The Crystal Structure of Plasmodium falciparum Glutamate Dehydrogenase, a Putative Target for Novel Antimalarial Drugs

doi:10.1016/j.jmb.2005.03.077

J. Mol. Biol. (2005) 349, 597–607

The Crystal Structure of Plasmodium falciparum Glutamate Dehydrogenase, a Putative Target for Novel Antimalarial Drugs Christof Werner1, Milton T. Stubbs1,3, R. Luise Krauth-Siegel2 and Gerhard Klebe1* 1

Institute of Pharmaceutical Chemistry, Philipps-University Marburg, Marbacher Weg 6 D-35037 Marburg, Germany 2 Centre for Biochemistry Ruprecht-Karls University Heidelberg, Im Neuenheimer Feld 504, D-69120 Heidelberg Germany 3

Institute for Biotechnology Martin-Luther-University Halle-Wittenberg Kurt-Mothes-Straße 3, D-06120 Halle, Germany

Plasmodium falciparum is the main causative agent of tropical malaria, the most severe parasitic disease in the world. Growing resistance of Plasmodia towards available drugs is an increasing problem in countries where malaria is endemic. As Plasmodia are sensitive to oxidative stress, augmenting this in the parasite represents a promising principle for the development of novel antimalarial drugs. The NADP-dependent glutamate dehydrogenase (GDH) of P. falciparum is largely responsible for the production of NADPH in the parasite, which in turn serves as electron source for the antioxidative enzymes glutathione reductase and thioredoxin reductase. As GDH does not occur in the host erythrocyte, GDH is a particularly attractive target for drug therapy. The three-dimensional structure of P. falciparum GDH in the unligated state has been determined ˚ . Compared to the by X-ray crystallography to a resolution of 2.7 A mammalian enzymes, two amino acid residues are exchanged in the putative active site of the parasite GDH. The most obvious differences between parasite and human GDH are the subunit interfaces of the hexameric proteins. In the parasite protein, several salt-bridges mediate contacts between the subunits whereas in the human enzyme these interactions are mainly of hydrophobic nature. Furthermore, P. falciparum GDH possesses a unique N-terminal extension that does not occur in any other GDH sequence so far studied. These findings might be exploited for the design of peptidomimetics capable of disrupting the oligomeric organisation of the parasite enzyme. q 2005 Elsevier Ltd. All rights reserved.

*Corresponding author

Keywords: Plasmodium falciparum; malaria; glutamate dehydrogenase; drug design; crystal structure analysis

Introduction Plasmodium falciparum is the causative agent of tropical malaria. The clinical manifestations are caused by the multiplication and differentiation of the protozoan parasite in human red blood cells. The parasite is exposed to multiple oxidative stress due to its high metabolic rate, the degradation of heme and reactive oxygen species imposed by the host immune system.1–3 As in other eukaryotes and prokaryotes, glutathione and thioredoxin systems represent the major antioxidant defence lines in Plasmodia.4–6 The two NADPH-dependent Abbreviation used: GDH, glutamate dehydrogenase. E-mail address of the corresponding author: [email protected]

enzymes glutathione reductase and thioredoxin reductase have received major interest as potential antimalarial drug targets. It is well established that glucose-6-phosphate-dehydrogenase deficiency (G6PDH-deficiency), the most frequent hereditary disease in the world, confers partial protection against P. falciparum malaria.2 Plasmodia possess a bifunctional glucose-6-phosphate dehydrogenase6-phosphogluconolactonase.7 Recently, it has been shown that expression of the gene reaches a maximum in the young trophozoite stage, 28 hours after infection, and then strongly declines in mature trophozoites (40 hours) and schizonts (48 hours).8 This expression pattern is not expected for a house-keeping enzyme and contrasts with that of structural protein genes such as those for myosin and actin, which show a continuous increase in expression during the cycle.8 Thus the contribution

0022-2836/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.

598

Crystal Structure of Glutamate Dehydrogenase

Figure 1. Sequence alignment of the P. falciparum, bacterial C. symbiosum and human GDH enzyme performed with the program CLUSTALW (Swissprot accession codes for the sequences O96940_PLAFA, P24295jDHE2_CLOSY, P00367jDHE3_HUMAN). The nucleotide-binding fold is indicated in red, the N-terminal domain in blue and the antenna region in the human enzyme in green. The N-terminal extension of PfGluDH is shown in violet. The residues underlaid in yellow indicate those that bind the glutamate and in pink those binding NADP.

of the parasite hexose monophosphate shunt to NADPH supply is still not clear.9 Glutamate dehydrogenases (GDHs) are ubiquitous enzymes that occupy an important branchpoint between carbon and nitrogen metabolism. They catalyse the reversible NAD(P)C-dependent oxidative deamination of L-glutamate to 2-oxoglutarate and ammonia: L-glutamate þ

NADðPÞþ þ H2 O %

þ 2-oxoglutarate þ NHþ 4 þ NADðPÞH þ H

NADP-dependent GDHs are usually concerned with ammonium assimilation, whereas the NAD-

specific enzymes are generally involved in glutamate catabolism. The dual-specific vertebrate GDHs such as the bovine or human enzyme can use both NADC and NADPC with comparable efficacy and the anabolic versus catabolic balance is tightly controlled by a complex network of allosteric effectors.10 It has long been recognized that malarial parasites exhibit an NADP-specific glutamate dehydrogenase activity11 and the enzyme can serve as a hexose monophosphate shunt-independent source of NADPH in P. falciparum.12,13 Remarkably, in contrast to glucose-6-phosphate dehydrogenase, glutamate dehydrogenase is present throughout the intraerythrocytic cycle of the parasite with the

599

Crystal Structure of Glutamate Dehydrogenase

highest concentration in both young and old trophozoites and the segmenter stage.14 Furthermore, GDH is a marker enzyme for the parasite compartment, as it is absent from human red blood cells.15,16 Glutamate dehydrogenase has been isolated from the intraerythrocytic stage of P. falciparum17 and the gene has been cloned and overexpressed in Escherichia coli.14 The parasite enzyme shows a low sequence identity of only 23% with human GDH (Figure 1).14,18 It is a homohexamer with a subunit Mr of 49,500,14 in accordance with bacterial and vertebrate NADPdependent enzymes, which are also homohexamers with a subunit Mr of 48,000 and 55,000, respectively.19 P. falciparum GDH possesses a unique N-terminal extension that is not observed in any other of the presently characterized GDH sequences and the function of which is as yet unknown. The N terminus of the enzyme is therefore 20 residues longer than the related bacterial enzyme and 30 residues longer than the mature human enzyme.14 Crystal structures of several hexameric GDHs have been determined, including those of Clostridium symbiosum,20,21 the enzymes of the hyperthermophilic archea Pyrococcus furiosus,22 Thermotoga maritima,23 Thermococcus litoralis24 and Thermococcus profundus25 as well as the GDHs of Bos taurus10 and Homo sapiens.26 The crystal structures of the clostridial and bovine enzymes contain the bound substrate glutamate and either NADP(H) or NAD(H) as cofactor, the latter in case of the mammalian enzymes (Table 1). The goal of the present study was a detailed elucidation of similarities and differences between the plasmodial and human enzymes in order to provide a basis for structure-based drug design of specific and selective inhibitors towards the P. falciparum GDH. Here, we report on the crystallisation and structural analysis of P. falciparum glutamate dehydrogenase.

Results and Discussion Purification and crystallisation The preparation of crystals of P. falciparum GDH suitable for diffraction experiments proved to be a major problem. Initial crystallisation attempts with protein purified using the established protocol14 yielded thin hexagonal plates. Subsequent X-ray crystallographic experiments showed that these crystals were of monoclinic space group C2 with a very long c-axis, which, together with a high mosaicity, led to overlap of reflections. Another major problem was the apparent disorder of the crystals. The diffraction patterns of initial X-ray crystallographic experiments showed the existence of multiple lattices, probably caused by the multilayered structure of the plate-like crystals. This led to difficulties in the indexing of the reflections and to a very low overall completeness of the data set. Introduction of an additional gel-filtration step improved the crystallisation properties of the prepared protein significantly. Two crystal forms, a plate-like crystal form in monoclinic space group C2 and a cubic-shaped crystal form in space group I23, were obtained. Although still slightly disordered, the plate-like crystals allowed recording ˚ resolution. The cubic crystal of a data set at 2.7 A ˚ and was thus form, however, diffracted only to 6 A not suited for further experiments. The crystal form in space group C2 could be solved using molecular replacement techniques with the clostridial enzyme as search model (1bgv, trigonal crystal form in H32, hexagonal setting). Refinement of the molecule indicated a breakdown of strict 32 local symmetry, so that symmetry averaging could not be applied. Subsequent refinement revealed that electron density is less well defined for three of the monomers A, C and F in their C-terminal portions. This is presumably a

Table 1. GDH-structures listed in the PDB PDB-code

Species

1AUP 1BGV 1K89 1HRD 1B26 1B3B 2TMG 1HWX 1HWY

C. symbiosum C. symbiosum C. symbiosum C. symbiosum T. maritima T. maritima T. maritima B. taurus B. taurus

1HWZ 1NQT 1NR7 1L1F 1NR1 1BVU 1EUZ 1GTM

B. taurus B. taurus B. taurus H. sapiens H. sapiens T. litoralis T. profundus P. furiosus

Ligands

Mutations

˚) Resolution (A

Mr monomer

K89L; S380V

2.50 1.90 2.05 1.96 3.00 3.10 2.90 2.50 3.20

49,165 49,165 49,165 49,165 45,689 45,689 45,689 61,512 61,512

2.80 3.50 3.30 2.70 3.30 2.50 2.25 2.20

55,561 55,561 55,561 61,398 61,398 46,608 46,699 47,114

Glutamic acid (Glu) K89L N97D; G376K Multiple Glu; GTP; NADH 2-Oxoglutaric acid, NAD phosphate Glu; GTP; NADPH ADP R463A Sulfate Sulfate

600

Crystal Structure of Glutamate Dehydrogenase

Figure 2. Stereo views of the P. falciparum GDH hexamer along (a) the 2-fold and (b) the 3-fold axes. The individual monomers are coloured as follows: A, dark blue; B, light green; C, green; D, light yellow; E, light blue; and F, yellow. The locations of the active sites are indicated by the red surface. The A monomer is chosen as reference.

result of the different crystalline environments of these regions. P. falciparum GDH structure P. falciparum GDH is a hexamer of 32 symmetry (Figure 2), composed of six identical subunits. Each monomer can be subdivided into two domains separated by a deep cleft in which the substrate and cofactor binding sites are accommodated. Domain I is composed of two sequential stretches involving residues 1–220 and 452–470, and contributes to the subunit–subunit contacts with neighbouring monomers. As in C. symbiosum GDH, the second domain (residues 221–451) corresponds to the classical nucleotide-binding fold with one of the b-strands in reverse orientation20 (Figure 3). GDHs are known to undergo large conformational changes upon substrate and/or cofactor binding to form the productive catalytic complex. Although the actual driving force for this transition remains to be resolved, the cleft between the two domains closes upon the bound glutamate to arrest both substrate and cofactor in an orientation productive for catalysis.27

Active site The glutamate-binding site in P. falciparum GDH can be inferred by comparison with the structure of C. symbiosum GDH.20 All residues in direct contact with the substrate in the crystal structure of the bacterial enzyme are conserved and occupy similar positions in P. falciparum GDH (Figure 4(a)). Residues Lys112 (C.s. 89) and Ser401 (C.s. 380) most likely bind the g-carboxylate group of the substrate, whilst Gln133 (C.s. 110) and Lys136 (C.s. 113) are assumed to bind the a-carboxylate group of glutamate; the amino group would then be coordinated by Gly187 (C.s. 164) and Asp188 (C.s. 165). Only two of these residues differ in the human enzyme: the a-carboxylate ligand Gln133 is substituted by Met115 and the amino group ligand Gly187 by Pro171. In addition to these differences in primary sequence, further differences to the mammalian enzymes may occur in the detailed binding of substrate. In bovine liver GDH, which displays 96% sequence identity with the human enzyme, the binding mode of glutamate differs slightly from that observed in C. symbiosum GDH.10 In the bovine enzyme, Lys126 interacts directly with the amino

601

Crystal Structure of Glutamate Dehydrogenase

crystal structures (bovine GDH is in a more closed conformation than C. symbiosum GDH).10). Cofactor-binding site The nature of the coenzyme-binding site clearly identifies P. falciparum GDH as an NADP-dependent enzyme. It displays greater similarity with the dual-specific mammalian enzymes than with the NAD-dependent protein of C. symbiosum variant. This is particularly clear in the proposed neighbourhood of the adenine-ribose ring, where the NADP and dual-specific enzymes accommodate the 2 0 -phosphate group of the ribose moiety (Figure 4(b)). Neighbouring residues Ser284, Lys156 and Lys306 in the plasmodial enzyme would allow charge compensation of the 2 0 -phosphate group, as observed in bovine GDH (residues Ser276, Lys134 and Lys295, respectively). Although the latter Lys306 (also present in the NADP-dependent GDH of E. coli.29) adopts a different conformation in the plasmodial and bovine enzymes, this is probably a result of the absence of cofactor in the parasite enzyme. In contrast, these NADP-specific residues are replaced in the clostridial enzyme by Arg285, Asn133 and Pro262, restricting access to the phosphate-binding site and resulting in a supposedly less positively charged environment. The N-terminal extension

Figure 3. (a) P. falciparum and (b) B. taurus (1hgz.pdb) glutamate dehydrogenases mutually aligned showing the nucleotide-binding fold in red, the N-terminal domain in blue and the antenna region in the bovine enzyme in green.

group of the substrate while the corresponding Lys125 in the bacterial enzyme binds to the amino function via a conserved interstitial water molecule.20 Furthermore, Asp168 in the bovine enzyme, equivalent to Asp165 in C. symbiosum GDH, is more ˚ away from the terminal glutamate amino than 4 A group in the bovine GDH–glutamate NADH complex structure.10 This is in clear contrast to the observation in the binary C. symbiosum GDH– glutamate complex where Asp165 forms a direct hydrogen bond to the substrate and is essential for enzyme catalysis.10,20,28 These findings could result in differences in the catalytic mechanism between the mammalian and bacterial enzymes and may also be effective for the closely related plasmodial catalyst. Nevertheless, additional information is required to substantiate whether the above-mentioned structural differences relate to deviating catalytic mechanisms between the plasmodial/ clostridial and bovine enzymes or simply result from two distinct conformations adopted in the

P. falciparum GDH exhibits a unique 20 residue long (compared to bacterial and yeast GDHs) or 30 residue long (compared to the mature human enzyme) amino-terminal extension.14 The extension, which ends at Tyr21 adjacent to the long helix 22–41, folds back as a random coil and resides on the remaining part of the N terminus, forming some internal polar interactions (Figure 5). The electron density is less well defined and the assigned B-factors correspond to the highest observed for this protein, indicating enhanced residual mobility or partial disorder. Arg10, which adjoins the C-terminal part of helix 447–465, plays an important role in this portion. It is hydrogen-bonded to Gln94 0 from a neighbouring subunit. Arg10 reveals lower B-factors compared to all other extensions, underlining its spatial arrest due to multiple interactions contributing to the packing against the N-terminal domain. The N-terminal peptide approaches the C-terminal helix, shielding it from solvent. Interestingly, the region covered by P. falciparum residues Leu4 to Ser20 is similarly buried in the clostridial and mammalian enzymes. In C. symbiosum, this is achieved via a short stretch of a-helix between residues Val35 and Leu48; in the bovine enzyme, an extension of two helices between residues Val24 and Ile48 is observed (Figure 5). Due to the inter-subunit contact formed by Arg 10 and Gln 94 0 , it is tempting to speculate that the N-terminal extension could play a role in the assembly of the GDH hexameric structure: According to the 32 symmetry of the hexamer, the

602

Crystal Structure of Glutamate Dehydrogenase

Figure 4. (a) Stereoview of the binding site in P. falciparum GDH with a modelled orientation of the substrate glutamate; the residues K112, Q133, K136, G187, D188, S401 form close contacts to the substrate. (b) NADPH binding site of bovine GDH with the residues involved in binding the 2 0 -phosphate group shown with grey carbon atoms. The residues shown with yellow and blue carbon atoms are a superposition of the corresponding residues present in the NADPH-dependent GDH of P. falciparum (yellow) and NAD-dependent GDH of C. symbiosum (blue).

N-terminal extension of one subunit is in contact with two neighbouring subunits. However, the total inter-subunit contact surface is rather small. As mentioned above, the extension folds towards the upper part of the N-terminal domain and remains largely solvent-exposed in the hexamer. Only residues Thr8 to Phe11 make short subunit– subunit contacts with two adjacent subunits, and

Arg10 bridges among the subunits via a hydrogen bond. In order to test the possible importance of the 20 residue extension in hexamer assembly, we produced a deletion mutant of P. falciparum GDH lacking the first 19 residues of this extension. However, the mutated protein shows unchanged activity and forms hexamers as revealed by

603

Crystal Structure of Glutamate Dehydrogenase

Figure 5. (a) The N-terminal peptide of P. falciparum GDH (pink), which is longer than that of the bacterial and mammalian enzymes by some 20–30 residues, shields the C-terminal helix from solvent and is involved in inter-subunit contacts. The side-chain of Arg10 in monomer A makes hydrogen bonds to the side-chains of Gln94 in monomer E and Tyr198 in monomer F, and points to the carboxy end of the C-terminal helix dipole. The region of the GDH molecule covered by the N-terminal extension is also shielded in C. symbiosum by (b) a short helical stretch from residues Val35 to Leu48 and (c) by the extension of two helices in B. taurus between residues Val24 and Ile48.

gel-filtration (data not shown). According to these data, the unique extension does not appear to be essential for catalysis and subunit assembly, and presumably fulfils some other yet unknown function.

Subunit–subunit interactions As described above, P. falciparum GDH forms a homohexamer with 32 symmetry. Each GDH monomer interacts in the overall packing with

604

Figure 6. Two examples for the charge-assisted network of hydrogen bonds in the subunit–subunit interface of P. falciparum GDH.

four subunits through 14 charge-assisted subunit– subunit contacts (Figure 6). Some of these cluster in small networks involving salt-bridges. Elaborate networks of charge-assisted hydrogen bonds between subunits have previously been observed in hexameric GDHs of hyperthermophilic species such as P. furiosus and T. litoralis, where they are believed to contribute to their enhanced thermostability.24,30,31 The subunit–subunit interactions in P. falciparum GDH differ significantly from those in the human enzyme. The most obvious difference results from the absence of the so-called antenna region in the parasite GDH. This area, exclusive to the allosterically regulated mammalian enzymes,10,32 results from a 48 amino acid residue insertion starting at residue 399 in human GDH26 and generates an enhanced interface between three adjacent subunits through clustering of the antenna regions along the local 3-fold axis of the hexamer. In addition, the inter-subunit contacts differ in regions where the two enzymes share a common fold. In contrast to the parasite GDH, the human enzyme exhibits only four residues at each monomer that are involved in charge-assisted interactions with a neighbouring subunit; no ionic hydrogen-bond networks are observed. Accordingly, only two out of the 13 charged residues (excluding the C-terminal carboxylate of Phe470) in

Crystal Structure of Glutamate Dehydrogenase

P. falciparum GDH are also present in the human enzyme. In consequence, the latter protein uses more non-charge-assisted contacts. This suggests that the allosterically regulated human enzyme may require less tightly bound monomers to achieve efficient inter-subunit communication. The numerous salt-bridges formed between the subunits in the parasite enzyme most likely explain its high thermostability: the enzyme resists denaturation when exposed for ten minutes to a temperature of 80 8C.17 Finally, the buried surface area (AD, determined as A monomer KA dimer/2 using the algorithm implemented in the program msms33) for each subunit along the 2-fold axis is larger in the plasmodial enzyme than in the human protein. AD ˚ 2 for the P. falciparum GDH and amounts to 1840 A 2 ˚ for the human enzyme, reflecting 9.5% 1560 A or 7% of the total surface of the respective monomers. The area AT buried about the 3-fold axis (AmonomerKAtrimer/3) is considerably larger in the human enzyme compared to the parasitic GDH due to the inserted antenna region. AT accounts ˚ 2 in the ˚ 2 in the human and 2320 A for 3500 A P. falciparum GDH. Interestingly, the buried dimerisation surface (AD) is also larger in the parasitic GDH compared to the bacterial C. symbiosum ˚ 2).27 This probably results from enzyme (1300 A the shorter N terminus in the bacterial protein. The trimer interface has about the same size in the ˚ 2 versus plasmodial and bacterial enzyme (2320 A 2 ˚ 2400 A ). Implications for drug design Apart from the large insertion of the antenna region in human GDH, the deviating nature of the subunit–subunit contacts is apparently the major structural difference between human and P. falciparum GDH. This difference in the parasite GDH subunit interface may be a target for exploitation for the development of specific drugs. Peptides and peptidomimetics have proven effective in disrupting protein subunit interfaces.34–37 Such compounds could be of interest as potential drug candidates. Peptidomimetics exhibiting negative charges could interact specifically with the positively charged Arg residues at the P. falciparum GDH subunit interface (Figure 5). This could be of particular relevance, since parasitised red blood cells are known to exhibit permeability for negatively charged compounds.38 In summary, the substrate-binding sites are highly conserved among human and plasmodial GDH. Although both structures have been crystallised in the uncomplexed state, their close relationship to structures determined in the presence of substrate and cofactor allow prediction of the putative closely related substrate binding mode. However, to fully exploit these differences for drug design it will be necessary to determine the structure of the parasitic and human GDH in

605

Crystal Structure of Glutamate Dehydrogenase

complex with glutamate or analogue and/or cofactor.

Materials and Methods

Table 2. Primers used for site-directed mutagenesis Mutagenesis primer (sense) CACAGAATTCATTAA AGAGGAGAAATTAAC ATGTATGAATCTTTA GTCGATCAAGAAATG AATAACG Mutagenesis primer (antisense) CGTTATTCATTTCTT GATCGACTAAAGATT CATACATGTTAATTT CTCCTCTTTAATGAA TTCTGTG

Expression, purification and crystallisation Plasmodium falciparum GDH was over-expressed and purified as described.14 A gel-filtration step on a Superdex 200 column was included to improve protein homogeneity. Plate-like crystals of monoclinic space group C2 were grown by vapour-diffusion using the sitting-drop method. Each 10 ml drop contained 5 ml of a 10 mg/ml P. falciparum GDH solution, which was dialysed against a 1 mM EDTA solution (pH 7) prior to crystallisation, and 5 ml of reservoir solution. The latter consisted of 0.1 M Tris–HCl buffer (pH 8.0) containing 30% (w/v) PEG 1000. Crystals appeared after two months at room temperature. A second cubic crystal form of space group I23 was obtained by vapour-diffusion in either a sitting or hanging-drop apparatus. The largest crystals grew from 10 ml drops composed of 5 ml of a 30% PEG 2000 containing reservoir solution which contained no buffer, and 5 ml of a 5 mg/ml protein solution in 1 mM EDTA (pH 7.0), 1 mM b-mercaptoethanol. The crystals grew to a size of 0.7 mm!0.7 mm!0.7 mm; however, their diffrac˚ resolution and no tion power was limited to about 6 A appropriate conditions to perform measurements under cryo-conditions could be established. A deletion mutant of P. falciparum GDH lacking the first 19 amino acid residues was produced by site-directed mutagenesis. A 67 bp sense and an anti-sense primer were produced, designed such that a 57 bp deletion formed upon annealing to the coding region of the pQE60/gdh plasmid and subsequent PCR reaction (Table 2). The site-directed mutagenesis was carried out using the Stratagene Quick Changew Kit. The product of the PCR reaction (i.e. the pQE60/gdh plasmid shortened by 57 bp) was then transformed into XL-10 goldw supercompetent cells and transferred onto agar plates after one hour incubation at 37 8C to be incubated overnight. Ten clones were picked, and each incubated in 5 ml of LB medium. The plasmid was then isolated and purified with the Qiagen Mini Prep-Kit and sequenced in both directions of the gene. The purified pQE/gdhD19 plasmids of four of the ten clones were subsequently transformed into competent E. coli PA340 cells, which are glutamate synthase and dehydrogenase-deficient. The mutant GDH was expressed as described for the wildtype protein14 and the recombinant E. coli cells were then subjected to sonication. The supernatant from IPTGinduced cells clearly showed GDH activity in contrast to the control supernatant from cells where expression was not induced. GDH activity was measured by following NADPH formation in the forward reaction: L-glutamate

þ NADðPÞþ þ H2 O /

þ 2-oxoglutarate þ NHþ 4 þ NADðPÞH þ H

at a wavelength of 340 nm in a buffer containing 100 mM KH2PO4 and 1 mM EDTA at pH 8.0 and at room temperature. Data collection and refinement A data set was recorded at the beamline BW6, DESY Synchrotron Hamburg, on a Mar Research CCD system.

Table 3. Data set statistics Cell constants ˚) a, b, c (A a, b, g (deg.) Space group ˚) Resolution limit (A Total number of reflections Number of unique reflections Rmerge overall (%) ˚ ) (%) Rmerge (2.75–2.7 A Completeness overall (%) ˚ ) (%) Completeness (2.75–2.7 A Number of frames recorded

167.6, 96.9, 196.2 90, 91.7, 90 C2 2.4 100,593 82,010 12 40 93.3 94.7 1800

For data collection, plate-like crystals of a size of about 0.3 mm!0.3 mm!0.05 mm were chosen. Prior to freezing in liquid nitrogen, crystals were soaked for a few seconds in a pH 8 cryobuffer containing 45% PEG 2000. Due to its high mosaicity, the crystal had to be rotated in the beam in small steps of 0.18 per frame in order to reduce the number of overlapping reflections and to achieve sufficient completeness of the data collection. This strategy resulted in a large data set containing 1800 frames to cover reciprocal space across 1808. The intensities were integrated using the program DENZO39 and scaled with the SCALEPACK40 facility. The crystal exhibited monoclinic symmetry in space ˚ , bZ group C2 with unit cell dimensions of aZ167.6 A ˚ , cZ196.2 A ˚ , bZ91.78. The crystal diffracted up to 96.9 A ˚ but reflection statistics suggested that data should 2.4 A ˚ resolution. Averaging over be evaluated only up to 2.70 A symmetry-related reflections resulted in an Rmerge of 0.40 for the highest resolution shell and 0.12 for the whole data set (Table 3). Solution and refinement The structure was solved by the method of molecular replacement using the structure of C. symbiosum GDH as phasing model (pdb code 1bgv). Crystal packing density calculations suggested that the asymmetric unit should be composed of a hexamer. Accordingly, the core of the hexamer (residues 1–192) of C. symbiosum GDH was used for phase determination. To find the correct rotational and translational operations and to perform the subsequent structural refinement, the routines implemented in the CNSsolve program suite were applied.40 After application of the obtained rotation and translation function, the solution with the highest correlation coefficient was subjected to rigid-body refinement. In this step, each monomer was allowed to refine independently. After rigid-body refinement, the Rwork and the Rfree dropped to 0.477 and 0.489, respectively. The programs O and Moloc were used for visualisation and manual rebuilding of the model.41,42 After the first round of rebuilding using 2FoKFc and FoKFc electron density maps, a simulated annealing refinement without applying non-crystallographic symmetry (NCS) constraints

606

Crystal Structure of Glutamate Dehydrogenase

Table 4. Refinement statistics ˚) Refinement resolution range (A I/s cutoff ˚) RMS bond error (A RMS angular error (deg.) Model Rwork (%) Model Rfree (%) Ramachandran % Favoured % Additional % Generous % Disallowed

100–2.7 3.0 0.006 1.23 23.3 27.6 85.2 13.3 1.2 0.2

was carried out. The Rwork dropped further to 0.40, however Rfree increased to 0.493. This indicated that the individual monomers of the hexamer should, at least in the beginning, be set to be identical. In the next refinement step, NCS constraints were applied and the Rfree value consequently dropped to 0.43. After six further cycles of simulated annealing with NCS constraints and group B-factor refinement Rwork and Rfree dropped to 0.29 and 0.30, respectively. After manual rebuilding of the model, the next refinement cycle was performed with tight NCS restraints instead of constraints and individual B-factors of the residues were calculated. SA refinement was no longer applied beyond this stage. This cycle of refinement further reduced Rwork and Rfree. Finally, the last refinement cycle including all the reflections, converged to the R-factor reported in Table 4. Figures Figures presented in this contribution were produced using PyMOL (http://pymol.sourceforge.net). Protein Data Bank accession code The coordinates of the structure have been deposited with the RCSB Protein Data Bank under accession number 2bma.

Acknowledgements The authors thank Drs Gleb Bourenkov and Hans Bartunik of the Max-Planck-Gesellschaft for providing beam-time at Beamline BW 6 of DESY Hamburg, and the Graduiertenkolleg “Protein function at atomic level” for financial support.

References 1. Hunt, N. H. & Stocker, R. (1990). Oxidative stress and the redox status of malaria-infected erythrocytes. Blood Cells, 16, 499–526. 2. Schirmer, R. H., Mu¨ller, J. G. & Krauth-Siegel, R. L. (1995). Disulfide reductase inhibitors as chemotherapeutic agents: the design of drugs for trypanosomiasis and malaria. Angew. Chem. Int. Ed. Engl. 34, 141–154. 3. Rahlfs, S. & Schirmer, R. H. (2002). The thioredoxin system of Plasmodium falciparum and other parasites. Cell. Mol. Life Sci. 59, 1024–1041.

4. Fa¨rber, P. M., Becker, K., Mu¨ller, S., Schirmer, R. H. & Franklin, R. M. (1996). Molecular cloning and characterization of a putative glutathione reductase gene, the PfGR2 gene, from Plasmodium falciparum. Eur. J. Biochem. 239, 655–661. 5. Kanzok, S. M., Schirmer, R. H., Turbachova, I., Iozef, R. & Becker, K. (2000). The thioredoxin system of the malaria parasite Plasmodium falciparum. Glutathione reduction revisited. J. Biol. Chem. 275, 40180–40186. 6. Krnajski, Z., Gilberger, T. W., Walter, R. D., Cowman, A. F. & Muller, S. (2002). Thioredoxin reductase is essential for the survival of Plasmodium falciparum erythrocytic stages. J. Biol. Chem. 277, 25970–25975. 7. Clarke, J. L., Scopes, D. A., Sodeinde, O. & Mason, P. J. (2001). Glucose-6-phosphate dehydrogenase-6-phosphogluconolactonase. A novel bifunctional enzyme in malaria parasites. Eur. J. Biochem. 268, 2013–2019. 8. Calvo, E., Rubiano, C., Vargas, A. & Wassermann, M. (2002). Expression of housekeeping genes during the asexual cell cycle of Plasmodium falciparum. Parasitol. Res. 88, 267–271. 9. Atamna, H., Pascarmona, G. & Ginsburg, H. (1994). Hexose-monophosphate shunt activity in intact Plasmodium falciparum infected erythrocytes and in free parasites. Mol. Biochem. Parasitol. 67, 79–89. 10. Peterson, P. E. & Smith, T. J. (1999). The structure of bovine glutamate dehydrogenase provides insights into the mechanism of allostery. Struct. Fold Des. 7, 769–782. 11. Walter, R. D., Nordmeyer, J.-P. & Ko¨nigk, E. (1974). NADP specific glutamate dehydrogenase from Plasmodium chabaudi. Hoppe-Seyler’s Z. Physiol. Chem. 355, 495–500. 12. Roth, E. F. J. (1987). Malarial parasite hexokinase and hexokinase-dependent glutathione reduction in the Plasmodium falciparum-infected human erythrocyte. J. Biol. Chem. 262, 15678–15682. 13. Roth, E., Jr (1990). Plasmodium falciparum carbohydrate metabolism: a connection between host cell and parasite. Blood Cells, 16, 453–460. 14. Wagner, J. T., Ludemann, H., Farber, P. M., Lottspeich, F. & Krauth-Siegel, R. L. (1998). Glutamate dehydrogenase, the marker protein of Plasmodium falciparum: cloning, expression and characterization of the malarial enzyme. Eur. J. Biochem. 258, 813–819. 15. Vander Jagt, D. L., Intress, C., Heidrich, J. E., Mrema, J. E. K., Rieckmann, K. H. & Heidrich, H. G. (1982). Marker enzymes of Plasmodium falciparum and human erythrocytes as indicators of parasite purity. J. Parasitol. 68, 1068–1071. 16. Vander Jagt, D. L., Hunsaker, L. A., Kibirige, M. & Campos, N. M. (1989). NADPH production by the malarial parasite Plasmodium falciparum. Blood, 74, 471–474. 17. Krauth-Siegel, R. L., Muller, J. G., Lottspeich, F. & Schirmer, R. H. (1996). Glutathione reductase and glutamate dehydrogenase of Plasmodium falciparum, the causative agent of tropical malaria. Eur. J. Biochem. 235, 345–350. 18. Amuro, N., Yamaura, M., Goto, Y. & Okazaki, T. (1988). Molecular cloning and nucleotide sequence of the cDNA for human liver glutamate dehydrogenase precursor. Biochem. Biophys. Res. Commun. 152, 1395–1400. 19. Britton, K. L., Baker, P. J., Rice, D. W. & Stillman, T. J. (1992). Structural relationship between the hexameric and tetrameric family of glutamate dehydrogenases. Eur. J. Biochem. 209, 851–859.

Crystal Structure of Glutamate Dehydrogenase

20. Stillman, T. J., Baker, P. J., Britton, K. L. & Rice, D. W. (1993). Conformational flexibility in glutamate dehydrogenase. Role of water in substrate recognition and catalysis. J. Mol. Biol. 234, 1131–1139. 21. Rice, D. W., Baker, P. J., Farrants, G. W. & Hornby, D. P. (1987). The crystal structure of glutamate dehydrogenase from Clostridium symbiosum at 0.6 nm resolution. Biochem. J. 242, 789–795. 22. Yip, K. S., Stillman, T. J., Britton, K. L., Artymiuk, P. J., Baker, P. J., Sedelnikova, S. E. et al. (1995). The structure of Pyrococcus furiosus glutamate dehydrogenase reveals a key role for ion-pair networks in maintaining enzyme stability at extreme temperatures. Structure, 3, 1147–1158. 23. Knapp, S., de Vos, W. M., Rice, D. & Ladenstein, R. (1997). Crystal structure of glutamate dehydrogenase from the hyperthermophilic eubacterium Thermotoga ˚ resolution. J. Mol. Biol. 267, 916–932. maritima at 3.0 A 24. Britton, K. L., Yip, K. S., Sedelnikova, S. E., Stillman, T. J., Adams, M. W., Ma, K. et al. (1999). Structure determination of the glutamate dehydrogenase from the hyperthermophile Thermococcus litoralis and its comparison with that from Pyrococcus furiosus. J. Mol. Biol. 293, 1121–1132. 25. Nakasako, M., Adachi, F. T., Kudo, S. & Higuchi, S. (2001). Large-scale domain movements and hydration structure changes in the active-site cleft of unligated glutamate dehydrogenase from Thermococcus profundus studied by cryogenic X-ray crystal structure analysis and small-angle X-ray scattering. Biochemistry, 40, 3069. 26. Smith, T. J., Schmidt, T., Fang, J., Wu, J., Siuzdak, G. & Stanley, C. A. (2002). The structure of apo human glutamate dehydrogenase details subunit communication and allostery. J. Mol. Biol. 318, 765–777. 27. Baker, P. J., Britton, K. L., Engel, P. C., Farrants, G. W., Lilley, K. S., Rice, D. W. & Stillman, T. J. (1992). Subunit assembly and active site location in the structure of glutamate dehydrogenase. Proteins: Struct. Funct. Genet. 12, 75–86. 28. Dean, J. L., Wang, X. G., Teller, J. K., Waugh, M. L., Britton, K. L., Baker, P. J. et al. (1994). The catalytic role of aspartate in the active site of glutamate dehydrogenase. Biochem. J. 301, 13–16. 29. McPherson, M. J. & Wootton, J. C. (1983). Complete nucleotide sequence of the Escherichia coli gdhA gene. Nucl. Acids Res. 11, 5257–5266. 30. Vetriani, C., Maeder, D. L., Tolliday, N., Yip, K. S., Stillman, T. J., Britton, K. L. et al. (1998). Protein thermostability above 100 8C: a key role for ionic interactions. Proc. Natl Acad. Sci. USA, 95, 12300–12305.

607 31. Britton, K. L., Baker, P. J., Borges, K. M., Engel, P. C., Pasquo, A., Rice, D. W. et al. (1995). Insights into thermal stability from a comparison of the glutamate dehydrogenases from Pyrococcus furiosus and Thermococcus litoralis. Eur. J. Biochem. 229, 688–695. 32. Smith, T. J., Peterson, P. E., Schmidt, T., Fang, J. & Stanley, C. A. (2001). Structures of bovine glutamate dehydrogenase complexes elucidate the mechanism of purine regulation. J. Mol. Biol. 307, 707–720. 33. Sanner, M. F., Olson, A. J. & Spehner, J. C. (1996). Reduced surface: an efficient way to compute molecular surfaces. Biopolymers, 38, 305–320. 34. Berezov, A., Chen, J., Liu, Q., Zhang, H. T., Greene, M. I. & Murali, R. (2002). Disabling receptor ensembles with rationally designed interface peptidomimetics. J. Biol. Chem. 277, 28330–28339. 35. Vassilev, L. T., Graves, B., Carvajal, D., Podlaski, F., Filipovic, Z., Kong, N. et al. (2004). In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science, 303, 844–848. 36. Oost, T. K., Sun, C., Armstrong, R. C., Al-Assaad, A. S., Betz, S. F., Deckwerth, T. L. et al. (2004). Discovery of potent antagonists of the antiapoptotic protein XIAP for the treatment of cancer. J. Med. Chem. 47, 4417–4426. 37. Loregian, A., Marsden, H. S. & Palu, G. (2002). Protein–protein interactions as targets for antiviral chemotherapy. Rev. Med. Virol. 12, 239–262. 38. Kutner, S., Baruch, D., Ginsburg, H. & Cabantchik, Z. I. (1982). Alterations in membrane permeability of malaria-infected human erythrocytes are related to the growth stage of the parasite. Biochim. Biophys. Acta, 687, 113–117. 39. Otwinowski, Z. (1993). Oscillation data reduction program. In Data Collection and Processing, pp. 56–62, Daresbury Laboratory, Warrington, UK. 40. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921. 41. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. (1991). Improved methods for binding protein models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110–119. 42. Gerber, P. R. & Mu¨ller, K. (1995). MAB, a generally applicable molecular force field for structure modelling in medicinal chemistry. J. Comput. Aided Mol. Des. 9, 251–268.

Edited by R. Huber (Received 26 November 2004; received in revised form 7 March 2005; accepted 9 March 2005)