as homology based modeling of PsCIPK and AtCIPK18 using Chk1 structure as template. Furthermore, T184D and deletion FISL mutants of PsCIPK were constructed and the interaction of PsCBL with PsCIPK mutants, were checked in vivo by yeast 2-hybrid system.

Computational Analysis of EF Hands of PsCBL Protein for Helix and Coil Analysis

The helix/coil tendency of EF-hand stretches was determined by computational analysis using a Chou and Fasman algorithm of ExPASy server (www.expasy.org). EF1 was analyzed, first for its helix motifs. The graphical analysis of EF1 depicts an upward peak for its first 10 amino acids (Fig. 1A) and 20–30 amino acids with an intervening deep peak for 10–20 amino acids, which strongly indicates the high tendency of amino acids (1–10 and 20–30) to form α-helix and a coiled loop region in between. When the same EF1 was viewed for its coil-forming propensity (Fig. 1B), amino acids 10–20, which showed a deep peak in case of helix analysis, showed a sharp upward peak in case of coil/loop analysis, indicating the high tendency for these amino acids to form a loop structure. Thus EF1 had the necessary helix-loop-helix motifs, which is the characteristic feature of EF hand. The same was also true for EF hand 2, 3 and 4 (Fig. 1A–D). Thus PsCBL had all the functional domains, which is a prerequisite for calcium binding activity for this protein.

Figure 1.

Computational analysis of EF hand motifs for helix and coil formation. (A–D) Prot scale study of the EF hands of PsCBL for helix and coil analysis. Computation was made according to Chou and Fasman algorithm. (A) The peaks above the mid line of graph represents the helical domains of EF1 and 2. (B) The same sequence was viewed for the tendency of amino acid residues to form coil (loop). Note that the residues which showed a downward peak when viewed for helix (graph 1) showed a high peak when viewed for coil analysis (B) indicating that these amino acids have a high coil forming tendency. (C and D) depicts EF hand 3 and 4 for their helix (C) and coil forming tendency (D).

Homology Based Modeling of PsCBL and OsCBL Using AtCBL2 Structure as a Template

The crystal structure of AtCBL2 has been determined at 2.1 A resolution (Nagae et al, 2003). PsCBL shares 89% identity at amino acid level with AtCBL2, therefore AtCBL2 was used as a template for modeling PsCBL as well as OsCBL, (92% identical with PsCBL). Structural predictions were made using Swiss-model server and visualized using Rasmol software. Figure 2A is the ribbon diagram of AtCBL2 and Figure 2B and C are the structural predictions of PsCBL and OsCBL, respectively. The polypeptide chain of all these predicted protein structures are folded into two globular domains (N and C terminal) composed of α-helical structure with 9 α-helices (shown in pink) and four short β-strands (in yellow). The loop region between the two α-helices is the calcium binding site and the β-strands stabilize the structure of protein. High degree of conservation is seen between the protein structures of AtCBL, PsCBL and OsCBL. There was no significant difference in the overall topology and protein folding, when these structures were overlapped. Figure 2D and E are the overlapped structures of AtCBL2 with PsCBL, and PsCBL with OsCBL, respectively.

Figure 2.

Homology modeling of PsCBL and OsCBL using AtCBL2 structure as a template. Structural predictions were made using Swiss-model server and visualized using Rasmol software. (A) Ribbon diagram of AtCBL2 (PDB ID code 1UHN). (B and C) Three dimensional structure of PsCBL and OsCBL obtained from Swiss-model. (D and E) Overlapping images of AtCBL2 and PsCBL, and PsCBL and OsCBL respectively. Overall a 2-domain structure, N and C terminal is seen. The alpha-helices are shown in pink, beta-sheets in yellow and the linker connecting the two domains in green.

Therefore, the homology based modeling represents high degree of conservation between AtCBL, PsCBL and OsCBL. This also reflects that calcium-binding proteins perform important function and necessitates the conservation of their overall 3 dimensional structures in different species. The number of calcium ions coordinated depends on the sequence of the protein and structural characteristics favor the coordination of calcium ions in particular EF hand motifs. AtCBL2 for example has two-calcium loaded EF hands, EF1 and EF4, whereas the second and third EF hand motifs are maintained in the open form by internal hydrogen bonding without coordination of calcium ions.¹² The amino acid sequences of all the EF hands of PsCBL exactly match with AtCBL2. Therefore, it is expected that PsCBL may also coordinate calcium ions in the second and fourth EF hand as seen for AtCBL2. However, this statement still awaits experimental validation.

Homology Based Modeling of PsCIPK and AtCIPK18 Using Chk1 Structure as Template

Check point kinase (Chk1) is a serine-threonine kinase that plays an important role in the DNA damage response, including G₂/M cell cycle control. The crystal structure of the Chk1 kinase domain has been determined.¹³ Part of CIPKs is similar to Chk1 and this similarity is interesting as sos2 mutants show cell cycle defect at the root meristem in the presence of Na⁺ stress.¹⁴ However, there has been no further report unraveling the fundamental behind this observation. Nevertheless, AtCIPKs are characterized by their unique NAF motif, required for CBL interaction. When PsCIPK sequence was submitted to Swiss-model server, predictions were made using Chk1 as a template. Structural prediction for only 275 amino acid residues of PsCIPK from the N terminal could be possible. Figure 3A is the ribbon diagram of Chk1 kinase domain. Chk1 kinase domain is composed of two subdomains, an 88-residue N-terminal domain and a 210-residue C-terminal domain. The N-terminal domain is composed largely of β-structure, the C-terminal domain is largely helical, and are connected by an extended linker region. The catalytic site is at the junction between these two domains.

Figure 3.

Homology modeling of PsCIPK and AtCIPK18 using Chk1 structure as a template. Structural predictions were made using Swiss-model server and visualized using Rasmol software. (A) Ribbon diagram of Chk1 kinase domain (PDB ID code 1NVS). (B and C) Three dimensional structure of PsCIPK and AtCIPK18 kinase domain obtained from Swiss-model. (D and E) Overlapping images of Chk1 and PsCIPK, and PsCIPK and AtCIPK18 respectively. The N-terminal domain is in yellow, the C-terminal domain is in pink, and the activation loop is in green.

The structural predictions of PsCIPK and AtCIPK18 were made using Chk1 structure as a template. Figure 3B and C are the predicted 3D structures of PsCIPK and AtCIPK18 respectively. High degree of conservation is seen between Chk1, PsCIPK and AtCIPK18 for the part for which the structure is known (till 275 amino acid residues of PsCIPK). These proteins were almost identical and there was no significant difference in the overall topology and protein conformation, when these structures were overlapped. Figure 3D and E are the overlapped structures of Chk1 with PsCIPK, and PsCIPK with AtCIPK18 respectively. The β-structure N-terminal domain is in yellow, largely helical C-terminal domain in pink and the activation loop is in green. The activation loop in all these proteins was between the conserved amino acids APE and DFG. The activation loop of Chk1 is stable and folded into an open and active-like conformation. On the basis of predicted model the activation loop of PsCIPK and AtCIPK is also folded into an open and active-like conformation in the absence of phosphorylation or ligand.

Direct Interaction of PsCBL with PsCIPK Mutants Via Yeast Two-Hybrid System

Further, T184D and deletion FISL mutants of PsCIPK were constructed. The interaction of PsCBL with PsCIPK mutants, were checked in vivo by yeast 2-hybrid system. Figure 4A is the template for panel B to D showing the clones streaked: Yeast (AH109) cells containing cotransformants of AD-PsCIPK (cloned under EcoR I and Xho I sites) and BD-PsCBL (cloned under Nco I and EcoR I), clone 1 (the positive control for this experiment), 2. cotransformants of AD-PsCIPK T184D plus BD-PsCBL, 3 AD-PsCIPKT184D and BD vector alone, 4. cotransformants of AD-PsCIPKΔFISL and BD-CBL, 5. cotransformants of AD-PsCIPKΔFISL and BD vector alone, 6. cotransformant of AD-PsCIPK T184DΔFISL and BD-CBL, 7. cotransformants of AD-PsCIPK double mutant and BD vector alone, 8. cotransformant of BD-CBL and AD vector alone, 9. cotransformants of AD-CIPK and BD vector. All these transformants grew well on the on SD-Leu^-Trp^- medium (2 drop out medium) (Fig. 4B). In a selection medium lacking Leu, Trp and His (SD-Leu^-Trp^-His^-+15 mM 3-AT) only selected clones of cotransformants (BD-PsCBL plus AD-PsCIPK and the AD-PsCIPKT184D plus BD-CBL), in which the HIS3 gene was transactivated, showed growth (Fig. 4C). The results from β-galactosidase filter lift assay for colonies of cotransformants giving blue color further confirmed the interaction between PsCBL and PsCIPK (full-length) and between PsCBL and PsCIPKT184D proteins (Fig. 4D, panel 1 and 2 respectively). This result shows that PsCIPK mutant with the deletion in its auto-inhibitory (FISL/NAF) motif, failed to interact with PsCBL thus confirming the authenticity of these proteins and revealing the significance of FISL domain as an important interaction module required for PsCBL and PsCIPK interaction.

Figure 4.

Interaction of PsCIPK mutants with PsCBL. (A) is a template for (B–D) showing the clones streaked: (B) Transformants on 2 drop-out media (-Leu,- Trp), (C), Phenotype on synthetic dextrose lacking Leucine, Trytophan, and Histidine containing 15 mM 3-AT (SD-Leu-Trp-His+3AT) here growth represent the interaction of full-length PsCIPK with PsCBL (clone 1) and T-D PsCIPK with PsCBL (clone 2) because of the activation of HIS3 reporter gene. PsCIPK with deletion in FISL domain and the double mutant failed to interact with PsCBL. (D) The results from b galactosidase filter lift-assay of colonies of cotransformants (AD-PsCIPK plus PsCBL-BD) and (TD PsCIPK-AD and PsCBL-BD) giving blue color further confirmed the interaction.

Conclusion and Perspectives

The essential role imparted by CBL-CIPK genes in stress tolerance, necessitates their detailed characterization from higher plants. Since almost all signal transduction processes in plants involve Ca²⁺ that serves as a vital second messenger, therefore understanding how Ca²⁺ mediates the cellular responses triggered by environmental signals is one of the most important goals for plant biologists in years to come. Our recent studies uncovered a new family of Ca²⁺ sensor that target a protein kinase (CIPK), establishing a novel paradigm for Ca²⁺ signaling in higher plants. Understanding the regulation of CBL and CIPK interacting genes will provide insight into the regulatory mechanisms of calcium signaling in general. How cells distinguish the Ca²⁺ signals produced by different kinds of stimuli is a challenge for us. As different isoforms of CBL protein show high identity at amino acid level but still behave differentially, so the work on UTR regions and promoter region of these genes requires a comprehensive analysis. The similarity of the three dimensional structure of PsCIPK with Chk1 may open an altogether new field and vision for dissecting the function of this kinase. All the above mentioned work and future directions should provide further insights into regulation of PsCBL and PsCIPK proteins and strengthens the basis of dissecting this pathway in higher plants. These studies will contribute to the understanding of signal transduction not only in plants but also in other organisms because calcium serves as a second messenger in all eukaryotes.

Addendum to: Mahajan S, Sopory SK, Tuteja N. Cloning and Characterization of CBL-CIPK Signaling Components from a Legume Pisum sativum. FEBS J. 2006;273:907–925. doi: 10.1111/j.1742-4658.2006.05111.x.