An Update on Development of Small-Molecule Plasmodial Kinase Inhibitors

Malaria control relies heavily on the small number of existing antimalarial drugs. However, recurring antimalarial drug resistance necessitates the continual generation of new antimalarial drugs with novel modes of action. In order to shift the focus from only controlling this disease towards elimination and eradication, next-generation antimalarial agents need to address the gaps in the malaria drug arsenal. This includes developing drugs for chemoprotection, treating severe malaria and blocking transmission. Plasmodial kinases are promising targets for next-generation antimalarial drug development as they mediate critical cellular processes and some are active across multiple stages of the parasite’s life cycle. This review gives an update on the progress made thus far with regards to plasmodial kinase small-molecule inhibitor development.


Introduction
Although significant progress has been made with regards to worldwide malaria control and eradication, this infectious disease continues to have devastating effects, especially in developing countries. An estimated 228 million malaria cases and 405,000 malaria-related deaths were reported globally in 2018 [1]. The ongoing coronavirus (COVID-19) pandemic could also have a profound negative impact on the progress made thus far in the fight against malaria [2]. Patients presenting with fever or requiring malaria treatment are now less likely to visit health care facilities out of fear of contracting COVID-19 [3]. Lockdown periods have disrupted the supply of malaria rapid diagnostic tests, antimalarial drugs and other interventions [3]. A number of antimalarial drugs (e.g., artemisinin, chloroquine and hydroxychloroquine) have also been repurposed for COVID-19 treatment, despite a lack of scientific evidence and approval [3]. This has resulted in a shortage of these drugs and, in the long-run, an increased risk of drug resistance due to improper use of antimalarial monotherapies [3].
Apart from the potential impact of the pandemic, recurring antimalarial drug resistance poses a major threat to malaria control and elimination. P. falciparum, the Plasmodium species responsible for the majority of malaria-related deaths, is resistant to most antimalarial drugs, including the current first-line artemisinin-based combination therapies [1,4]. The second most common human malaria species, P. vivax, has developed widespread resistance to chloroquine [5]. Malaria parasites can also develop cross-resistance to antimalarial drugs from the same chemical class or with the same mode of action, which further exacerbates the problem [6].
The life cycle of the malaria parasite is complex; it consists of an asexual stage which occurs in the human host and a sexual stage which occurs in the mosquito vector [7]. Existing antimalarial drugs are highly stage-specific, with the majority targeting blood-stage parasites for treatment of symptomatic malaria [7]. In order to eliminate and eventually eradicate malaria, the focus needs to shift from mainly providing curative treatment to blocking disease transmission [8]. Achieving this goal requires a new generation of cost-effective antimalarial agents that are safe and well-tolerated in a wide range of recipients, including vulnerable populations such as pregnant women and infants [8]. Since drug development is expensive and can take up to 15 years before reaching the market, it is important to have clear guidelines [8]. The Medicines for Malaria Venture (MMV) published defined criteria for the types of individual molecules (target candidate profiles (TCPs)) and drug formulations (target product profiles (TPP)) that would be ideal for new malaria therapy [8]. TPP-1 focusses on treating malaria infections and includes a combination of molecules with blood-stage activity (TCP-1), transmission-blocking activity (TCP-5) and activity against relapse causing liver stages (TCP-3) [8]. Alternatively, TPP-1 could also consist of rapid-acting TCP-1 molecules when treating severe/complicated malaria [8]. TPP-2 focusses on chemoprevention of travelers to endemic regions or during epidemics and includes molecules with TCP-1 and hepatic schizont activity (TCP-4) [8].
Plasmodial kinases have been explored as targets for next-generation antimalarial agents due to their involvement in various critical cellular processes throughout the life cycle of the parasite [9][10][11]. The P. falciparum kinome is predicted to encode 85 to 99 protein kinase genes [12,13] as well as a small number of lipid kinase genes. Overall, the P. falciparum kinome displays significant divergence from the eukaryotic kinome. The 65 plasmodial kinases that cluster within established eukaryotic protein kinase (ePK) groups (CAMK, AGC, CMGC, CK1 and TKL groups) often display structural and functional characteristics that are not seen in their mammalian counterparts [9,13]. The plasmodial kinome also contains protein kinases that have no mammalian orthologues (orphan kinases) or display homology with more than one of the established ePK groups (composite or hybrid kinases) [9]. These differences can be exploited for selective antimalarial drug development.
Progress on plasmodial kinase inhibitor development up until the beginning of 2018 was discussed in a detailed review by Cabrera and co-workers [14]. Herein, we give an overview thereof, as well as discussing additional kinases and new research related to plasmodial kinase inhibitor development.

Calcium-Dependent Protein Kinases (CDPKs)
Enzymes from the classical Ca 2+ /calmodulin-dependent protein kinase (CaMK) group seem to be rare in the P. falciparum kinome [15]. However, the kinome contains calcium-dependent protein kinases (CDPKs) which have a C-terminal calmodulin-like domain that is highly homologous to the CaMK group [16]. CDPKs comprise a unique family of serine/threonine kinases only found in plants, protozoans (including apicomplexan parasites) and some algae [17]. These enzymes play an important role in calcium signalling during the various life stages of the Plasmodium parasite [15]. Seven members of the CDPK family (Pf CDPK1 to Pf CDPK7) have been identified in P. falciparum [16]. Pf CDPK1 is expressed at all stages of the Plasmodium parasite life cycle. During asexual parasite development, Pf CDPK1 plays a role in parasite motility [18,19], microneme secretion and subsequent erythrocyte invasion [20], as well as merozoite egress from mature schizonts [19]. Previous studies have shown that Pf CDPK1 is likely to be essential for asexual development [19,21,22]; however, the parasite might have other mechanisms in place to compensate for loss of Pf CDPK1 activity [21,23,24]. During the sexual development of the parasite, Pf CDPK1 activity is indispensable for gametogenesis and subsequent infection of the mosquito vector [21]. In addition, the P. berghei homologue (PbCDPK1) is also involved in ookinete development [25].
Pf CDPK2, Pf CDPK3 and Pf CDPK4 are all predominantly expressed during the sexual stage of the parasite development [26][27][28]. Pf CDPK2 and Pf CDPK4 are essential for male gametocyte exflagellation and transmission to the mosquito vector [26,27]. In addition, Pf CDPK4 is also required for sporozoite invasion of hepatocytes [29]. Although the exact function of Pf CDPK3 is not yet known, its P. berghei orthologue (PbCDPK3) is expressed in ookinetes where it regulates motility required for invading the midgut of the mosquito [30,31].
Pf CDPK5 and Pf CDPK7 are expressed during the asexual erythrocytic stage. Pf CDPK5 acts synergistically with P. falciparum protein kinase G (Pf PKG) to regulate microneme secretion which is required for merozoites to egress mature schizonts [32,33]. Although Pf CDPK5 is essential, the parasite is able to compensate for loss of Pf CDPK5 activity through hyperactivation of Pf PKG [32]. While CDPK7 is not essential to parasite viability, it still plays an important role in the development of the erythrocytic parasite. The growth rate of CDPK7 knockout parasites is significantly reduced due to a delay in maturation of ring-stage parasites to trophozoites and the release of fewer merozoites from each schizont [34]. Little is currently known about CDPK6 of P. falciparum. The P. berghei orthologue (PbCDPK6) signals to sporozoites when to stop migration and initiate invasion of hepatocytes [35].

Inhibitor Development for the CDPK Group
The CDPKs are promising drug targets for the development of new antiplasmodial agents as there are no CDPK orthologues in the human host [16]. A unique structural feature of many parasitic CDPKs is the small gatekeeper residue at the hinge region [36]. A small gatekeeper residue results in enlargement of the hydrophobic pocket that accommodates the ATP purine group in the ATP-binding site [36]. Various medicinal chemistry campaigns have developed small-molecule inhibitors termed bumped kinase inhibitors (BKIs) that contain a bulky C3-aryl substituent that can occupy this enlarged hydrophobic pocket [36]. Most mammalian kinases have larger gatekeeper residues that block access to the bulky substituent of BKIs, therefore allowing better selectivity towards the parasitic kinases [36]. Amongst P. falciparum CDPKs, Pf CDPK4 has the smallest gatekeeper, which is a serine residue, followed by Pf CDPK1 with a medium threonine gatekeeper residue [37]. Pf CDPK2 (methionine), Pf CDPK3 (methionine) and Pf CDPK5 (leucine) all have bulky gatekeeper residues [37]. P. falciparum CDPK inhibitor development has mainly focussed on Pf CDPK1 and Pf CDPK4, and most of these inhibitors are BKIs [14].

Pf CDPK1
High-throughput screening campaigns have identified various scaffolds as Pf CDPK1 inhibitors (Figure 1), including 2,3,9-trisubstituted purines (1) [19], indolizines (2) [38] and imidazopyridazines (3)(4)(5) [23,[38][39][40][41][42]. Of all these scaffolds, imidazopyridazines have been studied more intently.  Imidazopyridazines are generally potent Pf CDPK1 inhibitors, with some compounds demonstrating low micromolar to submicromolar activity against recombinant Pf CDPK1 and P. falciparum erythrocytic parasites [23,[38][39][40][41][42]. However, discrepancies between the enzymatic and whole-cell activities are generally reported for these compounds, which may be due to off-target activity or permeability issues [38,39]. Considerable effort was made to improve the selectivity and ADME (absorption, distribution, metabolism, excretion) profiles of imidazopyridazines, which resulted in compounds with well-balanced Pf CDPK1 potency, permeability and in vitro activity against P. falciparum erythrocytic stage parasites [40][41][42]. Despite these efforts, studies still reported only modest in vivo activity in a P. berghei mouse model [39][40][41]. Further exploration of the mechanism of action of imidazopyridazines demonstrated that these compounds could be grouped into two classes based on the type of aromatic linker between the core and the R2 substituent ( Figure 2) [23]. Class 1 compounds (6) had a pyrimidine linker and inhibited P. falciparum parasite growth at the late schizont stage, while class 2 compounds (7) had a non-pyrimidine linker and inhibited the trophozoite stage of P. falciparum. Two additional parasitic targets were also identified for imidazopyridazines: class 1 compounds inhibited Pf PKG and class 2 compounds inhibited Pf HSP90 (a chaperone protein of P. falciparum). These results suggest that the activity of imidazopyridazines against erythrocytic P. falciparum parasites is primarily due to inhibition of Pf PKG and Pf HSP90, rather than Pf CDPK1 inhibition.  More recently, Flaherty and co-workers [43] designed a hydrocarbon constrained peptide that mimics the C-terminal helical region of the PfCDPK1 junction domain (J-domain). The autoinhibitory J-domain is located between the catalytic domain and the calmodulin-like domain and blocks the active site by acting as a pseudosubstrate when the kinase is in its inactive state. By mimicking the activity of the J-domain, the constrained peptide inhibits PfCDPK1 by locking the kinase in its inactive state. Uptake of the constrained peptide by P. falciparum-infected erythrocytes was highly stagespecific, as late-stage schizont erythrocytes demonstrated increased uptake relative to ring-stage and early trophozoite erythrocytes. The constrained peptide inhibited recombinant PfCDPK1 in the low micromolar range (IC50: 3.5 µ M) and caused a significant decrease in parasitemia at concentrations of ≥10 µ M.
Lima and co-workers [44] designed and developed shape-based and machine learning models of PfCDPK1, PfCDPK4 and PfPK6. These models were used for virtual screening of drug-like molecules to identify potent inhibitors with activity against multiple P. falciparum kinases. The computational hits were then evaluated in vitro against drug-sensitive (3D7) and multidrug-resistant (Dd2) P. falciparum erythrocytic parasites. Quinazoline derivatives (compounds 8-10, Figure 3) inhibited the growth of both drug-sensitive and multidrug-resistant P. falciparum strains in the nanomolar range. Compounds 8 and 10 also demonstrated good in vivo inhibition of P. berghei ookinete formation at a concentration of 10 µ M. Molecular docking studies indicated that compound 8 was able to interact with PfCDPK1, PfCDPK4 and PfPK6, thus highlighting its potential as a multikinase inhibitor.

AGC Group
Three of the five malarial kinases that cluster within this group, namely adenosine monophosphate (cAMP)-dependent protein kinase A (PKA), cyclic guanosine monophosphate (cGMP)-dependent protein kinase G (PKG) and protein kinase B (PKB) have been characterised [13]. cAMP-, cGMP-and calcium-mediated signalling pathways are closely linked within the malaria parasite. In merozoites, a rise in cytosolic cAMP levels leads to activation of PKA and an increase in cytosolic calcium levels via induction of the Epac (exchange protein directly activated by cAMP) pathway [55]. When activated by cGMP, PKG regulates phosphoinositide metabolism which produces inositol (1,4,5)-triphosphate (IP3), a messenger molecule that signals the release of intracellular calcium [56]. The release of calcium in turn activates stage-specific effector pathways, including CDPK signalling [29,56]. In contrast to the other two kinases, PKB is activated by More recently, Flaherty and co-workers [43] designed a hydrocarbon constrained peptide that mimics the C-terminal helical region of the Pf CDPK1 junction domain (J-domain). The autoinhibitory J-domain is located between the catalytic domain and the calmodulin-like domain and blocks the active site by acting as a pseudosubstrate when the kinase is in its inactive state. By mimicking the activity of the J-domain, the constrained peptide inhibits Pf CDPK1 by locking the kinase in its inactive state. Uptake of the constrained peptide by P. falciparum-infected erythrocytes was highly stage-specific, as late-stage schizont erythrocytes demonstrated increased uptake relative to ring-stage and early trophozoite erythrocytes. The constrained peptide inhibited recombinant Pf CDPK1 in the low micromolar range (IC 50 : 3.5 µM) and caused a significant decrease in parasitemia at concentrations of ≥10 µM.
Lima and co-workers [44] designed and developed shape-based and machine learning models of Pf CDPK1, Pf CDPK4 and Pf PK6. These models were used for virtual screening of drug-like molecules to identify potent inhibitors with activity against multiple P. falciparum kinases. The computational hits were then evaluated in vitro against drug-sensitive (3D7) and multidrug-resistant (Dd2) P. falciparum erythrocytic parasites. Quinazoline derivatives (compounds 8-10, Figure 3) inhibited the growth of both drug-sensitive and multidrug-resistant P. falciparum strains in the nanomolar range. Compounds 8 and 10 also demonstrated good in vivo inhibition of P. berghei ookinete formation at a concentration of 10 µM. Molecular docking studies indicated that compound 8 was able to interact with Pf CDPK1, Pf CDPK4 and Pf PK6, thus highlighting its potential as a multi-kinase inhibitor.
Another virtual screening campaign against a Pf CDPK1 homology model (PbCDPK1 crystal structure, PBD ID: 3Q5I), identified 18 compounds from the MyriaScreen Diversity Library II that complement the Pf CDPK1 ATP-binding site [45]. Two of these compounds, 11 (ST092793) and 12 (S344699) (Figure 4), significantly inhibited recombinant Pf CDPK1 and demonstrated in vitro activity against P. falciparum erythrocytic parasites. Interestingly, isothermal titration calorimetry and fluorescence spectroscopy showed that 11 preferentially binds to the inactive conformation of Pf CDPK1, thereby locking the enzyme in this state throughout the erythrocytic stage.  Overall, the results from these studies suggest that Pf CDPK1 may not be the most suitable target for P. falciparum blood-stage infections. It seems as though the pathways and/or enzymes that are able to compensate for the loss of Pf CDPK1 activity [21,23,24] greatly affect the potency of Pf CDPK1 inhibitors in vivo. Greater success may be achieved if future drug development focusses on Pf CDPK1 as a potential transmission-blocking target.
Based on pyrazolopyrimidine BKIs designed for Toxoplasma gondii CDPK1 (TgCDPK1) and Cryptosporidium parvum CDPK1 (CpCDPK1), a series of pyrazolopyrimidine derivatives (e.g., compounds 13-15, Figure 5) with potent activity against Pf CDPK4 were designed [49,50]. Minimal off-target activity was observed for some of these compounds when tested against human Src and Abl tyrosine kinases, which both have one of the smallest gatekeeper residues (threonine) in the human kinome [49,50]. However, when screened against human non-kinase targets, compound 14 also inhibited the human ether-a-go-go related gene potassium channel (hERG) which is critical for cardiac repolarisation [51]. Pyrazolopyrimidine compounds have been shown to block exflagellation of male gametocytes in P. falciparum parasites [49,50] and in P. berghei-infected mice [49] within the nanomolar range. When Anopheles stephensi mosquitoes were allowed to feed on P. berghei-infected mice treated with compound 13 (10 mg/kg, intraperitoneally), oocyst formation was blocked in the mosquito midgut. Similarly, infective sporozoite formation was inhibited in Anopheles stephensi mosquitoes that fed on Pf NF54-infected human blood containing 3 µM of compound 13 [49]. P. falciparum parasites expressing Pf CDPK4 with a mutated gatekeeper (small serine residue changed to a large methionine residue, S147M), were insensitive to pyrazolopyrimidine-based compounds and demonstrated normal exflagellation, which confirms Pf CDPK4 to be the target of these inhibitors [48,49]. Substituting the pyrazolopyrimidine scaffold with an imidazopyrazine core generally resulted in similar Pf CDPK4 selectivity and potency [50]. In silico studies revealed the structure-activity relationships of these pyrazolopyrimidine and imidazopyrazine compounds with the Pf CDPK4 target [52].
Another scaffold used for TgCDPK1 inhibitor development [53], 5-aminopyrazole-4-carboxamide, was also shown to potently inhibit Pf CDPK4 in the nanomolar range [47]. The most active 5-aminopyrazole-4-carboxamide derivatives (compounds 16 and 17, Figure 6) demonstrated potent inhibition of P. falciparum male gametocyte exflagellation at a concentration of 0.1 µM. The in vitro inhibition was much higher than the enzymatic assay predicted, which may indicate multiple targets for these compounds. In terms of selectivity over human kinases, these inhibitors demonstrated high selectivity over Src kinase and hERG. Members of the phenothiazine class were identified as possible non-ATP-competitive inhibitors of Pf CDPK4 [46]. Trifluoperazine (TFP) (18, Figure 7) was the most active of this class, with a binding affinity (K d ) of 134.5 µM and K i value of 150 µM for Pf CDPK4. The discrepancy between the enzymatic activity of TFP and the reported in vitro activity (EC 50 : 1.9 µM) against P. falciparum indicates that this compound modulates multiple targets. Homology modelling indicated that TFP possibly binds to the calmodulin-like domain of Pf CDPK4 which prevents repositioning of the autoinhibitory J-domain upon binding of Ca 2+ , thereby locking the kinase in its inactive state.

Pf CDPK5
To date, only one study has been published on inhibitor development for Pf CDPK5. Rout and Mahapatra predicted the three-dimensional structure of Pf CDPK5 through homology modelling using P. berghei CDPK1 as a template [54]. Possible inhibitors of Pf CDPK5 were then identified through virtual screening of five different sets of compounds with known antimalarial activity. MMV687246 (19, Figure 8), from the Malaria box assembled by The Medicines for Malaria Venture, demonstrated the highest binding affinity for Pf CDPK5 and was suggested as a possible lead for future experimental validation and inhibitor design.

AGC Group
Three of the five malarial kinases that cluster within this group, namely adenosine monophosphate (cAMP)-dependent protein kinase A (PKA), cyclic guanosine monophosphate (cGMP)-dependent protein kinase G (PKG) and protein kinase B (PKB) have been characterised [13]. cAMP-, cGMP-and calcium-mediated signalling pathways are closely linked within the malaria parasite. In merozoites, a rise in cytosolic cAMP levels leads to activation of PKA and an increase in cytosolic calcium levels via induction of the Epac (exchange protein directly activated by cAMP) pathway [55]. When activated by cGMP, PKG regulates phosphoinositide metabolism which produces inositol (1,4,5)-triphosphate (IP 3 ), a messenger molecule that signals the release of intracellular calcium [56]. The release of calcium in turn activates stage-specific effector pathways, including CDPK signalling [29,56]. In contrast to the other two kinases, PKB is activated by calmodulin in a calcium-dependent manner [57]. Phospholipase C has been identified as the upstream regulator responsible for releasing the calcium required for PKB activation [57].
During the asexual parasite stages, Pf PKA, Pf PKG and Pf PKB regulate different factors required for parasite invasion and egress. Pf PKA phosphorylates the P. falciparum apical membrane antigen 1 (Pf AMA1) which is critical for tight junction formation between the parasite and the host cell during erythrocyte invasion [58][59][60]. Pf PKA has also been implicated in regulation of parasite motility [61], microneme secretion [55], anion channel conductance at the erythrocytic plasma membrane [62] and the cell cycle of the intraerythrocytic parasite [63]. However, Patel and co-workers [60] reported that events prior to invasion, such as egress, rise in cytosolic calcium levels and microneme secretion, can all occur in the absence of cAMP and Pf PKA. Apart from merozoite invasion, they also did not observe any other critical role for cAMP and Pf PKA in the erythrocytic life cycle. Pf PKG controls invasion and egress of sporozoites (liver-stage parasites) [29,64] and merozoites (blood-stage parasites) [65][66][67] by regulating parasite motility and microneme secretion. Pf PKB also plays a role in merozoite invasion of erythrocytes by regulating parasite motility [68]. During the sexual parasite stages, Pf PKG regulates gametogenesis, which involves male gamete exflagellation and rounding up of female gametes, and ookinete motility required for mosquito midgut invasion [56,69,70]. Both Pf PKA and Pf PKG have been validated as essential kinases, Pf PKA being essential for blood-stage parasites [58,60] and Pf PKG being essential for multiple life cycle stages [67,69,71]. Pf PKB is regarded as likely essential to blood-stage parasite survival [22].

Inhibitor Development for the AGC Group
In terms of inhibitor development, Pf PKG is one of the plasmodial kinases that has been studied extensively thus far. It is regarded as a very attractive antimalarial drug target as its inhibition offers simultaneous prophylactic, curative and transmission-blocking potential [72]. As Pf PKG is an essential enzyme for multiple life cycle stages, there is a relatively low risk of the parasite developing high-grade resistance to Pf PKG-selective inhibitors [72]. The PKG enzyme is also highly conserved in all human malaria species, with an overall sequence identity of 90-92% and identical catalytic site and gatekeeper residues [73]. Thus, PKG inhibitors have the potential to be active against multiple malaria species. Although human PKG orthologues (cytosolic PKG-Iα and PKG-Iβ, membrane-associated PKG-II) exist, selective inhibitor development is still possible as there is significant structural divergence between plasmodial and mammalian PKGs [74].
A variety of scaffolds have been explored for Pf PKG inhibitor development, with the imidazopyridines [71,75,76] and thiazoles [73,77,78] being the most advanced inhibitors of Pf PKG. Compound 20 (Figure 9), a PKG inhibitor of Eimeria tenella [79], was used as a lead compound to develop potent imidazopyridine inhibitors of Pf PKG [71,75,76]. Compound 21 (Figure 9) was the most potent of a series of imidazopyridines synthesised by Baker and co-workers [71], with an IC 50 value of 0.16 nM against Pf PKG and an EC 50 value of 2.1 nM against P. falciparum blood-stage parasites. Compound 21 demonstrated no toxicity in vitro or in vivo, high selectivity over human kinases and moderate metabolic stability in vitro. Oral administration of compound 21 to P. falciparum-infected mice (twice-daily dose of 100 mg/kg) over a period of four days reduced parasitemia below detectable levels. Apart from blood-stage efficacy, compound 21 also inhibited transmission of mature P. falciparum gametocytes to Anopheles stephensi mosquitoes (IC 50 : 41.3 nM).
Subsequent studies focussed on improving the ADME properties of this chemical class while retaining potent inhibitory activity. By means of structure-activity relationship and modelling studies, Large and co-workers [75,76] systematically varied the substituents of compound 20 and explored other bicyclic cores. Compounds 22 [75] and 23 [76] (Figure 9) retained the potent Pf PKG and in vitro antimalarial activity of compound 20. Compound 22 also retained the LogD and lipophilic ligand efficiency of compound 20, and showed improved permeability, but demonstrated poor stability in mouse liver microsomes [75]. Compound 23 demonstrated an excellent balance of activity and physicochemical properties. In addition to good LogD and LLE, compound 23 showed improved microsomal stability and excellent selectivity over human kinases, including the two human PKG orthologues (PKG1α and PKG1β) [76]. A 2,3-diaryl-pyrrole inhibitor of PKG developed for Eimeria tenella (compound 24, Figure 10) [80] was shown to be a potent inhibitor of Pf PKG (IC 50 : 3.5 nM) [81]. However, compound 24 only demonstrated modest in vitro activity against P. falciparum and failed to reduce parasitemia in a P. berghei mouse model [81]. A scaffold-hopping approach performed on compound 24 lead to the identification of thiazoles (compound 25) as Pf PKG inhibitors [78]. Substitution of the thiazole scaffold was optimised to improve the enzymatic and in vitro activity of compound 25, which lead to compound 26 (IC 50 : 2 nM, EC 50 : 113 nM) [78]. Compound 26 demonstrated excellent selectivity over human kinases, good permeability and metabolic stability in mouse and human liver microsomes.
Molecules 2020, 25, x FOR PEER REVIEW 2 of 2 calmodulin in a calcium-dependent manner [57]. Phospholipase C has been identified as the upstream regulator responsible for releasing the calcium required for PKB activation [57].

PfPKG IC50
Pf3D7 blood stage EC50 Pf transmission-blocking IC50 In vivo Pf mouse model (% parasitemia reduction, oral dose)   During a high-throughput screening campaign, Penzo and co-workers [73] also identified several thiazole derivatives, such as compound 27 (Figure 11), with nanomolar potencies against Pf PKG [73]. The in vitro activity of the most potent compounds was studied at 48 and 72 h in wild-type and transgenic (Pf PKG gatekeeper mutant, T618Q) P. falciparum blood-stage parasites, and the EC 50 values against the two strains were very similar-indicating that the thiazole derivatives also inhibit targets other than Pf PKG. Additional targets identified for compound 27 included CDPK1, CDPK4, CDK-related kinase (Pfcrk-5), NIMA-related kinase (Pf nek-1), CK1 and an unnamed putative protein kinase (Pf3D7_0926100). Compound 27 also demonstrated good solubility, no toxicity against HepG2 cells, and selectivity over the human PKG orthologue (PKGIα), human lymphocyte-specific protein tyrosine kinase (LCK) and human Aurora B kinase. Besides blood-stage activity, compound 27 also had potent activity against male and female gametes. Matralis and co-workers [77] specifically focussed on developing a series of thiazole derivatives with a fast-killing profile similar to that of artemisinins. Compound 28 (Figure 11) was the most potent in this series, with in vitro nanomolar activity against Pf PKG, P. falciparum blood-stage parasites and gametocytes. It also possessed good physicochemical properties. Despite good selectivity over human enzymes, ion channels and receptors, this compound showed activity towards hERG. Parasite reduction ratio studies demonstrated that compound 28 had fast-killing properties similar to those of artesunate. CDPK1, CDPK4 and serine/arginine protein kinase 2 (SRPK2, also known as CLK2) were identified as additional targets of compound 28. The fast-killing activity of compound 28 was mainly attributed to SRPK2 inhibition. This study demonstrates the potential of simultaneously targeting Pf PKG and SRPK2 to develop fast-killing drugs with curative and transmission-blocking activity.
Vanaerschot and co-workers [72] identified Pf PKG as the primary target of the Medicines for Malaria Venture compound MMV030084 (29, Figure 12). MMV030084 had an IC 50 value of 0.4 nM against recombinant Pf PKG, and docking studies using the Pf PKG crystal structure (PDB ID: 5DYK) showed a strong interaction between MMV030084 and the ATP-binding site. When evaluated against the different life cycle stages of the parasite, MMV030084 inhibited sporozoite invasion of hepatocytes, merozoite egress from mature schizonts, and male gamete exflagellation. MMV030084 inhibited liver cell invasion by P. berghei parasites (IC 50 : 199 nM) with minimal toxicity against the host cells (CC 50 : 41.5 µM). The development of P. falciparum blood-stage parasites was halted at the schizont stage when treated with MMV030084 (drug-sensitive strain IC 50 : 109 nM; multidrug-resistant strain IC 50 : 120 nM). Male gamete exflagellation of the P. falciparum NF54 strain was inhibited by MMV030084 with an IC 50 value of 141 nM. In vitro MMV030084 resistance selection studies identified P. falciparum tyrosine kinase-like 3 (Pf TKL3) as an MMV030084-resistance mediator, which allows merozoite egress from erythrocytes when mutated. No mutation was identified for Pf PKG itself. A screening campaign by Mahmood and co-workers [82] identified isoxazole-based inhibitors with Pf PKG activity and selectivity over human PKG. Optimisation of this scaffold led to compounds 30-32 ( Figure 13) with IC 50 values <20 nM against Pf PKG. Further evaluation of the physicochemical properties and in vitro activity of these compounds against the whole-cell parasite has not yet been reported. 3.1.2. Pf PKA and Pf PKB Drug discovery efforts targeting Pf PKA and Pf PKB are limited. Buskes and co-workers [83] attempted to develop Pf PKA inhibitors using the commercially available PKA inhibitor 3-methylisoquinoline-4-carbonitrile (33, Figure 14) as a starting point. They studied the interactions of compound 33 and a series of substituted isoquinolines (e.g., 34, Figure 14) with a Pf PKA homology model. In vitro evaluation of this series demonstrated low micromolar activity against drug-sensitive (3D7) and -resistant (W2) P. falciparum strains. However, biochemical evaluation of this series showed minimal activity against PKA [84]. It was suggested that these compounds likely inhibit another kinase that is involved in parasitic cytokinesis and erythrocyte invasion [84]. As Pf PKB shares high sequence homology with the catalytic sites of mammalian PKB and protein kinase C (PKC), established PKB and PKC inhibitors were tested for activity against Pf PKB. Go 6983 (35, Figure 15), a PKC inhibitor, was shown to inhibit Pf PKB (IC 50 : ±1 µM) and significantly reduce P. falciparum parasite growth at the late schizont stage [85]. The mammalian PKB inhibitor A443654 (36, Figure 15) inhibited Pf PKB with an IC 50 value of 200 nM [57]. Incubation of P. falciparum schizonts with A443654 did not affect the morphology or the number of schizonts but dramatically reduced the number of ring-stage parasites formed after invasion. This indicates that A443654 blocks invasion, which corroborates the function of Pf PKB in erythrocytic invasion. A peptide inhibitor that corresponds to the pseudosubstrate motif of the Pf PKB N-terminal effectively inhibited Pf PKB activity and also reduced the formation of P. falciparum ring-stage parasites [57]. Crystal structures of Pf PKA's regulatory subunit (PDB ID: 5K8S, 5KBF and 5T3N) [86] and methods for expression and purification of active recombinant Pf PKAr (regulatory subunit) [86] and Pf PKAc (catalytic subunit) [87] are available for future drug development efforts for this target. The crystal structure of Pf PKB is not yet available; however, the full-length Pf PKB gene has been successfully expressed and purified as an active recombinant protein [57,68,85].
Pf PK5 is most related to mammalian CDK1 and CDK5 [90] and demonstrates sensitivity to mammalian CDK1/CDK2 inhibitors [91]. It has been proposed that Pf PK5 is likely involved in the regulation of DNA replication (S-phase of the cell cycle) during erythrocytic schizogony [90,92,93]. Pf mrk, a putative homologue of mammalian CDK7, is predominantly expressed in gametocytes and to a lesser extent in the asexual stages (trophozoites and schizonts) [94]. Pf mrk is localised in the nucleus of the parasite and is presumably involved in the regulation of DNA replication [95]. Pf crk-1 and Pf crk-3 display maximal homology to mammalian CDKs involved in transcriptional control [13,96]. Pf crk-1 is mainly expressed in gametocytes [97]; however, the P. berghei orthologue (Pbcrk-1) was found to be essential for erythrocytic schizogony [98]. Pf crk-3 has been demonstrated to be essential to erythrocytic parasites and presumably regulates gene expression via interaction with chromatin modification enzymes [96]. Pf crk-4 is another essential enzyme for asexual proliferation and also plays a critical role in ookinete formation and early oocyst development [22,99]. Two atypical CDKs, namely Pf PK6 and Pf crk-5, demonstrate both CDK and mitogen-activated protein kinase (MAPK) homology [100,101]. Nuclear and cytoplasmic localisation have been reported for Pf PK6 in trophozoite and schizont parasites [100]. Unlike Pf PK5 and Pf mrk which are activated by various cyclins in vitro [102,103], Pf PK6 appears to be a cyclin-independent kinase [100]. Pf crk-5 is a cyclin-dependent enzyme localised in the nuclear periphery [101]. Although parasites lacking Pf crk-5 are viable, they display decreased erythrocytic proliferation due to a lower number of merozoites released per schizont [101].

CDK Inhibitor Development
Pf PK5 Developing Pf PK5-selective inhibitors is quite challenging due to the high degree of homology between Pf PK5 and the human CDKs. Eubanks and co-workers [104] demonstrated that a target-based screening approach was more effective to identify Pf PK5-selective inhibitors than chemically modifying existing CDK inhibitors. The 4-methylumbelliferone analogues, compounds 37 and 38 (Figure 16), demonstrated a 2-fold binding affinity for Pf PK5 over human CDK2 (HsCDK2). No significant toxicity was observed against human hepatoma cell lines (HuH7 and HepG2) for either compound. However, both compounds failed to inhibit drug-resistant P. falciparum (Dd2 strain) asexual parasite growth in vitro, most likely due to poor physicochemical properties. Pf mrk Scaffolds studied for Pf mrk inhibition include quinolinones [105], oxindoles [106], chalcones [107], flavonoids [108] and sulfonamide-based compounds [109][110][111].
A series of quinolinones (e.g., compound 39, Figure 17) demonstrated Pf mrk activity with IC 50 values ranging from 18 to 539 µM [105]. Neither antimalarial evaluation against whole-cell parasites nor molecular docking has been performed for this chemical class. Based on the structure of commercially available indirubin-3 -monoxime, a moderate inhibitor of Pf mrk and in vitro P. falciparum parasite growth, oxindole-based compounds were explored as inhibitors of Pf mrk [106]. Several oxindoles selectively inhibited Pf mrk in the low micromolar range, with the most potent being compound 40 ( Figure 17). None of the oxindoles demonstrated any significant in vitro activity against whole-cell parasites. Homology modelling showed that these compounds had the same orientation in the active site of Pf mrk as in human CDK2, but with additional contact points which might be responsible for the Pf mrk specificity. The chalcone and sulfonamide-based scaffolds were identified by means of a three-dimensional structure-activity relationship (3D-QSAR) pharmacophore model [109,112]. Compound 41 (Figure 18) was the most active compound of a series of chalcones tested against Pf mrk (IC 50 : 1.3 µM) [107]. However, a weak correlation was observed between the Pf mrk activity and the in vitro activity against drug-sensitive (D6) and -resistant (W2) P. falciparum strains. As several mechanisms of action have been reported for the antimalarial activity of chalcones [113][114][115], Geyer and co-workers [107] proposed that Pf mrk inhibition might be an additional mechanism demonstrated by some chalcones. Several flavonoids isolated from Erythrina sp., were evaluated for activity against Pf mrk [108]. The most potent flavonoid, Abyssinone V (42), had an IC 50 value of 0.038 µM. Despite potent Pf mrk activity, the flavonoids demonstrated similar in vitro antimalarial activity to that of the chalcones, presumably due to the lower permeability of the flavonoids. Isoquinoline sulfonamides were generally weak inhibitors of Pf mrk, except compound 43 (Figure 19) which demonstrated an IC 50 value of 0.7 µM [111]. However, compound 43 failed to inhibit drug-sensitive (D6) and -resistant (W2) P. falciparum parasite growth. Thiophene sulfonamides exemplified by 44 and 45 ( Figure 19) were found to be potent inhibitors of Pf mrk with IC 50 values in the submicromolar range [110]. However, all of these compounds, except compound 45, are also potent inhibitors of human CDK7. Thiophene sulfonamides demonstrated minimal cytotoxicity and moderate in vitro activity against a multidrug-resistant P. falciparum (W2) strain.

Pf PK6 and CDK-Related Kinases
To our knowledge, no medicinal chemistry campaigns have focussed on drug development for Pf PK6 and the CDK-related kinases (Pf crk-1 to -5). Homology models have been developed for Pf PK6 [116,117] and Pf crk-4 [99], which can be used for virtual screening of small-molecule inhibitors against these targets. Methods for expression and purification of active recombinant Pf PK6 [100], Pf crk-1 (kinase domain) [118], Pf crk-3 [96] and Pf crk-5 [101] are also available in the literature.

Mitogen-Activated Protein Kinases (MAPKs)
The P. falciparum kinome encodes two MAPK homologues, namely Pf map-1 and Pf map-2. Interestingly, Pf map-2 is essential for the asexual stages of P. falciparum parasites [119]; however, the P. berghei orthologue (Pbmap-2) is only essential for male exflagellation in the mosquito midgut [120]. Pf map-1 also plays an important role during asexual development; however, parasites are able to compensate for loss of Pf map-1 activity by upregulating Pf map-2 [119].
Human p38 MAPK inhibitors have been shown to inhibit drug-sensitive (HB3) and -resistant P. falciparum strains in vitro; however, antiplasmodial activity has yet to be attributed to plasmodial MAPK inhibition [121].

Glycogen Synthase Kinase-3 (GSK-3)
Pf GSK-3 is one of three GSK3-related kinases identified in the P. falciparum parasite [13]. Although Pf GSK-3 is expressed throughout the erythrocytic stage, it is predominantly expressed during the early trophozoite stage [122]. After expression, Pf GSK-3 is rapidly transported to the cytoplasm of the erythrocyte where it appears to associate with membranous structures known as Maurer's clefts [122]. The exact biological functions of Pf GSK-3 remain to be elucidated; however, it has been demonstrated to be essential for the survival of asexual erythrocytic parasites [22].

CDK-Like Protein Kinases (CLKs)
Four members of the CLK family have been identified in P. falciparum, Pf CLK-1 to -4 [13]. All four enzymes are essential to the asexual erythrocytic parasites [22,128] as they regulate mRNA splicing through phosphorylation of serine/arginine-rich (SR) proteins [129]. Pf CLK1 and Pf CLK2 exhibit homology to the yeast SR protein, Sky1p, and are expressed throughout the erythrocytic stage and in gametocytes [128]. Both kinases are localised within the nucleus of the parasite, with Pf CLK2 also present in the cytoplasm [128]. Pf CLK3 is a closely related homologue of human pre-mRNA processing factor 4B (PRP4 or PRPF4B) [130] which regulates mRNA splicing through phosphorylation of proteins associated with the spliceosome complex [131]. Pf CLK4, also known as SR protein-specific kinase 1 (SRPK1), is expressed in erythrocytic parasites and abundantly in gametocytes [132,133]. Pf CLK4 negatively regulates mRNA splicing by phosphorylating a putative plasmodial SR protein (Pf SR1) in vitro [132,134]. Both Pf CLK4 and Pf SR1 are localised inside the nucleus of ring and early trophozoite stage parasites. As erythrocytic development progresses, the two proteins are exported to the nuclear periphery (mature trophozoites) and finally to the cytoplasm (schizonts and gametocytes) [132,134].

CLK Inhibitor Development
A high-throughput screening campaign identified TCMDC-135051 (51, Figure 21) as a highly potent and selective inhibitor of Pf CLK3 [135]. TCMDC-135051 demonstrated selectivity for Pf CLK3 over closely related human kinases (CLK2 and PRPF4B), the closest related plasmodial kinase (Pf CLK1) and two other plasmodial kinases (Pf PKG and Pf CDPK1). Pf CLK3 mutations were observed in P. falciparum parasites with reduced sensitivity to TCMDC-135051, which indicated that Pf CLK3 was the primary target of this compound. A recombinant Pf CLK3 variant with a G449P mutation and parasites expressing the G449P mutant Pf CLK3 both demonstrated reduced TCMDC-135051 sensitivity, confirming that the antimalarial activity of TCMDC-135051 was due to Pf CLK3 inhibition. TCMDC-135051 inhibition resulted in downregulation of 425 essential P. falciparum genes and upregulation of certain genes involved in RNA processing, which is consistent with the proposed mRNA splicing role of Pf CLK3. TCMDC-135051 was active against multiple stages of the parasite's life cycle, including liver-stage sporozoites, blood-stage parasites, gametocyte development and subsequent transmission to the mosquito vector. TCMDC-135051 also demonstrated activity against CLK3 of P. vivax (PvCLK3) and P. berghei (PbCLK3), as well as in vitro activity against P. knowlesi and P. berghei blood-stage parasites. A dose-dependent reduction in parasitemia was observed when TCMDC-135051 was administered intraperitoneally to P. berghei-infected mice (twice daily, 5-day period), with the maximal dose (50 mg/kg) reducing parasitemia below detectable levels.

Casein Kinase 1 (CK1) Group
The P. falciparum kinome encodes a single CK1 enzyme (Pf CK1) that is expressed throughout the erythrocytic stages. Pf CK1 is essential for blood-stage parasite survival [22] and is likely involved in cellular processes such as mRNA splicing, protein trafficking and erythrocyte invasion [136,137]. To our knowledge, no medicinal chemistry programs have targeted this enzyme thus far.

NIMA-and Aurora-Related Kinases
Four Never in Mitosis, gene A (NIMA)-related kinases or NEKs (Pf nek-1 to -4) [12,13] and three Aurora-related kinases (Pf ark-1 to -3) [138] have been identified for P. falciparum. Both groups of kinases are involved in the regulation of the parasitic cell cycle [88,139]. Pf nek-1 and the three Aurora-related kinases are likely essential for erythrocytic schizogony, while Pf nek-2 and Pf nek-4 are essential for sexual development of the parasite [139].
Human Aurora kinase inhibitors also demonstrated antiplasmodial activity in vitro; however, the activity is yet to be attributed to plasmodial Aurora-related kinase inhibition [143].

PI3K
To date, only two compounds have been identified for studying the functions of Pf PI3K. The mammalian PI3K inhibitors ( Figure 23) wortmannin (54) and LY294002 (55) both inhibit Pf PI3K and blood-stage P. falciparum parasite growth [146,147]. Pf PI3K was also identified as a target of dihydroartemisinin (56, Figure 23) during the early ring stages [148].
A phenotypic screening identified compound 57 (Figure 24), which demonstrated activity against P. falciparum blood-stage parasites but was inactive against P. yoelii (Py) and P. cynomolgi (Pc) liver-stage parasites [152]. Liver-stage activity was acquired by replacing the imidazopyridine core of compound 57 with an imidazopyrazine core (KAI407, 58, Figure 24). From a series of imidazopyrazines, compound 59 (KDU691, Figure 24) had optimal antimalarial activity (blood and liver stages) and physicochemical properties, which translated into in vivo efficacy against P. berghei-infected mice [152]. McNamara and co-workers [145] further demonstrated that compound 59 reduced liver-and blood-stage parasites, gametocyte viability and transmission to the mosquito vector for multiple Plasmodium species. Plasmodial PI4K was identified as the direct target of imidazopyrazine compounds [145].
Le Manach and co-workers [150] synthesised a series of imidazopyrazines based on structures of hits identified during a screening campaign. Compound 61 (Figure 24) was highly active against drug-sensitive (NF54) and -resistant (K1) P. falciparum strains and reduced parasitemia by 98% in P. berghei-infected mice (4 × 50 mg/kg). However, compound 61 failed to produce significant in vivo efficacy at lower doses, displayed poor solubility and showed activity towards hERG. Further optimisation of this scaffold led to compound 62 (Figure 24), which was completely curative in P. berghei-infected mice (4 × 50 mg/kg) and retained high in vivo efficacy at lower doses [151]. Compound 62 acted as a prodrug which was rapidly metabolised to the highly active sulfone (compound 63, Figure 24) in vivo.
Compound 64 ( Figure 25) was identified during a high-throughput screening campaign against drug-sensitive (NF54) and -resistant (K1) P. falciparum strains [155]. Optimisation of the 2-aminopyridine scaffold resulted in compound 65 (MMV390048 or MMV048, Figure 25), which demonstrated in vitro and in vivo activity against the liver, blood and sexual (gametocyte) stages of the parasite [155,158]. Whole-genome screening of MMV048-resistant P. falciparum strains and chemoproteomic profiling identified plasmodial PI4K as the target of MMV048 [158]. MMV048 showed high selectivity over human and other plasmodial kinases, a good ADME profile and an acceptable preclinical safety profile in various animal species (mice, rats, dogs and monkeys) [155,158]. Phase I clinical trials for MMV048 were recently completed (ClinicalTrials.gov: NCT02230579; NCT02281344; NCT02554799) [159]. A single oral dose of up to 120 mg was generally well tolerated in healthy volunteers and adverse events were mild to moderate. Treatment with 20 mg of MMV048 initially reduced parasitemia in volunteers with induced P. falciparum blood-stage malaria; however, recrudescence occurred 2 to 7 days after treatment. Formulation influences the pharmacodynamic profile of MMV048, with the tablet formulation resulting in significantly less variability than the powder-in-a-bottle formulation. MMV048 progressed to phase 2a clinical trials in 2017, where its activity was evaluated in Ethiopian adults with either uncomplicated P. falciparum or P. vivax infection (ClinicalTrials.gov: NCT02880241).  Younis and co-workers [156] replaced the 2-aminopyridine core of MMV048 with a 2-aminopyrazine ring (compound 66, Figure 25), which improved the in vitro antimalarial activity but had poor solubility. In an attempt to improve the aqueous solubility, the methyl sulfonyl group of compound 66 was replaced with a piperazinyl carboxamide (UCT943, 67, Figure 25) [154]. While retaining PI4K selectivity, UCT943 demonstrated improved solubility and potency against all life cycle stages compared to the clinical candidate, MMV048 [154,160]. The in vivo efficacy of UCT943 in a P. falciparum-infected NSG mouse model was also 2-fold more potent than that of MMV048 [160]. Gibhard and co-workers [153] also explored the option of using the more soluble sulfoxide analogue of compound 66 as a prodrug to improve drug exposure in vivo. In a P. falciparum-infected NSG mouse model, the sulfoxide was rapidly absorbed and converted to its sulfone analogue (compound 66), which resulted in higher exposure compared to when the sulfone was administered.
Hit compound 68 (Figure 26), discovered during a phenotypic screening campaign, displayed weak in vitro activity against P. falciparum blood-stage parasites (EC 50 : 3.9 µM) but had potent Pf PI4K activity (IC 50 : 7.7 nM) [157]. Systematic optimisation of the bipyridine sulfonamide scaffold led to compound 69 ( Figure 26) which was selective for Pf PI4K and had potent activity against several drug-sensitive and -resistant P. falciparum strains. Compound 69 showed in vivo blood-stage efficacy (99.9% reduction in parasitemia at 80 mg/kg for 7 days) in a P. yoelii-infected mouse model and liver-stage efficacy (1 mg/kg, single dose) in a P. berghei-infected mouse model. In silico screening campaigns against Pf PI4K homology models have identified a number of virtual hits that can be used as potential starting points for drug development [149,161,162].

Orphan Kinases
Orphan kinases are particularly attractive drug targets as they do not have orthologues in the human host [11]. Some of the orphan kinases that have been characterised for P. falciparum include protein kinase 7 (Pf PK7), protein kinase 9 (Pf PK9) and the FIKK family [163].
Pf PK7 is a composite kinase that displays homology to the mitogen-activated protein kinase (MAPKK) family in its C-terminal region and fungal PKA homology in its N-terminal region [164]. Despite its homology, Pf PK7 is unlikely to be a functional MAPKK orthologue as the typical MAPKK activation site is absent and it is unable to phosphorylate the two MAPK orthologues (Pf map-1 and Pf map-2) in vitro [164]. Pf PK7 is expressed in asexual liver-and blood-stage parasites as well as in gametocytes and is localised in the cytoplasm [164]. Disruption of the pfpk7 gene decreases the growth rate of erythrocytic parasites and drastically reduces the parasite's ability to produce oocysts during the sexual stage [165].
Pf PK9 clusters at the base of the CDPK and AGC family branches but does not associate with either of the two groups [13]. Pf PK9 is essential for P. falciparum parasite viability [22,166] and is expressed during the late ring stages as well as the schizont stage where it exhibits maximal expression [167]. During the ring stages, Pf PK9 is localised to the parasitophorous vacuolar membrane, which acts as the interface between the parasite and the cytoplasm of the erythrocyte. As the parasite matures into schizonts, the localisation of Pf PK9 shifts to the parasite's plasma membrane [167]. This suggests that Pf PK9 is involved in signal transduction between the cytosol of the parasite and the intraerythrocytic environment. Thus far, only one downstream target has been identified for Pf PK9, namely E2 ubiquitin-conjugating enzyme 13 (Pf UBC13) [167]. Pf UBC13 is an orthologue of eukaryotic UBC13 which is involved in the attachment of lysine 63 (K63)-linked polyubiquitin chains to target proteins. This modulates the activity of various cellular processes such as DNA repair and immune responses.
The FIKK family, a group of serine/threonine kinases specific to apicomplexan parasites, is named after the phenylalanine (F)-isoleucine (I)-lysine (K)-lysine (K) motif located in the N-terminal region of their kinase domains. While most Plasmodium species only have a single FIKK kinase, 20 FIKK kinase members have been identified for P. falciparum [168]. Although the biological functions of this group of kinases are still unclear, evidence suggests that most FIKK kinases are involved in erythrocyte remodelling during infection [169,170]. Studies have identified nine FIKK kinases that are exported via the Maurer's clefts to the erythrocytic membrane, where remodelling occurs [169,171]. Disruption of individual genes encoding for Pf FIKK4.2, Pf FIKK7.1 or Pf FIKK12 significantly altered erythrocytic membrane rigidity and phosphorylation of certain cytoskeletal membrane proteins [170,172]. One such erythrocytic cytoskeletal protein, dematin, was also identified as a potential substrate for Pf FIKK4.1 [173].

Inhibitor Development for Orphan Kinases
As Pf PK7 plays a role in both the erythrocytic and sexual stages, Pf PK7 inhibitors could possibly decrease parasite virulence and act as transmission-blocking agents. This possibility makes Pf PK7 an interesting target for drug development. A number of established kinase inhibitors, including a MAPKK inhibitor (U0126) and PKA inhibitors (H89 & PKI), had no activity against recombinant Pf PK7 (IC 50 >100 µM) [164]. A high-throughput screening campaign identified imidazopyridines (compounds 75 and 76, Figure 28) and pyrazolopyrimidine (compound 77, Figure 28) with Pf PK7 and in vitro activity in the low micromolar range [174]. However, these compounds also inhibited a number of other kinases in the low micromolar range. The crystal structures of Pf PK7 in complex with adenylylimidodiphosphate (an ATP analogue, PBD: 2PML), compound 75 (PBD: 2PMN) and hymenialdisine (PBD: 2PMO) were elucidated. These structures highlighted some atypical features that are specifically relevant to drug discovery: Firstly, an aspartic acid residue (D123) in the hinge region protrudes to block access to the C-terminal domain surface in the ATP-binding site ( Figure 29A). This structural impediment explains the inactivity of most established protein kinase inhibitors towards Pf PK7. Secondly, a hydrophobic pocket was identified in the back of the ATP-binding site, which can be exploited for designing Pf PK7-selective drugs ( Figure 29B). Another high-throughput screening campaign identified a number of imidazopyridazines (e.g., 78, Figure 30) as weak Pf PK7 inhibitors [175]. Using the crystal structure data from the previous study [174], Bouloc and co-workers [175] aimed to improve the potency of the imidazopyridazine scaffold by varying the aryl and amine substituents of 78. This ultimately led to compounds 79 and 80 which demonstrated improved Pf PK7 activity (IC 50 : 0.28 µM and 0.13 µM, respectively). These compounds also showed antiplasmodial activity against drug-sensitive (3D7) and -resistant (K1) strains of P. falciparum without significant cytotoxicity. However, these compounds were also unselective and inhibited several other kinases. Klein and co-workers [176] aimed to design Pf PK7 inhibitors that exploit the hydrophobic pocket in the ATP-binding site. They designed two series of pyrazolopyrimidines with 1-and 4-substituted triazole rings at the 3-position of the pyrazole ring. The substituted triazoles were designed with a "bent" geometry that would allow the inhibitor to interact with the hydrophobic pocket. Two compounds (81 and 82, Figure 31) demonstrated Pf PK7 activity in micromolar concentrations (IC 50 : 20 and 10 µM, respectively). Docking studies with compound 81 indicated binding interactions similar to those seen with the ATP analogue but with additional interaction of the 4-phenyl-(1,2,3-triazol-1-yl) moiety with the hydrophobic pocket.

Pf PK9
Only one study has been published thus far with regards to inhibitor development for Pf PK9. Screening of a kinase-targeted library against Pf PK9 identified takinib (83, Figure 32), which demonstrated low micromolar binding affinity (K d(app) : 0.46 µM) for Pf PK9 [177]. Takinib is a potent inhibitor (IC 50 : 9.5 nM, [178]) of human mitogen-activated protein kinase kinase kinase 7 (MAP3K7, or more commonly referred to as TAK1). TAK1 s activity is regulated through K63-ubiquitination by human ubiquitin-conjugating enzyme, UBC13 [179]. Treatment of P. falciparum-infected erythrocytes with takinib resulted in a dose-dependent reduction in K63-ubiquitin levels, confirming that Pf PK9 regulates the activity of Pf UBC13 in vivo [167]. In order to achieve Pf PK9 selectivity over TAK1, a series of takinib analogues was developed. From this series, compound 84 ( Figure 32) was identified as a Pf PK9-selective inhibitor (K d(app) : 4.1 µM, 8.9-fold less potent than takinib) with antiparasitic activity against liver-stage P. berghei parasites (EC 50 = 43 µM) and no significant hepatocyte cytotoxicity. Compound 84 also showed a similar decrease in K63-linked ubiquitin levels in P. falciparum-infected erythrocytes, as seen with takinib. Interestingly, takinib and compound 84 both induced an unusual phenotype in liver-stage parasites. Liver-stage drugs generally decrease the size and/or number of parasites. However, when liver-stage P. berghei parasites were treated with 10 µM of either takinib or compound 84, the parasite size increased while the number of parasites remained unchanged. Treatment with 30 µM of either takinib or compound 84 also increased the parasite size but simultaneously decreased the number of parasites. This suggests a unique mechanism of action for these inhibitors, which has potential in new antimalarial drug development.

FIKKs
To date, five P. falciparum FIKK kinase members have been identified as essential for parasite survival, three of which are exported to the erythrocytic membrane (Pf FIKK9.1, Pf FIKK10.1 and Pf FIKK10.2) and two of which are localised within the parasite (Pf FIKK3 and Pf FIKK9.5) [171]. The non-exported FIKK8 kinase was also demonstrated to be essential to P. berghei erythrocytic parasites [180]. FIKK8 is the only FIKK kinase member that is conserved in all Plasmodium species as well as other apicomplexan parasites [181].

Summary
Recurring antimalarial drug resistance necessitates the development of new antimalarial drugs with different chemical scaffolds and modes of action. Plasmodial kinases were identified as promising targets for next-generation antimalarial drug development. To date, significant progress has been made towards characterisation and small-molecule inhibitor development for plasmodial kinases. Various plasmodial kinases have been validated as essential for one or multiple stages of the parasite's life cycle. Therefore, targeting plasmodial kinases could result in new drugs for chemoprevention and transmission-blocking, which will contribute to malaria elimination.
Plasmodial kinase inhibitors have been successfully identified by means of phenotypic and target-based screening approaches. The expression of several plasmodial kinases as active recombinant enzymes has facilitated crystallography and target-based medicinal chemistry efforts. Knowledge gained from developing human kinase inhibitors has also significantly contributed to plasmodial kinase inhibitor development. A number of potential scaffolds were identified through either phenotypic or target-based screening of human kinase inhibitor libraries.
Inhibitor promiscuity is always a concern when dealing with a conserved group of targets such as the kinases. Therefore, it is important to cross-screen potential plasmodial kinase inhibitors against human kinase panels. Inhibitors that target multiple plasmodial kinases could be beneficial as this ability would limit the risk of drug resistance, provided that this promiscuity does not also affect host kinases. Overall, studies targeting plasmodial kinases have demonstrated that selectivity over host kinases is an achievable goal.
Although a significant amount of work still needs to be done in terms of fully understanding the functions, stage specificity and interactions of plasmodial kinases, this group shows promise for future antimalarial drug development.

Conflicts of Interest:
The authors declare no conflict of interest.