Molecular and Biochemical Parasitology, 41 (1990) 259-268 Elsevier

259

MOLBIO 01360

Expression, purification, biochemical characterization and inhibition of recombinant Plasmodium falciparum aldolase Heinz

D r b e l i 1, A r n o l d T r z e c i a k 1, D i e t e r G i l l e s s e n 1, H u g u e s M a t i l e 1, I n d r e s h S r i v a s t a v a 2, L u c H . P e r r i n 2, P e t e r E . J a k o b 1 a n d U l r i c h C e r t a 1

K.

1Central Research Units, F. Hoffmann-La Roche AG, Basel, Switzerland, and 2Division of Infections Diseases, Department of Medicine, University Cantonal Hospital, Geneva, Switzerland

(Received 4 December 1989; accepted 14 March 1990)

The energy metabolism of the blood stage form of the human malaria parasite Plasmodium falciparum is adapted to the host cell. Like erythrocytes, P. falciparum merozoites lack a functional citric acid cycle. Generation of ATP depends therefore fully on the glycolytic pathway. Aldolase is a key enzyme of this pathway and a high degree of sequence diversity between parasite and host makes it a potential drug target. We have expressed the enzyme in its tetrameric form in Escherichia coli and the catalytic constants Vm~xand Km of the recombinant enzyme correspond to the constants of parasite-derived aldolase. Rabbit antibodies against the recombinant P. falciparum aldolase inhibit the natural enzyme and no cross-reaction with human aldolase is detectable. Both the recombinant and the natural protein bind to the cytosofic domain of the band 3 membrane protein in vitro. A 19-residue synthetic peptide corresponding to the sequence of the binding domain of band 3 is an inhibitor when included in the binding assay. In addition, this peptide inhibits the catalytic activity of recombinant P. falciparum aldolase when assayed in a buffer system devoid of anions such as chloride or phosphate. The band 3-derived peptides compete with the aldolase substrate fructose-l,6-diphosphate for binding, suggesting that both reagents have a high affinity for the substrate pocket. A similar sequence motif exists in P. falciparum actin II. A 19-residue peptide corresponding to this sequence is also an inhibitor which could suggest that the P. falciparum aldolase can associate with the cytoskeleton of the parasite or of the host. Key words: Plasmodium falciparum; Recombinant aldolase; Cytoskeleton binding; Band 3 analogue peptide

Introduction T h e m a n i f e s t a t i o n of clinical s y m p t o m s of m a l a r i a caused by P l a s m o d i u m falciparum occurs during the erythrocytic cycle. A t this stage, the Correspondence address: H. Drbeli, Hoffmann-La Roche AG, 4002 Basel, Switzerland Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase, EC 1.2.1.12; GDH, glycerol-3-phosphate dehydrogenase, EC 1.1.1.8; PFK, phosphofructokinase, EC 2.7.1.11; PGK, 3-phosphoglycerate kinase, EC 2.7.2.3; TIM, triosephosphate isomerase, EC 5.3.1.1; rec.P.f., recombinant P. falciparum (aldolase) EC 4.1.2.13; FDP, fi'uctose-l,6diphosphate; F1P, fructose-l-phosphate; NADH, i~-nicotinamide-adenine dinucleotide; NTA, nitrilotriacetic acid; PMSF, phenylmethylsulfonyl fluoride; TLCK, tosyI-Llysylchloromethane-HCl; Tris, 2-amino-2-(hydroxymethyl)1,3-propanediol; RT, room temperature (approx. 20°C); RBC, red blood cells; IRBC, infected red blood cells

energy r e q u i r e m e n t of the parasites reaches a m a x i m u m due to R N A , D N A and p r o t e i n synthesis. Like the e r y t h r o c y t e , P. falciparum has no functional citric acid cycle, A T P p r o d u c t i o n therefore d e p e n d i n g fully on the glycolytic pathway. Consequently, erythrocytes infected by P. falciparum show dramatically increased c o n s u m p tion of glucose, which is c o n v e r t e d to lactate [1]. C o m p o u n d s inhibiting glycolytic e n z y m e s are therefore e x p e c t e d to interfere with parasite replication and function. E n z y m e s that catalyze a cascade of reactions often f o r m complexes. F o r e x a m p l e , v e r t e b r a t e phosphofructokinase (PFK), aldolase and glyceraldehyde-3-phosphate dehydrogenase ( G A P D H ) are believed to be associated with the cytoskeleton of the cell. T h e site of association is actin in muscle and the b a n d 3 p r o t e i n in erythro-

0166-6851/90/$03.50 (~) Elsevier Science Publishers B.V. (Biomedical Division)

260 cytes. When bound to the membrane, the catalytic properties of these enzymes are changed. Substrates or products, but not closely related metabolites, displace the enzymes from their binding sites. This suggests specific interaction either at the enzyme's active site or at a coupled allosteric site [2-6]. The ratio of soluble to bound enzymes is therefore variable, and depends on the actual metabolic state of the cell [2,6]. This matrix organization is not yet fully understood, but its high specificity suggests a key role, for example in the regulation of glycolysis or in the rapid conversion of the cytotoxic intermediate glyceraldehyde3-phosphate to 1,3-diphosphoglycerate. We therefore propose that disturbance of this equilibrium by reagents other than substrate or products could deregulate metabolism, which could ultimately lead to the accumulation of cytotoxic aldehyde. Here we report the expression and an efficient purification method of P. falciparum aldolase made in Escherichia coli. We establish that P. falciparum aldolase can associate in vitro with band 3 and that small peptides of that protein interact with the catalytic center. Materials and Methods

Cloning, expression and fermentation of recombinant P. falciparum aldolase. D N A of clone g41-D of P. falciparum [7] was digested with SspI and subcloned into the Sinai site of M13mp18 [8]. Double-stranded D N A of the subdone M13-41D was digested with BamHI and PstI and deletions were introduced as described [9] in order to remove an in-frame T A G stop codon. Clones were screened by D N A sequencing [10] and clone 2/13 had the stop codon removed. The insert was gel-purified after partial HindIII digestion. The sticky ends were filled in with Klenow polymerase followed by ligation of phosphorylated BamHI linkers (8-met, New England Biolabs). The fragment was gel purified and cloned for expression into the unique BamHI site of pDS78/ RBSII,6xHis [11]. A clone (p8/3) expressing P. falciparum aldolase was detected by immunoblotting using rabbit antibodies to P. falciparum aldolase [7]. D N A manipulations were carried out according to standard procedures. Transformed

E. coli cells of clone p8/3 were grown at 37°C in a 100-liter fermenter using LB medium containing 25 ~g m1-1 kanamycin and 100 ~g m1-1 ampicillin. At an absorbance at 600 nm of about 0.6, isopropyl-13-o-thiogalactopyranoside (IPTG) was added to a final concentration of 200 ~M and the ceils were fermented for an additional 3 h yielding 560 g biomass after centrifugation.

Purification of recombinant P. falciparum aldolase. The purification process is based on the inherent capability of aldolases to bind to phosphoceUulose resins and on the artificially introduced histidines having a high affinity to immobilized metal chelate resins [11,12]. 60 g E. coli cells containing recombinant P. falciparum (rec. P.f.) aldolase were suspended in 350 ml of 50 mM 2-amino-2-hydroxymethylpropane-l,3-diol-HC1 (Tris-HCl) + 50 mM KC1, pH 7 and disintegrated with a Manton Gaulin homogenizer. The cell debris was removed by centrifugation. 12 g CeUex P (BioRad) was added to the crude extract and stirred for I h at 4°C for adsorption. Cellex-beads with bound rec.P.f, aldolase were pelleted, resuspended and packed into a column (length 4 cm, diameter 5 cm). The elution was performed with a 2 h linear gradient from 0 to 1 M potassium hydrogen phosphate, pH 7, at a flow of 160 ml h -1. The pooled fractions containing rec.P.f, aldolase (as determined by SDSPAGE) were pumped on to a nitrilotriacetic acid nickel-chelate (NTA) column (synthesis of the gel, see ref. 12), equilibrated with 0.1 M TrisHC1, pH 7.5 + 0.5 M KCI, length 9 cm, diameter 1.6 cm, flow 156 ml h-i). Rec.P.f. aldolase was eluted with a linear gradient of 0-0.5 M imidazole for 2 h [11]. Rec.P.f. aldolase-containing fractions were pooled, concentrated by YM-10 ultrafiltration (Amicon) and rechromatographed on a Sephacryl S-200 column (Pharmacia; length 83 cm, diameter 2.6 cm, flow 15 ml h -1) using PBS (8 g NaCl/0.2 g KC1/0.2 g KH2PO4/2.9 g NaEHPO4.12H20 in 1000 ml H20, pH 7). The column was calibrated with rabbit muscle aldolase, bovine serum albumin and cytochrome c to check the molecular weight. After sterile filtration 20 mg of protein was recovered which fulfilled the following purity criteria: (a) no contamination visible on Coomassie Blue stained SDS-PAGE

261 (Fig. 1); (b) no E. coli contaminations detected by Western blotting (Fig. 1); and (c) endotoxin content less than 3 units (mg protein) -1 (determined with the Limulus amoebocyte lysate test).

Natural aldolases. Aldolase from P. falciparum blood stage was purified by immunoadsorption according to Perrin et al. [13]. The protein was homogeneous as judged by SDS-PAGE. Human aldolase A was isolated as a crude preparation from approx. 100 ml blood. The side fraction obtained during the preparation of ghosts after treating them with fructose-l,6-diphosphate (FDP) according to [4] was passed through a nickel-NTA column (diameter 5 cm, length 4.7 cm, flow 10 ml min-1; buffer 0.1 M sodium phosphate/0.5 M NaCI, pH 7). Aldolase was in the flow-through, and hemoglobin was adsorbed to the NTA resin. The aldolase was then further purified on a DEAE-anion exchange column (diameter 1.6 cm, length 10 cm, flow 1 ml min-1; buffer 50 mM Tris-HC1, pH 7.5; elution at approx. 0.1 M NaC1 in a linear gradient. Rabbit aldolase A was purchased from Boehringer (Mannheim, F.R.G.) Peptide synthesis and sequence analysis. Peptides were synthesized by the solid-phase technique [14] and purified by HPLC using a Lichrosorb RP-18 (10 V,) column in a 0.1% trifluoroacetic acid-ethanol gradient system. N-terminal amino acid sequences were determined with an Applied Biosystems 470 A protein sequencer. Fragmentation of recombinant parasite aldolase was performed with Staphylococcus aureus protease (Pierce) or trypsin from bovine pancreas (Serva) using 30 Ixg of protease per mg of substrate. The fragments were separated by RP-18 HPLC and identified by amino acid analysis using the experimental set-up described in ref. 15. Preparation of red cell membranes (ghosts). Erythrocyte ghosts were prepared as described [4]. Briefly: human blood was obtained from the Blutspendezentrum, Basel. 100-200-ml aliquots were washed in 5 mM sodium phosphate, pH 7/0.15 M NaCI. Erythrocytes were osmotically lysed in 5 mM sodium phosphate, pH 7 (5P7) at 37°C for 30 min. Membrane-bound aldolase was

depleted with 1 mM FDP. GAPDH and other proteins were removed by washing the ghosts with 0.15 M NaC1. Finally, the protein-depleted ghosts were washed with 5P7 until their color was faintly pink to white. After centrifugation at approx. 25 000 × g, the ghosts were resuspended in 5P7 to give the required percentage by volume for use in assays.

Aldolase activity assay. Aldolase activity was assayed by coupling the FDP cleavage to the triose phosphate isomerase (TIM)/a-glycerophosphate dehydrogenase (GDH) reaction and by measuring the consumption of NADH (continuous monitoring at 340 nm). The final reaction mixture contained 0.2 mM NADH/50 p,g m1-1 GDH/TIM (Boehringer, F.R.G.)/0-3 mM FDP and the sample [4]. For measuring the aldolase activity during the purification process or in the ghost binding experiment 0.1 M Tris-HCl, pH 7.5 was used. The effect of the band 3 peptides on the catalytic activity was assayed in 0.2 M glycine which was titrated with Tris base to pH 7.3. Since glycine is a zwitterion, this buffer system is devoid of pure anions. E. coli contains a class II aldolase with a divalent metal ion in the catalytic center, in contrast to P. falciparum or vertebrate aldolases. Addition of 0.1 M EDTA to the assay buffer allows discrimination between P. falciparum and E. coli aldolases during the initial purification steps [16]. Aldolase-binding experiments. Binding of rec.P.f., human A and rabbit A aldolases: the binding of aldolases to human erythrocyte ghosts was performed according to Strapazon and Steck [4]. All assays were performed in 5 mM sodium phosphate, pH 7.10 Ixg m1-1 of aldolase (dialyzed against 5P7) was incubated with varying concentrations of freshly prepared ghosts at 22°C for 30 min. Bound aldolase was separated from unbound by centrifugation and quantified by assaying the aldolase activity in the supernatant, using 0.1 M Tris HC1, pH 7.5 as buffer system (C1- reverses the effect of the peptides). Binding of natural P. falciparum aldolase (unpurified): washed and packed infected erythrocytes at a parasitemia of 15% were incubated with 50 volumes of 0.02% saponin in PBS. The erythrocyte-

262

free parasites were spun down at 2000 x g for 10 min at 4°C and washed twice with saponin PBS. The washed parasites contained in the pellet were extracted with 7 volumes of lysis buffer (5 mM Tris/5 mM EDTA/50 mM NaC1/1% Nonidet P40/2 mM PMSF/1 mM TLCK/0.5 mM iodoacetamide, pH 7.5). The lysate was centrifuged at 20000 x g for 25 min and clear supernatant was collected, aliquoted and frozen at 70°C. Before use the parasite extract was dialyzed overnight at 4°C against 5 mM sodium phosphate, pH 7 and then incubated for 2 h at room temperature (approx. 20°C) with a 10% (v/v) ghost suspension. The mixture was centrifuged at 20000 x g for 15 min. The pellet was resuspended at 10 mM NaCI and incubated for 5 min at RT and then centrifuged again at 20000 x g for 15 min. The supernatant was collected for the determination of the aldolase activity. Similarly, the pellet was further extracted with 50, 100 and 200 mM NaCI (Fig. 4). For Western blotting, 20 ~1 of each supernatant was separated by PAGE, with 9% SDS under reducing conditions, transferred to nitrocellulose [17] and incubated with anti-P, falciparum aldolase antibodies as described before [18].

Preparation of antibodies. A rabbit (Swiss rabbit KOHA) was immunized 3 × subcutaneously with 50 ~g of recombinant P. falciparum aldolase at intervals of 4 weeks using Freund's complete adjuvant the first time and then Freund's incomplete adjuvant. 3 weeks after the last immunization, the rabbit was bled and the serum kept at -20°C. The antibody titer was 1:40000 as measured by indirect immunofluorescence on acetonefixed infected erythrocytes.

~E.co(i 123

~1~_1_

stained

II J --92 --66 --65

6

--31

Fig. 1. Purification of recombinant P. falciparum aldolase. Efficiency of the different purification steps, monitored by SDS-polyacrylamide gel electrophoresis. Components of bacterial origin were visualized by immunoblotting using anti-E. coli antibodies (A). Recombinant P. falciparum aldolase was detected by immunoblotting using anti-native-P, falciparum aldolase antibodies (B) and by staining with Coomassie Brilliant Blue (C). The lanes represent: 1, E. coli lysate; 2, crude extract; 3, CeUex-P pool; 4, NTA-pool; 5, Sephacryl S-200 pool (this lane is highly overloaded, therefore dimers of P. falciparum aldolase are visible).

Results

Characterization of recombinant P. falciparum aldolase. P. falciparum aldolase is expressed in E. coli as a soluble, catalytically active tetramer. A renaturation step prior to or during purification is therefore not required. In order to facilitate the purification process, an affinity tail consisting of six tandem histidine residues was attached to the amino terminus of the P. falciparum aldolase sequence by genetic engineering [11]. The recombinant protein differs therefore from the natural parasite protein [22] by the following amino acids: Ac-AHCTEYMNA ......

Cultivation of parasites. Parasites of the P. falciparum K1 isolate (Thailand) were cultivated in

native

petri dishes as described in [19] or in an automated culture system as described [20]. Non-synchronous cultures were harvested when the parasitemia reached 15-20%, washed 2 × in serum-free culture medium and kept at -70°C until use. Some parasite preparations were further purified by saponin lysis [21].

recombinant MRGSHHHHHHGSELACQYMNA ...... The purified recombinant P. falciparum aldolase does not contain any detectable E. coli contaminations (Fig. 1) and the sequence is identical to the predicted sequence by the following criteria: (a) the amino acid composition matches that predicted from the DNA sequence. (b) Edman

263

degradation revealed the expected amino-terminal sequence M R G S H H H H H H G S E L A X Q Y M NAPKK ..... (c) Two peptides isolated by PR-18 HPLC after digestion with S. aureus V8 protease or trypsin were identified by amino acid analysis as having the compositions 2Lys+2Tyr+lVal (V8) and 2Tyr+ 1Val (trypsin). This matches the sequences Lys-Lys-Tyr-Val-Tyr (V8) and TyrVal-Tyr (trypsin) and confirms the integrity of the carboxy terminus. The tetrameric structure of the recombinant aldolase was confirmed by the following experiments: (a) on a Sephacryl S-200 column equilibrated with a physiological buffer the recombinant P. falciparum aldolase is eluted at the same position as rabbit muscle aldolase. (b) Quasi-elastic light-scattering [23] performed in 5 mM sodium phosphate buffer, pH 7, at a protein concentration of approx. 5 p,M revealed the expected hydrodynamic radius (Table I). (c) The circular dichroic spectrum of the recombinant protein is almost identical to that of rabbit muscle aldolase with a strong positive maximum at 190 nm. This indicates a high proportion of a-helices, which is in agreement with X-ray crystallographic data obtained from tetrameric rabbit muscle aldolase [24]. Independent evidence that the recombinant parasite aldolase has acquired a native-like structure is provided by immunochemical methods. Rabbit antibodies against the recombinant protein specifically inhibit the activity of the recombinant and of the natural enzyme but not of human or rabbit aldolase A (Fig. 2). On the other hand, antibodies against natural parasite aldolase recognize the recombinant protein in the Western blot, as shown in Fig. 1. TABLE I Evidence of tetramer structure

Enzyme and origin

R (A)

Oligomer Structure Mr

Recombinant P.

52-+1

160000

Tetramer

47---2 50-+1 29---2

160000 140000 52000

Tetramer Tetramer Dimer

falciparum aldolase Rabbit muscle aldolase Rabbit G A P D H Rabbit TIM

R, Hydrodynamic radius determined in 5 mM sodium phosphate.

E >. >

..

¢l

~

r

¢.

_-

--*

•

q) U~ n O q= m

100

-

-

17 27 -

34

6 27 12

>70 14

All experiments were performed in 0.2 M glycine-Tris, pH 7.3.To determine Km and K, values, the aldolase activity was measured with respect to varying FDP concentration at different inhibitor concentrations. The K= values were calculated from the linear regression line in the double-reciprocal plot. To calculate K,, 1/V was plotted against the inhibitor concentration. From the intercept of the curves deriving from different substrate concentrations - K , was obtained. The amino acid sequences are given in the single letter code. Negatively charged amino acids are printed in bold letters. The peptides with prefix B derived from human band 3 protein and A19 from P. falciparum actin II. -, Not tested.

100

100

80

80

60

60

40

40

20

20

ek.

Q.

e-

°~ 4.a

o

O 1D

0

I

l

I

I

I

I

0

2

4

6

8

10

Ghost

conc.

%

0 0

100 Peptide

conc.

200 jJg/ml

Fig. 3. Binding of recombinant P. falciparum aldolase to the band 3 protein. (Left) Binding of aldolases to ghosts. 10 p.g ml -~ of aldolase was incubated with varying concentrations of erythrocyte ghosts at 22°C for 30 min in 5 mM sodium phosphate, pH 7. Bound aldolase was then removed by centrifugation and the free aldolase determined by measuring the aldolase activity in the supernatant, e, P. falciparum; &, human A, a, rabbit A. (Right) Competition by band 3 analogue peptides. 10 p.g m l - ' of rec.P.f. aldolase was incubated with a constant amount of erythrocyte ghosts and a varying amount of peptides. The ghost concentration was 1.25% (v/v). Under these conditions about 15% P. falciparum aldolase molecules are free and about 85% bound (see arrow in (A)). The amount of rec.P.f, aldolase released from the membrane depends on the concentration and the sequence of the peptides: O , B19; l , B l l ; e, B9; &, Ac-QEEYED (sequences in Table II).

265

trostatic nature of the interaction. Furthermore, rabbit muscle G A P D H and rabbit PFK compete for the binding site. We therefore expected that the binding site is the band 3 protein as it is for the vertebrate aldolases [2,3]. To test this hypothesis, we synthesized a set of peptides corresponding to the putative binding domain of band 3 [3]. These were expected to block the binding site of P. falciparum aldolase when included in the assay. Indeed, a 19-residue peptide of band 3 (Table II, peptide B19) turned out to be an efficient inhibitor. The ll-residue peptide B l l is drastically less efficient, and shorter peptides like B9, Ac-QEEYED, (Ac)2-KEEYEDPGHHHHHH or peptides not related to the band 3 sequence like GDDDDK-13-naphthylamide (substrate of enterokinase) do not inhibit the association to band 3. However, a peptide with the sequence Ae-MEELQDDYEDDMEENLK(Ae)K ( A c ) E E Y E D P G H H H H H H is an inhibitor with

the same potency as B19. The peptides B l l and B9 have a higher percentage of acidic residues than B19, and the two hexahistidine peptides show the opposite behavior. Therefore, we conelude that the inhibitory effect of B19 is sequencespecific and not merely electrostatic in nature. As expected, natural P. falciparum aldolase shows the same association behavior to the erythrocyte membrane (Fig. 4). This confirms that the properties of the natural and the recombinant enzymes are highly similar, if not entirely identical.

Inhibition of the catalytic activity by peptides. Inhibition of the enzymatic activity is observed in buffer systems containing zwitterions but no small anions like chloride or phosphate. None of the peptides are inhibitors in Tris-HCl, but FDPdependent competitive inhibition occurs in 0.2 M glyeine-Tris (Table II). Anions such as chloride or

MW 12

>,

?

4d

:> °~ 4~

8

:/

0 co

O)

--

100

--

69

--

46

--

30

c'

4

¢D O "0


100 IxM. On the other hand, phosphate does not influence KI or Vmaxin the absence of B19. It is known that vertebrate glycolytic enzymes are bound to cytoskeleton proteins such as actin [6]. Because P. falciparum actin II contains a similar sequence motif as band 3 [25] we synthesised a 19-residue peptide (A19) corresponding to this homologous region. However, the inhibition of A19 is four times lower than B19 (Table II) which confirms that the interaction is sequence-specific.

Inhibition by suramin. The inhibitory peptides as well as the natural enzyme substrate contain clusters of two or three negatively charged groups. Therefore, we extended our search for potential aldolase inhibitors to molecules with such properties. One reagent is Suramin, a sulfonic dye with a rod-like structure bearing an accumulation of three negative charges at either end. Suramin has been in use for 65 years as an antiparasitic drug to treat trypanosomal infections. A major limitation, however, is its severe side-effects. Binding of recombinant P. falciparum aldolase to erythrocyte ghosts is inhibited by Suramin with a slightly higher efficiency than by the peptide B19. This clearly demonstrates that the site of Suramin interaction is related to the band 3 association site which in turn overlaps with the substrate-binding pocket. As expected, Suramin is also an inhibitor of the catalytic activity of P. falciparum aldolase and rabbit muscle aldolase (Table II). An effect on TIM and GDH, which are present in the coupled assay system, is also possible, but suppressed by the experimental set up (high excess of TIM and GDH over aldolase and variation in FDP concentration). Interestingly, there is a difference in inhibition characteristics between the rabbit and the parasite aldolase. Rabbit aldolase is inhibited predominantly competitive to FDP, the Km is shifted to higher values but Vmaxremains almost constant. On the other hand Suramin affects both Km and Vmaxof P. falciparum aldolase. This is a further example that the catalytic site of P. falciparum and rabbit aldolases are different.

Discussion

We have expressed, purified and characterized the recombinant P. falciparurn aldolase. For ease of purification we extended the amino terminus by six histidine residues. According to crystallographic data [24], the carboxy-terminal region of rabbit muscle aldolase is part of the active site pocket, and removal or addition of extra amino acids could thus interfere with enzyme function. Comparison of natural with recombinant P. falciparum aldolase reveals virtually no enzymatic or immunological differences. Therefore, the recombinant aldolase can substitute for the natural enzyme in the initial search for novel antimalarial drugs. A requirement for an antimalarial drug is parasite specificity. However, the structure of the catalytic centre of enzymes is generally highly conserved because it is designed for the invariant substrate structure. All amino acids present in the catalytic center of vertebrate, Drosophila, maize, Trypanosoma brucei and P. falciparum aldolase are almost identical [26,27]. On the other hand, the organization of the metabolic pathway is different in respect to the association to other proteins. For instance, glycolytic enzymes of trypanosomes associate in distinct organelles termed glycosomes [28]. Three dimensional structure analysis of TIM, GAPDH and PGK of T. brucei has revealed that all these enzymes have in common two clusters of positively charged residues about 40 A apart. These clusters are proposed to be important for the import into these glycosomes [28,29]. As Suramin has two clusters of three negatively charged groups, it may interfere in glycolysis by interacting with these 'hot spots' [30]. A homologous organization may exist in vertebrate cells: PFK, aldolase and GAPDH associate with actin in muscle cells [6] or with the band 3 in erythrocytes [2-5]. Based on our in vitro experiments, we suggest that a similar organization of glycolytic enzymes exists in P. falciparum. First, the parasite aldolase has a high specific affinity to band 3 and secondly, a P. falciparum actin II-derived peptide which resembles the binding region of band 3 is an inhibitor of the recombinant parasite aldolase. The common elements of these sequence motifs

267 are tyrosine residues which are flanked by negatively c h a r g e d a m i n o acids. T h e s e sequence elements are present in t a n d e m in the b a n d 3 protein and in actin [5,25]. A t least tyrosine 8 is a p h o s p h o r y l a t i o n site [31]. Tyrosine phosphorylation and d e p h o s p h o r y l a t i o n is correlated to metabolic regulation and m a y play an i m p o r t a n t role in the infectious cycle o f P. falciparum: The cytoskeletal proteins o f erythrocytes are d e p h o s p h o r y l a t ed after infection by m e r o z o i t e s [32]. If it turns out that the glycolytic m a c h i n e r y of the malaria parasite d e p e n d s on an association of e n z y m e s to a matrix we expect to find inhibitors that interfere with this assembly reaction in vivo.

Acknowledgements W e wish to t h a n k Y v o n n e Burki, Patricia G u i c h a r d , K a r e n H o l l a n d e r and Daniel R o t m a n n for excellent technical assistance. W e are further grateful to Dr. W. B a n n w a r t h for synthesizing oligonucleotides, to Dr. B. W i p f for large scale ferm e n t a t i o n s , to D r . H . - W . L a h m , Dr. W . Vetter, M r U. R6thlisberger and M r M. M a n n e b e r g for s e q u e n c e c o n f i r m a t i o n and to Drs. M. Z u l a u f and K. N o a c k for m e a s u r i n g the physical p a r a m e t e r s o f P. falciparum aldolase. W e w o u l d also like to t h a n k Dr. J. Sygusch and Dr. R. Pink for helpful discussions. T h e w o r k in the l a b o r a t o r y o f D r Pern n is s u p p o r t e d by the Swiss N a t i o n a l R e s e a r c h F o u n d a t i o n (grant 3.923.0.87).

References 1 Sherman, I.W. (1984) Antimalarial drugs, Chapter 2: Metabolism. In: Handbook of Experimental Pharmacology 68/I, (Peters, W. and Richards, W.H.G., eds), pp. 31-81. Springer-Verlag, Berlin. 2 Jenkins, J.D., Madden, D.P. and Steck, T.L. (1984) Association of PFK and aldolase with the membranes of the intact erythrocyte. J. Biol. Chem. 259, 9374-9378. 3 Tsai I.-H., Murthy, S.N.P. and Steck, T.L. (1982) Effect of red cell membrane binding on the catalytic activity of GAPDH. J. Biol. Chem. 257, 1438-1442. 4 Strapazon, E. and Steck, T.L. (1977) Interaction of the aldolase and the membranes of human erythrocytes. Biochemistry 16, 2966-2971. 5 Low, S.P. (1986) Structure and function of the cytoplasmic domain of band-3: center of erythrocyte membrane-peripheral protein interaction. Biochim. Biophys. Acta 864, 145--167. 6 Masters, C. (1984) Interactions between glycolytic

enzymes and components of the cytomatrix. J. Cell Biol. 99,222s--225s. 7 Certa, U., Ghersa, P., D6beli, H., Mafile, H., Kocher, H.P., Srivastava, I.K., Shaw, A.R. and Perrin, L.H. (1988) Aldolase activity of a Plasmodiumfalciparum protein with protective properties. Science 240, 1036-1038. 8 Vieira, J. and Messing, J. (1982) The pUC plasmids, an M13 mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19,259268. 9 Henikoff, S. (1984) Unidirectional digestion with exonuclease III creates targeted breakpoints for DNA sequencing. Gene 28, 351-359. 10 Sanger, F., Nicklen, S. and Coulson, A.R. (1977) DNA sequencing with chain terminating inhibitors. Proc. Natl. Sci. USA 74, 5463-5467. 11 Hochuli, E., Bannwarth, W., D6beli, H., Gentz, R. and Stiiber, D. (1988) Genetic approach to facilitate purification of recombinant proteins with a novel metal chelate adsorbent. Biotechnology 6, 1321-1325. 12 Hochuli, E., D6beli, H. and Schacher, A. (1987) New metal chelate adsorbent selective for proteins and peptides containing neighbouring histidine residues. J. Chromatogr. 411,177-184. 13 Pert'in, L.H., Merkli, B., Gabra, M.S., Stocker, J.W., Chizzolini, C. and Ricble, R. (1985) Immunization with a Plasmodiumfalciparum merozoite surface antigen induces a partial immunity in monkeys. J. Clin. Invest. 75, 17181721. 14 Atherton, E. and Sheppard, R.C (1987) In: The Peptides: Analysis, Synthesis, Biology, Voi. 9. (Udenfriend, S. and Meienhofer, J., eds.), pp. 1-38, Academic Press, New York. 15 D6beli, H., Gentz, R., Jucker, W., Garotta, G., Hartmann, D.W. and Hochuli, E. (1988) Role of the carboxyterminal sequence on the biological activity of human immune interferon (IF'N-T). J. Biotechnol. 7, 199-216. 16 Rutter, W.J., Rajkumar, T., Penhoet, E., Kochman, M. and Valentin, R. (1968) Aldolase variants: structure and physiological significance. Ann. NY Acad. Sci. 151, 102117. 17 Dunn, S.D. (1986) Effect of the modification of transfer buffer composition and the renaturation of proteins in gels on the recognition of proteins on western blots by monoclonal antibodies. Anal. Biochem. 157,144-153. 18 Srivastava, I.K., Schmidt, M., Certa, U., D6beli, H. and Perrin, L.H. (1990) Specificity and inhibitory activity of antibodies to Plasmodium falciparum aldolase. J. Immunol. 144, 1497-1503. 19 Trager, W. and Jensen, J.B. (1976) Human malaria parasite in continuous culture. Science 193,673-675. 20 Ponnudurai, T., Lensen, A.H.W., Meis, J.F.G.M. and Meuwissen, J.H.E.Th. (1986) Synchronization of Plasmodium falciparum gametocytes using an automated suspension culture system. Parasitology 93, 263-274. 21 Goman, M., Langsley, G., Hyde, J.E., Yanofsky, N.K., Zolg, J.W. and Scaife, J. (1982) The estabfishment of genomic libraries for human malaria parasite P. falciparum and identification of individual clones by hybridisation. Mol. Biochem. Parasitol. 5, 391-400.

268 22 Ghersa, P., Srivastava, I.K., D0beli, H., Becherer, D., Matile, H. and Certa, U. (In Press) Initiation of translation at a U A G stop codon in the aldolase gene of Plasmodium falciparum. EMBO J. 23 Cuillel, M., Zulauf, M. and Jacrot, B. (1983) Selfassembly of Brome mosaic virus protein into capsids. Initial and final states of aggregation. J. Mol. Biol. 164, 589--603. 24 Sygusch, J., Beaudry, D. and Allaire, M. (1987) Molecular architecture of rabbit skeletal muscle aldolase at 2.7 A resolution. Proc. Natl. Acad. Sci. USA 84, 7846-7850. 25 Wesseling, J.P., Smits, M.A. and Schoenmakers, J.G.G. (1988) Extremely diverged actin proteins in Plasmodium falciparum. Mol. Biochem. Parasitol. 30, 143-154. 26 Kelly, P.M. and Tolan, D.R. (1987) The complete amino acid sequence for the anaerobically induced aldolase from maize derived from cDNA clones. Plant Physiol. 82, 10761081. 27 Marchand, M., Poliszczak, A., Gibson, W.C., Wierenga, R.K., Opperdoes, F.R. and Michels, P.A.M. (1988) Characterization of the genes for fructose-biphosphate

28

29

30 31

32

aldolase in Trypanosoma brucei. Mol. Biochem. Parasitol. 29, 65-76. Wierenga, R.K., Swinkels, B., Michels, P.A.M., Osinga, K., Misset, O., Van Beeumen, J., Gibson, W.C., Postma, J.P.M., Borst, P., Opperdoes, F.R. and Hol, W.G.H. (1987) Common elements on the surface of glycolytic enzymes from Trypanosoma brucei may serve as topogenic signals for import into glycosomes. EMBO J. 6, 215-221. Hol, W.G.J. (1986) Proteinkristallographie und Computer-Graphik: auf dem Weg zu einer planvollen Arzneimittelentwicklung. Angew. Chem. 98, 765-777. Hart, D., Langridge, A., Barlow, D. and Sutton, B. (1989) Antiparasitic drug design. Parasitol. Today 5, 117-120. Murray, M.C. and Perkins, M.E. (1989) Phosphorylation of erythrocyte membrane and cytoskeleton proteins in cells infected with Plasmodium falciparum. Mol. Biol. Parasitol. 34, 229-236. Dekowski, S.A., Rybicki, A. and Drickamer, K. (1983) A tyrosine kinase associated with the red cell membrane phosphorylates band 3. J. Biol. Chem. 258, 2750-2753.