2017 Cyclic nucleotide signalling in malaria parasites PDF

Cyclic nucleotide signalling in malaria parasites rsob.royalsocietypublishing.org David A. Baker1, Laura G. Drought1,†, Christian Flueck1, Stephanie D. Nofal1, Avnish Patel1, M...

Cyclic nucleotide signalling in malaria parasites rsob.royalsocietypublishing.org David A. Baker1, Laura G. Drought1,†, Christian Flueck1, Stephanie D. Nofal1, Avnish Patel1, Maria Penzo1,2 and Eloise M. Walker1 1 Faculty of Infectious and Tropical Diseases, London School of Hygiene & Tropical Medicine, Keppel Street, London WC1E 7HT, UK 2 Review Tres Cantos Medicines Development Campus, Diseases of the Developing World, GlaxoSmithKline, Severo Ochoa 2, Tres Cantos, 28760, Madrid, Spain Cite this article: Baker DA, Drought LG, DAB, 0000-0002-5490-8933; CF, 0000-0001-9145-9324; AP, 0000-0002-3608-3405 Flueck C, Nofal SD, Patel A, Penzo M, Walker The cyclic nucleotides 30 , 50 -cyclic adenosine monophosphate (cAMP) and EM. 2017 Cyclic nucleotide signalling in 30 , 50 -cyclic guanosine monophosphate (cGMP) are intracellular messengers malaria parasites. Open Biol. 7: 170213. found in most animal cell types. They usually mediate an extracellular http://dx.doi.org/10.1098/rsob.170213 stimulus to drive a change in cell function through activation of their respect- ive cyclic nucleotide-dependent protein kinases, PKA and PKG. The enzymatic components of the malaria parasite cyclic nucleotide signalling pathways have been identified, and the genetic and biochemical studies of Received: 15 September 2017 these enzymes carried out to date are reviewed herein. What has become Accepted: 28 November 2017 very clear is that cyclic nucleotides play vital roles in controlling every stage of the complex malaria parasite life cycle. Our understanding of the involvement of cyclic nucleotide signalling in orchestrating the complex biology of malaria parasites is still in its infancy, but the recent advances in our genetic tools and the increasing interest in signalling will deliver Subject Area: more rapid progress in the coming years. biochemistry/cellular biology/molecular biology/microbiology 1. Introduction Keywords: Malaria parasites are fascinating organisms to study; research efforts, however, malaria parasites, anopheles, cyclic nucleotides, are primarily motivated by their devastating impact on humanity each year. phosphodiesterase, cyclase, Plasmodium Although the numbers of deaths and clinical cases have fallen dramatically over the last 15 years, mainly due to successful mosquito control programmes and effective drug treatments, there are still around 200 million cases of malaria resulting in over 400 000 deaths each year. Around 70% of these deaths are in Author for correspondence: children under 5 years old in Africa infected with Plasmodium falciparum David A. Baker who have not built up sufficient natural immunity. Worryingly, the spread of e-mail: [email protected] insecticide-resistant mosquito vectors and drug-resistant parasites threatens to reverse the promising recent trend. The rate at which these organisms develop resistance vastly exceeds our ability to develop and bring new interventions to the field. There are a small number of very promising new antimalarial drugs that are undergoing clinical testing, some of which have been identified by phe- notypic screening of compound libraries. However, the attrition rate in drug discovery is very high and so the pipeline needs to remain productive. Trends in Pharma drug discovery seem to be shifting back towards target-based approaches, which emphasizes the importance of understanding the biology of malaria parasites at the molecular level and the identification of essential bio- chemical pathways that can be targeted with novel, small molecule inhibitors. The malaria parasite life cycle is complex, with extended periods in the human host and also in the Anopheles mosquito vector which transmits disease † Present address: Malaria Programme, between people (figure 1). When an infected female mosquito bites a person, sporozoites are injected into the bloodstream which travel to the liver where Wellcome Trust Sanger Institute, Cambridge they invade hepatocytes. Here they divide asexually and eventually the hepato- CB10 1SA, UK cyte ruptures to liberate thousands of merozoites from each infected cell. These merozoites must then invade red blood cells where they again replicate & 2017 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.  2 rsob.royalsocietypublishing.org ocyte gamet ility gam etoc ab deform ytog ene sis PKA sis ene P KA tog ame G g PK ss n re tio eg lla PDEd ite Open Biol. 7: 170213 ge o 2 G roz fla PK PK a me ex 5 CD GC PK CD gony a 4 Ed GC zation PK zo PD schi CD fertili b PDE blood stage asexual cyclic nucleotide and development GCb ookinete development Ca2+ signalling in the CDPK7 malaria parasite life and motility CDPK3 ACb PKG cycle PDEd PKA ACb G 1 PK PK inva ocyte a CD GC sion oo CD thr cys evelo PK A ery PK 4 t fo pm d PD rm ent Eg ati on ACa r and ive G dl PK n a t CDPK4 ion en spo vas opm ro in vel gla zoit yte de nd e sa PKG o c e inv liv PKA at ag asi ary on hep st sporozoite motility and liver cell invasion Figure 1. Stage-specific involvement of cyclic nucleotide signalling components and calcium-dependent protein kinases. A schematic depicting all the key stages of the complex P. falciparum life cycle. The inner circles depict which life-cycle stages have been associated with specific cyclic nucleotide signalling components and calcium-dependent protein kinases. asexually by a process called schizogony whereby multiple haploid male gametocyte within around 10 min of activation, daughter cells develop within a vacuole inside the infected allowing them to fertilize female gametes in the insect erythrocyte. This blood stage cycle is responsible for all the midgut. The resulting zygote develops into a motile ookinete pathology associated with malaria and, in the case of P. falci- which burrows through the mosquito midgut epithelium to parum, takes 48 h. The synchronous rupture of mature form an oocyst. Asexual replication produces thousands of schizonts to release the invasive merozoites causes the sporozoites, which when mature are released from the characteristic regularly spaced fevers associated with this oocyst and migrate to the mosquito salivary glands to con- infection. A subpopulation of the asexual blood stage para- tinue the cycle. It is clear that this complex life cycle must sites develops into male and female sexual precursors, be tightly regulated to allow appropriate development and called gametocytes, which mature in around 10 days. To con- survival of all these highly specialized parasite forms in the tinue the life cycle, gametocytes must then be taken up by a varied environments they encounter. This has led researchers feeding mosquito where, following activation, they round up to investigate the signalling pathways that might be respon- and gametes emerge from their host red blood cells. Male sible for coordinating the timely progression of this intricate gametogenesis involves a visually and metabolically specta- life cycle. This review will focus on current knowledge of cular process known as exflagellation. Here, eight highly the role of cyclic nucleotide signalling in regulating malaria motile, flagellated gametes are generated from a single parasite development.  3 red blood cell rsob.royalsocietypublishing.org parasitophorous vacuole GCa PDEb PDEa cAMP cGMP GMP GMP AMP ACb ARO PI Open Biol. 7: 170213 PKG PKG PI4K rhoptries Epac Ac PI4P PK Ar nucleus ER PK micronemes PIP5K PKAc /NCS exonemes PKAr PIP2 CDPK5 IP3R Ca2+ PLC IP3 K1 CDP DAG Figure 2. Localization of the key Plasmodium cyclic nucleotide signalling components in a merozoite prior to egress. A schematic showing the key cyclic nucleotide signalling players in a merozoite within a P. falciparum blood stage schizont. Some of the depicted cellular locations of pathway components are speculative (such as PDEa, PDEb and GCa). Evidence suggests that a subpopulation of PKG is localized in the ER membrane, but the means of attachment is unknown. A rhoptry localization is depicted for ACb based on data obtained from Toxoplasma gondii, where the location of the ACb orthologue in the rhoptries is known to be dependent on the presence of ARO. PKA is depicted as being rhoptry localized based on P. falciparum epitope tagging studies. Also speculative is the existence of an (as yet) unidentified IP3-gated calcium channel (IP3R) for which there is pharmacological evidence. Cyclic nucleotides are intracellular messenger molecules. approaches (figure 2 and table 1). Although genetic manipu- Cyclic AMP (cAMP) and cyclic GMP (cGMP) are synthesized lation of P. falciparum remains a significant challenge, the by adenylyl and guanylyl cyclase from ATP and GTP, respect- tools available have improved greatly in recent years and so ively. Cyclic nucleotide signalling plays diverse functions advances in our understanding of these parasite signalling across the phylogenetic tree. In mammals, for example, pathways will accelerate in the coming years. Functional cAMP has multiple roles ranging from auditory function to studies of the key enzymatic components also facilitate our mediating hormone action. cGMP is a central player in pro- interpretation of pharmacological studies and in some cases cesses such as cardiac function and light detection in the eye. necessitate a re-evaluation of the data. In single-celled eukaryotes, again cyclic nucleotides have myriad roles ranging from differentiation in Dictyostelium to motility in ciliates. cAMP and cGMP are hydrolysed 2. Cyclase enzymes are distinct from those by cyclic nucleotide phosphodiesterases (PDEs), which inacti- vate the messenger molecules. When cellular cyclic nucleotide of other organisms concentrations reach threshold levels, following activation It has been reported that the properties of adenylyl cyclase of a cyclase/inactivation of a PDE, they stimulate cAMP- activity derived from P. falciparum and the infected human dependent protein kinase (PKA) or cGMP-dependent protein red blood cell are distinct; for example, the host enzyme kinase (PKG). The only other apparent cyclic nucleotide is activated by forskolin, whereas that of the blood stage effector in malaria parasites is a molecule with putative parasite is not. Malaria parasites have two adenylyl cAMP-binding sites (PF3D7_1417400), designated PfEpac cyclases and two guanylyl cyclases that are topologically differ- [5,6]. Prior to the availability of the full P. falciparum genome ent to those of other organisms outside of the alveolate sequence, identification of some of these key enzymatic com- superphylum. Although it is a predicted integral mem- ponents in malaria parasites was a laborious process, but brane protein, adenylyl cyclase alpha (ACa) is quite distinct bioinformatic approaches have facilitated the task. Subsequent from the membrane-associated mammalian G protein- steps involved demonstration of biochemical activity and also dependent adenylyl cyclases. It has an apparently bifunctional functional analysis, which is still under way using genetic structure with a C-terminal catalytic domain containing all the Table 1. Phenotypes of mutants corresponding to pathway components. Details of the mutants that have been generated to date for the key players in malaria 4 parasite cyclic nucleotide signalling (n.a., not applicable; n.d., not determined). rsob.royalsocietypublishing.org type of blood stage sexual stage sporozoite/liver protein description mutant phenotype phenotype stage phenotype references adenylyl cyclase a P. falciparum no mutant n.a. n.a. n.a. (PF3D7_1404600) P. berghei gene deletion none none dramatically reduced apical (PBANKA_1037500) exocytosis in response Open Biol. 7: 170213 to liver cell transit adenylyl cyclase b P. falciparum no mutant n.a. n.a. n.a. (PF3D7_0802600) P. berghei gene deletion essential * n.a. n.a. (PBANKA_1227600) guanylyl cyclase a P. falciparum no mutant n.a. n.a. n.a. (PF3D7_1138400) P. berghei gene deletion slow * [10,11] (PBANKA_0910300) guanylyl cyclase b P. falciparum gene deletion none none n.d. (PF3D7_1360500) P. berghei gene deletion slow * ookinete motility and [6,11,13] (PBANKA_1136700) mosquito midgut invasion defect phosphodiesterase a P. falciparum gene deletion none n.d. n.d. (PF3D7_1209500) P. berghei gene deletion none (PBANKA_1019350) phosphodiesterase b P. falciparum no mutant n.a. n.a. n.a. (PF3D7_1321500) P. berghei gene deletion essential * (PBANKA_1419800) phosphodiesterase g P. falciparum gene deletion none none n.d. (PF3D7_1321600) P. berghei no mutant n.a. n.a. n.a. (PBANKA_1419900) P. yoelii gene deletion peak blood stage none reduction in salivary gland (PY17X_1421600) parasitaemia sporozoites (55%) and reduced by 50% substrate-dependent gliding defect phosphodiesterase d P. falciparum gene deletion none gamete emergence n.a. (PF3D7_1470500) defect (Continued.) Table 1. (Continued.) 5 rsob.royalsocietypublishing.org type of blood stage sexual stage sporozoite/liver protein description mutant phenotype phenotype stage phenotype references P. berghei gene deletion none ookinete motility and n.a. (PBANKA_0835600) mosquito midgut invasion defect cAMP-dependent protein kinase catalytic subunit P. falciparum no mutant n.a. n.a. n.a. Open Biol. 7: 170213 (PF3D7_0934800) P. berghei gene deletion essential * (PBANKA_0835600) cAMP-dependent protein kinase regulatory subunit P. falciparum over-expression growth rate reduced stiffness of [5,16] (PF3D7_1223100) reduced immature by 78% gametocytes P. berghei no mutant n.a. n.a. n.a. (PBANKA_1438000) cGMP-dependent protein kinase P. falciparum inhibitor- block in schizont block in gamete egress n.d. [17,18] (PF3D7_1436600) resistant rupture mutant ** P. falciparum FKBP DD block in schizont n.d. n.d. (PF3D7_1436600) knockdown rupture P. berghei inhibitor- block in schizont block in gamete egress, n.d. (PBANKA_1008200) resistant rupture ookinete motility mutant ** defect P. berghei conditional gene n.a. n.a. block in late liver (PBANKA_1008200) excision stage development Epac-like protein P. falciparum truncation none n.a. n.a. (PF3D7_1417400) P. berghei no mutant n.a. n.a. n.a. (PBANKA_1025300) *PlasmoGem phenotype **in conjunction with specific kinase inhibitor canonical residues required for cAMP synthesis and six pre- deletion, nor was subsequent development in the mosquito, dicted transmembrane domains, which have significant up to and including sporozoite growth and gliding. How- sequence identity to voltage-gated Kþ channels. Expression in ever, mutant sporozoites lacking this gene were unable to Dictyostelium and Xenopus confirmed the predicted adenylyl undergo apical regulated exocytosis in response to transit of cyclase activity , while activity of the predicted voltage- sporozoites through hepatocytes (a prerequisite for infection gated Kþ channel has so far not been confirmed. However, to be established). Mutant sporozoites also showed a 50% in the related ciliate Paramecium, which has a gene with the reduction in invasion of a mouse hepatoma cell line. same configuration, it was demonstrated prior to gene sequen- The study also reported that pharmacological elevation of cing that a purified native adenylyl cyclase has the properties of cAMP levels and inhibition of PKA activity, respectively, a voltage-gated Kþ channel. It will be important to under- stimulated or reduced apical regulated exocytosis in both P. stand the physiological significance of a malaria parasite yoelii and P. falciparum sporozoites. Interestingly, in light of adenylyl cyclase enzyme which might transport Kþ ions to the predicted voltage-gated Kþ channel domain of ACa, regulate cAMP synthesis. this study found that extracellular Kþ is also required for ACa (PBANKA_1037500) has been deleted in the P. ber- sporozoite exocytosis. Together these results point to a role ghei rodent malaria parasite. Blood stage parasite growth for cAMP signalling in apical regulated exocytosis in sporo- and gametocyte production was not affected by gene zoites. The involvement of intracellular messengers in albumin-stimulated sporozoite gliding was examined using presence of an interacting partner (AIP) of the armadillo 6 pharmacological agents in a subsequent study. It was con- repeats only protein (ARO). The latter is also known to be rsob.royalsocietypublishing.org cluded that cAMP and PKA activity stimulated gliding and targeted to the rhoptries in P. falciparum, where it is attached that elevation of cAMP levels could bypass the need for to the cytosolic face of the rhoptry membrane by dual acyla- addition of albumin. Using the PDE inhibitors IBMX tion. It has also been observed that the P. falciparum and Zaprinast, the authors found that only the former stimu- PKA regulatory subunit has a similar location to PfARO lated sporozoite gliding and concluded that cAMP rather (PF3D7_0414900) and is also probably targeted to the rhop- than cGMP signalling is involved in sporozoite motility. Cur- tries. Interestingly, PfARO is phosphorylated in a PKG- iously, it has previously been shown that IBMX does not dependent manner at Ser33 and Ser36 , suggesting this inhibit PDE activity in P. falciparum blood stage parasites may be a focal point for crosstalk between these two [16,29], but it is possible that it can inhibit PDE activity in pathways during merozoite invasion. sporozoites from rodent malaria parasites. The negative The two malaria parasite guanylyl cyclases, GCa and Open Biol. 7: 170213 result obtained with Zaprinast (which is expected to raise GCb, have unusual bifunctional structures , which are cGMP levels) is seemingly at odds with the more recent only shared by apicomplexans and ciliate relatives such as observation that PKG activity is required for sporozoite glid- Paramecium and Tetrahymena. They comprise a pair of ing , but it is possible that the PDE activity in sporozoites C-terminal cyclase catalytic domains and two sets of six is insensitive to Zaprinast. The authors next measured the predicted transmembrane helices, reminiscent of the top- effects of agents that influence mammalian heterotrimeric G ology of mammalian G protein-dependent adenylyl proteins, which in mammals regulate transmembrane adeny- cyclases. Key purine-binding residues in their catalytic lyl cyclases (cholera toxin: stimulatory and mastoparan: domains, however, are clearly characteristic of guanylyl inhibitory), to investigate whether this could be the mechan- cyclases, and heterologous expression of the P. falciparum ism that regulates cAMP synthesis during sporozoite gliding. GCb (PF3D7_1360500) catalytic domain and the whole Although the use of both of these agents suggested a role for Paramecium enzyme in insect cells have demonstrated heterotrimeric G proteins in sporozoite gliding, the apparent that they are functional guanylyl cyclases. Curiously, amino absence of the encoding genes in the parasite genome makes acid motifs characteristic of the mammalian C1 AC catalytic the results difficult to interpret. A more recent study investi- domain are found in the C2 domain of the alveolate enzymes gated the effects of cholera toxin and mastoparan on P. and vice versa. Evidence for membrane association was falciparum sexual commitment. It was surprisingly found obtained using immunoelectron microscopy in both of these that mastoparan led to elevated cAMP levels and that cholera studies. The N-terminal region of all the alveolate GCs has toxin had no effect, and it was concluded that the effects of sequence and structural homology with type IV P-type both agents were unlikely to be via interaction with heterotri- ATPases [4,11,36], but no functional activity has so far been meric G proteins. demonstrated for this domain. Interestingly, proteolytic The second adenylyl cyclase, ACb, is an orthologue of the processing of the region linking the two functional domains bicarbonate-sensitive soluble adenylyl cyclases. So far, of the Paramecium GC was observed in Sf9 cells, but whether this gene has proved refractory to deletion in both P. berghei this happens in vivo is not known. It will be intriguing and P. falciparum, suggesting that it may be essential in the to uncover the significance of the genetic linkage of these asexual blood stage. This is consistent with the fact that two distinct enzyme activities in apicomplexan parasites high levels of ACb mRNA were measured in P. falciparum and their free-living relatives. late blood stages. Also, an antibody raised to Early pharmacological studies [37 –39] suggested cGMP P. falciparum PfACb (PF3D7_0802600) reacted with both late as a possible regulator of exflagellation. Xanthurenic acid, schizonts and free merozoites. Two distinct classes of a product of tryptophan catabolism and present in the soluble AC inhibitors (KH7 and 2CE) were found to kill mosquito midgut, is known to stimulate exflagellation P. falciparum blood stage parasites with LD50 values of [40,41]. Later it was shown that addition of xanthurenic 8.5 mM and 60 mM, respectively , further suggesting that acid to crude membrane preparations from P. falciparum PfACb is essential in blood stages. Experiments using syn- gametocytes led to increased levels of cGMP, suggesting chronized cultures showed that parasites died when KH7 that xanthurenic acid may stimulate guanylyl cyclase acti- was added during the first 24 h of the cycle, but were most vity (directly or indirectly) , again linking this pathway sensitive between 24 and 31 h post-invasion. Expression of with exflagellation. the N-terminal 785 amino acids of PfACb (containing the pre- Transcriptomic studies have shown that the patterns of dicted catalytic domains) in insect cells did not allow mRNA expression for PfGCa (PF3D7_1138400) and PfGCb investigation of bicarbonate sensitivity of the enzyme, but it are very different (http://plasmodb.org), suggesting they each did reveal a distinct pH sensitivity (optimum 7.5), which have roles in distinct life-cycle stages. Two studies have reported led the authors to hypothesize that PfACb acts as a pH deletion of GCb (PBANKA_1136700) in P. berghei which sensor during blood stage development. Exposure of mero- revealed a role in ookinete motility and mosquito midgut inva- zoites to a buffer containing low levels of Kþ, as would be sion [6,13]. Exflagellation occurred at a similar rate in mutant encountered on entering the plasma post egress, triggered and wild-type lines, indicating that GCb is not essential for calcium release and discharge of AMA1 (PF3D7_1133400) this process in P. berghei. Similar findings were obtained follow- from the micronemes. The buffer also triggered bicarbonate ing disruption of the P. falciparum GCb in both the cyclase ion-sensitive cAMP synthesis that was reduced by addition domain and the P-type ATPase domain. Gametogenesis was of KH7 suggesting a key role for cAMP synthesis by PfACb able to proceed normally despite the fact significantly reduced in these events [5,34]. It has been shown in Toxoplasma that guanylyl cyclase activity was measured in mutant gametocyte ACb is located on the outer face of the rhoptries facing the membrane preparations. It was deduced that the residual cytosol and its rhoptry association is dependent on the GC activity was due to the presence of GCa, which may also be expressed in gametocytes, and that this was sufficient for extremely valuable tool (table 2) in a number of subsequent 7 gametogenesis to occur normally. GCa has proved refractory studies on cyclic nucleotides in apicomplexan parasites, and rsob.royalsocietypublishing.org to deletion in both P. berghei and P. falciparum, suggesting has helped provide proof of concept that malaria parasite an essential role in blood stage development [6,11]. PDEs are valid drug targets. The effects of analogues of the human PDE5 inhibitor, Tadalafil, on malaria parasites have also been reported, with EC50 values down to the sub-micro- 3. The parasite phosphodiesterases are all molar level. Progress towards achieving selectivity for the parasite PDE by reducing activity against human PDE5 membrane-associated was also made by the same team by modifying the piperonyl There are four cyclic nucleotide phosphodiesterase enzymes group. Another sub-micromolar inhibitor of malaria (PDEa-d) encoded in the malaria parasite genomes, but parasite PDEs, BIPPO, derived from a human PDE9 series, curiously there are potentially 18 in Toxoplasma gondii has an EC50 of 0.4 mM against P. falciparum proliferation Open Biol. 7: 170213 (TGGT1 strain; http://toxodb.org/toxo/). The malaria para- and an IC50 of 150 nM against recombinant PDEa, but still site PDEs are predicted to have up to six transmembrane retains an IC50 of 30 nM against human PDE9. Nonetheless, domains, suggesting they are integral membrane proteins. this compound will also be a very useful tool for studying The mammalian PDE superfamily is much more complex cyclic nucleotide signalling in the parasite and could be an than that of Plasmodium. The mammalian PDEs are divided important lead for drug discovery. into 11 functionally divergent but structurally similar gene PDEa has been deleted in P. falciparum, showing that it is families, most containing several different genes. These genes not essential for blood stage growth in vitro. Phenotype are frequently spliced or transcriptionally processed to give analysis suggested that ring-stage parasites appeared earlier over 100 PDE isoforms. One family (PDE3) contains a in synchronized cultures of mutant parasites compared single transmembrane domain, but, unlike Plasmodium PDEs, with wild-type, but it was not clear whether this was related no mammalian PDE contains multiple membrane spanning to the absence of the PDEa gene. There was no evidence of domains. Mammalian PDEs contain 15 conserved residues, upregulation of mRNA levels of the other three PDEs to com- thirteen of which are conserved in all four PfPDE enzymes pensate for the deleted gene. The study also highlighted the. A number of transcriptional studies (available via existence of two splice variants of PfPDEa (A and B), one http://plasmodb.org) have shown that the four P. falciparum of which encodes a protein with four predicted TM domains PDEs exhibit life cycle stage-specific patterns of expression. rather than six, but the significance of this is not known. PDEa (PF3D7_1209500) and PDEb (PF3D7_1321500) are cGMP-PDE activity in membrane fractions of the PDEa the predominant isoforms expressed in blood stage parasites mutant was 20% lower than those from wild-type parasites, and peak at the late trophozoite and schizont stages, while with no effect on cAMP-PDE activity. This is consistent PDEg (PF3D7_1321600) and PDEd (PF3D7_1470500) are with the findings of the recombinant expression studies that predominantly expressed in sporozoite and gametocyte PDEa is cGMP specific. Interestingly, the authors also stages, respectively. found that addition of excess cGMP to blood-stage membrane A modelling study of the malaria parasite PDEs has been fractions from wild-type and PDEa knockout parasites greatly carried out using the human PDE9A crystal structure as a reduced cAMP hydrolytic activity in PDE assays, suggesting template. Interestingly, the model could accommodate that one of the other PDEs expressed in blood stages is likely cGMP binding, but not cAMP binding, but as the authors to be dual specific. PDEb, which has proved refractory to say, roles for cAMP and PKA in the parasite have been estab- deletion in P. falciparum and P. berghei , is the most likely lished and so at least one of the four PfPDEs is likely to be candidate for this activity, but this is yet to be demonstrated. able to hydrolyse cAMP and perhaps be dual specific. To PDEd has been deleted in both P. falciparum and P. date, it has been possible to measure catalytic activity for berghei , showing that this gene is dispensable for blood only PDEa (PfPDE1) by expressing the C-terminal catalytic stage replication in vitro. Deletion of PDEd in P. falciparum domain in Escherichia coli [14,29]. This has shown that the led to a 50% reduction in native cGMP hydrolytic activity enzyme specifically hydrolyses cGMP. The catalytic specifici- in gametocytes compared with wild-type, with a correspond- ties of PDEg and PDEd for cGMP have been deduced from ing twofold increase in cellular cGMP levels in late-stage phenotype analysis of null mutants (but require confirmation; gametocytes. cAMP-PDE activity levels were consistently see below) while that of PDEb remains unknown as it is the low, and indistinguishable between wild-type and mutant only one refractory to deletion. Expression of PDEa also parasites. Interestingly, addition of Zaprinast to wild-type allowed analysis of its sensitivity to some commonly used gametocyte membrane fractions reduced cGMP hydrolytic PDE inhibitors. Importantly, it was shown that both IBMX activity by around 75%, but it was totally ablated in PDEd2 and theophylline, two PDE inhibitors that have been used mutants. This suggests that the residual PDE activity is in a number of studies on malaria parasites and have little more sensitive to Zaprinast than PDEd. Together these results or no activity against Plasmodium PDEs with EC50. indicate that PDEd has cGMP hydrolytic activity and that at 100 mM [14,16,29]. It is important to note, in this respect, least one other cGMP-PDE is active in gametocytes. Although that some studies involving parasite PDEs use preparations PDEd mRNA and protein are present in gametocytes, their in which the host red cell has not been removed and so development up to maturity is not affected by the absence human PDEs will be present, whereas some studies use sapo- of this gene. However, upon stimulation of gametogenesis nin lysis to remove the host cell material. Yuasa and with xanthurenic acid, the ability of mutant parasites to colleagues were able, however, to identify Zaprinast as a round up was significantly reduced, and exflagellation of low micromolar inhibitor of recombinant PDEa and native male gametocytes was reduced dramatically. Examination PDE activity in P. falciparum blood stage parasites as well of mutant and wild-type parasites by immunofluorescence as blood stage growth. Zaprinast has proved an microscopy using an antibody against the human red cell Table 2. Pharmacological agents used to study cyclic nucleotide signalling in malaria parasites. Details of compounds that have been used to target Plasmodium proteins involved in cyclic nucleotide signalling. The parasite and host targets are indicated, if known, along with the potency reported for parasite or host cell and biochemical/phenotypic effects (n.a., not applicable; n.d., not determined). published potency P.falciparum published potency host compound P.falciparum target host target target cell target biochemical effect published phenotypic effect Compound 1 cGMP-dependent protein n/a [IC50] 0.49 mM asexual blood stage no data avaliable inhibition of parasite cGMP- inhibtion of blood stage schizont rupture , kinase growth dependent protein kinase inhibition of XA stimulated rounding up of gametocytes and gamete emergence Compound 2 cGMP-dependent protein n/a [EC50] 395.0 nM + 21.9 asexual no data avaliable inhibition of parasite cGMP- inhibtion of blood stage schizont rupture , kinase blood stage growth dependent protein kinase inhibition of XA stimulated rounding up of gametocytes and gamete emergence Zaprinast phospho-diesterases phospho-diesterase 5 [IC50] 33.7 + 1.3 mM for [IC50] 1.3-3.2 mM PDE5 elevation of cellular cyclic induction of blood stage egress , induction of gametocytes and 3.0 + 1.2 mM isoforms from cell lysate nucleotide levels rounding up of gametocytes and for schizonts gamete emergence Sildenafil phospho-diesterases phospho-diesterase 5 [IC50] 75.6 mM + 1.6 for [IC50] 3.5 nM PDE5 33 nM elevation of cellular cyclic induction of rounding up of gametocytes and and 6 gametocytes and 22.5 PDE6 nucleotide levels gamete emergence mM + 1.8 for schizonts BIPPO phospho-diesterases n/a [IC50] 0.4 mM + 0.14 asexual no data avaliable elevation of cellular cyclic induction of blood stage egress blood stage growth nucleotide levels IBMX n/a phospho-diesterases [IC50].200 mM asexual blood [IC50] 5.8-26.2 mM elevation of host cell cyclic n/a stage growth nucleotide levels? H89 cAMP-dependent protein cAMP-dependent protein [IC50] 2.9 mM + 1.6 asexual blood Ki 0.05 mM inhibition of cAMP- inhibition of blood stage schizont development kinase catalytic domain kinase catalytic stage growth dependent protein kinase , inhibition of blood stage merozoite domain invasion , induces deformability in stage III gametocytes KT5720 cAMP-dependent protein cAMP-dependent protein [IC90] 10 mM asexual blood stage Ki 60 nM inhibition of cAMP- inhibition of blood stage merozoite invasion kinase catalytic domain kinase catalytic growth dependent protein kinase , induces deformability in stage III domain gametocytes KH7 adenylyl cyclase b soluble adenylate cyclase [IC50] 8.5 mM asexual blood stage [IC50] 10 mM inhibition of cAMP synthesis inhibiton of blood stage schizogony , inhibition growth of blood stage microneme release Forskolin n/a soluble adenylate cyclase no effect on the blood stage [EC50] 2-10 mM stimulation of host cell cAMP n/a enzyme synthesis 8 Open Biol. 7: 170213 rsob.royalsocietypublishing.org membrane protein Band 3 showed that although most high sporozoite doses resulted in successful transmission to 9 mutant parasites could partially round up following stimu- a proportion of mice, to establish a blood-stage infection. rsob.royalsocietypublishing.org lation of gametogenesis, very few parasites had emerged The authors demonstrated that this low level of transmission from the encapsulating red cell. This suggests that elevated was due to the fact that the sporozoites lacking PDEg are levels of cGMP, and perhaps premature PKG activation in defective in substrate-dependent gliding motility. Tran- gametocytes lacking PDEd, have deleterious effects on scriptome analysis of PDEg2 mutant sporozoites showed gamete emergence. However, premature activation of game- upregulation of PDEb and PDEd and downregulation of sev- tocytes would be expected if this were the case, and so eral mRNAs encoding proteins involved in sporozoite further work is needed to understand this phenotype. The motility and hepatocyte invasion. Mutant sporozoites results also indicate that PDEd is important for regulation were found to have an almost 20-fold higher cGMP concen- of cGMP levels during gametogenesis, and that both male tration than wild-type sporozoites, strongly suggesting this and female gamete emergence is affected. This study also to be the cause of the defective gliding phenotype. Consistent Open Biol. 7: 170213 reported that PDEg, which is also expressed in gametocytes with this assertion, the PDE inhibitor Zaprinast reduced spor- at the mRNA level (PlasmoDB.org), was dispensable for P. ozoite gliding in a dose-dependent manner. The authors falciparum blood-stage growth and gametocyte development point out that according to these and previous observations, in vitro, but in contrast to PDEd2 mutants, there was no it is possible that cGMP and cAMP may have opposing effect on gametogenesis. effects on sporozoite function and that crosstalk between Perhaps surprisingly, the phenotype was different in P. the two pathways could occur. berghei; deletion of PDEd had no effect on gamete emergence. This suggests that another PDE might be able to compen- sate for its absence in P. berghei by maintaining cGMP 4. PKG is essential in all of the key stages concentrations in mature gametocytes below the level required for PKG activation. However, deletion of PbPDEd of the malaria parasite life cycle had a dramatic effect on ookinete development. During the The apicomplexan PKGs have both structural and biochemi- first 12 h of development, PDEd2 parasites progressed to cal properties that distinguish them from mammalian the expected banana-shaped ookinetes, but by 24 h they isoforms. For example, apicomplexan PKGs have four recog- had ‘dedifferentiated’ into rounded or stumpy forms. This nizable cGMP-binding domains (A, B, C, D); only three of resulted in a more than 94% decrease in the number of these are functional and domain C is degenerate. By contrast, oocysts that developed subsequently, compared with wild- mammalian PKGs have only two cGMP-binding domains. In type parasites. Interestingly, if these mutant ookinetes were the absence of cGMP, PKG adopts a globular configuration injected into the haemocoel of the mosquito, transmission where a pseudo-substrate, autoinhibitory sequence occupies was restored to wild-type levels. Therefore, the mutant ooki- the active site and so the enzyme is kept in an inactive state netes were viable, but could not cross the epithelium of the (figure 3). On binding cGMP, the conformation becomes mosquito midgut due to a lack of forward gliding movement. more open and extended, and the autoinhibition is released, This abnormality was explained by some marked ultrastruc- thereby exposing the active site to allow phosphorylation of tural abnormalities (gaps in the inner membrane complex substrate proteins. In P. falciparum, cGMP-binding site D and misorientation of the underlying subpellicular microtu- has by far the highest affinity for cGMP and selectivity over bules) rather than a lack of motor activity. Interestingly, cAMP. Crystallography carried out on this binding site the dedifferentiation phenotype could be reversed by has identified the key determinants of cGMP binding and addition of a specific PKG inhibitor, confirming that inap- identified a distinct capping mechanism. A conditional propriate activation of this enzyme had taken place reverse genetic approach was taken to show that substitution following PDEd deletion resulting in an elevation of cGMP of either of these capping residues is lethal to blood-stage levels. This conclusion was further strengthened by the fact parasites. that GCb2/PDEd2 double mutant ookinetes also had Two PKG isoforms have been identified in the coccidian normal morphology. Another interesting finding was that a parasites Eimeria and Toxoplasma. One of these is dually acy- CDPK32/PDEd2double mutant glided normally ( prior to lated at the N-terminus; modifications that allow membrane dedifferentiation), thereby reversing the CDPK32 mutant association. Mammalian PKGII is membrane associated gliding phenotype. This indicated that elevated signalling by myristoylation. By contrast, in Plasmodium PKGs, through cGMP and PKG could compensate for the absence the acylation motifs are absent and although a subpopu- of this calcium-dependent kinase, which suggests there may lation of the P. falciparum PKG (PfPKG; PF3D7_1436600) be overlapping substrates for these two protein kinases. apparently becomes associated with the ER membrane in Analysis of the localization of PDEg using an epitope late schizonts, the means of attachment is unknown. tagged P. yoelii transgenic line showed low-level expression Mutational analysis to examine the properties of the individ- in blood stages that partially co-localized with a marker of ual apicomplexan cGMP-binding domains has revealed that, the endoplasmic reticulum (ER), but strong expression in unlike mammalian PKGs, they are relatively insensitive to 8- sporozoites. Although deletion of PDEg caused a 50% substituted cGMP analogues [63,64]. Progress with studying reduction in peak blood-stage parasitaemia during the the malaria parasite PKG has been greatly facilitated by the second half of a 20-day mouse infection, the gene is not essen- existence of specific inhibitors which were generated through tial for blood-stage replication. Furthermore, gametocyte and pioneering work by a group at Merck who identified a tri- mosquito-stage development was unperturbed, including substituted pyrrole (Compound 1) in a whole cell screen of sporozoite formation in oocysts. However, there was a 55% E. tenella, a related apicomplexan parasite of chickens. reduction in the number of salivary gland sporozoites in Using a range of biochemical approaches, they found that mutant parasites compared with wild-type, and only very the target of Compound 1 is the E. tenella PKG and a  (a) mammalian PKG (b) Plasmodium PKG 10 autoinhibitory hinge catalytic autoinhibitory hinge catalytic rsob.royalsocietypublishing.org sequence domain domain sequence domain domain CNB-A CNB-B CNB-A CNB-B CNB-C CNB-D dimerization regulatory ATP binding substrate binding regulatory ATP binding substrate binding domain domain domain lobe domain domain lobe C C R inactive inactive PKG R PKG R C Open Biol. 7: 170213 + cGMP + cGMP ATP C ATP R C active active PKG R PKG R C ATP (c) mammalian PKA (d) Plasmodium PKA autoinhibitory ATP binding substrate binding autoinhibitory ATP binding substrate binding sequence domain lobe sequence domain lobe CNB-A CNB-B CNB-A CNB-B dimerization regulatory catalytic regulatory catalytic docking domain domain domain domain domain C C R1 inactive inactive PKA R PKA R2 C + cAMP + cAMP ATP C active ATP R1 C PKA R active PKA R2 ATP C Figure 3. Cartoon of PKG and PKA domain organization and structure in mammals and Plasmodium. (a) Mammalian PKGs are homodimers encoded by single genes where dimerization is mediated by the dimerization domain. The inactive form of the enzyme adopts a conformation such that the autoinhibitory sequence interacts with the substrate-binding lobe, preventing access of substrates to the catalytic site. Binding of four cGMP molecules to the homodimer causes a conformational change, exposing the catalytic the domain to allow phosphorylation of substrates. (b) Plasmodium PKG is also encoded by a single gene; however, it lacks a dimerization domain and evidence suggests it forms a monomer. Plasmodium PKG contains four consensus cyclic nucleotide-binding domains (CNBs). One of these domains (CNB-C) is degenerate, therefore, the binding of only three cGMP molecules is required for activation of the enzyme. (c) Mammalian PKA is a heterotetramer consisting of two regulatory domains and two catalytic domains; however, unlike mammalian PKG, the regulatory and catalytic subunits are encoded by separate genes. The autoinhibitory sequence within the regulatory subunit binds to the substrate-binding lobe in the inactive form. The binding of four cAMP molecules results in the dissociation of the catalytic subunits from the regulatory subunits and activation of PKA. (d) Plasmodium PKA regulatory and catalytic subunits are also encoded by two separate genes; however, the regulatory subunit lacks a dimerization docking domain and the inactive enzyme is thought to form a heterodimer. The binding of two cAMP molecules results in dissociation of the subunits and allows binding of substrate proteins to the catalytic subunit. subsequent medicinal chemistry effort led to the generation it had a very similar structure and properties to the coccidian of a more potent imidazopyridine PKG inhibitor, Compound enzyme [64,67]. The two Merck PKG inhibitors were sub- 2. At a similar time, another study carried out a bio- sequently reported to be potent inhibitors of this malaria chemical analysis of the P. falciparum PKG and showed that parasite PKG [18,45], and this proved crucial to functional  wild-type gatekeeper mutant 11 PKG PKG rsob.royalsocietypublishing.org ATP-binding ATP-binding pocket gatekeeper pocket gatekeeper pocket pocket T Q Open Biol. 7: 170213 Compound 2 binds Compound 2 can’t bind PKG inhibition PKG remains active Figure 4. Mutation of the PKG gatekeeper residue confers inhibitor insensitivity. The ATP-binding pocket of PKG contains a small hydrophobic pocket. Inhibitors such as the imidazopyridine, Compound 2, compete with ATP for binding to PKG, with the fluorophenyl group of the inhibitor interacting with the hydrophobic pocket. Substitution of the threonine gatekeeper residue with a bulkier glutamine residue prevents binding of Compound 2 but the kinase remains functional with ATP still able to bind. studies when used in conjunction with an inhibitor-resistant genetic approach, a recent study showed that CDPK4 is the transgenic line. This chemical genetic approach relies on the major effector of calcium signalling downstream of PKG acti- fact that the apicomplexan PKG has a structural feature, vation in male gametogenesis. It dissected multiple windows which is distinct from that of the human PKG and (impor- of kinase activity needed for DNA replication, mitotic spindle tantly) the majority of serine threonine kinases, that allows assembly and completion of cytokinesis to produce eight inhibitor binding. The parasite PKG has a relatively small motile, flagellated male gametes. The same study also ident- amino acid residue (threonine) in the so-called gatekeeper ified distinct CDPK4 substrates that translate each of the position (Thr618 in PfPKG). This is thought to allow access of three bursts of CDPK4 activity into an appropriate cellular the fluorophenyl group, shared by Compound 1 and Com- response. Another recent study implicated an additional pound 2, to a small hydrophobic pocket which adjoins the calcium-dependent protein kinase (CDPK2) in male gameto- kinase ATP-binding pocket. Substitution of the small gate- genesis. While female CDPK2 knockout gametocytes were keeper residue with one with a bulkier side chain (such as able to emerge from the host cell normally, a proportion of fla- glutamine or methionine) prevents access of the inhibitor to gellated male gametes were trapped inside the host red blood the gatekeeper pocket and renders the enzyme insensitive cell. This suggests that these CDPKs might act in concert to it (figure 4). Introducing the gatekeeper mutant PKG into to regulate different aspects of gamete production. the parasite by complementation (Toxoplasma and Eimeria Again using the gatekeeper mutant/inhibitor strategy, [68,69]) or allelic replacement (P. falciparum [17,18] and PKG activity was shown to be required for the rupture of P. berghei ) confers inhibitor resistance to the transgenic blood stage schizonts. Conditional knockdown of CDPK5 line (the degree of resistance conferred varies with species (PF3D7_1337800) caused a similar phenotype where schizonts and life-cycle stage). Comparison of the effects of the inhibitor could not rupture, but whereas PKG inhibition prevented pro- on wild-type and gatekeeper mutant parasites first allowed teolytic processing of key merozoite proteins by the essential demonstration that PKG is the primary target of the com- protease PfSUB1 (PF3D7_0507500) , the CDPK5 pheno- pounds, but importantly also allowed determination of the type was downstream of this proteolytic activity and cGMP role of PKG in specific cellular processes in Toxoplasma signalling. A subsequent study examined the underlying and Plasmodium [17,18,20,30]. In the presence of PKG inhibitor, mechanism by which PKG inhibitors block blood stage schizont wild-type P. falciparum gametocytes did not undergo rounding rupture and merozoite egress. It revealed that inhibition of PKG up and emergence from red cells when stimulated by xanthure- blocked release of PfSUB1 from the apical organelles known as nic acid or the PDE inhibitor Zaprinast (coupled with a exonemes, indicating that cGMP signalling and PKG activity reduction in temperature), whereas gatekeeper mutant para- are required for release of PfSUB1 from these organelles. Fur- sites (expressing the PKG Thr618Gln substitution) could thermore, in PKG inhibitor-treated schizonts, AMA1 could undergo rounding up normally, but exflagellation does not not be released from the micronemes onto the merozoite surface proceed, suggesting the involvement of a second, inhibitor-. A conditional knockdown approach using a fusion with sensitive kinase in this process. This demonstrated a role for the destabilization domain (DD) showed that depletion of PKG in gametogenesis. It is known that calcium signal- PKG led to a block in P. falciparum blood stage egress , a ling is essential for exflagellation , but interestingly, phenotype that is consistent with that obtained with the rounding up could occur normally in the presence of the cal- chemical genetic approach. cium chelator BAPTA-AM, suggesting that PKG activity is Another vital cellular process that was shown to be PKG- upstream of calcium signalling and governs events up to dependent using the gatekeeper mutant approach is calcium and including emergence from the host cell. Using a chemical release from internal stores (figure 2). Elevation of cGMP levels stimulated by zaprinast, which triggers merozoite CDPK5 and CDPK6 (PF3D7_1122800) measured by the 12 egress, was found to stimulate calcium release. This was authors in the mutant line probably indicate that these closely rsob.royalsocietypublishing.org measured by loading schizonts with the calcium-sensitive related kinases compensate for the reduced CDPK1 activity fluorescent dye, Fluo-4 AM. Addition of PKG inhibitor. CDPK7 has been shown to be important for early blocked calcium release in a dose-dependent manner. Analy- blood-stage development and is therefore likely to operate sis of phospholipids by mass spectrometry in P. falciparum independently of PKG. This kinase is unique among the wild-type and gatekeeper mutant schizonts treated simul- CDPKs in having a pleckstrin homology domain that has taneously with zaprinast and Compound 2 showed PKG- been shown to bind to PIP2 and is proposed to guide its sub- dependent changes in phosphoinositide levels. For example, cellular localization. levels of PIP2 (which in higher eukaryotes would be hydro- PKG has also been shown to play an essential role in P. lysed by phospholipase C (PLC) to generate IP3, which berghei ookinetes. The PKG inhibitor Compound 1 prevented would in turn lead to activation of IP3-stimulated calcium ookinete gliding with an EC50 of 100 nM. PKG fused to the Open Biol. 7: 170213 channels leading to calcium release) were reduced by PKG fluorescent reporter GFP indicated a largely cytosolic location inhibition. This study indicated that the process of PKG- in ookinetes. A subsequent study examined in more dependent merozoite egress is mediated by calcium release detail how cGMP signalling and PKG activity might regulate which is linked to phosphoinositide metabolism. ookinete gliding. The study used quantitative mass spec- To date it is not known what acts upstream of cGMP signal- trometry to analyse cGMP-dependent phosphorylation ling to trigger egress, but there is evidence of events prior events using two parallel genetic approaches. The first used to PKG activation, with permeabilization of the parasito- a GCb knockout line and the second used the PKG inhibitor phorous vacuole membrane occurring 10 –30 min before its Compound 2 in conjunction with a P. berghei gatekeeper PKG-dependent breakdown. Interestingly, in Toxoplasma mutant parasite line (expressing a Thr619Gln substitution). it has been shown that the plant hormone abscisic acid Changes in phosphorylation levels in either of the approaches triggers egress in a calcium-dependent manner , but it detected key sites in PKG itself as well as in GCb and PDEd, remains to be seen whether this molecule has a role in showing potential mechanisms for self-regulation of the path- malaria parasites. way. Interestingly, regulation of proteins involved in The gatekeeper mutant and PKG inhibitor have also phosphoinositide metabolism and vesicle trafficking was been used in conjunction with global phosphoproteomics to measured in both of these two approaches in P. berghei. For identify potential PKG substrates in P. falciparum schizonts. example, phosphopeptides identified in three of the four Peptides derived from schizont proteins from PKG inhibi- PI4 K (PBANKA_110940; PBANKA_072200) and PIP5 K tor-treated and untreated cultures of wild-type and (PBANKA_020310) isoforms were reduced significantly, gatekeeper mutant parasites were labelled with isobaric suggesting that phosphoinositide metabolism is required for tags and analysed using quantitative mass spectrometry. ookinete motility and that it is regulated by PKG. The lack A total of 107 phosphosites from 69 individual proteins of any change in phosphosites on PLC suggests that regu- were regulated in a PKG-dependent manner, revealing invol- lation is at the level of sequential phosphorylation of vement of cGMP signalling in a range of cellular processes phosphatidylinositol (PI) rather than at the level of PIP2 including egress/invasion, motor function, signal transduc- hydrolysis. Interestingly, P. berghei ookinetes containing tion, ion and protein transport, and chromatin regulation. a PI4 K (PBANKA_110940) phosphomutant (Ser534Ala) Interestingly, PKG-dependent phosphosites were identified showed a reduced gliding speed compared with wild-type on GCa, PDEb and ACb suggesting feedback loops involving parasites. Furthermore, a P. berghei line expressing HA- cGMP and possible crosstalk with cAMP signalling. Work is tagged PIP5 K (PBANKA_020310) showed a change in distri- under way to determine the physiological significance of a bution from the cell periphery to the cytosol when treated number of these phosphosites in key proteins. This approach with PKG inhibitor. Intriguingly, this PIP5 K has a predicted identified PKG-dependent phosphorylation events in schi- bifunctional structure with an N-terminal domain with hom- zonts, but in many cases it is likely that the proteins are ology to the EF-hand-containing neuronal calcium sensors, a actually phosphorylated by kinases that act downstream of subfamily of which activate guanylyl cyclase in mammals. PKG. One protein that was found to be a direct substrate of The P. falciparum orthologue (PfPIP5 K/NCS; PF3D7_ PKG in schizonts is CDPK1 (PF3D7_0217500), suggesting 0110600) has been shown to be activated by ADP-ribosylation that PKG may play a dual regulatory role in calcium signal- factor (ARF1) to generate PIP2 , providing another poten- ling, stimulating calcium release and direct phosphorylation tial link between phosphoinositide metabolism and vesicular of calcium-dependent kinases. A recent study reported trafficking. Both of these observations strongly suggest a link the successful generation of a P. falciparum CDPK1 gatekeeper between PKG and phosphoinositide metabolism. Analysis of mutant (harbouring a Thr145Met substitution). This line had a ookinete phospholipids by quantitative mass spectrometry in significantly reduced kinase activity, but interestingly had a the presence and absence of PKG inhibitor revealed reduced greater sensitivity to the PKG inhibitor Compound 2 than levels of PI4P and PIP2 synthesis following PKG inhibition. wild-type parasites. This was interpreted as indicating that This in turn suggested a potential link to calcium release PKG was probably compensating for the reduced activity of via hydrolysis of PIP2 (by PLC) to give IP3 (and diacylgly- the mutant CDPK1 enzyme. We believe it is more likely cerol) that in other systems activates calcium channels. An that the increased sensitivity of this line to Compound 2 is IP3-gated calcium channel sequence has not been identified due to the fact that CDPK1 is also a target of this inhibitor in Plasmodium genomes, but it is likely that a protein exists and that because the gatekeeper mutant CDPK1 cannot in the parasite which performs this role. A transgenic P. ber- bind Compound 2, the amount of free compound available ghei line expressing a fluorescent calcium reporter showed to block PKG function is elevated, resulting in a lower EC50 clearly that inhibition of PKG led to rapidly reduced levels value than in wild-type. The increased transcript levels of of cytosolic calcium in ookinetes. Furthermore, the PKG- mediated stimulation of gametogenesis by reduced tempera- genes (figure 3). The inactive enzyme is made up of a cataly- 13 ture and xanthurenic acid was also shown to stimulate tic subunit (PKAc) and a regulatory subunit (PKAr) which rsob.royalsocietypublishing.org calcium release, and this was detectable just 10 s after stimu- contains a pair of cAMP-binding sites. When cAMP binds lation. The P. berghei line expressing gatekeeper mutant to PKAr, the complex dissociates and PKAc is thereby acti- PKG was used to demonstrate that the effects of PKG inhi- vated and can bind to substrate proteins which are then bition were specific. phosphorylated. When bound to PKAr, the PKAc enzyme In addition to its role in blood-stage egress and invasion, is inactive, with part of PKAr binding to the active site via gametogenesis and ookinete motility, PKG has also been a substrate-like autoinhibitory sequence. In mammals, there shown to be essential for late-liver-stage development in are four regulatory subunits (two type R1 and two type P. berghei. As it was not possible to delete the gene, an elegant R2), and a complex of four polypeptides made up of two cat- conditional knockout approach was used, whereby gene exci- alytic subunits (three C types) and two regulatory subunits sion was triggered during mosquito stage development by a constitutes the inactive heterotetrameric state. By con- Open Biol. 7: 170213 drop in temperature. Gene excision was detectable in the trast, in Plasmodium, there are single genes encoding both sporozoite stage. Injection of mice with these sporozoites did PKAc (PF3D7_0934800) and PKAr (PF3D7_1223100) sub- not lead to a subsequent blood-stage infection. Examination units. A unique feature of the malaria parasite PKA of liver stages in HepG2 cell cultures showed that development regulatory subunit sequence is that the dimerization could apparently proceed normally up to 40 h post-infection; domain characteristic of PKAr in other species is absent however, after 72 h, cultures appeared blocked at the late. The sequence of Plasmodium PKAr suggests it may be liver schizont stage with few merosomes (vesicles containing a hybrid of the human R1 and R2 types , and crystallo- merozoites) released. A follow-up study using the PKG inhibi- graphic analysis confirmed that it has features that are tor Compound 1 showed that it blocked sporozoite gliding, characteristic of both human R1 and R2 types. invasion and infection, but in the light of the previous findings In mammals, the localization and function of PKA is this was interpreted as off-target activity. However, a more determined by interaction with a family of A kinase activat- recent study used the gatekeeper mutant approach in P. berghei ing proteins (AKAPs). AKAPs ensure that PKA activity for both PKG and CDPK4 (PF3D7_0717500) in conjunction is targeted to the correct cellular compartment and ensures with either a PKG inhibitor or a bumped kinase inhibitor its specificity of response to discrete cellular stimuli. How- (BKI 1294), used as a CDPK4 inhibitor, to demonstrate that ever, the dimerization/docking domain of PKAr that in PKG is in fact required for sporozoite gliding motility (and other species allows tethering to AKAP is absent in the Plas- that Compound 1 was on target), microneme exocytosis and modium PKAr sequence. To date only one gene encoding hepatocyte invasion, with CDPK4 also having a role in hepato- an AKAP-like protein (PfAKAL; PF3D7_0512900) has been cyte invasion. The findings of this paper are seemingly at identified in P. falciparum, and while no direct interaction odds with the earlier work described above, which suggested with the parasite PKAr could be detected, there was evidence that PKG is not required for sporozoite function. The authors that it binds to Pf14 –3-3 (PF3D7_0818200; which in turn pointed out that this discrepancy is likely to be due to residual binds to PKAr) and also to AMP and cAMP , suggesting PKG protein synthesized prior to conditional excision carried a divergent AKAP mechanism in Plasmodium species. Intrigu- over into the sporozoites. ingly, P. falciparum Rab5A (PF3D7_0211200) and Rab7 It is clear from the studies described above that PKG plays (PF3D7_0903200) proteins have also been identifie