Dissection of Mechanisms Involved in the Regulation of Plasmodium falciparum Calcium-dependent Protein Kinase 4^,

Abstract

Recent studies have demonstrated that calcium-dependent protein kinases (CDPKs) are used by calcium to regulate a variety of biological processes in the malaria parasite Plasmodium. CDPK4 has emerged as an important enzyme for parasite development, because its gene disruption in rodent parasite Plasmodium berghei causes major defects in sexual differentiation of the parasite (Billker, O., Dechamps, S., Tewari, R., Wenig, G., Franke-Fayard, B., and Brinkmann, V. (2004) Cell 117,503 -514). Despite these findings, it is not very clear how PfCDPK4 or any other PfCDPK is regulated by calcium at the molecular level. We report the biochemical characterization and elucidation of molecular mechanisms involved in the regulation of PfCDPK4. PfCDPK4 was detected on gametocyte periphery, and its activity in the parasite was regulated by phospholipase C. Even though the Junction Domain (JD) of PfCDPK4 shares moderate sequence homology with that of the plant CDPKs, it plays a pivotal role in PfCDPK4 regulation as previously reported for some plant CDPKs. The regions of the J-domain involved in interaction with both the kinase domain and the calmodulin-like domain were mapped. We propose a model for PfCDPK regulation by calcium, which may also prove useful for design of inhibitors against PfCDPK4 and other members of the PfCDPK family.

Plasmodium falciparum is an intracellular protozoan parasite, which propagates in human erythrocyte and hepatocytes. It causes millions of deaths annually, which occur mostly in third world countries. It is very important to understand the molecular events involved in parasite development and growth in detail to develop effective drugs against this disease. In the human host, the parasite from an infected mosquito bite enters the bloodstream as a sporozoite. The sporozoite migrates and resides in the hepatocytes and replicates to yield merozoites. Upon rupture of the host cells, released merozoites invade red blood cells (RBCs),3 and the asexual development takes place inside the RBC resulting in the formation of ring and trophozoite stages. During the schizont stage, the parasite undergoes nuclear division resulting in the formation of ∼24 merozoites, which infect fresh RBCs upon release. Some parasites commit to sexual differentiation resulting in the formation of male and female gametocytes. Upon ingestion into the mosquito gut, gametocytes are activated in response to environmental cues like lower temperature and xantheurenic acid (2, 3). The male gametocyte undergoes endomitotic division followed by exflagellation. After fertilization, the zygote gives rise to a motile ookinete. Subsequently, an oocyst is formed that attaches to the outside wall of the gut.

It is clear that calcium controls a wide variety of processes in the parasite (4, 5), like invasion (6-8), migration (9), gametogenesis (1, 10), and circadian rhythms (11). Detailed understanding of calcium-mediated signaling pathways could provide insights into novel molecular mechanisms involved in parasite development. For instance, calcium regulation of a protein kinase B-like enzyme (12) via calmodulin seems to be important for erythrocyte invasion (8). Plasmodium contains calcium-dependent protein kinases (CDPKs) (13), which have been found only in plants and some protists but are absent from almost all metazoans (14). These enzymes have a catalytic domain at their N terminus and a C-terminal calmodulin-like domain (CLD) composed of four calcium-binding EF-hand motifs (14), a junction domain (JD) separates the catalytic domain from the CLD. Biochemical studies done on plant CDPKs have revealed that the JD regulates CDPK activity by interacting with both the kinase domain and the CLD (14, 15). Work reported here reveals that, despite modest homology between the amino acid sequences of JD of plant CDPKs (like AtCPK1 and Soybean CDPKα) and PfCDPK4, there are some similarities between the mechanism of regulation between these kinases.

Plasmodium has at least five CDPKs (13), and the importance of CDPKs in Plasmodium signaling has been highlighted by recent gene disruption studies. PfCDPK1 seems to be an essential gene as its disruption has not been possible in P. falciparum (7). Studies performed using an inhibitor against this kinase suggest that it may be important for RBC invasion and egress (7). Reverse genetic studies done in P. berghei suggest that CDPK3 may play an important role in ookinete gliding motility and invasion (9, 16). The gene disruption of CDPK4 in P. berghei caused severe defects in sexual reproduction and mosquito transmission (1). These studies highlighted the importance of CDPK4 in Plasmodium biology and suggested that it may serve as a target for transmission-blocking drugs. The biochemical mechanisms involved in the regulation of this enzyme have remained unknown. In this study, we have explored the molecular and cellular mechanisms involved in the regulation of PfCDPK4.

EXPERIMENTAL PROCEDURES

P. falciparum Cultures—P. falciparum 3D7 strain was cultured at 37 °C in RPMI 1640 medium, supplemented either with 0.5% Albumax II (Invitrogen) or 10% AB+ human serum as described earlier (12, 17). For sexual stage studies, 3D7A, a variant of P. falciparum strain 3D7, was used to obtain gametocyte-enriched culture as described previously (18).

Molecular Cloning and Site-directed Mutagenesis of PfCDPK4—Initially, PfCDPK4 gene sequence was obtained by using either TgCDPK1 or the published sequence of other CDPKs to BLAST search the P. falciparum genome sequence. Subsequently, PlasmoDb annotation (19) appeared in the public domain, and the gene sequence PF07_0072, which matched the PfCDPK4 sequence obtained by our studies. For PCR amplification, primers based on the nucleotide sequence of the PfCDPK4 gene were used. Total RNA from asynchronous P. falciparum cultures was used for reverse transcription (RT) along with random hexamers provided with the RT-PCR kit (Invitrogen). Both cDNA or genomic DNA were used as template for PCR. The reaction was carried out using Hi-fi Platinum Taq polymerase (Invitrogen) with the following cycling parameters: 94 °C for 2 min initial denaturation followed by 30 cycles at 94 °C for 30 s, 45 °C for 30 s, 68 °C for 2 min, and final extension at 72 °C for 10 min. Following sets of primers were used for amplification of the full-length gene PfCDPK4 forward (CDPK4F): 5′-ATGGGACAAGAGGTATCGAGTGTTAACAA-3′ and PfCDPK4 reverse (CDPK4R): 5′-TTAATAATTACAAAGTTTGACTAGCATAT-3′. PCR products were cloned in pGEM-T easy vector (Promega, Madison, WI), and the sequence for the cloned PfCDPK4 gene was obtained by automated DNA sequencing. For cloning in the expression vector pGEX4T1, PfCDPK4 or its variants were amplified using primers with overhangs containing restriction sites for SmaI and XhoI. All site-directed mutagenesis studies were performed using the QuikChange kit (Stratagene) following the standard protocol described by the manufacturer. Primer sets used for making mutants are provided in the supplemental materials.

Expression and Purification of Recombinant Proteins—Plasmid DNA was transformed in Escherichia coli BL21-RIL (Stratagene) strain for the expression of GST-PfCDPK4 and its mutants. Protein expression was induced by overnight incubation of cells with 0.1 mm isopropyl 1-thio-β-d-galactopyranoside at 18-20 °C. Subsequently, cell pellets were suspended in ice-cold lysis buffer, containing 50 mm Tris, pH 7.4, 2 mm EDTA, 1 mm dithiothreitol, 1% Triton X-100, and proteases inhibitors (1 mm phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml pepstatin), and sonication was performed for 6 cycles of 1 min each. The resulting cell debris was removed by centrifugation at 20,000 × g for 40 min at 4 °C. Fusion proteins from the cell lysates were affinity-purified using glutathione-Sepharose resin as described previously (20). Briefly, the resin was washed with lysis buffer, and bound proteins were eluted with 50 mm Tris, pH 8.0, with 10 mm glutathione. Finally, purified proteins were dialyzed against 50 mm Tris, pH 7.4, 1 mm dithiothreitol, and 10% glycerol. Protein concentration was determined by densitometry analysis of Coomassie-stained gels.

For expression of PfCDPK1, PfCDPK1 (PFB0815w) was amplified from the asexual stage cDNA of P. falciparum and cloned in BamHI and XhoI pET-28a vector (see Table S1 in supplemental material for primer sequences) to facilitate its expression as a His₆-tagged protein. Briefly, plasmid transformed BL21(DE3)RIL strain of Escherichia coli was grown in LB media containing 50 μg/ml kanamycin and 35 μg/ml chloramphenicol, and protein expression was induced using 1 mm isopropyl 1-thio-β-d-galactopyranoside at 18 °C for 16 h. Cells were harvested by centrifugation at 6000 rpm for 10 min at 4 °C, resuspended in lysis buffer (50 mm potassium phosphate, pH 7.4, 150 mm NaCl, 0.1% Nonidet P-40, and 1 mm dithiothreitol), sonicated for 8 cycles of 1 min each. The soluble protein was incubated with nickel-nitrilotriacetic acid-agarose (Qiagen) with end-to-end shaking for 6 h at 4 °C, and the protein was eluted with 50 mm potassium phosphate, pH 8.0, 500 mm NaCl, 0.1% Nonidet P-40, and 1 mm dithiothreitol containing 300 mm imidazole and dialyzed against 50 mm potassium phosphate, pH 7.4, 1 mm dithiothreitol, and 10% glycerol.

Assay of PfCDPK4 Activity—The catalytic activity of PfCDPK4, PfCDPK1, or their variants was assayed in a buffer containing 50 mm Tris, pH 7.5, 10 mm magnesium chloride, 1 mm dithiothreitol, and 100 μm [γ-³²P]ATP (6000 Ci/mmol). Either 6 μg of myelin basic protein (MBP) or 150 μm Syntide-2 (PLARTLSVAGLPGKK, custom synthesized by Peptron Inc., South Korea) was used as phosphate-acceptor substrate. Reactions were performed in the presence of 2 mm calcium chloride or 2 mm EGTA (0 mm Ca²⁺) for 40 min at 30 °C. When MBP was used as the substrate, reactions were stopped by boiling the assay mix for 5 min followed by SDS-PAGE. Phosphate incorporation was adjudged by autoradiography of SDS-PAGE gels. When Syntide-2 was used as substrate, reactions were stopped by spotting the reaction mix on P81 phosphocellulose paper (Millipore), followed by washing of the paper strips with 75 mm ortho-phosphoric acid. Phosphate incorporation was assessed by scintillation counting of P81 paper. In PfCDPK4 inhibition assays, peptide inhibitors were preincubated with proteins in a kinase assay buffer at 25 °C for 30-60 min prior to the addition of substrate and ATP.

Generation of Anti-PfCDPK4 Serum, Immunoblotting, Immunoprecipitation, and Immunofluorescence—A synthetic peptide (KMMTSKDNLNIDIPS) based on PfCDPK4 sequence was custom synthesized (Peptron Inc.), conjugated to keyhole limpet hemocyanin via an additional N terminus cysteine residue, and was used to raise antisera against PfCDPK4. Cell-free protein extracts were prepared from specific parasite stages as described previously (8). After separation on 10% SDS-PAGE gel, lysate proteins were transferred to a nitrocellulose membrane. Immunoblotting was performed using anti-PfCDPK4 antisera, and blots were developed using West-pico chemiluminescence (Pierce) reagent following the manufacturer's instructions. For immunofluorescence assays, thin blood smears of parasite cultures were fixed with cold methanol, and a previously published protocol was followed (20). The microscopy was performed on a Zeiss Axio Imager fluorescence microscope, and images were processed using the AxioVision software.

Inhibitor Treatment and Immunoprecipitation—Gametocytes were distributed in six-well culture dishes and treated with various inhibitors for 30 min. Subsequently, protein lysates were prepared in a buffer containing 10 mm Tris, pH 7.4, 100 mm NaCl, 5 mm EDTA, 1% Triton X-100, phosphatase inhibitors (20 μm NaF, 20 μm β-glycerophosphate, and 100 μm sodium vanadate) and protease inhibitor mixture (Roche Applied Science). 100 μg of soluble lysate protein was incubated with PfCDPK4 anti-sera for 6 h at 4 °C. Subsequently, Protein A+G-Sepharose (Amersham Biosciences) was added to the antibody-protein complex and incubated on a end-to-end shaker for 2 h. After washing with phosphate-buffered saline-Sepharose, beads were suspended in the lysis buffer. PfCDPK4-IP was used for kinase assays as described above for the recombinant protein.

Homology Modeling—The CLD-J domain shares ∼51% similarity with the CDPK from Arabidopsis thaliana AtCPK-1. The homology model of CLD-JD was determined using Swiss Model from EMBL. The template model used was CLD-JD of AtCPK-1, which was crystallized as a dimer. The J-domain helices from the two monomers were swapped with each other in this structure (21). Therefore, the initial homology model generated for the complementary CLD-J domain for PfCDPK4 was also a dimer. To understand the interaction of this helix (Gln³⁵⁸-Lys³⁷¹) with CLD of the monomer, this helix was rotated and translated keeping residues 372-375 as the flexible linker region and superimposed on to the helix from the other monomer, which resulted in the initial model for the CLD-J domain monomer. Initially, these flexible linker residues (372-375) were locally minimized using COOT (22), and the overall structure was refined with slow cooling using annealing of CNS (23) to remove all the short contacts. Finally, the model quality was checked with the Procheck software (24).

CD—CD spectra were obtained on a Jasco J-710 spectropolarimeter with a constant dispersion of 1 nm. Spectra were measured with a time constant of 1s, scan speed of 100 nm/min. Signals were averaged 10 times before measurement. The peptide concentration was between 185 and 350 μm, and a 0.1-cm path length cell was used.

RESULTS

Molecular Cloning of PfCDPK4 Gene—Gene-specific primers were used to amplify PfCDPK4 from genomic DNA and cDNA. The larger size of the PCR product obtained from genomic PCR suggested the presence of introns (data not shown) in PfCDPK4. Sequencing of RT-PCR product confirmed that PfCDPK4 contains an intron of 347 nucleotides (Fig. 1A), which was similar to the predictions made by PlasmoDb. The deduced amino acid sequence suggested that PfCDPK4 possesses a calmodulin-like domain at the C terminus and an N-terminal serine/threonine kinase domain, which are characteristic of CDPKs (14). PfCDPK4 catalytic domain possesses all 11 sub-domains that are representative of most protein kinases (supplemental Fig. S1). A glycine residue at the second position suggests that it has a putative myristoylation signal, a similar myristoylation signal in PfCDPK1 is important for its membrane targeting (25). The CLD of PfCDPK4 consists of an N and C lobe and each lobe comprises of two EF-hand motifs (Fig. 1B). A ∼34 amino acid (aa) junction domain links the CLD and the kinase domain.

FIGURE 1.

A, gene structure of PfCDPK4. PfCDPK4 gene was amplified by RT-PCR using specific primers. DNA sequencing revealed that it has a two-exon structure. B, the domain architecture of PfCDPK4. Deduced amino acid sequence of PfCDPK4 (supplementary Fig. S1) indicated that it possesses a CLD at the C terminus, which has four EF-hand motifs. The kinase domain is present near the N terminus and a small JD separates the two domains. A putative myristoylation signal present at its N-terminal end is highlighted.

PfCDPK4 Is Expressed in Sexual Stages of the Parasite Life Cycle—Even though previously published transcriptome in PlasmoDb and proteome (26) analyses suggested gametocyte-specific expression of PfCDPK4, localization of CDPK4 in the parasite had remained unknown. PfCDPK4-specific antisera were raised against a synthetic peptide corresponding to a unique motif in the J-domain, and Western blots were performed using protein lysates from different asexual stages as well as the gametocytes. A band corresponding to ∼60 kDa, which was consistent with the predicted molecular mass of PfCDPK4, was observed only in the gametocyte lysates (Fig. 2A). These data confirmed that PfCDPK4 is expressed mainly in the sexual stages. To further confirm this, immunofluorescence studies were performed, which revealed PfCDPK4 staining mainly on gametocyte periphery (Fig. 2B). Control assays performed with pre-immune sera did not show any staining (Fig. 2A, left panel). It is likely that the presence of a myristoylation signal in this kinase (Fig. 1B) is responsible for targeting this kinase to the parasite periphery.

FIGURE 2.

Stage-specific expression and localization of PfCDPK4. A, Western blot was performed using PfCDPK4 antisera and protein lysates from different parasitic stages: R, ring; T, trophozoite; S, schizont; and G, gametocyte. A ∼60-kDa band corresponding to the predicted size of PfCDPK4 was observed mainly in the gametocyte stages. Antiserum prepared from preimmune bleeds was used as a control (left panel). B, immunofluorescence assays were performed on gametocyte smears using anti-PfCDPK4 antisera and Alexa-594-anti rabbit IgG. Parasite nucleus was stained with Hoechst 33224 (blue).

Calcium Stimulates Autophosphorylation and Catalytic Activity of Recombinant PfCDPK4—To characterize PfCDPK4, it was expressed as a GST fusion protein in E. coli. The recombinant PfCDPK4 was active only in the presence of calcium as adjudged by its ability to phosphorylate MBP (Fig. 3A). Recombinant PfCDPK4 also exhibited calcium-dependent autophosphorylation (Fig. 3A). A small peptide, syntide-2, which has been used as an in vitro substrate for several CDPKs (15, 27, 28) was effectively phosphorylated by PfCDPK4 in a calcium-dependent manner (Fig. 3B).

FIGURE 3.

Recombinant PfCDPK4 is autophosphorylated and activated by calcium. A, PfCDPK4 was expressed as a GST fusion protein and purified by affinity chromatography. Protein kinase assays were performed using recombinant GST-PfCDPK4 in the presence of 2 mm CaCl₂ or 2 mm EGTA (absence of calcium) with MBP as phosphor-acceptor substrate. Subsequently, the kinase assay mix was electrophoresed on a SDS-PAGE gel, and phosphorimaging was performed. PfCDPK4 phosphorylates MBP only in the presence of calcium (lane 1). PfCDPK4 also exhibited calcium-dependent autophosphorylation. B, kinase assay was performed as described in panel A except syntide was used as a substrate instead of MBP. PfCDPK4 effectively phosphorylated syntide in the presence of calcium. The activity is measured in nanomoles/mg/min.

Phospholipase C Acts as an Upstream Regulator of PfCDPK4 in the Parasite—It has been demonstrated that the release of calcium from intracellular stores in Plasmodium is controlled by PLC (4, 29), which is used by the parasite for various purposes. A strong correlation between increase in the levels of PLC hydrolysis product, inositol 1,4,5-trisphosphate, and gametogenesis has also been reported (30), which is suggestive of a role for PLC in this important parasitic process. Given the dependence of recombinant PfCDPK4 on calcium and its role in gamete formation, it was worth exploring whether PLC regulates its activity in the parasite. Gametocytes were treated with PLC inhibitor, which have been used successfully in Plasmodium (8, 12, 29), and the activity of immunoprecipitated PfCDPK4 was assayed. In one of the experiments, an intracellular chelator, 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis (acetoxymethyl ester), was included (Fig. 4A). This inhibitor caused a significant decrease in PfCDPK4 activity, which emphasized the importance of intracellular calcium on PfCDPK4 activation. Treatment with the PLC inhibitor, U73122, caused a significant decrease in PfCDPK4 kinase activity. In contrast, U73343, the inactive analogue of this inhibitor, failed to alter the activity of this kinase. The levels of PfCDPK4 did not change significantly upon inhibitor treatment (Fig. 4B). These data suggest that PLC acts as a regulator of PfCDPK4, which is most likely a result of its ability to control levels of free calcium in the parasite.

FIGURE 4.

Phospholipase C acts as an upstream regulator of PfCDPK4. A, gametocytes were treated with DMSO, 30 μm U73122 or U73343, 100 μm 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis (acetoxymethyl ester) (BAPTA-AM). Subsequently, PfCDPK4 was immunoprecipitated from protein lysates (see “Experimental Procedures”) and PfCDPK4-IP associated kinase activity was assayed using syntide as substrate. B, immunoblotting was performed on protein lysates from the experiments described in panel A using anti-PfCDPK4 antisera.

J-domain Is Responsible for PfCDPK4 Activation—Deletion and truncation mutants of PfCDPK4 domains were created to understand its regulation (Fig. 5A), and the activity of recombinant mutant proteins was determined by performing in vitro kinase assays. To evaluate the role of the J-domain in the activation of PfCDPK4 by calcium, a deletion mutant, ΔJ, lacking the majority of this domain was generated (Fig. 5A). Unlike the wild-type PfCDPK4, ΔJ exhibited significant catalytic activity even in the absence of calcium (Fig. 5B). Because the level of activity of this mutant did not change in the presence of calcium, it is reasonable to propose that the JD negatively regulates PfCDPK4 activity in a calcium-dependent manner (Fig. 5B). A similar effect of JD deletion was also observed on the autophosphorylation of PfCDPK4. Although the wild-type enzyme exhibited autophosphorylation only in presence of calcium, the ΔJ mutant was autophosphorylated in both the presence or absence of calcium (Fig. 5B, lower panel), which fits in well with the activity data.

FIGURE 5.

A, schematic diagram illustrating PfCDPK4 mutants created for biochemical studies. Truncation mutants (T) were named based on the number of their C-terminal amino acid. All mutants were expressed as GST fusion proteins as described for wild-type PfCDPK4. “ΔJ” is a deletion mutant of PfCDPK4 that lacks the J-domain (aa 350-379). B, the kinase activity of equal amounts of PfCDPK4 or ΔJ mutant was assayed in the presence or absence of calcium as described below for truncation mutants. Lower panel, a kinase assay was performed for assessing the autophosphorylation of PfCDPK4 (lanes 1 and 2) or ΔJ (lanes 3 and 4) in the presence (lanes 1 and 3) or absence (lanes 2 and 4) of calcium followed by the autoradiography (upper panel) of the SDS-PAGE gel (lower panel). C, equal amounts of recombinant PfCDPK4 or its truncation mutants were incubated with syntide in a kinase assay mix containing either 2 mm CaCl₂ or 2 mm EGTA (0 mm Ca²⁺). The kinase activity was determined by measuring phosphate incorporation in syntide as described under “Experimental Procedures.” A representative of more than three independent experiments is shown, and error bars represent ±S.E. between replicates of the same experiment.

The JD of plant CDPKs like AtCPK1 and GmCDPK controls their activation (14, 15) and shares very similar amino acid sequence (Fig. 7A). In comparison, PfCDPK4 exhibits an only average sequence homology (∼50%) with plant CDPKs. To understand PfCDPK4 regulation in detail, biochemical characterization of this enzyme was needed. Firstly, truncation mutants lacking the CLD and the portions of the J-domain were generated (Fig. 5A) to identify the location of regulatory elements on this domain. A truncation mutant, T369, lacking the CLD, failed to exhibit kinase activity either in the presence or in the absence of calcium (Fig. 5C). A further deletion of the preceding 10 aa (T359) also rendered an almost inactive form of PfCDPK4. Strikingly, removal of additional 10 amino acids (T349) from the J-domain caused a dramatic increase in the catalytic activity. In addition, the activity exhibited by this mutant was independent of calcium (Fig. 5C). These data suggested that the stretch between aa 349 and 359 may be responsible for sequestering PfCDPK4 in an inactive state. The interaction of J-domain with the catalytic domain was one possibility via which PfCDPK4 was inhibited as demonstrated for AtCPK1 (15, 31). To analyze this further, a peptide (Peptide I) corresponding to residues 346-364, which spans almost the entire J-domain, was synthesized and included in kinase assays (Fig. 6A). Peptide I inhibited the activity of PfCDPK4 confirming the inhibitory nature of the J-domain (Fig. 5B). Significantly, the activity of T349 mutant, which was constitutively active (Fig. 5B), was also inhibited by this peptide (Fig. 6C). Because this mutant lacks the CLD, it is reasonable to attribute this inhibition to a possible pseudo substrate motif in this region. A smaller peptide, Peptide II, corresponding to aa 350-358, was used to further narrow down the inhibitory region. This peptide inhibited the activity of a T349 mutant (Fig. 6D) as efficiently as peptide I with IC₅₀ ∼ 100 μm suggesting that the segment 350-358 may act as a pseudo substrate inhibitor. Collectively, these and truncation mutant data (Fig. 5C) suggest that these N-terminal residues (aa 350-358) of the J-domain may operate as a pseudosubstrate inhibitory motif.

FIGURE 7.

A, ClustalW sequence alignment of J-domains of PfCDPK4 and AtCPK-1 and Soya bean CDPKα (GmCDPKα). The C-terminal region of AtCPK-1 JD, which forms an α-helix (21), and the corresponding region of PfCDPK and GmCDPKα are shaded in red. Phe⁴³⁶ of AtCPK1 and a corresponding Leu³⁶⁰ of PfCDPK4 are indicated by an asterisk. The arrow indicates the start of the CLD of these kinases. B, firstly, a homology model of PfCDPK4 CLD and its J-domain was obtained using the coordinates for the domain-swapped crystal structure of the corresponding domains of AtCPK-1 (21). Based on this structure, a structural model for intermolecular interaction of the J and CLD of PfCDPK4 was constructed (see “Experimental Procedures”), which is shown here. Key residues in the C-lobe (green) of the CLD (magenta) that may form a binding pocket for Leu³⁶⁰ (yellow) of the J-domain (red), are indicated in ball and stick. C and D, Pep III (C and D) or Pep II (D) was incubated with PfCDPK4 (C) or T349 (D) in a kinase assay mix. Subsequently, 150 μm syntide-2 was added, and the assay was initiated by the addition of ATP. Activity was assessed by measuring syntide phosphorylation as described earlier. E, kinase activity of equal amounts of recombinant PfCDPK4 or L360A mutant was measured in the presence or absence of calcium by measuring syntide phosphorylation. A representation of three independent experiments is shown. Error bars represent ±S.E. between replicates of the same experiment. F, CD of Pep III. CD spectra for Pep III was acquired in water (blue) or increasing concentration of TFE. The TFE concentration used is indicated in the inset.

FIGURE 6.

A, amino acid sequence of the J-domain and peptides corresponding to different regions of this domain are indicated. B-D, kinase assays were performed using PfCDPK4 (B) or T349 mutant (C and D). Recombinant enzymes were preincubated with different concentrations of Pep I (B and C) or Pep II (D) prior to the addition of 150 μm syntide and ATP in a kinase assay mix. Phosphate incorporation in syntide was determined as described above. Representative of three experiments is shown in the figure.

C Terminus of the J-domain Controls PfCDPK4 Activation via CLD—We next investigated if the C-terminal region of the J-domain plays a part in PfCDPK4 regulation. For this purpose, peptide III corresponding to residues 357-368 of the J-domain was used (Fig. 6A). When added to kinase assays, peptide III inhibited the activity of full-length PfCDPK4 (Fig. 7C). Strikingly, it failed to influence the activity of T349 mutant (Fig. 7D, third bar). Because T349 lacks the CLD, it was possible that this peptide regulates PfCDPK4 via the CLD.

While these studies were in progress a crystal structure of the CLD and the J-domain of A. thaliana CDPK1 (AtCPK1) was reported (21). Even though AtCPK1 does not form a dimer in solution, the CLD-J domain of this kinase existed as a dimer in this structure. This structure revealed that the C terminus of the J-domain forms a α-helix, which interacts with the C-lobe of the CLD of the other monomer via a “domain-swap” mechanism. Based on these findings, the C terminus of the AtCPK1 J-domain (aa 433-446) was proposed to interact with its CLD (21). Sequence alignment and homology modeling of the CLD-J-domain of PfCDPK4 suggested that T357-S370 may correspond to the helical region of AtCPK1 J-domain (Fig. 7, A and B).

Although the CLDs of the two kinases are highly similar, the smaller JD possesses limited sequence similarity. For instance, only 7 of the 14 aa of the putative CLD interacting region are similar (Fig. 7A). Therefore, this model was “low resolution” and served as a guide for further experiments. The first support for the predictions made by the model was obtained from experiments done with peptide III, which spans most of the CLD-interacting segment (Fig. 6A). The inhibition of PfCDPK4, and not T349, by this peptide could be attributed to its interaction with the CLD (Fig. 7, C and D). To extend these studies further, CD studies were performed to elucidate the conformation of peptide III. The CD spectra suggested a random structure for this peptide in water as a large negative ellipticity was observed close to 200 nm (Fig. 7F). It is known that the addition of trifluoroethanol (TFE) to an aqueous solution of peptide, which otherwise has a random structure, may stabilize nascent, secondary structural elements (32). The addition of TFE to peptide III aqueous solution resulted in an increase in ellipticity at ∼204 and 220 nm and a peak at 190 nm (Fig. 7F), which was indicative of stabilization of the helical conformer (33). Therefore, it is reasonable to state that the Thr³⁵⁷-Met³⁶⁸ region has a propensity to form α-helix in a hydrophobic environment, which supports the homology model (Fig. 7B).

The modeling studies also suggested that Leu³⁶⁰ of PfCDPK4, which corresponds to Phe⁴³⁶ of AtCPK-1, anchors the J-domain to the CLD by interacting with a group of C-lobe hydrophobic residues. To confirm this, a L360A mutant of PfCDPK4 was generated. In comparison to PfCDPK4, L360A exhibited significantly reduced catalytic activity even in the presence of calcium (Fig. 7E). These results support the proposed model in which Leu³⁶⁰ seems to be critical for exerting the control of CLD over J-domain. A F436A mutation in AtCPK1 was also reported to exhibit reduced catalytic activation of the kinase (34). Taken together with the peptide III inhibition results (Fig. 7, C and D), these data suggest that the CLD interaction with the C terminus of the J-domain is a key step in PfCDPK4 regulation. Based on these findings, the J-domain can be divided into two segments: a pseudo-substrate region, which resides between aa 350 and 358 and interacts with its catalytic site and a CLD-interacting region (aa 357-368).

Autophosphorylation of Thr²³⁴ Is Crucial for PfCDPK4 Activation—Some CDPKs are not totally dependent on autophosphorylation for their activation (35). The observations made in Fig. 3 suggested that PfCDPK4 is autophosphorylated in the presence of calcium. To assess the role of autophosphorylation in its regulation, it was important to identify the autophosphorylation site on this enzyme. It is well known that the phosphorylation of the activation loop of most protein kinases results in their activation (36). The role of the activation loop phosphorylation in PfCDPK4 was explored. Sequences of the activation loop, which reside between subdomains VII and VIII of PfCDPK4 (supplemental Fig. S1), were compared with the corresponding activation loops of kinases in which the regulatory phosphorylation sites are known (data not shown). Three possible regulatory autophosphorylation sites in the loop region (Ser²¹⁹, Thr²²⁰, and Thr²³⁴) (supplemental Fig. S1) emerged as likely sites from this exercise. To evaluate this experimentally, two mutants of PfCDPK4 were created: a double mutant S219/T220A and T234A a single point mutant. The S219/T220A double mutant exhibited both autophosphorylation (Fig. 8A) as well as kinase activity (Fig. 8B) as observed for the wild-type enzyme ruling out a role of Ser²¹⁹ and Thr²²⁰ in PfCDPK4 regulation. In comparison, Thr²³⁴ mutation to A resulted in almost complete loss of PfCDPK4 autophosphorylation (Fig. 8A), which was accompanied by an inhibition of its catalytic activity (Fig. 8B). Therefore, it is reasonable to propose that autophosphorylation of Thr²³⁴ is essential for PfCDPK4 catalytic activation.

FIGURE 8.

Regulation of PfCDPK4 by autophosphorylation. A, equal amounts of PfCDPK4, single point mutant T234A or S219/T220A double mutant were incubated in the presence or absence of CaCl₂, and the kinase assay was performed as described earlier. The kinase assay mix was electrophoresed, and autoradiography was performed to detect autophosphorylation of recombinant PfCDPK4. Although PfCDPK4 and S219/T220A exhibited autophosphorylation in the presence of calcium, T234A remained unphosphorylated. B, recombinant PfCDPK4, a single point mutant T234A, or double mutant S219/T220A were incubated in a kinase assay mix with syntide in the presence or absence of calcium. Kinase activity was assessed by measuring syntide phosphorylation. The mutation of Thr²³⁴ to Ala resulted in almost complete loss of PfCDPK4 activity.

Based on the above biochemical data, we propose the following mechanism for the activation of PfCDPK4 (Fig. 9): when calcium binds to the CLD of PfCDPK4, it results in its interaction with the C terminus of the J-domain. As a result, constraints are imparted on the pseudo substrate region of the J-domain resulting in its dissociation from the catalytic domain (Figs. 6 and 7). These events facilitate the autophosphorylation of the activation loop at Thr²³⁴, which ultimately results in PfCDPK4 activation (Fig. 8).

FIGURE 9.

A model for calcium-mediated regulation of PfCDPK4 by its JD and CLD. The JD of PfCDPK4 contains two segments: an N-terminal pseudosubstrate motif and a C-terminal CLD interacting region. In the absence of calcium, the pseudo substrate motif occupies the catalytic site keeping the kinase in an inactive state (I). The binding of calcium to the CLD promotes interaction between the C-lobe of the CLD and the C terminus of J-domain (II), which may impart conformational constrains on the pseudo-substrate region resulting in its release from the catalytic cleft (III). Once free of the J-domain, T234 may get autophosphorylated, and the catalytic site may interact with its substrates and facilitate their phosphorylation.

Identification of Regulatory Elements in PfCDPKs—It was worth investigating the presence of regulatory motifs in the JD of other PfCDPKs using the information obtained from PfCDPK4 studies. When sequences of J-domains of PfCDPK1-5 in P. falciparum were compared using ClustalW (Fig. 10A), a KLXXΦAΦXXΦAXXΦ (Φ= hydrophobic aa) motif was found conserved in almost all PfCDPKs with some minor differences in PfCDPK 1/4 (Fig. 10A). This motif corresponds to the CLD-interacting region of PfCDPK4 (Fig. 7A). The residues separating KL and the ΦAΦ segments are basic in PfCDPK2, -3, and -5 but are neutral in PfCDPK1 and -4. The other difference between PfCDPK1/4 and other PfCDPKs is in the central -ΦAΦ-segment, whereas the first residue is hydrophobic in PfCDPK2, -3, and -5, it is replaced by an Ala in PfCDPK1 and -4. The core NΦR/KXF pseudosubstrate motif is highly conserved among all PfCDPKs. Sequence comparison indicates that two hydrophobic residues, which may provide additional interaction with the catalytic site, precede this motif in PfCDPK2, -3, and -5. A small variation was observed for PfCDPK1 and -4, and this core motif is preceded by only one hydrophobic residue. The subtle differences between PfCDPK1 and -4 and other PfCDPKs may suggest that these enzymes are closest members of this group.

FIGURE 10.

Prediction of regulatory elements in PfCDPKs. A, sequences of the J-domain of PfCDPK1-5 were aligned using ClustalW software. Based on the data obtained for PfCDPK4 in this work, a putative pseudo substrate motif (black box) and the CLD interacting region (orange box) of CDPKs were predicted. The consensus sequence for these motifs, which emerged from this analysis, is also indicated below. B and C, wild-type PfCDPK1 (B) or its truncation mutant PfCDPK1_T341 (C) were preincubated with peptides I, II, or III, and kinase assays were performed using syntide as substrate as described above for PfCDPK4 in Fig. 7.

Peptides I, II, and III, which inhibited PfCDPK4 activity by interacting with the kinase domain and/or CLD, were tested against PfCDPK1 to validate the predictions. These three peptides effectively inhibited PfCDPK1 (Fig. 10B) supporting the predictions made above. PfCDPK1_T341, a truncation mutant of PfCDPK1 that ends at aa 341, was generated by deleting most of the portion of JD and CLD. Like the T349 mutant of PfCDPK4, this PfCDPK1 truncation was also active both in the presence and the absence of calcium (not shown here). Although peptides I and II inhibited this mutant, peptide III did not alter its activity (Fig. 10C). These studies corroborate well with the results obtained for PfCDPK4 and its T349 mutant (Fig. 7, C and D). Therefore, it is reasonable to suggest that these peptides or the corresponding segments in the J-domain may regulate PfCDPK1 via mechanism similar to the one proposed for PfCDPK4.

DISCUSSION

Plasmodium genome codes for several signaling molecules like protein kinases, which includes five members that belong to the calcium-dependent protein kinase family. The CDPKs are typically absent from animals or fungi; interestingly, they are present in protozoan parasites (14). However, there are ∼42 CDPK isoforms in A. thaliana that are crucial for various aspects of plant physiology (37). Calcium regulates a wide variety of important functions in the parasite life cycle, and recent studies have demonstrated that the activity of CDPKs may be critical in carrying out several calcium-regulated processes (1, 7, 9, 38, 39). PbCDPK4 gene disruption stalls cell cycle progression in P. berghei male gametocytes and results in reduced sexual reproduction and mosquito transmission (1). Another kinase, cGMP-dependent protein kinase from P. falciparum, was recently shown to play a role in gametogenesis (40) of P. falciparum indicating the importance of signaling events in sexual development of the parasite.

The immunofluorescence studies suggested that PfCDPK4 is mainly present near the gametocyte surface. The N-terminal myristoylation signal may be responsible for its cell surface targeting, which has been observed for PfCDPK1 (25).

It was demonstrated previously that the products of PLC hydrolysis inositol 1,4,5-trisphosphate and diacylglycerol may be involved in exflagellation (30, 41). It should be noted that a PLC homologue is present in the parasite. The dependence of PfCDPK4 activity on PLC in gametocytes correlates well with proposed role of PLC and its products in mobilizing parasite calcium (29, 42) and that of CDPK4 in gametogenesis (1). Given these findings, it will be interesting whether PLC serves as an upstream activator for other CDPKs.

It was clear from the activity assays that PfCDPK4 activity was dependent on autophosphorylation. The mechanism of autophosphorylation and dependence on it for catalytic activation can vary among different CDPKs. A phosphoproteomic study involving CDPKs suggested that these kinases can be autophosphorylated in at least five different motifs. Interestingly, several CDPKs do not exhibit autophosphorylation of the activation loop (43), which indicates the differences in the mechanisms through which these kinases may be regulated. Our results suggest that PfCDPK4 is regulated by autophosphorylation of Thr²³⁴, which resides in its activation loop. Although autophosphorylation of Ser²¹⁹ and Thr²²⁰ of PfCDPK4 seems to be non-existent, complementary sites in some CDPKs do get autophosphorylated (43). It was interesting to note that AtCPK1, which shares similarities with PfCDPK4 in J-domain regulation, did not exhibit autophosphorylation of the activation loop (43). Even though PfCDPK1 was shown to be autophosphorylated at a site complementary to Thr²³⁴ (43), the role of autophosphorylation in its catalytic activation has not been demonstrated.

Biochemical studies revealed that the J-domain of PfCDPK4 exerts control over its timely stimulation by calcium. It was evident from the calcium-independent activity of the ΔJ mutant that the J-domain of CDPK4 may regulate PfCDPK4 activation. Studies performed with the J-domain truncation mutants resulted in identification of two functionally distinct regions. The N terminus of the J-domain possesses an NIRQFQS motif (aa 350-358), which acts as a pseudosubstrate. The pseudosubstrate region in AtCPK1 and GmCDPKα is RM/LKQFQS (15, 31) (Fig. 7A), and it has an extra basic residue at first position, which is missing from PfCDPK4 (Fig. 7A). Interestingly, this extra basic residue is absent from all PfCDPKs except PfCDPK5 (Fig. 10). Peptide II, which corresponds to this motif, inhibited the activity of the constitutively active Thr³⁴⁹ mutant confirming its pseudosubstrate nature. It has remained unknown how CDPKs interact with its pseudosubstrate motifs or substrates. Preliminary data from our laboratory suggests that the basic residues in its substrates may form key interactions with active site residues in the kinase.4

In contrast to Pep II, Pep III, which is complementary to the C terminus of the J-domain, only inhibited the wild-type PfCDPK4 and not the constitutively active T349 mutant. Because T349 lacks the CLD, we conclude that peptide III may regulate PfCDPK4 via its CLD.

The crystal structure of the AtCPK1 J-domain along with the CLD revealed that the C-terminal portion of the JD may interact with the CLD (21). A homology model was generated for the PfCDPK4 CLD-JD, which suggested that aa Gln³⁵⁸-Lys³⁷¹ may form a α-helix, which may interact with the CLD. The biochemical results obtained with peptide III, which constitutes a large part of this segment, were consistent with this observation as this peptide inhibited PfCDPK4 and not the constitutively active T349 mutant. Pep III exhibited the propensity to form α-helix, which provided further support to the model. It is worth appreciating that, despite only modest homology between the CLD interacting segment of PfCDPK4 and AtCPK1, this motif adopts a helical conformation, which facilitates interaction with CLD via residues like Leu³⁶⁰.

These findings led us to propose a model for PfCDPK4 activation, which may also be relevant for other PfCDPKs: the J-domain uses a bipartite motif to interact with the CLD as well as the kinase/catalytic domain. Firstly, calcium binding to CLD results in interaction with the C-terminal region of the J-domain. As a result, the N-terminal pseudosubstrate region may be forced to dissociate from the catalytic domain resulting in PfCDPK4 autophosphorylation and catalytic activation.

These studies on PfCDPK4 aided the prediction of regulatory motifs in the J-domain of other PfCDPKs. PfCDPK1 and PfCDPK4 appeared to be the closest among this group. The inhibition of PfCDPK1 by PfCDPK4-JD peptides confirmed that this kinase may also be regulated by the predicted regulatory motifs via a mechanism similar to that of PfCDPK4.

Given the importance of CDPK4 (1) and other CDPKs in the development of malaria parasite, this information may prove useful for design of inhibitors against CDPKs. It is evident from the present studies that blocking the catalytic site of PfCDPK4 and/or preventing the interaction of the CLD with the J-domain by targeting the CLD may be very useful to achieve this goal. To understand PfCDPK4 function in the parasite it is important to identify cellular targets of this enzyme.