|Home | About | Journals | Submit | Contact Us | Français|
Calcium controls many critical events in the complex life cycles of apicomplexan parasites including protein secretion, motility, and development. Calcium levels are normally tightly regulated and rapid release of calcium into the cytosol activates a family of calcium-dependent protein kinases (CDPKs), which are normally characteristic of plants. CDPKs present in apicomplexans have acquired a number of unique domain structures likely reflecting their diverse functions. Calcium regulation in parasites is closely linked to signaling by cyclic nucleotides and their associated kinases. This review summarizes the pivotal roles that calcium-and cyclic nucleotide-dependent kinases play in unique aspects of parasite biology.
The Apicomplexa is an ancient phylum of some 5,000 diverse eukaryotic species that are largely parasitic on marine invertebrates, insects, and vertebrates, where they have gained notoriety because of their role in animal and human diseases. Five Plasmodium species are responsible for significant mortality and morbidity due to malaria, the most serious form of which is caused by P. falciparum (Snow et al., 2005). Related organisms such as Toxoplasma gondii (Joynson and Wreghitt, 2001) and Cryptosporidium spp. (Tzipori and Widmer, 2008) cause opportunistic infections of considerable importance in immunocompromised individuals. Highly divergent from well-studied organisms like yeast, flies, and worms, apicomplexans are most closely related to ciliates and dinoflagellates (Baldauf, 2003). As a result of this extreme evolutionary divergence, much of their basic biology is distinct from what we know of model organisms. Apicomplexans often contain plant-like features, owing to two events: 1) very early branching that likely predates the animal-plant split (Baldauf, 2003), 2) acquisition of a secondary endosymbiont derived from engulfment of an algal cell (Waller and McFadden, 2005). As such, signaling pathways in apicomplexans contain both conserved and unique features.
Apicomplexans have highly polarized cells that are specialized for regulated secretion and directed entry into their host cells (Cowman and Crabb, 2006; Sibley, 2004). Calcium controls a number of critical events in the life cycle including secretion of adhesins, gliding motility, cell invasion, and egress (Moreno and Docampo, 2003). Calcium also influences developmental processes that occur at distinct stages in their complex life cycles (Alano and Billker, 2005). In this review, we focus on the role of calcium signaling in two important apicomplexan parasites: Toxoplasma gondii, and Plasmodium spp. Aside from its significance as a pathogen, T. gondii has gained importance as model for cellular and biochemical studies of an otherwise intractable group of parasites (Sibley, 2004). In contrast, cellular studies are more challenging in Plasmodium spp., yet application of genetics has been useful to explore the role of stage-specific genes (Carvalho and Menard, 2005). Together, studies conducted in these two systems provide a framework for understanding how calcium regulates several important classes of protein kinases, which in turn control much of the unique biology of these organisms. In this review, we outline findings that have been developed separately in each of these systems, calling attention to cases where parallel studies have been informative and highlighting areas where further comparative studies are needed.
Apicomplexans are united by a conserved set of features that define their apically polarized cellular forms, which are specialized for migration and cell invasion (Morrissette and Sibley, 2002). The apical pole is defined by a unique microtubular organizing center, which in Toxoplasma is composed of a conoid and two polar rings, from which a group of singlet microtubules emanates and subtends the membrane of the elongated organism (Morrissette and Sibley, 2002). Beneath the plasma membrane, an inner membrane comprised of flattened vesicles derived from the endoplasmic reticulum surrounds the parasite. A meshwork of filaments formed from articulin-like proteins (Gould et al., 2008) provides structural support for the inner membrane complex that extends beneath the entire surface of the cell, except at its very apex. Three sets of secretory organelles discharge their contents from the apical end of the parasite during cell invasion (Carruthers and Sibley, 1997). These features are conserved among sporozoites of most apicomplexans including malaria, except for the absence of a conoid; and while they are less well defined in merozoites, these forms also share a common mechanism of cell invasion.
Apicomplexans rely on an actin-myosin motor for gliding motility, tissue migration, and cell invasion (Cowman and Crabb, 2006; Sibley, 2004). Motility and invasion are highly dependent on assembly of filamentous actin in the parasite (Dobrowolski and Sibley, 1996). The force for motility is generated by a small myosin, best studied in T. gondii but also conserved in Plasmodium, which is anchored in the inner membrane (Gaskins et al., 2004; Meissner et al., 2002). Cell invasion is also critically dependent on secretion of adhesins, which are stored in apical secretory organelles called micronemes. Among the best studied of these are TRAP, which is essential for gliding motility and invasion of Plasmodium (Sultan et al., 1997), and its orthologue MIC2, which is likewise essential in T. gondii (Huynh and Carruthers, 2006). TRAP and MIC2 contain several extracellular domains that interact with the substratum, a transmembrane domain, and a short cytoplasmic tail. The cytoplasmic domains of TRAP and MIC2 bind specifically to aldolase through a combination of charge and hydrophobic interactions, thereby linking to filamentous actin, and coupling adhesion with actin-based motility (Bosch et al., 2007; Buscaglia et al., 2003; Jewett and Sibley, 2003; Starnes et al., 2009; Starnes et al., 2006). Following apical secretion, adhesins are translocated by the actin-myosin motor, traveling along the cell surface to the posterior end, thus driving net forward movement. This process culminates with trimming from the surface by a family of rhomboid proteins that clip the adhesins within their transmembrane domains (Brossier et al., 2005; O’Donnell et al., 2006). Actin-myosin based gliding motility is both conserved within the group and unique to apicomplexans, which otherwise lack locomotory organelles such as cilia and flagella (other than in the male gamete).
The requirements for intracellular calcium have been extensively studied in T. gondii due to the availability of quantitative in vitro assays for motility, protein secretion, host cell invasion and egress (Fig. 1). Calcium levels in T. gondii are maintained at very low levels (i.e. 100 nM) in the cytosol of resting parasites, as is typical of eukaryotic cells (Moreno and Docampo, 2003). Intracellular storage sites of calcium include the endoplasmic reticulum (ER), mitochondrion, and acidocalcisomes (Moreno and Docampo, 2003). The primary mobilizable store for calcium is likely the ER and intracellular calcium provides the key signal for triggering a number of important events (Lovett and Sibley, 2003). Apicomplexans contain a conserved sacroplasmic-endoplasmic reticulum calcium ATPase (SERCA) that pumps calcium into the lumen of the ER, hence refilling this store and decreasing cytoplasmic levels (Fig. 2). The SERCA pump in T. gondii has been localized to the ER and is sensitive to the plant compounds thapsigargin and artemisinin (Nagamune et al., 2007). Thapsigargin and artemisinin also affect the activity of the Plasmodium SERCA when expressed in Xenopus (Eckstein-Ludwig et al., 2003), and this molecular target may partially explain the effectiveness of artemisinins as anti-malarial agents. In addition, T. gondii, contains several plasma membrane Ca2+ ATPases, one of which has been genetically disrupted leading to decreased infectivity (Luo et al., 2005). Apicomplexans also contain two PMR1-like transporters, which in yeast are Golgi-type Ca2+ ATPases, and a single Ca2+/H+ exchanger (Nagamune et al., 2008b). Collectively, these pumps control resting calcium levels and affect changes in calcium in response to environmental cues.
Recording of calcium transients from Fluo-4 labeled tachyzoites of T. gondii during gliding motility revealed an oscillating pattern of intracellular calcium that was abruptly dampened during cell invasion (Lovett and Sibley, 2003). The pattern of calcium waves correlates with periods of microneme secretion, which also abruptly terminate following successful invasion of host cells. How these calcium transients are controlled, and how they are shut off is presently unclear. However, agents that alter the pattern or amplitude of oscillations also affect microneme secretion and cell invasion. For example, treatment with calmidazolium increases the frequency of calcium transients while decreasing the amplitude of the waves, resulting in enhanced microneme secretion and gliding (Wetzel et al., 2004). Other compounds, such as thapsigargin and artemisinin, oppose this pattern by increasing the amplitude but decreasing the frequency (Nagamune et al., 2007). Hence, the timing of release and spatial pattern of calcium may control activation of downstream effectors important in mediating secretion and / or activating the motor complex.
Cytoplasmic calcium in T. gondii controls microneme secretion as shown by pharmacological agents such as ionophores that raise calcium and stimulate secretion of micronemal proteins; conversely, chelation of intracellular calcium using BAPTA-AM prevents microneme secretion and disrupts motility and cell invasion (Carruthers et al., 1999; Carruthers and Sibley, 1999). While not as amenable to experimental analyses, similar studies suggest a similar requirement for calcium in the control of microneme discharge from sporozoites of Cryptosporidium parvum (Chen et al., 2004), Eimeria tenella (Wiersma et al., 2004), and P. berghei (Gantt et al., 2000). Biochemical studies in T. gondii reveal the presence of both cyclase and hydrolase activities involved in production of the second messenger cyclic ADP ribose (cADPR) (Chini et al., 2005). A separate pathway generates IP3, presumably through phospholipase C (Lovett et al., 2002), which has been characterized in vitro using recombinant protien (Fang et al., 2006). Pharmacological evidence also indicates the presence of calcium release channels that respond to IP3 and cADPR (Lovett et al., 2002) (Fig. 2). Similar to plant cells, the molecular basis of these calcium-release channels has not been defined and the genomes of apicomplexans do not contain obvious orthologues of the IP3 or ryanodine-responsive channels that have been characterized in mammalian cells (Nagamune et al., 2008b). Additionally, the receptors that sense environmental changes to generate IP3 or cADPR have not been identified, despite biochemical and pharmacological evidence that both second messengers are required for motility and invasion (Lovett et al., 2002) (Fig. 2).
Elevated calcium also participates in activating egress of T. gondii from the cell, a process that depends on microneme secretion and active parasite motility (Fig 1). Artificial elevation of calcium using calcium ionophore triggers egress, a process that also relies on active motility (Endo et al., 1982). Treatment with calcium ionophore will trigger egress of T. gondii throughout the process of intracellular development, perhaps due to its unusual mode of binary cell division (Morrissette and Sibley, 2002), which leaves the parasite ready to reinvade throughout the entire cell cycle. Recent studies indicate that a plant-like pathway for controlling cADPR production through abscisic acid (ABA) controls this pathway in T. gondii (Nagamune et al., 2008a). This process is similar to the role of ABA in controlling cADPR production and elevated calcium in guard cells in plants (Wu et al., 1997). In T. gondii, the signaling pathway resembles quorum sensing in that ABA accumulates during intracellular replication, ultimately triggering increases in cADPR and resulting in elevated calcium that stimulates cell egress (Nagamune et al., 2008a). ABA is derived from the isoprenoid pathway, presumably present in apicomplexans due to their ancestral acquisition of a secondary endosymbiont of algal origin (Ralph et al., 2004). Blockade of this pathway with the plant herbicide fluoridone (FLU) prevents cell egress and results in a developmental switch to slow-growing tissue cysts (Fig. 1). Egress also requires microneme secretion, in part due to release of a perforin-like protein called TgPLP1 that participates in lysis of the parasite-containing vacuole (Kafsack et al., 2009). TgPLP1 is found in the micronemes and requires elevated levels of intracellular calcium for its release. Prior to egress of T. gondii, the parasite may also sense damage to the host cell plasma membrane, which results in a decrease in potassium levels in the host cytosol that is sensed by the parasite (Moudy et al., 2001). Several orthologues of voltage-gated calcium channels present in T. gondii are candidates for controlling this pathway (Nagamune et al., 2008b), although their roles have yet to be explored experimentally.
Cell attachment by apicomplexans involves regulated secretion of adhesive proteins from the micronemes. The release of microneme proteins is tightly controlled, perhaps to prevent serum antibodies from coating them prematurely and blocking attachment. Furthermore, releasing microneme proteins at the apex of the cell assures apical attachment, which is a necessary step for cell invasion. Downstream of elevated cytoplasmic calcium levels, protein kinases have been implicated as effectors in mediating protein secretion in T. gondii (Fig. 2). Early studies indicated that broad spectrum inhibitors such as staurosporine block microneme secretion, motility, and cell invasion by T. gondii (Carruthers et al., 1999). Subsequent studies revealed that at least two different kinases control these events. Cyclic GMP-dependent kinase (PKG) was identified as the target of a selective parasite inhibitor called compound (Cmpd) 1 (Gurnett et al., 2002). Cpmd 1 is a trisubstituted pyrrole pyridine that inhibits replication of T. gondii as well as related apicomplexans. Cmpd 1 blocks microneme secretion, motility, and cell invasion by T. gondii with similar efficacy as staurosporine, suggesting PKG controls these phenotypes (Wiersma et al., 2004). PKG was validated as the primary target of Cmpd 1 using genetic complementation with a mutant that was engineered to be resistant based on molecular modeling (Donald et al., 2002). Since gliding motility and cell invasion both rely on microneme secretion, the simplest interpretation of this effect is that protein phosphorylation by PKG is needed for either docking of micronemes to the membrane or inducing fusion with the apical end of the parasite, although the molecular targets of PKG in vivo are unknown. PKG has also been implicated in calcium signaling in Plasmodium (Baker and Deng, 2005), as discussed further below.
The protein kinase inhibitor KT5926 (an inhibitor of calmodulin dependent kinase II (CaMKII) and myosin light chain kinase (MLCK) in mammalian cells) was also shown to block microneme secretion, gliding motility and cell invasion by T. gondii (Dobrowolski et al., 1997). KT5926 has little reactivity against PKG in vitro (Wiersma et al., 2004), indicating that these two pathways are independent and implying that in addition to PKG, another kinase is required for microneme secretion (Fig. 2). KT5926 has been shown to inhibit one member of a family of calcium-dependent protein kinases in T. gondii, called CDPK1, in vitro, and it was suggested that activity against the parasite might be due to this molecular target (Kieschnick et al., 2001). However, this study was conducted prior to the genome project for T. gondii, hence the complexity of the CDPK family in apicomplexans was under appreciated. As such, the activity of KT5926 against the parasite may reflect its inhibition of CDPK1 or another related target(s).
Different groups of eukaryotic protein kinases serve as important effectors of calcium signaling. Those that bind to calmodulin via a C-terminal association domain are referred to as calmodulin-dependent kinases (CaMK) (Hook and Means, 2001). Classical CaMKs are typical of animal cells, but rare in plants, which instead contain several related classes of calcium-dependent protein kinases (CDPKs). In plants, CDPKs are abundant, for example Arabidopsis has been estimated to have more than 40 CDPK isoforms (Harper and Harmon, 2005). Not surprisingly given this diversity, CDPKs control a diverse array of functions including transcription, metabolism, ion pumps and channels, and the cytoskeleton (Harmon et al., 2000). Canonical CDPKs contain a number of calmodulin-like EF-hands fused directly to the kinase domain and linked by a short inhibitory domain (Harmon et al., 2000). The typical domain structure involves four C-terminal EF hand domains, however other configurations are also known. The autoinhibitor domain is thought to regulate CDPKs by interacting with the kinase domain and acting as a pseudosubstrate; autoinhibition is relieved by calcium binding to the EF hand domains. The junction domain of PfCDPK4 exerts an inhibitory effect on kinase activity in vitro (Ranjan et al., 2009), suggesting that the autoinhibitor model first developed for plant CDPKs, can be extended to apicomplexan CDPKs, Titration of calcium activation of plant CDPKs reveals widely different sensitivities in vitro, and the threshold for activation is also dependent on the substrate used (Harper et al., 2004). Hence, CDPKs are tuned such that activity on preferred substrates is more easily activated at low calcium levels. The calcium binding properties of the different EF hand domains also likely differ and mutational studies have shown that the two N-terminal domains are more important than the C-terminal domains in controlling the activity of P. falciparum CDPK1 in vitro (Zhao et al., 1994). CDPKs respond not only to absolute levels of calcium but to oscillating patterns, as shown by the involvement of a calcium-calmodulin dependent kinase (CCaMK) in development of rhizobial bacterial symbionts (Levy et al., 2004). Hence, a variety of features make CDPKs optimally adapted to control diverse responses to elevated calcium including differences in calcium responsiveness, localization, and substrate preference.
Similar to higher plants, CaMK are rare in protozoan parasites, while they express diverse families of CDPKs (Nagamune et al., 2008b). A single example of a classical CaMK in malaria (i.e. PfPK2 (XP_001350783) contains a calmodulin binding domain and requires both calcium and calmodulin for activation (Kato et al., 2008a). This gene has a single orthologue each in T. gondii (EEE26021.1) and C. parvum (XP_627434.1). In addition, apicomplexans contain CamK-related kinases that lack either an obvious calmodulin binding domain or EF hands (Nagamune et al., 2008b; Ward et al., 2004); whether these are regulated by calcium is uncertain. By far the most abundant class of calcium-dependent kinases in apicomplexans is the CDPKs. A phylogenetic comparison based on the kinase domains of CDPKs in apicomplexans and plants also defines several common domain architectures. Apicomplexan CDPKs fall into five major classes that have different domain structures (Fig. 3). Although historical names have been retained in Fig. 3, we have also suggested a uniform system for naming families of CDPKs based on the grouping of orthologues, as described below and further detailed in Table S1. The canonical CDPK structure containing an N-terminal kinase domain and four EF hands is represented by several examples from plants and ciliates (Fig. 3; green) and two groups in the Apicomplexa (Fig. 3; blue). The first of these apicomplexan groups contains relatively short N-termini that are predicted to be acylated and includes the previously studied proteins designated as PfCDPK1 and PfCDPK4 (Fig. 3). These have orthologues in T. gondii and C. parvum, although the numbering is different due to historical reasons (i.e. PfCDPK1 is orthologous to TgCDPK3, while PfCDPK4 is orthologous to TgCDPK1). A second group of canonical CDPKs has somewhat longer N-termini of unknown function and are not predicted to be acylated (PfCDPK2 is an exception to this profile); this group is typified by PfCDPK3 and PfCDPK5, as well as a number of orthologues in C. parvum and T. gondii (Fig. 3). Several other domain architectures are also apparent in apicomplexan CDPKs. For example, T. gondii and C. parvum contain CDPKs with only three C-terminal EF hands similar to CCaMK in plants (Fig. 3; purple). All three genera contain members of a group with one or more N-terminal EF hands followed by a kinase domain and three or four C-terminal EF hands (Fig. 3; red), such as PfCDPK6. Finally, P. falciparum and T. gondii contain a class of CDPKs with two or more N-terminal EF hands followed by a plekstrin homology domain and a C-terminal kinase domain, referred to here as CDPK7 (Fig. 3; yellow). The arrangement of these N-terminal EF domains is unusual and their role in regulation of kinase activity has not been examined. Several canonical CDPKs in parasites contain consensus motifs for N-myristoylation or palmitoylation, a feature also seen in plant CDPKs, many of which show membrane localization (Dammann et al., 2003; Hrabak et al., 1996). This suggests that association with membranes may be important in localization, as has been shown for PfCDPK1, which is modified by both palmitate and myristate (Möskes et al., 2004). The role of such modifications in regulation of activity has not been explored in parasites.
Plasmodium has a complex life cycle involving a number of different motile stages that are essential in either the vertebrate and invertebrate hosts (Fig. 4). This complexity, combined with some inherent limitations in available genetic tools, has imposed some restrictions on how calcium-signaling pathways have been studied in Plasmodium. For example, only the asexually replicating blood stages of malaria parasites can be propagated continuously to select genetically modified parasites. In the absence of robust techniques for conditional mutagenesis, genes with essential functions in erythrocyte invasion are not easily studied by experimental genetic analysis. Several recent alternatives for regulated (Armstrong and Goldberg, 2007) or stage-specific (Combe et al., 2009) expression show some promise for alleviating this limitation. However, to date, chemical and biochemical approaches have primarily been used to characterize the function of CDPK1 and other protein kinases in erythrocyte invasion by the merozoite. In contrast, experimental genetics in both the human pathogen P. falciparum and the model rodent parasite P. berghei have helped identify many genes with stage-specific functions in signaling during sexual development and in the mosquito stages of the life cycle (Fig. 4). Collectively, such studies have shown that Ca2+ regulates diverse biological processes in Plasmodium including erythrocyte invasion by merozoites, motility and invasion by ookinetes and sporozoites, discharge of secretory organelles, and sexual differentiation in the mosquito vector (Billker et al., 2004; Gantt et al., 2000; Ishino et al., 2006; Ono et al., 2008; Vaid et al., 2008).
One of the first protein kinases identified in Plasmodium asexual blood stages was PfCDPK1 (Zhao et al., 1993), yet the gene encoding this protein has remained refractory to genetic analysis (Kato et al., 2008b), suggesting it is essential in the blood stages. Dual N-terminal acylation and a cluster of basic amino acids near the N-terminus help PfCDPK1 to associate with membranes (Möskes et al., 2004). Recent studies have generated three lines of evidence that PfCDPK1 may regulate some aspect of the motor complex that powers erythrocyte invasion by the merozoite. Firstly, PfCDPK1 is co-expressed with other genes of the motor complex in late schizonts (Kato et al., 2008b), and the bulk of the PfCDPK1 protein is found near the periphery of the mature schizont and merozoite, most likely associated with the plasma membrane (Green et al., 2008), where it would have access to the motor complex. Secondly, recombinant PfCDPK1 phosphorylates in vitro both the myosin light chain (called myosin tail interacting protein, MTIP, in Plasmodium) and the glideosome-associated protein 45 (GAP45) (Green et al., 2008). Mass spectrometry indicates that at least one of the same sites on GAP45 is phosphorylated in vivo, although another modification seen in vivo was not detected by in vitro phosphorylation with PfCDPK1, possibly suggesting the participation of another kinase(s) (Green et al., 2008). This role might be fulfilled by PfPKB, a protein kinase B-like enzyme present in schizonts and merozoites. which is regulated by calcium/calmodulin and is also capable of phosphorylating GAP45 in vitro (Vaid and Sharma, 2006; Vaid et al., 2008). Thirdly, the trisubstituted purine, purfalcamine, which emerged from a screen as a potent inhibitor of recombinant PfCDPK1 in vitro, blocks P. falciparum development at the stage when schizonts segment into merozoites (Kato et al., 2008a). While the expression level of PfCDPK1 peaks in late schizonts, it is also expressed well before mature merozoites are formed (Green et al., 2008), suggesting it may be involved in motor complex assembly as well as potentially in regulation of motor activity. Consistent with this there is evidence that phosphorylation-dephosphorylation of GAP45 controls assembly of the motor complex in T. gondii (Gilk et al., 2009).
Although the combined evidence makes PfCDPK1 a candidate for regulation of the motor complex, the precise role(s) of this kinase in vivo is still uncertain. Moreover, this example serves to illustrate some of the challenges to defining the role of protein kinases. Serine threonine kinases are notoriously permissive and they will phosphorylate many substrates in vitro, while their precise activities are often regulated by location or other binding-scaffolding partners in vivo (Raman et al., 2007; Tasken and Aandahl, 2004). Hence, results of in vitro phosphorylation reactions only provide a starting point for testing modifications in vivo. A further complication is that even authentic modifications that do occur in vivo often do not appear to control essential processes. Deciphering which modifications are functionally important requires generation of mutant targets that are no longer phosphorylatable (or modified to mimic constitutive activation) and then testing their respective phenotypes in vivo. Additionally, since most kinase inhibitors are ATP analogues, they often have secondary activity against related or even seemingly unrelated kinases (McInnes and Fischer, 2005). In this regard, although purfalcamine clearly in inhibits CDPK1 in vitro (Kato et al., 2008b), its precise target(s) in vivo remain uncertain. Despite these concerns, chemical genetic approaches offer the potential to define the precise function of protein kinases in vivo (Blair et al., 2007; Blethrow et al., 2004; Greenbaum, 2008). Additionally, the function of such essential genes could be analyzed using stage-specific (Combe et al., 2009) or regulated expression systems (Armstrong and Goldberg, 2007) that have recently been developed in Plasmodium, or systems for regulated expression that have proven reliable in Toxoplasma (Herm-Gotz et al., 2007; Meissner et al., 2002).
Malaria transmission to the mosquito relies on specialized sexual precursor stages, the gametocytes (Fig. 4). Circulating gametocytes reside inside erythrocytes, where they are developmentally arrested until ingested by a mosquito. Once in the blood meal they rapidly resume development in response to three environmental signals: (1) a drop in temperature, (2) a small rise in extracellular pH and (3) xanthurenic acid (XA), which in insects is a major catabolite of tryptophan (Billker et al., 1998; Billker et al., 1997; Garcia et al., 1998). At a permissive temperature, P. berghei gametocytes respond to stimulation by XA with a sharp rise in intracellular Ca2+, which has been detected in a transgenic reporter parasite expressing the calcium-dependent luciferase, aequorin (Billker et al., 2004). Calcium is released from internal stores after a characteristic lag phase of 8–10 seconds and then persists for at least 1 min. Mobilization of Ca2+ causes male gametocytes to quickly enter the cell cycle, and after three rapid rounds of replication and mitosis, eight flagellated microgametes are released (Alano and Billker, 2005). The rise in cytosolic calcium is also required for male and female gametocytes to emerge from their host cells. Targeted deletion of PbCDPK4 showed that this male-expressed protein kinase is responsible for coupling the increase in cytosolic Ca2+ to cell cycle progression (Billker et al., 2004) (Fig. 4). PbCDPK4 substrates among the early cell cycle machinery have not been identified. However, recombinant PfCDPK4 phosphorylates and activates PfMap2 in vitro (Fig. 4). PfMap2 is an atypical mitogen activated protein kinase that is required at a later stage of differentiation, just before male gametes become motile and are released (Tewari et al., 2005). Consistent with its consensus motif for N-terminal myristoylation, PfCDPK4 localizes to the periphery of non-activated gametocytes (Ranjan et al., 2009). Additional Ca2+ effector pathways must be present in gametocytes since some cellular events during gametogenesis rely on Ca2+ but not on CDPK4. These include emergence of male and female gametocytes from the host cells and the lifting of a translational block on mRNAs stored in the female gamete (Mair et al., 2006). Ca2+ is not the only second messenger controlling gametogenesis. Early pharmacological experiments also suggested a role for cGMP (Kawamoto et al., 1993a) and a recent study confirmed this, showing that in P. falciparum gamete formation is blocked by the PKG inhibitor Cmpd1 and stimulated by zaprinast, an inhibitor of cGMP phosphodiesterase activity, which is predicted to raise cGMP levels and thereby activate PKG (McRobert et al., 2008). The chemical genetics strategy employed for T. gondii PKG (Wiersma et al., 2004) and described above, provided strong evidence in support of PKG as an early regulator of gametogenesis when applied to P. falciparum (Fig. 4). Rounding up due to the loss of longitudinal microtubules, an early morphological response unique to P. falciparum gametocytes is not inhibited by Cmpd1 in gametocytes expressing a Cmpd1-resistant PKG allele instead of wild type PKG. In contrast, other signs of gametocyte activation, including cell cycle progression in the male, remained sensitive to Cmpd1, revealing off-target effects of the inhibitor. How Ca2+ and cGMP dependent pathways are linked is not yet clear, but two lines of reasoning indicate that cGMP functions upstream of Ca2+ in P. falciparum gametocytes. Unlike all other aspects of gametocyte activation, the XA-induced rounding up of gametocytes is resistant to the intracellular Ca2+ chelator BAPTA-AM (0,0’ -Bis (2-aminophenyl) ethyleneglycol-N,N,N’,N’-tetraacetic acid, acetoxymethyl ester), yet requires PKG, which must therefore become active independently of Ca2+ . On the other hand, eliciting cGMP signaling through zaprinast is sufficient to trigger all constituent events of gametogenesis, including those that also require Ca2+ (McRobert et al., 2008), while raising calcium with an ionophore is not sufficient to activate gametocytes (Kawamoto et al., 1993b).
A parasite receptor for xanthurenic acid has not been identified, but intriguingly, XA was found to stimulate guanylyl cyclase (GC) activity in membrane preparations from P. falciparum gametocytes (Muhia et al., 2001), suggesting a short link into cGMP signaling. The major source of GC activity in gametocyte membranes was shown to be GCβ, but surprisingly, this cyclase is dispensable for gametocyte activation in both P. falciparum (Taylor et al., 2008) and P. berghei (Hirai et al., 2006). The critical source of cGMP in gametocytes must therefore be GCα, the other guanylyl cyclase encoded in the Plasmodium genome (Baker and Kelly, 2004). Negative regulation of cGMP in P. falciparum gametocytes is achieved by a cGMP-specific phosphodiesterase, PDE, the targeted deletion of which leads to degeneration of late stage gametocytes, probably due to premature activation of PKG (Taylor et al., 2008) (Fig. 4).
Following fertilization, the zygote undergoes meiosis and differentiates into a motile stage, the ookinete, which starts gliding through the blood meal (Fig. 4). It eventually penetrates the peritrophic matrix and the midgut epithelium before differentiating into an oocyst. Gliding motility in P. berghei ookinetes does not require extracellular calcium but is blocked by chelating intracellular calcium with BAPTA-AM (Ishino et al., 2006). Gliding speed and transmission are reduced in ookinetes lacking PbCDPK3 (Ishino et al., 2006; Siden-Kiamos et al., 2006) (Fig. 4). PbCDPK3 appears to be cytosolic in ookinetes (Ishino et al., 2006) and lacks the acylation motif that targets PfCDPK1 to the plasma membrane, although both kinases may regulate gliding differently. In fact, an analysis of the ookinete microneme proteome identified PbCDPK1 and PbCDPK4 (Lal et al., 2009), suggesting that different CDPKs may each regulate different aspects of ookinete biology (Fig. 4). A phenotype very similar to the PbCDPK3 mutant was observed in a mutant lacking the cGMP producing guanylyl cyclase GCβ(3 (Hirai et al., 2006) (Fig. 4). Similar interactions between calcium and cGMP dependent pathways may thus regulate gametocyte activation and ookinete gliding, although some of the enzymes involved are clearly stage specific.
Once they have reached the hemocoel side of the midgut epithelium ookinetes form oocysts, which grow, undergo meiosis and eventually release thousands of haploid sporozoites (Fig. 4). Sporozoites are the most versatile life cycle stage of Plasmodium. They require a repertoire of different behaviors to complete a long journey from the salivary glands of the mosquito into the dermis and eventually into the liver of the vertebrate host (Fig. 4). Sporozoites delivered into the skin by an infected mosquito start gliding in response to environmental stimuli, which are thought to include serum albumin and a rise in temperature (Vanderberg, 1974) (Fig. 4). Gliding and the ability to traverse cells in the dermis by means of putatively pore-forming, secreted proteins is important for migration out of the dermis and entry into the bloodstream, from whence they are carried to the liver (Amino et al., 2006; Ishino et al., 2006). To differentiate further, sporozoites need to switch their behavior from disrupting cell membranes to forming a tight moving junction with a hepatocyte, as the sporozoite migrates into a nascent parasitophorous vacuole (reviewed in (Menard et al., 2008)). Sporozoites have long been known to bind heparan sulfate proteoglycans (Frevert et al., 1993), but only recently has it been shown that the particularly high level of sulfation typically found in the liver provides an environmental signal that switches parasite behavior from traversal to invasion (Coppi et al., 2007). Two signaling genes were identified in P. berghei, neither of which are required for sporozoite gliding, nor cell traversal, but instead are involved in switching between these modes. One is PbCDPK6, an atypical Ca2+ dependent protein kinase with a large N-terminal extension and EF hands N- and C-terminal to the kinase domain (Fig. 3). PbCDPK6 has putative orthologues in other apicomplexa and is expressed from the oocyst stage onward (Coppi et al., 2007). The other mutant with a switching phenotype lacks a membrane bound adenylyl cyclase, ACα (Ono et al., 2008), which was first described from P. falciparum gametocytes (Muhia et al., 2003). Together with pharmacological evidence, the analysis of gene knockout mutants supports the hypothesis that Ca2+ and cAMP dependent pathways act together to regulate sporozoite infection in the liver.
Several potential mechanisms to fine-tune CDPK responses in vivo may have given rise to the independent expansion of this family of protein kinases in plants and apicomplexans, thus providing means to decode calcium signals in a temporal and spatially specific manner. As such, CDPKs play a role similar to the protein kinase C family, which is a major group of calcium and lipid responsive kinases in other eukaryotes. The versatile CDPK family allows changes in a single second messenger, intracellular Ca2+, to regulate a multitude of diverse cellular events in Apicomplexa. CDPKs likely control calcium-dependent processes such as microneme secretion, cytoskeletal dynamics, and regulation of motor complexes, and hence influence gliding motility, cell invasion and egress. CDPKs also regulate developmental transitions, as shown best in the malaria cycle, although the precise cellular pathways that they control remain to be defined. This is further enhanced by crosstalk between calcium-dependent kinases and cyclic nucleotide-dependent pathways, and the tandem nature of these two pathways is an emerging theme throughout apicomplexan life cycles.
Future studies will be important to define the biochemical features of individual CDPKs including responsiveness to calcium, cellular localization, stage specific expression, and substrate preferences. It is possible that individual CDPKs play multiple roles throughout the life cycle, and this could be further addressed using regulated expression systems. Although much of the underlying biology is conserved in apicomplexans, they also differ substantially in some attributes. Hence, further studies will be necessary to validate the extent of similarity vs. divergence in calcium-regulated pathways. Chemical genetic approaches may aid in untangling additional activities of CDPKs, provided that selective inhibitors can be developed and their specificity validated in vivo. Structural information about the conformational changes that occur between inactive and active forms of the kinase would be a major advance in understanding the regulation of this unique family of protein kinases. Identification of targets of these kinases in vivo, as well as mapping downstream signaling pathways, will ultimately be necessary to complement the genetic approaches that have been thus far been insightful in revealing the importance of CDPKs in apicomplexan biology. Finally, untangling the complex relationship between cyclic nucleotide signaling and calcium dependent kinases will be necessary to complete the picture of how calcium regulates complex biological pathways in apicomplexans.
All three authors contributed to writing of the text, S.L. and L.D.S. performed the phylogenetic analysis, and S.L. produced the illustrations. Partially supported in part by NIH grant AI34036 to LDS. SL was partially supported by a Morse Berg Predoctoral Graduate Student Fellowship.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.