Purine and Pyrimidine Pathways as Targets in Plasmodium falciparum

Abstract

Malaria is a leading cause of morbidity and mortality in the tropics. Chemotherapeutic and vector control strategies have been applied for more than a century but have not been efficient in disease eradication. Increased resistance of malaria parasites to drug treatment and of mosquito vectors to insecticides requires the development of novel chemotherapeutic agents. Malaria parasites exhibit rapid nucleic acid synthesis during their intraerythrocytic growth phase. Plasmodium purine and pyrimidine metabolic pathways are distinct from those of their human hosts. Thus, targeting purine and pyrimidine metabolic pathways provides a promising route for novel drug development. Recent developments in enzymatic transition state analysis have provided an improved route to inhibitor design targeted to specific enzymes, including those of purine and pyrimidine metabolism. Modern transition state analogue drug discovery has resulted in transition state analogues capable of binding to target enzymes with unprecedented affinity and specificity. These agents can provide specific blocks in essential pathways. The combination of tight binding with the high specificity of these logically designed inhibitors, results in low toxicity and minor side effects. These features reduce two of the major problems with the current antimalarials. Transition state analogue design is being applied to generate new lead compounds to treat malaria by targeting purine and pyrimidine pathways.

INTRODUCTION

Malaria causes 300 to 500 million clinical cases and 1.5 to 2.7 million deaths per year in the broad tropical regions across all continents [1]. Malaria is caused by protozoan parasites of the Plasmodium genus, of which Plasmodium falciparum is responsible for most of the fatal cases. P. falciparum infection of the human host is initiated by injection of sporozoites into the bloodstream by an infected female Anopheles mosquito. The sporozoites gain access to the liver and invade hepatocytes, where asymptomatic asexual multiplication (exoerythrocytic schizogony) occurs, leading to the production of several thousand merozoites. These are released into the bloodstream and invade erythrocytes. Erythrocytes are the site of major parasite expansion resulting from several rounds of asexual multiplication, where each parasite produces 8 to 24 new merozoites. This erythrocytic schizogony phase of the infection is responsible for malaria pathogenesis and therefore is the target for most antimalarial compounds. The erythrocyte is the metabolic warehouse for all essential metabolites for Plasmodium parasites to cause clinical disease. During a mosquito-borne infection, the parasite expansion is many billion-fold, requiring extraordinary resources from the host to support parasite growth.

Malaria has been treated for many years through chemotherapeutic and vector control strategies. These approaches have not prevented widespread disease occurrence and resurgence from the ecologically devastating use of insecticides. Currently, an alarming increase in both the resistance of malaria parasites to drug treatment and in mosquito vectors to insecticides renders the development of novel chemotherapeutic agents an urgent task [2, 3]. Advances in the knowledge of the metabolic and nutritional needs of the parasite offer new potential routes for chemotherapy. The rapid rate of nucleic acid synthesis during the intraerythrocytic growth phase makes purine and pyrimidine metabolic pathways promising targets for novel drug development.

P. falciparum is a purine auxotroph, salvaging host cell purines for synthesis of cofactors and nucleic acids [4, 5]. Erythrocytes contain millimolar concentrations of ATP in equilibrium with ADP and AMP hence the parasite has had no evolutionary pressure to retain de novo purine synthetic pathways. Purine nucleosides and nucleobases can be transported across the parasite plasma membrane by the P. falciparum NT1 transporter Fig. (1). The mechanisms by which Plasmodium salvage purines during the intraerythrocytic cycle are diverse with regard to primary sources and to routes of interconversion. Hypoxanthine is the key precursor for all purine synthesis in Plasmodium metabolism and is commonly used as a nutritional supplement in malarial culture media. A key source of hypoxanthine in vivo is from the erythrocyte purine pool, where ATP is in dynamic metabolic exchange with hypoxanthine via ADP, AMP, IMP, inosine and adenosine [5]. Human erythrocytes also lack de novo purine biosynthesis and maintain their adenine nucleotide pools by adenosine salvage from plasma by the action of adenosine kinase (AK, EC 2.7.1.20). The catalytic efficiency of adenosine kinase keeps erythrocytic adenosine at low concentrations and Plasmodium has not retained an adenosine kinase activity. Instead, P. falciparum can salvage adenosine by conversion to hypoxanthine using the sequential activities of adenosine deaminase (ADA, EC 3.5.4.4) and purine nucleoside phosphorylase (PNP, EC 2.4.2.1). Transport of most purine bases and nucleosides into P. falciparum are facilitated by the NT1 Fig. (1). Hypoxanthine is then converted to IMP by hypoxanthine-guanine-xanthine phosphoribosyltransferase (HGXPRT, EC 2.4.2.8 and EC 2.4.2.22). IMP serves as the metabolic precursor for all purine nucleotides and deoxynucleotides needed for nucleic acid synthesis. ADA, PNP and HGXPRT are highly expressed proteins in the Plasmodium [6, 7]. Since no adenosine kinase (AK) gene has been found in the P. falciparum genome [8] and no adenosine kinase activity is detected by metabolic labeling [9], the parasite can not directly convert adenosine to AMP. However, metabolic analysis revealed that P. falciparum can use AMP synthesized in the erythrocyte cytosol when high concentrations of exogenous adenosine are used to force erythrocytic AMP production [9].

Fig. (1).

Purine and pyrimidine metabolism in P. falciparum infected-erythrocytes. Purine pathway: AMP, adenosine 5′-monophosphate; ADP, adenosine 5′-diphosphate; ATP, adenosine 5′-triphosphate; IMP, inosine 5′-monophosphate; XMP, xanthosine 5′-monophosphate; GMP, guanosine 5′-monophosphate; hADA, human adenosine deaminase; MTA, methylthioadenosine; MTI, methylthioinosine; AdS, adenoylsuccinate; hPNP, human purine nucleoside phosphorylase; hHGPRT, human hypoxanthine-guanine phosphoribosyl transferase; hAK, human adenosine kinase; hAMPDA, human adenosine 5′-monophosphate deaminase; hAPRT, human adenine phosphoribosyl transferase; PfADA, P. falciparum adenosine deaminase; PfPNP, P. falciparum purine nucleoside phosphorylase; PfHGXPRT, P. falciparum hypoxanthine-guanine-xanthine phosphoribosyl transferase; PfAMPDA, P. falciparum adenosine 5′-monophosphate deaminase; PfIMPDH, P. falciparum inosine 5′-monophosphate dehydrogenase; PfGMPs, P. falciparum guanosine 5′-monophosphate synthase; PfAdSS, adenylosuccinate synthetase; PfAdSL, adenylosuccinate lyase. Pyrimidine pathway: OMP, orotidine 5′-monophosphate; UMP, uridine 5′-monophosphate; UTP, uridine 5′-triphosphate; CTP, cytidine 5′-triphosphate; PfCPSII, P. falciparum carbamoyl phosphate synthetase II; PfATCase, P. falciparum aspartate carbamoyltransferase; PfDHOase, P. falciparum dihydroorotase; PfDHODH, P. falciparum dihydroorotate dehydrogenase; PfOPRT, P. falciparum orotate phosphoribosyltransferase; PfODC, P. falciparum orotidine 5′-monophosphate decarboxylase; PfCTPs, P. falciparum cytidine 5′-triphosphate synthase. Bold arrows on reversible steps, indicates the metabolically favored direction. Nucleoside/nucleobase transporters are indicated on each membrane: human erythrocyte nucleoside transporter (brown), P. falciparum NT1 transporter (purple) and yet-to-be characterized adenosine 5′-monophosphate transporter (red).

Unlike the abundance of purines in human erythrocytes, pyrimidines exist at only small concentrations. Therefore, P. falciparum has retained the capacity for de novo synthesis of pyrimidines and does not have active pathways for the salvage of pyrimidines from the host. De novo synthesis from carbamoyl phosphate and aspartic acid follows essentially the same metabolic steps found in other eukaryotes, including the human host. In this six-step process, orotic acid is formed by dihydroorotase (DHOase; EC 3.5.2.3) and dihydroorotate dehydrogenase (DHODH; EC 1.3.3.1) followed by addition to 5′-phospo-D-ribosyl-α-1-pyrophosphate to form orotidine 5′-monophosphate (OMP) by orotate phosphoribosyltransferase (OPRT; EC 2.4.2.10). OMP is subsequently decarboxylated to uridine 5′-monophosphate (UMP), the precursor of all other pyrimidine nucleotides and deoxynucleotides needed for nucleic acid synthesis Fig. (1). Following phosphorylation to UTP, CTP synthase (CTPs; EC 6.3.4.2) converts UTP to CTP in a rate-limiting step, the only known route for de novo synthesis of cytosine derivatives. Conversion of purine and pyrimidine nucleotides to deoxyribonucleotides, as in most eukaryotes, takes place at the level of nucleoside diphosphates, catalyzed by ribonucleotide diphosphate reductase (RNR; EC 1.17.4.1). Formation of thymine nucleotides requires methylation of dUMP to produce dTMP by thymidylate synthase (TS; EC 2.1.1.45), another step that parallels human metabolism and makes the parasite succeptible to dihydrofolate reductase inhibitors that disrupt the folate pathway.

P. falciparum cell growth and division demand robust purine and pyrimidine supplies, in particular dATP and dTTP, since the parasite contains the most (A+T)-rich genome sequenced to date (approximately 80%). Recent applications of enzymatic transition state theory to the practical problem of transition state analogue design has permitted the design and synthesis of unique and powerful inhibitors of ribosyltransferases, phosphoribosyltransferases and deaminases, enzymes essential to the pathways of nucleic acid production in P. falciparum. Inhibition of the purine salvage pathway with transition state analogue (TSA) inhibitors of both human and Plasmodium PNP, such as Immucillin-H (Imm-H) [10] and DADMeImm-G [9], is lethal for P. falciparum in vitro. Coformycin is a picomolar, transition state analogue inhibitor of both human and P. falciparum ADAs [11]; 2′-deoxycoformycin, a related ADA inhibitor, have been reported to have antimalarial potential in primates [12, 13]. Phosphorylated Immucillins including Immucillin-H 5′-phosphate and Immucillin-G 5′-phosphate bind tightly to P. falciparum HGXPRT [14]. Traditional drug discovery applied to malaria has relied on screening for biological activity, with many compounds being used in humans without knowledge of their mechanism of action. Quinine was the first such example and this discovery approach continues into the 21^st century, where the artemisinins have gained wide use [15, 16]. Transition state analogue design is a departure by using transition state information for specific enzymes to provide powerful inhibitors as lead molecules with specific modes of action. Here we briefly review how TSA design is being applied to generate new lead compounds targeting purine and pyrimidine pathways related to the infective process of malaria parasites.

ENZYMATIC TRANSITION STATES AND ANALOG DESIGN

Enzymatic transition state analysis provides an understanding of catalysis from the perspective of bond changes at the catalytic site. Knowledge of the transition state permits the design of transition state inhibitors. All chemical transformations pass through the transition state, the point of equal probability for reactants on the reaction coordinate to be converted to product or to return to reactant. This structure is a characteristic of both the reactant and the catalyst; specifically for enzymes, involves the instantaneous geometry of the catalytic site at the transition state Fig. (2). Transition states have lifetimes similar to single bond vibrations, near 10⁻¹⁴ sec; methods are not available to directly observe transition state structures for enzymatic reactions. Transition state structure can be approached experimentally by measuring intrinsic kinetic isotope effects for the specific reaction. Computational chemistry is used to generate a quantum chemical structure of the transition state that matches the intrinsic isotope effects. This theoretical structure provides the blueprint for the transition state analogue design. Excellent accounts of this approach and its development are available [17–19].

Fig. (2).

(A) Substrate, transition state and transition state analogue inhibitors for P. falciparum PNP. Atoms of critical importance in the reaction and in the transition state analogue design are colored in blue. (B) Molecular electrostatic potential surfaces of inosine, PfPNP transiton state, Immucillin-H and DADMe-Immucillin-H. Blue is electron deficient and red is electron rich in this diagram. Note that the stick figure of inosine in (A) is rotated 180° in the vertical axis for presentation in (B).

Transition state theory proposes that a perfectly designed, chemically stable mimic of the transition state will bind better than the substrate to the target enzyme by a factor equivalent to the catalytic rate enhancement imposed by the enzyme. Since enzymes typically accelerate solution reactions by factors of 10¹⁰ to 10¹⁵, it is theoretically possible to design inhibitors binding with unprecedented affinity [19]. In practice, dissociation constants in the range of 10⁻¹³ M have been achieved [20]. Transition state analogues capture the simultaneous dynamic molecular excursions used by enzymes to generate the transition state. This state is linked to the dynamic path of protein conformational change that the enzyme has evolved to favor catalysis. In the presence of a transition state analogue, the protein structure collapses around the analogue into a static, thermodynamically equilibrated potential energy well. Release of the analogue requires the slow and energetically unfavorable conformational expansion of the protein. Release rates of transition state analogue inhibitors are very slow, from minutes to days, resulting in tight binding [21]. The combination of tight binding with the high specificity of these “custom-made” inhibitors, results in low toxicity and minor side effects, two of the major problems with the current antimalarials. A more detailed explanation about enzymatic transition states and transition state analogues can be found in the literature [22, 23].

PURINE SALVAGE METABOLISM ENZYMES

Adenosine Deaminase

Mammalian adenosine deaminases (ADA, EC 3.5.4.4) catalyze the physiologically irreversible hydrolysis of adenosine or 2′-deoxyadenosine to the corresponding (deoxy)inosine and ammonia. P. falciparum ADA also deaminates 5′-methylthioadenosine (MTA) in addition to adenosine, Fig. (1) [24]. Thus, P. falciparum ADA serves the dual functions of adenosine salvage and recycling MTA formed from the synthesis of polyamines [24]. Mammalian ADAs do not deaminate MTA and instead express a specific MTA phosphorylase for recycling of MTA [24]. Mammalian erythrocytes do not synthesize polyamines. Thus, an intact polyamine synthetic pathway is required for the viability of malaria parasites [24, 25]. In P. falciparum, MTA is deaminated by malarial ADA to 5′-methylthioinosine (MTI), a metabolite that has not been reported in mammalian metabolism [26]. The Plasmodium PNP also serves a dual purpose by converting both inosine and MTI to hypoxanthine for conversion to IMP and incorporation into nucleic acids [24].

The unique substrate specificity of the P. falciparum ADA led us to synthesize 5′-methylthio-coformycin (MT-coformycin) as a specific transition state analogue inhibitor of plasmodial ADAs [11]. MT-Coformycin is a sub-nanomolar inhibitor of malarial ADA and demonstrates >20,000 fold selectivity for Plasmodium ADA relative to human ADA. This selectivity is remarkable since coformycin, a transition-state analogue, and 2′-deoxy-coformycin are powerful picomolar inhibitors of both human and P. falciparum ADAs [11]. Thus, coformycin, inhibits P. falciparum and human ADA with dissociation constants of 60 and 100 pM, respectively. The chemical structure of the P. falciparum and human ADA transition states resembles the sp³-hybridized Meisenheimer intermediate, rather than the reactant adenosine and the product inosine [27]. Although transition state features are similar for the reaction centers in human and malarial ADAs, distinct substrate specificity permits the design of parasite-specific transition state analogues. The tight binding of coformycin and 2′-deoxycoformycin with adenosine deaminases support a capture of transition state similarity as inhibitor binding energy.

The crystal structure of ADA from Plasmodium vivax in complex with MT-coformycin revealed an unprecedented binding geometry for 5′-methylthioribosyl groups in the malarial ADAs [13]. Compared to malarial ADA complexes with adenosine or 2′-deoxycoformycin, the 5′-methylthioribosyl groups are rotated 130°. A hydrogen bonding network between Asp172 and the 3′-hydroxyl of MT-coformycin is essential for recognition of the 5′-methylthioribosyl group. Water occupies the 5′-hydroxyl binding site when MT-coformycin is bound. Mutagenesis of Asp172 abrogates the substrate specificity for MTA and MT-coformycin. Coformycin alone does not inhibit parasite growth in cultured erythrocytes, a demonstration that host and parasite ADAs are not essential for growth in culture media [9]. However, when MTA was used as the sole purine source in cultured P. falciparum, we demonstrated that P. falciparum growth is inhibited by coformycin or MT-coformycin [13]. One indication that ADA may act as a target for drug design comes from a report that 2′-deoxycoformycin (Pentostatin) causes decreased parasitemia in P. knowlesi-infected primates [12]. Studies of MT-coformycin in Plasmodium-infected primates would test the hypothesis that the parasite ADA is important in establishing a viable infection.

Purine Nucleoside Phosphorylase

Mammalian purine nucleoside phosphorylases (PNP, EC 2.4.2.1) convert inosine, deoxyinosine, guanosine and deoxyguanosine into hypoxanthine and guanine, in each case releasing α-D-(deoxy)ribose 1′-phosphate. Daddona and coworkers first described the activity of P. falciparum PNP using protein extracts of parasites grown in erythrocytes from PNP-deficient donors. They were able to show that infection restores PNP-activity to high levels and that the parasite enzyme showed a approximately four-fold lower K_M for both inosine and guanosine than human PNP [28]. Two decades later we expressed recombinant P. falciparum PNP and confirmed the high affinity of P. falciparum PNP for its substrates and also showed the decreased catalytic efficiency for 2′-deoxyinosine and 2′-deoxyguanosine. These findings support a metabolic role for P. falciparum PNP being hypoxanthine release from the host cell inosine obtained from the erythrocyte ATP pool [29]. In contrast, the role of human PNP is suppression of blood deoxyguanosine, which otherwise leads to a specific T-cell deficiency autoimmune disorder in humans [30].

The crystal structure of P. falciparum PNP bound to the transition state analogue Imm-H and a sulfate ion revealed a large solvent-filled cavity in the malarial PNP active site and is located near the 5′-hydroxyl group of Imm-H. This discovery distinguished the parasite enzyme from its human counterpart and indicated that P. falciparum PNP can accept bulky substitutions at the 5′-position. Kinetic investigations and structural analysis identified 5′-methylthioinosine as a good substrate for P. falciparum PNP [7]. Further metabolic investigation of the parasite identified a second role (see above) for malarial PNP and its partner, adenosine deaminase, in recycling of 5′-methylthiopurines formed in polyamine metabolism Fig. (1) [24].

Inhibitor design for PNP began with the transition state structure of PNP prepared from calf spleen. The transition state model showed a very early S_N1-like transition state with an N-ribosyl bond distance of 1.8 Å, the ribosyl group exhibiting partial carbocation character, an increased pK_a of the purine N7; no significant participation of the attacking anion nucleophile (arsenate was used to replace phosphate) [31]. Later, a comparative transition state analysis of the human and P. falciparum PNPs demonstrated these PNPs to have nearly identical transition states, even though their steady-state kinetic properties were distinct [32]. These transition states differed from bovine PNP in one key respect. The distance between ribose and hypoxanthine at the transition state was 1.8 Å in bovine PNP but was 3.0 Å for both human and P. falciparum PNPs. The chemically large difference in transition state structures distinguishes both the geometry and the charge of these transition states. Human and P. falciparum PNPs both generate a fully-developed ribocation at the transition state while bovine enzyme does not.

Using the electrostatic potential map of the transition state model of the bovine PNP transition state as a guide, we designed the transition state mimic Imm-H, which binds over 700,000 times tighter than substrate to the bovine enzyme [33]. Imm-H also shows strong, time-dependent inhibition of P. falciparum PNP to give a dissociation constant of 0.6 nM, 9,000-fold greater than the K_M value for substrate [29]. The biological availability of Imm-H was demonstrated by the robust killing of P. falciparum in cell culture assays without excess hypoxanthine in the growth medium [10]. This killing was termed “purine-less death” since inhibition of PNP with Imm-H induced purine starvation. The Imm-H blocked both human and malarial PNPs and was rescued only by supraphysiological levels of hypoxanthine.

Design of Plasmodium-specific inhibitors was based on the dual substrate specificity found for P. falciparum PNP. Since P. falciparum PNP is active on inosine and 5′-methylthioinosine, 5′-methylthio-Immucillin-H (MT-Imm-H) was synthesized and tested. Disruption of the purine pathways with MT-Imm-H inhibited parasite growth in cell-culture assay with an IC₅₀ of 50 nM similar to that of Imm-H. Parasite growth is slowed at low nanomolar concentrations; however, complete inhibition of growth was not achieved until MT-Imm-H reached micromolar concentrations [24]. The two-phase response curve to MT-Imm-H indicated that while P. falciparum PNP is important for normal growth of the parasite, human PNP must also be inhibited to completely block purine salvage in the cultured parasites.

Second-generation Immucillins, the DADMe-Immucillins, incorporate the specific features of the transition state structures specific for human and P. falciparum PNPs. A methylene linker between the iminoribitol and purine ring more accurately mimics the geometry, electrostatics and relative positions of the ribocation group and the hypoxanthine leaving group [34]. DADMe-Immucillin-H and DADMe-Immucillin-G inhibit human PNP with K_i values of 16 pM and 7 pM [21] and P. falciparum PNP with K_i values of 500 pM and 890 pM, respectively. As a proof of concept of the applicability of transition state analogues in vivo, a single oral dose of DADMe-Immucillin-H is able to completely inhibit erythrocyte PNP activity in mice for the entire duration of the erythrocyte life cycle [21].

Hypoxanthine-Guanine-Xanthine Phosphoribosyltransferase

Given the central role of hypoxanthine in the metabolism of P. falciparum, hypoxanthine/guanine (EC 2.4.2.8); xanthine phosphoribosyltransferase (EC 2.4.2.22) activities are critical for purine salvage. In P. falciparum, both activities are found in a single enzyme called HGXPRT. The malarial HGXPRT catalyzes the transfer of the phosphoribosyl moiety of PRPP to the 6-oxopurine nucleobases creating the respective monophosphorylated nucleotide and inorganic pyrophosphate. The human enzyme does not use xanthine as substrate, marking the divergence of the host and parasite purine salvage pathways. In humans, formation of xanthine constitutes irreversible loss from the purine pool since the only fate of xanthine in humans is oxidation to uric acid by xanthine oxidase. Purine metabolites earlier than xanthine in human and parasite metabolic pathways freely exchange between the host and parasite cytosols via the P. falciparum NT1 [35]. Other transporters are also likely to be involved since NT1 is the only characterized of four purine transports predicted to be present in the Plasmodium genome [36]. After the phosphoribosyl transfer, the purine nucleotides are sequestered in the parasite cytosol by the 5′-phosphate charges. Early studies of HGXPRT isolated from whole-cell lysates identified the unique ability of P. falciparum HGXPRT to use xanthine, hypoxanthine and guanine as substrates [37, 38]. Studies with recombinant HGXPRT were complicated by the low catalytic activity of P. falciparum enzyme expressed in Escherichia coli. Detailed analysis of preincubation conditions solved this problem of converting the inactive conformation to achieve an appreciable increase in catalytic activity [39]. While there are differing reports on the specific activity of P. falciparum HGXPRT, most agree that the K_M of the enzyme for the nucleobase hypoxanthine is submicromolar, well below that of the human HGPRT [37–41]. This high affinity is indicative of the metabolic role for P. falciparum HGXPRT of binding free hypoxanthine, guanine and xanthine and preventing diffusional release until the enzyme can convert these bases to the nucleotide monophosphates, thus sequestering them into the parasite.

Tight binding of substrates to P. falciparum HGXPRT prevent the measurement of intrinsic kinetic isotope effects; hence, no transition state of a purine phosphoribosyltransferase has been solved. However, enzymes catalyzing similar reactions have had their transition states solved Fig. (3). Transition state structures have been solved for the orotate phosphoriboyltransferases from human, P. falciparum and Salmonella typhimurium sources, using a slow substrate analogue in place of pyrophosphate [42, 43]. These transition states show a highly-dissociative ribocation character with weak participation of the attacking nucleophiles. Other ribosyltransferases are similar to the reaction of HGXPRT, in that the transition state involves bond breaking between ribose and a 6-oxypurine leaving group. Specifically, PNP reactions have this type of transition state. As these ribosyltransferases all have ribocation character at their transition states, the transition state of HGXPRT may be similar.

Fig. (3).

Comparison of the HGXPRT reaction (A) to similar reactions in which the transition state structures have been solved by the combination of kinetic isotope effects and computational chemistry described in the text. The reaction in (B) is for purine nucleoside phosphorylase and in (C) for orotate phosphoribosyltransferase. The pyrophosphate products are not showed.

Transition state analogue design elements from the PNP and OPRT reactions were used to design inhibitors of P. falciparum HGXPRT. The 5′-phosphorylated Immucillins, Immucillin-H 5′-P (Imm-HP) and Immucillin-G 5′-P (Imm-GP) are low nanomolar inhibitors for both the human HGPRT and P. falciparum HGXPRT [14]. Crystal structures of the human and parasite enzymes bound to Imm-GP and Imm-HP, respectively, did not reveal significant differences in their active sites that can be readily exploited [44, 45]. However, recent work from John de Jersey’s group has reported a family of linear-phosphonate compounds linked to purine bases that are selective for the parasite relative to the human purine phosphoribosyltransferases [46]. A significant barrier for inhibitor design against HGXPRT is the apparent need for the 5′-phosphate group for tight binding to the enzyme. Nucleotides are not easily transported into cells and are readily hydrolyzed in the gut, making oral delivery problematic. Prodrug protecting group chemistries are possible, but these increase the cost of goods to beyond useful ranges for most tropical disease markets.

Downstream Enzymes in the Purine Salvage Metabolism

Although the discussion above has focused on early steps in purine salvage pathways, all downstream enzymes (IMP → nucleic acid precursors) are also viable targets. Of these, only the guanosine monophosphate synthase (GMPs, EC 6.3.5.20) [47] adenylosuccinate synthetase (AdSS, EC 6.3.4.4) [48, 49] and adenylosuccinate lyase (AdSL, EC 4.3.2.2) [50] had been characterized in P. falciparum. GMPs catalyses the last step in GMP formation in the purine salvage pathway Fig. (1). The irreversible conversion of XMP into GMP is a reaction that requires ATP and involves the transfer of an amino group from glutamine to the C-2 carbon of XMP via an adenyl-XMP intermediate. Adenylosuccinate lyase is known to catalyze two distinct but chemically similar reactions in purine biosynthesis. One reaction is exclusive to the de novo pathway and involves cleavage of 5-aminoimidazole-4-(N-succinylcarboxamide) ribonucleotide (SAICAR) to 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) and fumarate but this pathway is absent in P. falciparum. The second reaction catalyzed by this enzyme is required downstream of IMP and involves the cleavage of succinyladenosine monophosphate (SAMP) to AMP and fumarate. P. falciparum AdSL catalyzes the cleavage of SAICAR with a K_M of 40.0 ± 8 μM and binds AICAR with affinity similar to that of AMP [50]. Binding of AICAR has no function in P. falciparum metabolism, but this residual memory of its evolutionary history can be explored as a potential target for drug design.

DE NOVO PYRIMIDINE METABOLISM ENZYME

Carbamoyl Phosphate Synthetase II

Carbamoyl phosphate synthetase II (CPSII, EC 6.3.5.5) catalyzes the formation of carbamoyl phosphate in the cytosol from bicarbonate, glutamine and ATP. This is the first step of de novo pyrimidine biosynthesis. CPSII is activated by ATP and PRPP and inhibited by UTP [51]. P. falciparum CPSII is a monofunctional protein, while human CPSII activity is included in a trifunctional protein named CAD [5]. The mRNA transcript of P. falciparum CPSII possesses two inserted sequences that are absent in other homologues. These insertions are located between junctions of the glutamine aminotransferase and synthetase domains [52]. A synthetic ribozyme with specificity for the P. falciparum CPSII gene has been shown to have potent inhibitory effect on growth of cultured P. falciparum parasites and no toxicity to mammalian cells [53]. Encoded amino acid insertions are common in Plasmodium transcripts, are unique to the parasite and provide potential targets for nucleic acid therapy. Hendry and colleagues have redesigned and tested a series of miniribozymes with high specificities against P. falciparum CPSII gene in vitro [54]. These ribozyme derivatives with chemical modifications exhibit improved cleavage activities and metabolic stabilities. Although ribozyme approaches are technologically amusing, the barriers to application in tropical medicine are significant, including target accessibility, stability, specificity and delivery efficiency [55].

Dihydroorotase

Dihydroorotase (DHOase, EC 3.5.2.3), the third enzyme in the pyrimidine biosynthesis pathway, catalyzes the reversible cyclization of N-carbamoyl-L-aspartate (CA-asp) to L-dihydroorotate (DHO). P. falciparum DHOase is a mono-functional protein with sequence most similar to Escherichia coli. Mammalian DHOase is the central sequence of the CAD trifunctional protein [56]. 5′-Substituted orotate derivatives are weak competitive inhibitors of P. falciparum DHOase [57]. L-6-Thiodihydroorotate (TDHO) is a moderate inhibitor of malarial DHOase with a K_i of 6.5 μM. Inhibition of DHOase in P. falciparum by TDHO or its methyl ester results in accumulation of CA-asp in cultured parasites [58]. Several dihydropyrimidine analogues have been synthesized and exhibit inhibition against mammalian DHOase at nanomolar concentrations [59–61].

Dihydroorotate Dehydrogenase

Dihydroorotate dehydrogenase (DHODH, EC 1.3.5.2) is a flavin mononucleotide (FMN)-dependent catalyst for the conversion of dihydroorotate (DHO) to orotate. There are two types of DHODH. Type I is a cytosolic enzyme found in Saccharomyces and certain protozoa and requires fumarate or NAD⁺ as electron acceptor. P. falciparum and humans express type II DHODH, a membrane-bound protein linked to ubiquinone (CoQ) as the electron acceptor [62, 63]. Malarial and human DHODH reside in the outer side of the inner mitochondrial membrane. The first half of reaction catalyzed by P. falciparum and human DHODHs involves oxidation of DHO to orotate coupled with electron transfer to FMN. The electron pair is then transferred to CoQ in the second half of reaction, with concomitant recycling of FMNH₂ to FMN.

Human DHODH is a validated drug target. Leflunomide has been used for the treatment of rheumatoid arthritis [64, 65]. Its active metabolite A77-1726 exhibits potency against human DHODH activity by acting as CoQ analogue [66]. Brequinar, a potent 4-quinolinecarboxylic acid inhibitor of human DHODH, has been used as an anti-tumor and immunosuppressive agent [67, 68].

By employing structure-based computational drug design, Davies and colleagues have synthesized a variety of A77-1726 derivatives with good inhibitory potential against P. falciparum DHODH [69]. From a library containing 220,000 drug-like molecules, Baldwin and colleagues identified a competitive inhibitor of P. falciparum DHODH with an IC₅₀ value of 16 nM and 12,500-fold selectivity over the human enzyme [70]. Chemical library screening of >200,000 compounds by Patel and colleagues has identified potent inhibitors with different scaffolds for malarial DHODH [71]. These species-specific inhibitors display submicromolar IC₅₀ values against cultured P. falciparum and have low toxicity for human cells. Phillips and colleagues have identified a triazolopyrimidine-based compound with potent antimalarial activity and high specificity for P. falciparum DHODH [72]. This P. falciparum DHODH-specific inhibitor arrests parasite growth in vitro with an IC₅₀ of 0.14 μM for the Dd2 strain. However, this compound has no activity in vivo due to its metabolic instability. In contrast, phenyl-substituted triazolopyrimidines are metabolically stable and possess similar potent inhibitory and selectivity on P. falciparum DHODH. A p-trifluoromethylphenyl exhibits prolonged exposure in vivo and suppresses the parasitemia in P. berghei infected mice [73]. Tricyclic aromatic amines have also been identified as potent inhibitors against malarial DHODH with low nanomolar binding affinities [74]. In vitro studies show that these inhibitors block the growth of cultured parasites at submicromolar concentrations.

Atovaquone, a CoQ analogue, has an IC₅₀ of 1.0 nM against cultured P. falciparum [75, 76]. By forming a covalent bond with a protein from Complex III in the parasite mitochondria, atovaquone collapses the membrane potential and causes cell death and the indirect inhibition of P. falciparum DHODH activity [77–79]. The selective action of atovaquone on the parasite electron transport chain results from the structural difference in the CoQ complexes of humans and parasites [5]. Malarone^™, an antimalarial approved by the U.S. Food and Drug Administration (FDA), is a combination of atovaquone and proguanil. Proguanil is a prodrug of cycloguanil, a dihydrofolate reductase inhibitor and therefore brings a synergistic effect to atovaquone action [80].

Orotate Phosphoribosyltransferase

Orotate phosphoribosyltransferase (OPRT, EC 2.4.2.10) is the fifth enzyme in the de novo pyrimidine biosynthesis pathway. It catalyzes the formation of orotidine 5′-monophosphate (OMP) from α-D-phosphoribosylpyrophosphate (PRPP) and orotate. In most prokaryotes OPRT and OMP decarboxylase (ODC) activities are encoded in separate proteins. However, in eukaryotes, uridine 5′-monophosphate (UMP) synthase is bifunctional and contains both enzyme activities [62, 81, 82]. Transition state structures of P. falciparum and human OPRT were analyzed by KIEs, computational chemistry and substrate specificity [42]. Both P. falciparum and human OPRT form late associative transition states with complete orotate loss and partially associative nucleophile. p-Nitrophenyl β-D-ribose 5′-phosphate is a poor substrate of both enzymes, but is a nanomolar inhibitor. The use of these agents as antimalarials has not been reported.

5′-Fluoroorotate is an alternative substrate for P. falciparum OPRT with the same kinetic parameters as orotate [83]. It inhibits the in vitro growth of P. falciparum with an IC₅₀ less than 10 nM [84, 85]. The potent antimalarial activity of 5′-fluoroorotate is related to the inactivation of malarial thymidylate synthase by 5′-fluoro-2′-deoxy-UMP metabolite through covalent binding to methylene tetrahydrofolate at the active site. 5′-Fluoroorotate clears parasitemia from mice infected with P. berghei [86]. Other 5′-substituted orotate analogues display competitive inhibitory effects against P. falciparum OPRT activity, but the antimalarial activities of those compounds are at least 100-fold less than 5′-fluoroorotate [83]. Uracil and 5′-fluorouracil are weak competitive inhibitors of P. falciparum OPRT and reduce P. falciparum in vitro growth with IC₅₀ values of 5 to 10 μM [84, 87].

Pyrazofurin, a C-riboside antibiotic, is a modest inhibitor of P. falciparum OPRT. Its antimalarial activity exhibits IC₅₀ values of 6 to 24 μM by blocking the maturation of trophozoites to schizonts [84, 88]. Since Plasmodium lack enzymes for pyrimidine salvage, the toxicity was not relieved by the addition of uracil or uridine to the culture. Pyrazofurin does not inhibit the OPRT activity in mammalian cells [89]. However, its 5′-monophosphate derivative is a potent inhibitor against both mammalian and malarial ODC [58, 90, 91].

Orotidine 5′-Monophosphate Decarboxylase

Orotidine 5′-monophosphate decarboxylase (ODC, EC 4.1.1.23), has received considerable attention as being one of the most proficient enzymes. It catalyzes the decarboxylation of orotidine 5′-monophosphate (OMP) to uridine 5′-monophosphate (UMP) with a rate enhancement in excess of 10¹⁷-fold [92]. Unlike most decarboxylases, ODC activity does not require the presence of a cofactor or metal ion [93]. ODC-catalyzed conversion of OMP to UMP in the final step of de novo pyrimidine biosynthesis pathway. The ODC and orotate phosphoribosyltransferase (OPRT) activities in most prokaryotes and lower eukaryotes are distinct proteins but are expressed as a bifunctional UMP synthase in higher eukaryotes [82].

Nucleotide 5′-monophosphate analogues of P. falciparum ODC, including 6-aza-UMP, allopurinol 5′-monophosphate, pyrazofurin 5′-monophosphate and xanthosine 5′-monophosphate (XMP) are competitive inhibitiors with tighter binding than OMP [94]. Of these, XMP is highly selective with a K_i of 4.4 nM for P. falciparum ODC and 670 nM for human ODC, a 150-fold preference for the malarial enzyme. 6-Aza-UMP has a K_i of 12 nM for P. falciparum ODC and 510 nM for human ODC, 43-fold selectivity P. falciparum ODC. Pyrazofurin 5′-monophosphate is a potent inhibitor against P. falciparum ODC with K_i of 3.6 nM.

6-Iodouridine 5′-monophosphate (6-iodo-UMP) and 6-azidouridine 5′-monophosphate (6-N₃-UMP) are potent, covalent inhibitors of P. falciparum ODC [95, 96]. Mass spectral analyses and crystal structures of enzyme-inhibitor complexes reveal that both inhibitors form covalent adducts to Lys138. 6-Iodouridine, the nucleoside form of 6-iodo-UMP, shows antimalarial activities with an IC₅₀ of 4.4 μM and 6.2 μM against P. falciparum ItG strain (chloroquine-resistant) and 3D7 strain (chloroquine-susceptible). Cytotoxicity against CHO cells gives an IC₅₀ of 366 μM for 6-iodouridine, 60-fold lower than toxicity against parasites. Unlike 6-iodouridine, 6-azidouridine (nucleoside form of 6-N₃-UMP) does not inhibit cultured P. falciparum growth, possibly from inefficient cellular phosphorylation.

Barbiturate 5′-monophosphate (BMP) is a powerful inhibitor of human and yeast ODC with K_i of 9 pM [97, 98] but its complex inhibitory effect on P. falciparum ODC activity has not been mechanistically resolved [94]. Other 6-substituted UMP derivatives are competitive inhibitors of P. falciparum ODC with submicromolar K_i values [96]. For the nucleoside forms of these 6-substituted derivatives, only 6-N-methylamino uridine and 6-N,N-dimethylamino uridine exhibit moderate antimalarial activities with IC₅₀ of 28 μM and 32 μM, respectively. Other derivatives have no inhibitory effect against cultured P. falciparum. Despite the inhibitory potential for the pyrimidine nucleoside and nucleotide analogues listed above, concerns remain for side effects with those analogues capable of being incorporated into host nucleic acids.

Deoxyuridine 5′-Triphosphate Nucleotide Hydrolase

Deoxyuridine 5′-triphosphate nucleotidohydrolase (dUT-Pase, EC 3.6.1.23) catalyzes the hydrolysis of deoxyuridine 5′-triphosphate (dUTP) to deoxyuridine 5′-monophosphate (dUMP) and inorganic pyrophosphate. Because most DNA polymerases can not distinguish between dUTP and dTTP, misincorporation of dUTP into DNA depends on low levels of dUTP. Loss of this enzyme activity generates cells with high mutation rates [99]. The product dUMP is the precursor for dTMP synthesis. In many organisms dCMP deamination also generates dUMP, but Plasmodium only encode dCTP deaminases [99]. Thus, dUTPase appears to be essential for dUMP synthesis. The expression of P. falciparum dUTPase is associated with parasite proliferative stages [100]. P. falciparum dUTPase is strictly specific for dUTP with discrimination against ribosyl and base changes [101]. It functions as homotrimer with similar structural fold and subunit organization to human dUTPase [100].

Triphenylmethane derivatives of deoxyuridine are reported to be selective inhibitors against P. falciparum dUT-Pase and exhibit moderate inhibitory effects on in vitro growth with IC₅₀ in the low micromolar range [100]. Nguyen and colleagues have identified a number of 5′-tritylated deoxyuridine analogues with antimalarial activities and more than 200-fold selectivities relative to human dUTPase [102]. Acyclic nucleoside derivatives also show inhibitory effects against P. falciparum dUTPase [103]. The most potent acyclic analogues (containing triphenylsilyloxy groups) arrest P. falciparum in vitro growth with an IC₅₀ of 0.4 μM and display low toxicity to mammalian cells. Tritylated uracil acetamide derivatives exhibit selective action on P. falciparum dUTPase and moderate antimalarial activities [104].

Dihydrofolate Reductase

Dihydrofolate reductase (DHFR, EC 1.5.1.3) is indispensable for de novo folate biosynthesis, salvage of folate derivatives and recycling of dihydrofolate (DHF) formed in dTMP synthesis [105]. P. falciparum DHFR activity is the N-terminal portion of the bifunctional protein thymidylate synthase-dihydrofolate reductase (TS-DHFR) [106, 107]. Human DHFR and TS activities are present in two separate proteins. In P. falciparum the TS-DHFR catalyzes two essential reactions, dTMP synthesis and the conversion of DHF to methylenetetrahydrofolate (CH₂H₄folate). DHFR is one of the best established, multiply-targeted enzymes in malaria.

Pyrimethamine is a potent inhibitor of P. falciparum DHFR with K_i of 0.2 nM [108]. In combination with sulfadoxine, an inhibitor of dihydropteroate synthase (DHPS), pyrimethamine is marketed as Fansidar^™. The combination of DHFR and DHPS inhibitors is synergistic since Plasmodium also synthesizes the dihydropteroate group needed for folate synthesis, a process not found in humans. Cycloguanil, the active metabolite of proguanil, shares the 2, 4-diamino scaffold with pyrimethamine and exhibits potency against P. falciparum DHFR (K_i = 0.3 nM) [108]. Lapdap^™ is a combination of chloroproguanil (chlorcycloguanil prodrug, an inhibitor of DHFR) and dapsone (an inhibitor of DHPS) [109]. This antimalarial is effective against pyrimethamine-resistant P. falciparum strains [110]. The diamino triazine derivative WR99210 has a flexible trichlorophenoxypropyloxy side chain and is effective against the pyrimethamine-resistant parasites [111, 112]. The K_i value of WR99210 is in the subnanomolar range for native P. falciparum TS-DHFR as well as for double and quadruple mutant strains [108]. The cytotoxicity and poor oral bioavailability of WR99210 prevent its clinical use [113, 114]. PS-15, the phenoxypropoxybiguanide precursor of WR99210, shows improved bioavailability and tolerance by the host cells [115].

Analogues of pyrimethamine and cycloguanil have been developed against drug-resistant strains. Several of these are low nanomolar inhibitors of P. falciparum DHFR from mutant strains and show low toxicity to mammalian cells [116]. Trimethoprim (TMP) derivatives have also been developed and show antimalarial activities against parasites with mutant DHFR enzymes [117]. Through in silico screening, Dasgupta and colleagues identified three inhibitors of P. falciparum DHFR with IC₅₀ of 20 μM for wild type or antifolate-resistant P. falciparum strains [118]. These biguanide compounds exhibit highly specific action on parasites. Antisense oligodeoxynucleotides with sequences targeting mRNA transcripts of P. falciparum DHFR also show selective growth inhibition on cultured parasites (IC₅₀ ~ 0.05 μM) [119].

Thymidylate Monophosphate Kinase

Thymidylate monophosphate kinase (TMK, EC 2.7.4.9) catalyzes the reversible phosphorylation of dTMP to deoxythymidine diphosphate (dTDP), an essential step for cellular DNA synthesis. TMK from Mycobacterium tuberculosis is a target for antituberculosis agents with bicyclic thymidine derivatives providing micromolar binding affinities to M. tuberculosis TMK [120]. Thymidine analogues with 5′-thiourea-substituents inhibit M. tuberculosis TMK at nanomolar concentrations. These compounds also act on malarial enzyme TMK, display moderate antituberculosis activities, and exhibit 600-fold selectivity relative to human TMK [121]. P. falciparum TMK has broad substrate specificity including dTMP, dUMP and purine nucleotides [122]. This specificity distinguishes malarial TMK from other homologues including the human enzyme and makes the Plasmodium enzyme a viable target. A few nucleoside analogues display low micromolar inhibition activity against P. falciparum TMK [123].

Cytidine Triphosphate Synthetase

Cytidine triphosphate synthetase (CTPs, EC 6.3.4.2) catalyzes the conversion of UTP to CTP. The de novo pyrimidine biosynthesis pathway-catalyzed formation of CTP is the only route leading to cytosine derivatives [5]. CTPs is actively expressed in P. falciparum parasites [124]. P. falciparum CTPs protein contains two unique peptide segments, differing from human CTPs, suggesting a potential target for drug design [125]. Although there are no potent inhibitors against this enzyme, 3-deazauridine and cyclopentenyl cytosine are converted to triphosphate derivatives in mammalian cells and become potent inhibitors of mammalian CTPs [126, 127]. These nucleoside analogues have shown anti-tumor and anti-HIV activities [128–131].

SUMMARY

Targeting the catalytic machinery of Plasmodium species in the design of antimalarials has only begun. Differences in protein structure and specificity provide one approach. Differences in the host and parasite metabolism provide another. Finally, the trend toward drug combinations for malaria is expected to continue and broaden as powerful new inhibitors are brought to bear on novel targets.