Malarial Dihydroorotate Dehydrogenase

The malarial parasite relies onde novo pyrimidine biosynthesis to maintain its pyrimidine pools, and unlike the human host cell it is unable to scavenge preformed pyrimidines. Dihydroorotate dehydrogenase (DHODH) catalyzes the oxidation of dihydroorotate (DHO) to produce orotate, a key step in pyrimidine biosynthesis. The enzyme is located in the outer membrane of the mitochondria of the malarial parasite. To characterize the biochemical properties of the malarial enzyme, an N-terminally truncated version of P. falciparum DHODH has been expressed as a soluble, active enzyme in E. coli. The recombinant enzyme binds 0.9 molar equivalents of the cofactor FMN and it has a pH maximum of 8.0 (k cat 8 s−1, K m app DHO (40–80 μm)). The substrate specificity of the ubiquinone cofactor ( CoQ n ) that is required for the oxidation of FMN in the second step of the reaction was also determined. The isoprenoid (n) length of CoQ n was a determinant of reaction efficiency; CoQ4, CoQ6 and decylubiquinone ( CoQD) were efficiently utilized in the reaction, however cofactors lacking an isoprenoid tail (CoQ0 and vitamin K3) showed decreased catalytic efficiency resulting from a 4 to 7-fold increase in K m app . Five potent inhibitors of mammalian DHODH, Redoxal , dichloroallyl lawsone ( DCL ), and three analogs of A77 1726 were tested as inhibitors of the malarial enzyme. All five compounds were poor inhibitors of the malarial enzyme, with IC50's ranging from 0.1–1.0 mm. The IC50 values for inhibition of the malarial enzyme are 102-104-fold higher than the values reported for the mammalian enzyme, demonstrating that inhibitor binding to DHODH is species specific. These studies provide direct evidence that the malarial DHODH active site is different from the host enzyme, and that it is an attractive target for the development of new anti-malarial agents.

Malaria afflicts between 500 and 900 million people worldwide and causes greater than 2 million deaths per year (1,2), making this parasitic disease an enormous public health problem throughout the developing world. Currently, several medicinal therapies are available for use as prophylaxis and treatment for at-risk or infected individuals. However, widespread drug resistance against these agents (e.g. chloroquine (3), atovaquone (4), pyrimethamine (5), and sulfadoxine (6)), has compromised the effectiveness of these treatments resulting in the pressing need for the development of new anti-malarial compounds. The identification of targets that exploit the unique biology of the parasite is an essential step in the development of new therapeutics.
Pyrimidines are essential metabolites in all cells. They are required not only for DNA and RNA biosynthesis, but also for the biosynthesis of phospholipids and glycoproteins. The de novo pyrimidine biosynthetic pathway is intact in most organisms, including Plasmodium (7)(8)(9). However, unlike mammalian cells, the human malaria parasite, Plasmodium falciparum, cannot salvage preformed pyrimidine bases or nucleosides and utilizes de novo pyrimidine biosynthesis exclusively to meet its metabolic requirements (7, 10 -12). The importance of the pyrimidine biosynthetic pathway to the survival of the malarial parasite is demonstrated by the finding that dihydrofolate reductase is a validated target for the treatment of the parasite (13). Furthermore, a number of inhibitors directed at thymidylate synthetase, such as 5-fluoroorotate (5-FO), 1 have nanomolar anti-malarial activity in vitro and cure malaria without toxicity in vivo (14 -17). As expected from the inability of the parasite to use nucleosides, the antiproliferative effects of thymidylate synthetase inhibitors on the malarial parasite cannot be reversed by thymidine, whereas mammalian cells cultured in the presence of thymidine were resistant to these inhibitors (16,17). These studies suggest that additional enzymes in the pyrimidine biosynthetic pathway should be studied for their potential as drug targets in the malarial parasite.
Dihydroorotate dehydrogenase (DHODH) catalyzes the fourth step in the pyrimidine biosynthetic pathway. In human cells, DHODH is localized in the mitochondria and is the ratelimiting enzyme in UMP formation (18). Although human cells can both salvage and synthesize pyrimidines de novo, activated T-and B-lymphocytes require de novo synthesis. A potent inhibitor of human DHODH (A77 1726) has been shown to be the active metabolite of a recently approved treatment for rheumatoid arthritis (leflunomide, Arava), and strong evidence has been reported that the mechanism of action of leflunomide is inhibition of de novo pyrimidine biosynthesis in these cells (19 -21). In addition to A77 1726, a number of high affinity inhibitors of human DHODH have been reported along with extensive structure-activity analysis (22)(23)(24)(25). Furthermore, high affinity inhibitors of Helicobacter pylori DHODH and the Escherichia coli enzyme have also been described that display considerable species selectivity (26,27). Thus, the known species differences in DHODH inhibitor binding, and the established clinical pharmacology of DHODH inhibitors in humans, suggest that selective DHODH inhibitors may also be developed for malaria chemotherapy.
DHODH catalyzes the oxidation of dihydroorotate to orotate utilizing the flavin cofactor FMN in the first of two half reactions (Scheme 1). In the second step, the enzyme catalyzes the re-oxidation of FMNH 2 using one of several cofactors. Two forms of DHODH have been described, cytoplasmic and membrane-bound (28). The cytoplasmic enzymes utilize fumarate or NAD ϩ to oxidize FMNH 2 , whereas the membrane-bound enzymes, which are mitochondrial in eukaryotes, require respiratory quinones as their physiological oxidant (18,29). Mammals, plants, and most Gram-negative bacteria have membrane-bound enzymes. There are, however, no unifying rules. Although some yeast, eubacteria, and protozoa utilize fumarate or NAD ϩ , others require respiratory quinones (28). Sequence analysis of the malarial DHODH gene (Fig. 1) demonstrates that it belongs to the mitochondrial-type enzymes (28). The enzyme has also been localized to the mitochondria by studies in parasites (30,31).
X-ray structures of truncated human DHODH have been solved (32) in complex with orotate and FMN, plus the antiinflammatory compounds brequinar and A77 1726 (the active metabolite of leflunomide). The enzyme is a ␤/␣ barrel, and the orotate and FMN stack against each other in the center of the barrel. Brequinar and A77 1726 are thought to bind to the ubiquinone binding site. Consistent with this model, brequinar is a competitive inhibitor of ubiquinone, however, the kinetics of A77 1726 are more complex (33). In the x-ray structure, the inhibitors bind in a channel between two ␣-helices at the N terminus of the molecule that extend away from the ␤/␣ domain (Fig. 2). This slot forms the entrance to a tunnel that ends at FMN, and presumably extends from the membrane allowing ubiquinone to diffuse into the site during the catalytic cycle. This binding pocket is present only in enzymes of the mitochondrial type.
Biochemical characterization of malarial DHODH has been limited to the study of partially purified enzyme preparations from cultured parasite cells, in which detailed characterization of the enzyme was limited by both purity and quantity of the preparation (30,31). In this study the truncated soluble form of DHODH from P. falciparum has been overexpressed in E. coli and purified to near homogeneity. The biochemical properties of the recombinant DHODH were examined to determine pH dependence, substrate specificity, and inhibition profiles for a series of compounds that are effective against the human enzyme.

Methods
Cloning of the P. falciparum pyrD Gene-The polymerase chain reaction was used to amplify a truncated segment of the pyrD gene from P. falciparum strain C2B genomic DNA encoding the enzyme dihydroorotate dehydrogenase. Primers 1 (CCTGAATTTTTTTCCATGGATAT-ATTTTTAAAATTC) and 2 (CACTTATGTGTCGACCGTGTTTAATTA-ACTTTTGC), which introduce NcoI and SalI restriction sites, respectively (shown in boldface), were used to generate a 1244-bp DNA fragment. The PCR product was ligated into the pProEX HTa prokaryotic expression vector (Invitrogen, Carlsbad, CA) that produces protein fused to an N terminus His 6 sequence. The cloned gene was sequenced to verify that no unintentional mutations were introduced.
The 42-amino acid insert unique to P. falciparum was verified by PCR amplification of a pyrD gene fragment encoding this region using the cDNA library from strain Dd2 (MR4/ATCC, Manassas, VA). Primers 3 (GGAAGATACGCTGATTATATAGC) and 2 were used to amplify a fragment of ϳ750 bp that was directly ligated into ZeroBlunt TOPO vector (Invitrogen, Carlsbad, CA). The cloned pyrD fragment was sequenced with the M13/reverse primer following amplification and purification of plasmid DNA. SCHEME 2. Structures of DHODH inhibitors. For A77 1726 and analogs: R ϭ CF 3 and the cyclopropyl group is replaced with methyl for A77 1726; R ϭ NO 2 for CCHNP; R ϭ CF 3 for CCHTFP; R ϭ CN for CCHCP. SCHEME 1. Reactions catalyzed by DHODH. Substrate L-DHO is oxidized by the FMN cofactor. FMN is re-oxidized by coenzyme Q n , where the length of the isoprenoid unit (n) is variable.
Protein Expression and Purification of pfDHODH-Chemically competent E. coli DH5␣ cells were cotransformed with the His 6 -DHODH expression vector and the RIG plasmid that encodes the rare tRNAs encoding Arg, Ile, and Gly residues (34). The overexpressed enzyme of ϳ45-kDa molecular mass lacks 168 amino acids from the N terminus thereby removing the hydrophobic membrane-associated domain ( Fig.  1). Cells were grown in LB medium to an A 600 nm of 0.5, supplemented with 0.1 mM flavin mononucleotide and induced with 0.6 mM isopropyl ␤-D-thiogalactopyranoside (Fisher Scientific, Fair Lawn, NJ) and harvested by centrifugation 3 h after induction, followed by freezing in liquid nitrogen. Typically, 5 g of cell paste was obtained per liter of liquid culture.
Fractions were analyzed by SDS-PAGE (35) for purity, and the appropriate fractions were pooled and concentrated as described previously. Protein concentration was determined using the Bradford assay with bovine serum albumin as a standard (36). The yield of purified protein is generally 1.5-2 mg per gram of cell paste for the 45-kDa DHODH. All measurements were performed on enzyme preparations containing the His 6 tag, except where noted. In such cases, the tagged enzyme was incubated with TEV-protease immobilized on glutathioneagarose overnight at 4°C as described (37). DHODH was eluted from the reduced glutathione beads, and the fraction of enzyme with the His 6 tag remaining was removed by additional chromatography on Ni 2ϩagarose resin.
Cloning and Expression of Human DHODH-The gene encoding an N-terminal truncation of human DHODH (Met 30 -Arg 396 ) was amplified from a cDNA library derived from human pituitary gland (Clontech, Palo Alto, CA) by PCR using primers hD1 (GCCTCCTACCATATGGC-CACGGGAG) and hD2 (ACGCTGGAATTCCTCCGATGATCTGCTCC), which introduce NdeI and EcoRI restriction sites, respectively. The 1125-bp PCR product was ligated into pCR-Blunt II-TOPO vector (Invitrogen, Carlsbad, CA). Recombinant plasmids were digested with NdeI and EcoRI restriction enzymes and separated by electrophoresis. The PCR product was extracted using a gel purification kit (Qiagen, Valencia, CA) and ligated into a similarly restricted and purified pET22b (Novagen, Madison, WI) protein overexpression vector to generate a construct with a His 6 C-terminal tag. This construct is similar to previously described expression constructs for truncated human DHODH (33,40,41). Transformed BL21(DE3) cells were grown at 37°C in rich LB medium (35 g of Tryptone, 20 g of yeast extract, 5 g of NaCl, and 10% glycerol per liter) to an A 600 nm of 0.6, and 100 M FMN was added followed by 200 M isopropyl-1-thio-␤-D-galactopyranoside. Cells were harvested after 4 h of growth at 30°C. All other details of the purification are identical to the procedure for isolation of pfDHODH outlined above. Approximately 2 mg of purified hDHODH was obtained per liter of medium.
Enzymatic Assays for DHODH Activity-Steady-state measurements to determine the pH dependence of enzyme activity were performed by varying the L-DHO concentration (5-500 M) and measuring the reduction of 2,6-dichloroindophenol (DCIP) at 600 nm (⑀ ϭ 18.8 mM Ϫ1 cm Ϫ1 ) as described (38). The pH range of the reaction was varied by employing the following buffers (100 mM): MES, pH 6 -7; HEPES, pH 7-8; and Tris, pH 8 -8.5. DCIP was added to a final concentration of 60 M in the buffer indicated above at saturating Q D (100 M). Inhibition studies on 5-FO and IC 50 data (Table II) were also performed with this assay. Reactions were initiated by addition of enzyme to a final concentration of 5-50 nM, and the temperature was maintained at 25°C with a circulating water bath.
For analysis of the CoQ substrate specificity and inhibition patterns by DCL, Redoxal, and CCHNP, formation of L-orotate was measured by the direct assay (39) with a Beckman DU-650 spectrophotometer. The form of the enzyme used in these studies lacked the N-terminal histidine tag. For inhibitor studies decylubiquinone (⑀ ϭ 4.3 mM Ϫ1 cm Ϫ1 ) was used as the cosubstrate, and assays were performed at 296 nm. For quinone specificity determinations, the CoQ concentration was varied (1-150 M) and the production of orotic acid was measured at 287 nm (⑀ ϭ 6.04 mM Ϫ1 cm Ϫ1 ) for CoQ 0 , 293 nm for CoQ 4/6 (⑀ ϭ 4.7 mM Ϫ1 cm Ϫ1 ), and 282 nm (⑀ ϭ 6.6 mM Ϫ1 cm Ϫ1 ) for vitamin K 3 . Assays were performed in buffer containing 100 mM HEPES, pH 8.0, 150 mM NaCl, 0.1% Triton X-100, and 10% glycerol. L-DHO and CoQ were added as 100ϫ stocks to the final concentrations. For inhibition studies, DCL (NCI no. 126771) and Redoxal (NCI no. 73735) were dissolved in Me 2 SO as 100ϫ stocks to determine IC 50 values. CCHNP, Redoxal, and DCL inhibition patterns were determined by separately varying L-DHO (5-500 M) and CoQ D (20 -200 M) concentrations. The synthesis and testing of the A77 1726 derivatives as inhibitors of mammalian DHODH has been previously described (24). Reactions were initiated by addition of enzyme to a final concentration of 5-50 nM, and the temperature was maintained at 25°C with a circulating water bath. SigmaPlot Synthesis of A77 1726 Derivatives-The derivatives of A77 1726 were synthesized in two steps as described (24). In step one the respective para-substituted aniline (Scheme 2, R ϭ NO 2 , CF 3 , or CN) was condensed with cyanoacetic acid to form the corresponding cyanoacetamide. In step 2, the product of this reaction was acylated with cyclopropanecarbonyl chloride to form the corresponding ␤-hydroxy enamide. The three resulting compounds (CCHNP, R ϭ NO 2 ; CCHTFP, R ϭ CF 3 ; CCHCP, R ϭ CN, numbered 49, 17, and 38, respectively, in the original publication (24)) were purified by high-performance liquid chromatography as previously described (24). The identity of each compound was verified by liquid chromatography-MS, MS-MS, IR, and NMR analysis.

RESULTS AND DISCUSSION
Cloning and Expression of Malarial DHODH in E. coli-Malarial DHODH has a significantly longer N-terminal sequence (129 amino acids) than the human enzyme (Fig. 1). The N-terminal sequence contains the membrane-binding domain, which localizes the enzyme to the mitochondria. Soluble overexpression of human DHODH was achieved by truncating 30 amino acids from the N-terminal membrane anchor (40). The truncation was demonstrated not to affect inhibitor binding affinity, nor enzyme activity (41). The x-ray structure of human DHODH (32) also provides evidence that this truncated protein contains the functional catalytic domain (Fig. 2). To obtain soluble malarial DHODH for study, an expression plasmid encoding a truncated enzyme of similar length to the soluble human enzyme was constructed. Genomic DNA from P. falciparum strain C2B served as the template for PCR amplification of a fragment of the pfDHODH gene (42), which encodes an enzyme lacking 168 amino acids from the amino terminus. The truncated gene product was ligated into the prokaryotic expression vector pProEXHTa (Invitrogen), which contains an His 6 -TEV protease cleavage site upstream of the cloning site.
Initial attempts to express malarial DHODH in DH5␣ cells using the pProEXHTa vector resulted in low protein yields. Expression of P. falciparum proteins in E. coli is often hindered by the high AT content of the genome (82%), which results in the utilization of codons that are rarely used by bacteria (43). To overcome this problem the RIG plasmid was cotransformed with the vector containing the pfDHODH gene to provide rare tRNAs encoding Arg, Ile, and Gly (34). In the presence of the RIG plasmid P. falciparum DHODH was expressed in E. coli DH5␣ cells in good yield. The recombinant pfDHODH was purified using a combination of Ni ϩ2 -agarose and gel filtration chromatography. Typical yields were 1.5-2 mg of purified protein (95% as estimated by SDS-PAGE) per gram of cell paste (Fig. 3A). The UV-visible spectrum of the purified enzyme (Fig.  3B) contains absorption bands at 380 and 454 nm that are characteristic of a flavin cofactor. FMN is tightly bound to the enzyme and copurifies with the protein during all steps. The purified recombinant pfDHODH has ϳ0.9 equivalents of FMN present, based on extinction coefficients of FMN that range from 11 to 12 mM Ϫ1 cm Ϫ1 at 454 nm.
pH Dependence of the Steady-state Reaction Catalyzed by pfDHODH-The pH dependence on activity was examined using the coupled DCIP assay. The pH of the reaction buffer was varied from 6.0 to 8.5, and the rates were measured as a function of increasing L-dihydroorotate (at constant decylubiquinone). The data were fitted to a single proton transition (Equation 4) for both the acidic and basic limbs (Fig. 4). Based on this model, maximal catalysis (k max ) of 8 s Ϫ1 was achieved in the pH range from 7.5 to 8.0. This activity is 90-fold higher than that reported for the partially purified native enzyme from the malarial parasite (30). Activity diminishes above or below the optimal pH range, however, the value of the apparent Michaelis constant, K m app (40 -80 M) for DHO remained largely unaffected. The pH dependence was repeated with the direct assay, and identical results were obtained (not shown). Based on these results, HEPES-containing reaction buffer at pH 8.0 was chosen for the standard assay conditions. Subsequent characterization employing either assay gave similar kinetic parameters with the value of k cat consistently between 2 and 6 s Ϫ1 and a K m app for DHO ranging from 30 to 60 M depending on the preparation. The form of the enzyme in which the His 6 tag was removed had a k cat that was ϳ1.5-fold lower, however, the K m app for L-DHO at saturating CoQ D was unchanged.
Substrate Specificity of pfDHODH for the Ubiquinone Cofactor-The length of the hydrophobic tail of CoQ is a determinant of substrate specificity in the mammalian and bacterial DHODHs. The H. pylori enzyme will donate electrons not only to Q 6 , but also to Q o , whereas the mammalian enzymes are significantly less active with Q o (26,39). Furthermore, the H. pylori enzyme utilizes menaquinone and menadione (vitamins K 2 and K 3 ) for the re-oxidation of FMNH 2 , but these coenzymes are poor substrates of the human enzyme. To determine the substrate specificity of pfDHODH for the ubiquinone cofactor, several quinone cosubstrates were examined to determine the effect of the isoprenoid unit length on enzymatic activity using the direct assay.
The concentration of CoQ was varied (1 to 150 M) at saturating concentration of L-DHO (500 M). CoQ 4 , CoQ 6 , and CoQ D (a soluble synthetic analog of Q 10 ) functioned with similar efficiency as indicated by the calculated specificity constants (k cat /K m app ) of 0.13 M Ϫ1 s Ϫ1 (CoQ 6 and CoQ D ) and 0.10 M Ϫ1 s Ϫ1 (CoQ 4 ) summarized in Table I. The substrate L-DHO was varied at saturating CoQ concentration (100 M), and the determined catalytic rates were 1.8 s Ϫ1 (CoQ 4 and CoQ D ) and 2.0 s Ϫ1 (CoQ 6 ). Cosubstrates that lack the isoprene moiety (CoQ o and vitamin K) had similar catalytic rates to substrates with the hydrophobic chain (k cat ϳ 2 s Ϫ1 ), however K m app was increased 4to 7-fold reducing the catalytic efficiency with these substrates (Table I).
Comparative Inhibitor Analysis for P. falciparum and Human DHODH-To determine if selective inhibitor binding to P. falciparum DHODH would be feasible, the inhibition profiles for several of the reported inhibitors of the human enzyme were measured against recombinant pfDHODH. Brequinar, A77 1726, and a number of its derivatives are nanomolar inhibitors of the human enzyme (24,25). Two structural analogs of brequinar, including Redoxal and DCL, are all also reported to be high affinity inhibitors of the human enzyme (23,44). These inhibitors are thought to exploit the unique binding site that is found in the membrane-bound type DHODH enzymes, which is the presumed site of CoQ binding. Inhibition studies on pfDHODH were conducted with 5-FO, an analog expected to compete directly with the substrate dihydroorotate, with the brequinar analogs available from the NCI/NIH compound collection, Redoxal and DCL, and with three A77 1726 derivatives (CCHNP, CCHTFP, and CCHCP) synthesized for this study.
The substrate analog 5-FO was included in the standard DCIP reduction assay, and the inhibitory effect on enzymatic   FIG. 4. Effect of pH. The catalytic rates were measured in buffer that varied from pH 6.0 to 8.5. Data were fitted to Equation 4 yielding the following parameters: k max ϭ 8.0 Ϯ 1.2 s Ϫ1 and ionization constants, pK a ϭ 6.5 Ϯ 0.2 and pK b ϭ 8.4 Ϯ 0.2. Errors represent the standard error of the fit.  activity was measured as a function of increasing L-DHO concentration. A double-reciprocal plot of the data showed a pattern typical of a competitive inhibitor (Fig. 5). The kinetic parameters were calculated from a global fit of the entire data set to Equation 1, describing competitive inhibition. The values of k cat (2.9 Ϯ 0.05 s Ϫ1 ) and K m app (22 Ϯ 2 M) obtained from the fit were consistent with previous measurements, and the inhibition constant (K I ) of 85 Ϯ 10 M is of similar magnitude to the K m app for L-DHO. CCHNP, CCHTFP, and CCHCP were synthesized as described under "Experimental Procedures." The IC 50 values for CCHNP, CCHTFP, and CCHCP were measured as 0.5, 1.1, and 0.7 mM, respectively, against the recombinant malarial DHODH using the DCIP-based assay (Table II). Recombinant truncated human DHODH was prepared as described under "Experimental Procedures" for comparative purposes. Steadystate constants were determined for the purified recombinant human enzyme (k cat ϭ 11 Ϯ 0.4 s Ϫ1 and K m app (L-DHO) ϭ 32 Ϯ 4 M) using the DCIP assay; these values are similar to previously reported values for this enzyme (40,41). The IC 50 values for inhibition of the malarial DHODH by CCHNP, CCHTFP, and CCHCP are 2200-, 5700-, and 2000-fold higher then the values measured for the recombinant truncated human enzyme under the same assay conditions (Table II). Previously, IC 50 values for these compounds had been reported for the rodent and human enzymes (24,41). The values for the human enzyme are in good agreement with these previous reports. The IC 50 values for human DHODH are typically 4to 10-fold higher then those values measured on the rodent enzymes (41).
The inhibition characteristics of CCHNP against pfDHODH were further characterized as a function of L-DHO and CoQ D concentration by following the formation of orotate using the direct assay. Double-reciprocal plots of the data sets showed that CCHNP is a competitive inhibitor of Q D (Fig. 6). Global fitting of each data set was performed using the appropriate equation ("Experimental Procedures"). Global analysis of the inhibition pattern versus Q D shows that the data are best fit by the competitive inhibition model and the calculated K I app value for CCHNP is 400 Ϯ 90 M (Table III). In contrast, global analysis of the inhibition pattern versus L-DHO shows that the data are equally well fitted by either a competitive or non-competitive binding model, and thus the mechanism of inhibition with regard to DHO cannot be conclusively determined by the kinetic analysis. Analysis of the x-ray structure (Fig. 2) demonstrates that the binding sites for DHO and A77  1726 are separated by the FMN binding site, thus CCHNP can not be a competitive inhibitor of both DHO and Q D unless it has two separate binding modes. Given all available data, CCHNP is likely to be a non-competitive inhibitor of L-DHO with an intercept on the double-reciprocal plot that is close to the y-axis. The IC 50 values for Redoxal and DCL against the recombinant malarial DHODH were measured to be 71 and 220 M, respectively, using the DCIP-based assay (Table II). These values are 5500-and 3400-fold higher, respectively, than the IC 50 data determined with the human enzyme. Redoxal and DCL inhibition of pfDHODH were further characterized by analysis of the inhibition as a function of both L-DHO and CoQ D concentration using the direct assay. Double-reciprocal plots of both data sets showed DCL is an uncompetitive inhibitor with respect to both L-DHO and CoQ D (Fig. 7), whereas Redoxal is a non-competitive inhibitor with respect to L-DHO and an uncompetitive inhibitor relative to CoQ D (data not shown). Global fitting of each data set was performed using the appropriate equation ("Experimental Procedures"). For DCL the calculated K I app values were 160 Ϯ 40 M and 250 Ϯ 50 M for the uncompetitive inhibition of L-DHO and CoQ D , respectively. For Redoxal inhibition constants of (K I ) of 59 Ϯ 11 M and 86 Ϯ 6 M for the non-competitive and uncompetitive series, were measured, respectively (Table III). These values are similar to the determined IC 50 value. Comparing the results for DHO, these binding constants for Redoxal and DCL are 120-and 4000-fold weaker, respectively, than the K I values reported for the human enzyme (41,44,45).
Potent inhibition of mammalian DHODH has been achieved with two classes of chemical scaffolds (e.g. A77 1726 and brequinar) that have received significant pharmacological interest for the treatment of human autoimmune disease (19 -21). The data we report here demonstrate that analogs of both A77 1726 and of brequinar are poor inhibitors of malarial DHODH. Thus, significant species-selective inhibitor binding can be achieved between the parasite and host DHODH, supporting the hypothesis that selective inhibition of malarial DHODH is feasible.
X-ray structure analysis of brequinar and A77 1726, bound to human DHODH, found a common binding mode for both inhibitors (Fig. 2); this binding site is thought to also bind the cofactor ubiquinone, although direct structural evidence for the ubiquinone binding site has not been obtained (32). For human DHODH, both brequinar and DCL are competitive inhibitors of CoQ analogs consistent with the hypothesis that they bind at the ubiquinone binding site (33,45). However, both A77 1726 and Redoxal, display non-competitive inhibition patterns with all CoQ analogs tested (33,44). With the malarial enzyme, CCHNP, which is a structural analog to A77 1726, displays competitive inhibition patterns with Q D . These data demonstrate that an A77 1726 analog binds to the same site as CoQ, and they suggest that these analogs bind to the malarial enzyme at the inhibitor binding site that has been defined in the x-ray structure analysis of human DHODH (Fig. 2). For malarial DHODH, Redoxal and DCL are uncompetitive inhibitors of CoQ D , suggesting that the binding of inhibitor is dependent on the binding of substrate for the interaction. This analysis is consistent with a binding site in which CoQ and inhibitor bind to separate but adjacent sites.
The brequinar and A77 1726 binding site that is observed in the x-ray structure of human DHODH is composed of amino acids that are highly variable between the human and malarial enzymes ( Figs. 1 and 2). Comparison of the structure with the sequence alignment shows that nine of the fifteen residues that are within van der Waals contact (defined as within 4.2 Å) of A77 1726 in the structure are variable between the malarial and human enzymes. These structural differences provide an explanation for the observed differences in the potency of binding of these inhibitors to the two enzymes. The striking differences in inhibitor potency between the malarial and human enzymes strongly suggest that species-selective inhibitors of malarial DHODH can be developed to exploit these structural differences. Finally, recent RNA interference studies, in which malarial parasites were treated with double-stranded RNAencoding DHODH, correlated a reduction in mRNA levels to reduced parasite growth (46). These studies provide further evidence that DHODH is essential to parasite growth. Thus, the potential to develop selective inhibitors for an essential metabolic enzyme in the malarial parasite suggests that malarial DHODH is an attractive target for the development of new anti-malarial agents.