The Plasmodium falciparum Bifunctional Ornithine Decarboxylase, S -Adenosyl- L -methionine Decarboxylase, Enables a Well Balanced Polyamine Synthesis without Domain-Domain Interaction*

In the human malaria parasite Plasmodium falciparum ( Pf ), polyamines are synthesized by a bifunctional enzyme that possesses both ornithine decarboxylase (ODC) and S -adenosyl- L -methionine decarboxylase (AdoMetDC) activities. The mature enzyme consists of the heterotetrameric N-terminal AdoMetDC and the C-terminal dimeric ODC, which results in the formation of a heterotetrameric complex. For the native bifunctional protein a half-life longer than 2 h was determined, which is in contrast to the extreme short half-life of its mammalian monofunctional counterparts. The biological advantage of the plasmodial bifunctional ODC/ AdoMetDC might be that the control of polyamine synthesis is achieved by only having to regulate the abundance and activity of one protein. An interesting feature in the regulation of the bifunctional protein is that putrescine inhibits Pf ODC activity (cid:1) 10-fold more efficiently than the mammalian ODC activity, and in contrast to the mammalian AdoMetDC the activity of the Pf AdoMetDC domain is not stimulated by the diamine. To analyze post-translational processing, by saponin lysis (16). For the determination of AdoMet, samples of isolated P. falciparum were deproteinized by adding 0.2 M perchloric acid for 12 h at 4 °C. After centrifugation (10,000 (cid:3) g, 4 °C, 20 min), aliquots of the supernatant were analyzed by reversed phase HPLC on a Spherisorb ODS II column (5 (cid:1) m, 250 (cid:3) 3 mm, Machery-Nagel) according to Guattari (17) with minor modifications. Separation was performed isocratically at a flow rate of 0.55 ml/min using 40 m M ammonium dihydrogenphosphate, 6 m M heptanesulfonic acid sodium salt monohydrate, and 6% (v/v) methanol, pH 4.2, as mobile phase. The effluent was monitored by UV detection at 254 nm (diode array detector 440, Kontron). The ornithine content of P. falciparum was determined by the method of Patchett et al. (18) with minor modifications. Isolated P. falciparum cells were deproteinized by heating at 95 °C for 5 min before being centrifuged at 10,000 (cid:3) g for 20 min at 4 °C. Aliquots of the supernatant were derivatized with equal volumes of o -phthalaldehyde reagent (40 m M o -phthalaldehyde, 0.2 M potassium borate buffer, pH 9.4, 0.14 M 2-mercaptoethanol, and 10% methanol) and subjected to a Spherisorb ODS II (5 (cid:1) m, 125 (cid:3) 3 mm, Machery-Nagel) reversed phase HPLC for analysis. The mobile phase consisted of solvent A (10 m M potassium dihydrogenphosphate, pH 5.9) and solvent B (acetonitrile/ methanol/water, 4:3:3, v/v/v). Separation was performed at a flow rate of 0.9 ml min (cid:4) 1 applying the following gradient (percentage of solvent

In the human malaria parasite Plasmodium falciparum (Pf), polyamines are synthesized by a bifunctional enzyme that possesses both ornithine decarboxylase (ODC) and S-adenosyl-L-methionine decarboxylase (AdoMetDC) activities. The mature enzyme consists of the heterotetrameric N-terminal AdoMetDC and the Cterminal dimeric ODC, which results in the formation of a heterotetrameric complex. For the native bifunctional protein a half-life longer than 2 h was determined, which is in contrast to the extreme short half-life of its mammalian monofunctional counterparts. The biological advantage of the plasmodial bifunctional ODC/ AdoMetDC might be that the control of polyamine synthesis is achieved by only having to regulate the abundance and activity of one protein. An interesting feature in the regulation of the bifunctional protein is that putrescine inhibits PfODC activity ϳ10-fold more efficiently than the mammalian ODC activity, and in contrast to the mammalian AdoMetDC the activity of the PfAdoMetDC domain is not stimulated by the diamine. To analyze post-translational processing, polymerization, and domain-domain interactions, several mutant proteins were generated that have single mutations in either the PfODC or PfAdoMetDC domains. The exchange of amino acids essential for the activity of one domain had no effect on the enzyme activity of the other domain. Even prevention of the post-translational cleavage of the AdoMetDC domain or ODC dimerization and thus the interference with the folding of the protein hardly affected the activity of the partner domain. In addition, inhibition of the activity of the PfODC domain had no effect on the activity of the PfAdoMetDC domain and vice versa. These results demonstrate that no domain-domain interactions occur between the two enzymes of the bifunctional PfODC/AdoMetDC and that both enzymatic activities are operating as independent catalytic sites that do not affect each other.
The polyamines putrescine, spermidine, and spermine are ubiquitous and essential cellular components involved in various metabolic processes including the proliferation and differentia-tion of bacteria, plants, and animals (1,2). The two rate-limiting enzymes of polyamine synthesis are ornithine decarboxylase (ODC, 1 EC 4.1.1.17) and S-adenosyl-L-methionine decarboxylase (AdoMetDC, EC 4.1.1.50), which provide putrescine and decarboxylated S-adenosyl-L-methionine (AdoMet) for the synthesis of spermidine. Interference with polyamine biosynthesis is considered as antitumor and antiparasitic strategy, although the success of polyamine-related cancer therapeutic approaches has been modest (3)(4)(5). However, 2-difluoromethylornithine (DFMO), an inhibitor of ODC originally designed as an anticancer agent, has proved to be clinically successful in the treatment of African sleeping sickness caused by the protozoan Trypanosoma gambiense (6). The selective toxicity of DFMO on T. gambiense is complex and discussed to depend on various metabolic differences between parasite and mammalian host including the presence of trypanothione, altered levels of AdoMet, and also the longer half-life of the parasite ODC when compared with the mammalian enzyme (7). This success has reinforced the view that the polyamine metabolism of parasitic protozoa might be suitable for chemotherapeutic exploitation. We suggest that because of a higher polyamine requirement of rapidly proliferating cells, the interference with their synthesis has more severe consequences for the parasite compared with the host organism.
Malaria is the most important tropical disease in the developing countries; rough estimates indicate 500 million clinical cases per annum causing more than 1 million deaths. Malaria is undergoing resurgence, and the control of Plasmodium falciparum malaria has become a severe problem. Because of rapidly spreading drug resistance there is an urgent and pressing need for new antimalarial drugs.
ODC and AdoMetDC are usually derived from separate genes and act individually as highly regulated monofunctional enzymes. In P. falciparum, however, ODC and AdoMetDC activities are located on a single polypeptide with the AdoMetDC domain in the N-terminal part connected to the C-terminal ODC domain by a hinge region (8). Corresponding to the respective dimeric and heterotetrameric structures of the mammalian ODC and AdoMetDC (9,10), in Plasmodium a heterotetrameric enzyme complex is formed consisting of two bifunctional polypeptides post-translationally cleaved in the AdoMetDC domain. Here we report on the expression and mutational analyses of the recombinant bifunctional P. falciparum (Pf)ODC/AdoMetDC and address possible functional consequences and the potential significance of the bifunctional nature of this parasite protein, which may help to validate this protein as a target for a chemotherapeutic intervention of malaria.
To determine the half-life of the PfODC/AdoMetDC, a synchronized Plasmodium culture at the trophozoite stage was incubated with cyclohexamide (50 g/ml, Sigma) to prevent protein de novo synthesis. The decay of ODC and AdoMetDC activities was measured after 30, 60, and 120 min in infected erythrocytes enriched with gelafundin (B. Braun Melsungen) (12,13). Synchronization of the culture was performed by incubation of cells in 2 volumes of 0.3 M alanine, 10 mM HEPES, pH 7.4, for 5 min at 37°C (14). The percentage of infected erythrocytes was determined by light microscopy of Giemsa-stained thin smears, and the number of the erythrocytes was determined in a Coulter Max M cell counter.
Recombinant Expression of the PfODC/AdoMetDC-The entire coding region of the PfODC/AdoMetDC wild type and mutants cloned into the vector pASK-IBA3 (Institut fü r Bioanalytik, Göttingen, Germany) were expressed in the AdoMetDC-and ODC-deficient Escherichia coli line EWH331, kindly provided by Dr. H. Tabor (19), and subsequently purified as described previously (8). The purified recombinant protein was subjected to fast protein liquid chromatography on a calibrated Superdex S-200 column (2.6 ϫ 60 cm) equilibrated with 40 mM Tris/ HCl, pH 7.5, containing 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, and 0.02% Brij-35 at a flow rate of 2 ml/min. The homogeneity of the enzyme preparation was analyzed by SDS-polyacrylamide gel electrophoresis. Proteins were revealed by Coomassie staining (20). The concentration of the purified recombinant protein was determined according to Bradford (21).
Cloning, Expression, and Purification of the rPfAdoMetDC Domain-The AdoMetDC domain comprises the amino acids 1-660 including 131 amino acids of the hinge region of the bifunctional protein ( Fig. 1). Base pair 1-1980 of the coding region of the ODC/AdoMetDC was amplified by polymerase chain reaction (PCR) from genomic P. falciparum DNA using the sense oligonucleotide 5Ј-GCGCGCGGTCTCCGCGCAT-GAACGGAATTTTTGAAGG-3Ј and the antisense oligonucleotide 5Ј-GCGCGCGGTCTCCTATCATTAATCTTTCTCATTTGTTTGTACC-3Ј.
The PCR product was cloned into pCR R 2.1-TOPO using a TOPO TA cloning kit (Invitrogen) and transformed into E. coli DH5␣ cells. Recombinant plasmid DNA was isolated and digested with BsaI (5 units/ l, New England Biolabs). The insert was gel-purified and cloned into BsaI-cut pASK-IBA7 expression vector (Institut fü r Bioanalytik). The cloned PfAdoMetDC was expressed in E. coli BL21-CodonPlus TM (DE3)-RIL (Stratagene) and purified as described previously (8).
Enzyme Assays-Both AdoMetDC and ODC activities were assayed by trapping released 14 CO 2 from S-adenosyl-L-[methyl-14 C]methionine and L-[1-14 C]ornithine (57 and 52 mCi/mmol, respectively, Amersham Pharmacia Biotech) as described previously (22,23). The K m values of the bifunctional rPfODC/AdoMetDC and rPfAdoMetDC domains for the respective substrates were determined by varying the concentrations of ornithine and AdoMet from 10 to 200 M. The inhibition constant (K i ) for putrescine to inhibit ODC activity was determined under standard assay conditions with varying concentrations of ornithine and the addition of 50, 100, and 200 M putrescine. K m and K i values were calculated by Lineweaver-Burk plots using the software GraphPad Prism 1.02 (GraphPad Software).
An additional assay was used for a functional analysis of the activity rates of both active sites of the bifunctional ODC/AdoMetDC. Both reactions were performed simultaneously, and the reaction mixture contained (in a final volume of 250 l) 40 mM KH 2 PO 4 or Tris/HCl buffer, pH 7.5, for determination of AdoMetDC and ODC activities, respectively, 1 mM dithiothreitol, 1 mM EDTA, 40 M pyridoxal 5-phosphate, 100 M ornithine, and 100 M AdoMet. The activity rates of ODC and AdoMetDC were determined by trapping 14 CO 2 derived from the respective added labeled substrates 50 nCi [ 14 C]ornithine or [ 14 C]AdoMet. The competitive inhibitors CGP40215A and CGP54169A are from a new series of AdoMetDC and ODC inhibitors synthesized and provided by Novartis Pharma (24,25). Both inhibitors, when used at a concentration of 100 M, totally inhibited the respective domain activity (Ref. 26). 2 Oligonucleotides and Site-directed Mutagenesis-14 oligonucleotides were designed to replace amino acid residues potentially involved in PfODC/AdoMetDC activity (Table I). The putative active site residues of the ODC domain Lys 868 , Lys 970 , Cys 1355 , Asp 1356 , and Asp 1359 were exchanged by alanine, and Gly 1382 was mutated into tyrosine. In the AdoMetDC domain, Ser 73 was replaced by alanine. In addition, the amino acid residue Gly 721 (equivalent to Gly 1382 ) of the PfHinge-ODC (26) was mutated into tyrosine. 35 ng of the double-stranded supercoiled expression plasmid pASK-IBA3PfODC/AdoMetDC and 100 ng of mutagenic sense and antisense primers were used in a 50-l PCR containing deoxyribonucleotides, reaction buffer, and Pfu DNA polymerase according to manufacturer recommendations (Stratagene). The cycling parameters were 95°C for 50 s, 55°C for 60 s, and 68°C for 12 min for 17 cycles. The linear amplification product was treated with endonuclease DpnI (10 units/l, New England Biolabs) for 1 h to eliminate the parental template. An 8-l aliquot from each PCR was used for the transformation of competent E. coli DH5␣ cells. All mutations were verified by nucleotide sequencing using the Sanger dideoxy chain termination reaction for double-stranded DNA (20). Approximately 95% of the analyzed colonies contained the desired mutation, and one clone of each construct was transformed for the expression in competent E. coli EWH331 cells. The expressed proteins were purified by Strep-tactin affinity chromatography according to manufacturer recommendations (Institut fü r Bioanalytik).
Dot Blot Analyses-The recombinant wild-type and Gly 1382 -Tyr mutant PfODC/AdoMetDC as well as wild type and the equivalent Gly 721 -Tyr mutant PfHinge-ODC were subjected to gel filtration, and dot blot analyses of the collected fractions were performed. The fractions were transferred to nitrocellulose using a BioDot apparatus (Bio-Rad). The membrane was blocked in 3% low fat milk powder in PBS (20) overnight at 4°C. Subsequently the membrane was incubated with polyclonal rabbit antiserum raised against the rPfODC/AdoMetDC protein (1: 5000) for 1 h. After three washes with 0.05% Tween 20 in PBS, the membrane was incubated for 1 h with 1:10,000 diluted alkaline phosphatase-coated anti-rabbit IgG (Dianova) in PBS containing 1% low fat milk powder. After washing the membrane three times in 0.05% Tween 20 in PBS, the proteins were visualized with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Sigma) in 0.1 M Tris/HCl, 0.1 M NaCl, and 20 mM MgCl 2 , pH 9.5.

Characterization of the rPf ODC/AdoMetDC and AdoMetDC
Domain-To characterize and determine the steady-state kinetic parameters of ODC and AdoMetDC activities, the bifunctional PfODC/AdoMetDC was recombinantly expressed as a Strep-tactin fusion protein in the E. coli strain EWH331 and purified according to Mü ller et al. (8). The yield was 250 g liter Ϫ1 of bacterial cell culture.
Likewise, the AdoMetDC domain of the bifunctional P. falciparum ODC/AdoMetDC consisting of the amino acids 1-529 and additional 131 amino acids of the hinge region ( Fig. 1) was separately expressed as a Strep-tactin fusion protein in the E. coli BL21-CodonPlus TM (DE3)-RIL and purified according to that described in Ref. 8. 1 liter of culture yielded ϳ350 g of recombinant PfAdoMetDC domain. The rPfAdoMetDC was found to be catalytically active, just as it has been reported for the separately expressed rPfODC (26). The P. falciparum ODC/ AdoMetDC exhibited ODC and AdoMetDC activity optimal between pH 7.0 and 8.5, peaking at pH 8.0 and 7.5, respectively, when tested with potassium phosphate and Tris/HCl buffer. Interestingly the AdoMetDC activity of the bifunctional protein was not inhibited by Tris/HCl as it has been reported for the mammalian AdoMetDC (27).
The K m values of the rPfODC/AdoMetDC for ornithine and AdoMet are 41 and 58 M, respectively. These data confirm the values determined for the native PfODC/AdoMetDC (8) and are comparable with those obtained when expressing the ODC and AdoMetDC domains separately (26). The kinetic characteristics are also similar to those of the monofunctional mammalian enzymes (Table II).
The specific activities of the plasmodial AdoMetDCs (bifunctional and domain) were determined to be 20 and 51 nmol min Ϫ1 mg Ϫ1 , and the k cat for both enzymes are similar with 3.3 and 4.0 min Ϫ1 , respectively. These data are comparable with the specific activity of the mammalian AdoMetDC, reported to be 24 nmol min Ϫ1 mg Ϫ1 (Ref. 28 , Table II).
However, the specific ODC activity of the bifunctional rPfAdoMetDC/ODC protein differs considerably from that of the separately expressed ODC domain and from values reported for the mammalian ODC. The specific ODC activity of the bifunctional protein was determined to be 38 nmol min Ϫ1 mg Ϫ1 , whereas those for both separately expressed ODC domains were found to be 2.9 and 4.3 nmol min Ϫ1 mg Ϫ1 , depend-ent on whether they were expressed with a part of the preceding hinge region (26). The specific activity of the bifunctional protein, although nearly 10-fold higher than that of the separately expressed ODC/hinge domain, is still much lower than the value of 12,800 nmol min Ϫ1 mg Ϫ1 reported for the mammalian ODC (Ref. 29, Table II).
Attributable to the bifunctional nature of the Plasmodium protein, the similar specific activities of 38 and 20 nmol min Ϫ1 mg Ϫ1 of PfODC and PfAdoMetDC, respectively, might lead to a balanced production of putrescine and decarboxylated AdoMet, depending on the availability of their respective substrates. To relate the enzyme activities of PfODC and PfAdoMetDC to their substrate concentrations, we determined the ornithine and AdoMet levels in P. falciparum. The analyses revealed a 6-fold higher concentration for ornithine compared with AdoMet. The values obtained were 9.5 Ϯ 0.9 and 1.5 Ϯ 0.2 nmol 10 10 cells Ϫ1 (n ϭ 3), respectively. It is difficult to calculate precisely the molarity of both in the Plasmodium cell, but based on the relative volume of parasite to host erythrocyte (volume of host compartment 2/3, volume of parasite compartment 1/3; Ref. 30), nonsaturated substrate levels of 28.5 and 4.5 M for ornithine and AdoMet, respectively, were estimated. Considering enzyme kinetics and substrate levels, we expect that the reaction leading to putrescine is favored, which should lead to higher putrescine levels in Plasmodium such that the availability of decarboxylated AdoMet might be the limiting factor for spermidine synthesis as it has been discussed for mammalian cells (31).
Of some interest is that the PfODC activity of P. falciparum is inhibited by putrescine. The K i value of the PfODC was determined to be 68 M, which is in the comparable range of the K m value for ornithine and allows feedback control of putres-  cine synthesis considering intracellular concentrations of 81 M for putrescine 2 and 29 M for ornithine in Plasmodium. In contrast, the mammalian ODC is hardly inhibited by putrescine; the K i value of 600 M (32) is about 10-fold higher, a value that makes it unlikely that putrescine at a physiological concentration exerts an effect on the mammalian ODC activity. Another important feature in the regulation of polyamine synthesis in Plasmodium is that PfAdoMetDC is not stimulated by putrescine in a concentration range between 10 M and 2 mM, which shows that P. falciparum lacks the regulatory mechanism proposed for mammalian cells to relate putrescine abundance to the synthesis of spermidine (33).
To investigate whether the active sites of the bifunctional ODC/AdoMetDC affect each other and/or depend on a catalytically active partner site for catalysis, we used an assay system in which both reactions were performed simultaneously. The mixture contained the substrates of both enzyme reactions as described under "Experimental Procedures." The activity rates of the individual enzymes were determined by the trapping of 14 CO 2 derived from radiolabeled ornithine and AdoMet, respectively. As shown in Table III, the specific activities of ODC and AdoMetDC under these assay conditions were 34.4 and 20.6 nmol min Ϫ1 mg Ϫ1 , respectively. When the substrate of the "partner site reaction" was omitted, the respective activity rates of ODC and AdoMetDC were not affected. Further, the addition of the specific ODC inhibitor CGP52623A at a concentration that leads to total loss of ODC activity (26) did not influence the AdoMetDC activity; similarly, the inhibition of the AdoMetDC site by CGP40215A did not result in alterations of the ODC activity. These data clearly suggest that both active sites of the bifunctional ODC/AdoMetDC act independently of each other. In contrast to our finding on the plasmodial bifunctional ODC/AdoMetDC, a domain-domain interaction was reported for the leishmanial bifunctional dihydrofolate reductase/thymidylate synthase, in which the DHFR activity is enhanced when the TS site is occupied (34).
Mutagenic Analyses of the rPfODC/AdoMetDC-The comparison of the deduced amino acid sequence from the bifunctional PfODC/AdoMetDC with those of the respective mammalian proteins revealed that the amino acids involved in catalytic activities, protein processing, and cofactor binding are conserved in the Plasmodium sequence (8). To confirm that these amino acids are essential for processing of the AdoMetDC domain and for ODC domain activity and polymerization, mutagenic analyses of the residues Ser 73 , Lys 868 , Lys 970 , Cys 1356 , Asp 1356 , and Asp 1359 were performed (their positions are shown in Fig. 1).
The PfAdoMetDC domain contains the putative cleavage site 70 Leu-Ser-Glu-Ser-Ser 74 . Processing of the proenzyme into a small ␤-subunit and a large ␣-subunit was expected to occur between Glu 72 and Ser 73 (8). The mutation of Ser 73 into alanine prevented the post-translational cleavage into the ␤and ␣-subunit. As shown by SDS-polyacrylamide gel electrophore-sis analysis, the size of the unprocessed mutant compared with the wild-type protein was increased by 9 kDa (Fig. 2). In addition, the mutant PfAdoMetDC was inactive. The cleavage of the proenzyme is essential to form the pyruvoyl prosthetic group derived from Ser 73 and is essential for the enzymatic activity of PfAdoMetDC, similar to the mammalian AdoMetDC (35). Interestingly the activity of the PfODC domain was not affected by this mutation (Table IV) despite the fact that part of the recombinant protein did not fold properly anymore.
In the ODC domain of the bifunctional protein the amino acids Lys 868 and Lys 970 , which are thought to be involved in binding of the cofactor pyridoxal 5-phosphate and to be necessary for the conformation of the active site (29, 36), were ex-

TABLE III
Functional analysis of the active sites of the bifunctional rPfODC/AdoMetDC The activity rates of ODC and AdoMetDC were determined in the presence of the enzymatically active partner sites, respectively. The complete assay mixture contained the substrates for both enzyme reactions. Omission of the respective substrates AdoMet and ornithine led to nonactive partner sites, and likewise, the inhibition by CGP40215A and CGP52623A at 100 M concentration. Results are the means of duplicate determinations of three independent recombinant expressions under assay conditions as described under "Experimental Procedures." Spec. act., specific activities.   (29,37) were exchanged by alanine. The ODC activity of the Cys 1355 , Asp 1356 , and Asp 1359 mutants were found to be reduced by 91, 92, and 97%, respectively, a result consistent with the former reports on the mammalian ODC. Again, AdoMetDC activities were not significantly affected by these mutations (Table IV). In conclusion, these data support the results presented in Table III, which show that there is no functional interaction between the two enzymatic activities and strongly suggest that no domain-domain interactions between the PfODC and PfAdoMetDC exist. Both domains are independently acting enzymatic sites of a bifunctional protein.
Gly 387 (equivalent to Gly 1382 in PfODC/AdoMetDC), although not located at the dimer interface, has been reported to be important for the dimerization of the mouse ODC monomer (9,38). To investigate the role of this residue in the organization of the heterotetrameric ODC/AdoMetDC, we changed Gly 1382 into tyrosine, expecting that this mutation would prevent dimerization of the ODC domain and possibly interfere with the formation of the heterotetrameric complex. The mutation resulted in the total inactivation of ODC activity, whereas the AdoMetDC activity was reduced by only 27% (Table IV). Interestingly the mutation of Gly 1382 did not prevent the formation of the 330-kDa heterotetrameric ODC/ AdoMetDC complex as shown in Fig. 3A. Recombinant PfODC/ AdoMetDC wild type and the Gly 1382 -Tyr mutant protein were analyzed by gel filtration and subsequent dot-blot analyses. Should the mutation of Gly 1382 obstruct the formation of this complex structure, one would expect the formation of heterodimers of about 160 kDa. However, as shown by dot-blot analysis, the mutant protein eluted from the gel-sizing column in the range of 330 kDa, clearly consistent with the size of a heterotetrameric complex. This result does not exclude the participation of Gly 1382 in the dimerization of the ODC domains but indicates that additional residues take part in the polymerization process, possibly located in the AdoMetDC domain or the hinge region. To show that Gly 1382 is essential for the dimerization of the ODC domains, we mutated this residue in the separately expressed rPfHinge-ODC domain (26) and analyzed the organization of the resulting mutant protein. Gel filtration of the rPfHinge-ODC domain wild type and the mutant Gly 721 -Tyr (equivalent to Gly 1382 in PfODC/AdoMetDC) revealed proteins in the range of 160 and 80 kDa, which dem-onstrates dimeric and monomeric organization of the expressed ODC mutant proteins, respectively (Fig. 3B). This result confirms that this glycine residue is involved in the dimerization process of the plasmodial ODC.
It has been discussed that the organization of bifunctional compared with monofunctional proteins might have a biological advantage. For the leishmanial DHFR/TS, which catalyzes sequential reactions, substrate channeling and domain-domain interactions (34,39) have been shown. In the case of the P. falciparum ODC/AdoMetDC, substrate channeling can be excluded because the spermidine synthase is not part of the bifunctional enzyme complex (8). Another possible advantage of a bifunctional enzyme could be domain-domain interactions, which would facilitate the synthesis of the respective products or possibly exhibit regulatory functions on the activity of the partner domain. Shallom et al. (39) have shown for the plasmodial bifunctional DHFR/TS that physical interaction between the DHFR and TS domains is necessary to obtain a catalytically active TS. Data obtained from the mutagenesis experiments suggest, however, that PfODC and PfAdoMetDC domains do not interact directly and operate independently of each other. Inactivation of the PfODC and PfAdoMetDC domains by specific enzyme inhibitors and exchange of essential residues did not deplete or enhance the activity of the other  The turnover of the bifunctional protein was determined by measuring the decay of ODC and AdoMetDC after the addition of 50 g/ml cycloheximide to the culture to inhibit protein synthesis. Aliquots were drawn at the time points indicated, and the ODC and AdoMetDC activities were determined in the lysate of the isolated parasitized erythrocytes as described under "Experimental Procedures." OE, cyclohexamide-treated culture; f, control culture. domain in the mutant compared with the wild-type protein.
Our results with the separately expressed PfODC and PfAdoMetDC domains support the data obtained with the bifunctional mutant proteins. Steady-state kinetic analyses of both separately expressed domains are in the same range as those determined for the entire bifunctional protein (26). A biological advantage of the plasmodial bifunctional ODC/ AdoMetDC might be that the control of polyamine synthesis is achieved by only having to regulate the abundance and activity of one protein.
Turnover of P. falciparum ODC/AdoMetDC-An important aspect when thinking about targeting a protein with drugs is its half-life in the cell. In mammalian cells ODC and AdoMetDC have a very short half-life (40 -43). The availability of DFMO, the specific inhibitor of ODC, did not have the anticipated success in cancer therapy, possibly because the target enzyme is so rapidly degraded within the cell. The rapid degradation of mammalian ODC is conferred by its C-terminal PEST region, a domain that is characterized by the high abundance of the amino acids proline, glutamic acid, serine, and threonine (44). This PEST region is not found in T. brucei ODC, which accordingly has a longer half-life than its mammalian counterpart (45). Interestingly ODC of the related insect parasite Crithidia fasciculata possesses a PEST region, and the half-life of the protein is as short as that of the mammalian enzyme (46). Analysis of the C-terminal part of the plasmodial ODC/AdoMetDC sequence reveals a relatively high abundance of these amino acids, which we previously considered to be a potential PEST region (8). Incubating trophozoite-infected erythrocytes with cycloheximide for up to 2 h and determining the specific activities of ODC and AdoMetDC after the respective incubation periods showed, however, that the enzyme activities remained almost unchanged during the incubation time (Fig. 4). These data indicate that PfODC/AdoMetDC is a rather stable protein in these developmental stages of the parasites and seems to have a half-life of several hours similar to the T. brucei ODC (45).
Even though the half-life of the plasmodial protein is rather long, DFMO has only moderate effects on the survival of the erythrocytic stages of P. falciparum in vitro and on the development of Plasmodium berghei in an experimental animal model (47). There are several explanations for this inefficiency of DFMO including the possible poor uptake into the infected cells; after all, the drug has to cross three membranes to reach its target. Another possibility could be the occurrence of alternative pathways that supply the required putrescine. To evaluate the precise role of the ODC in the synthesis of polyamines in P. falciparum and to validate its potential as a drug target, a selective knockout using gene replacement techniques will be necessary.