Crystal Structures of a Schistosomal Drug and

JMB—MS 321 Cust. Ref. No. PEW 166/94

[SGML] J. Mol. Biol. (1995) 246, 21–27

COMMUNICATION

Crystal Structures of a Schistosomal Drug and Vaccine Target: Glutathione S-Transferase from Schistosoma japonica and its Complex with the Leading Antischistosomal Drug Praziquantel Michele A. McTigue, DeWight R. Williams and John A. Tainer* The Scripps Research Institute, Department of Molecular Biology-MB4 10666 North Torrey Pines Road, La Jolla, CA 92037 U.S.A.

*Corresponding author

Glutathione S-transferase (GST), an essential detoxification enzyme in parasitic helminths, is a major vaccine target and an attractive drug target against schistosomiasis and other helminthic diseases. Crystal structures of the 26 kDa GST from the helminth Schistosoma japonica (SjGST) have been determined for the unligated enzyme (resolution = 2.4 Å, R-factor = 19.7%) and for the enzyme bound to the leading antischistosomal drug praziquantel (resolution = 2.6 Å, R-factor = 21.2%). The protein, recombinantly expressed using the Pharamacia PGEX-3X vector for production of GST fusion proteins, contains all 218 residues of SjGST and an additional 13 residues at the C terminus. The structure of unligated SjGST shows that the glutathione binding site pre-exists unchanged in the ligand-free enzyme and is conserved between parasitic and the mammalian class m enzymes. At therapeutic concentrations the leading antischistosomal drug praziquantel (PZQ) binds one drug per enzyme homodimer in the dimer interface groove adjoining the two catalytic sites. This establishes a protein target for PZQ, identifies the GST non-substrate ligand transport site, and implicates PZQ in steric inhibition of SjGST catalytic and transport for large ligands. Thus, increased expression or mutagenesis of SjGST by the parasite may confer resistance to PZQ. Differences in the xenobiotic binding region between parasitic and mammalian GSTs reveal a distinct substrate repertoire for SjGST and, together with the newly identified PZQ binding site, provide the basis for design of novel antischistosomal drugs. Due to the widespread use expression systems based on SjGST fusions, the atomic structure of SjGST should also provide an important tool for phasing fusion protein structures by molecular replacement. Keywords: schistosomiasis; helminth; X-ray structure; drug design; glutathione S-transferase

Schistosomiasis, a parasitic disease infecting more than 275 million people worldwide and resulting in 200 thousand deaths annually, is caused by helminth worms of the genus Schistosoma, which live in the blood vessels of mammalian hosts (Lucey & Maguir, 1993). Host immune response, mainly to worm eggs, causes the formation of granulomata, resulting in severe tissue damage (Shekhar, 1991). The predomiPresent address: D. R. Williams, University of Hawaii, Department of Zoology, 2538 The Mall, Edmondson Hall 465, Honolulu, HI 96822, U.S.A. Abbreviations used: GSH, glutathione; GST, glutathione S-transferase; SjGST, GST from Schistosoma japonica; r.m.s., root-mean-square; PZQ, praziquantel. 0022–2836/95/060021–07 $08.00/0

nant species infecting humans (S. mansoni, S. haematobium and S. japonicum) differ widely in tissue localization and clinical manifestations (Lucey & Maguir, 1993). As an effective drug without serious side-effects, praziquantel (PZQ) is the single major treatment for schistosomiasis and several other trematode and cestode infections (Andrews et al., 1983), although its target and mechanism of activity are not known (Shekhar, 1991). Recent reports of schistosome strains resistant to PZQ and other antihelminths (Katz et al., 1994; Coli et al., 1993; Hayes & Wolf, 1990) indicate that new therapies must be identified to fight this debilitating disease successfully, which remains a major global health problem despite concerted efforts in endemic regions. 7 1995 Academic Press Limited

JMB—MS 321 22 The detoxification enzyme glutathione S-transferase (GST), which catalyzes the nucleophilic addition of the tripeptide glutathione (GSH) to endogenous and xenobiotic electrophilic toxins, is an important target for the development of antischistosomal vaccines and drugs (Smith et al., 1986; Sexton et al., 1990; Brophy & Barrett, 1990; Balloul et al., 1987; Davern et al., 1990). As helminths contain very low levels of other detoxification enzymes, such as catalase, superoxide dismutase and cytochrome P450, GST may provide the worm’s primary defense against electrophilic and oxidative damage (Brophy & Barrett, 1990). Helminths contain two distinct classes of GSTs of Mr 26,000 and 28,000 with differing substrate specificity (Walker et al., 1993). These GSTs are believed to prevent toxin accumulation in the worm by catalyzing the conjugation of glutathione to lipid peroxidation products and other toxins (Walker et al., 1993), and by facilitating export of toxic host hydrophobic metabolites such as heme (Davern et al., 1990). These helminth GSTs show promising vaccine potential: GST injections induce partial (52 to 67%) protection against infection by S. japonica and S. mansoni in rodents (Balloul et al., 1987; Sher et al., 1989; Davern et al., 1990; Smith et al., 1986) and by Fasciola hepatica in sheep (Sexton et al., 1990). Here we present the crystal structure of a ligand-free, recombinant, 26 kDa GST from S. japonica (SjGST) and compare it to structures for ligand-bound forms of human a (Sinning et al., 1993), pig p (Reinemer et al., 1992), and rat m (Ji et al., 1992) class GSTs. Whereas the glutathione binding site pre-exists in the ligand-free enzyme and is conserved between parasitic and mammalian enzymes, the xenobiotic site is distinct in the parasitic SjGST, thus providing a structural basis for selective drug design. As a step toward understanding the action of the major schistosomal drug PZQ and toward the design of improved drugs, we report ˚ structure of SjGST complexed with PZQ, the 2.6 A show that PZQ binds to SjGST dimers at therapeutic concentrations, and identify the GST ‘‘non-substrate’’ hydrophobic ligand transport binding site as a channel linking the two xenobiotic sites in the dimer.

SjGST protein structure The SjGST crystal structure was determined by multiple isomorphous replacement (MIR) and molecular replacement methods, and refined to a ˚ resolution (Table 1). The R-factor of 19.7% at 2.4 A refined model has continuous main-chain electron density, with almost all carbonyl oxygen and side-chain positions well-defined, and clearly agrees with electron density omit maps, calculated to remove model bias (Figure 1). The 218 residues of each SjGST subunit fold to form a small N-terminal a/b domain (residues 1 to 76), a short linker region (residues 77 to 84), and a larger C-terminal a domain (residues 85 to 218). The N-terminal domain contains three a helices and a

Communication

four-stranded, predominantly antiparallel b-sheet (Figure 2). The C-terminal domain contains five a-helices and a long coil (residues 195 to 218) after the final helix (a8). The open hairpin loop formed by the last eight amino acids (residues 211 to 218, Gly-Gly-Gly-Asp-His-cisPro-Pro-Lys) is not present in the known structures of mammalian GSTs, and is selectively recognized by mouse monoclonal antibodies raised against SjGST. The surface exposure of this epitope should allow selective recognition by host antibodies. The overall fold of SjGST otherwise resembles mammalian GSTs, as shown by least squares superposition of the 117 Ca positions from a helices and b strands of SjGST onto those rat m, pig p, and human a GSTs, yielding respective r.m.s. ˚ . The areas of deviations of 1.35, of 1.35 and 1.75 A greatest structural difference between SjGST and the mammalian enzymes occur at residues 33 to 41 and 200 to 218, which form part of the xenobiotic substrate binding site (discussed below). Two SjGST subunits assemble to form a dimer of ˚ × 47 A ˚ × 44 A ˚ (Figure 3). The 2-fold dimensions 57 A rotation axis relating the subunits within the dimer lies approximately 11 degrees from the helical axis of a4, such that the symmetry-related a4 helices cross at an angle of approximately 22°. Residues involved in intersubunit contacts include: 50 to 53 (b-turn), 63 to 66 (b4), 67 to 76 (a3), 77 to 85 (loop), 88 to 109 (a4), and 129 to 136 (a5). The dimer contact has two salt bridges (Asp77–Arg89 and Glu51–Arg136). The most prominent hydrophobic contact is the penetration of the side-chain of Phe52 of one subunit between residues 91 to 94 (a4) and 129 to 133 (a5) of the other subunit. This dimer assembly creates a long ˚ ), narrow (approximately 6 to 10 A ˚ ) cleft lined (40 A by primarily by Ser, Gln, Tyr, Asp and Arg polar side-chains but also including Leu100 and Met69.

Praziquantel binding to SjGST To test the possibility that SjGST is a protein target for the major schistosomal drug PZQ (2-(cyclohexylcarbonyl)-1,2,3,6,7-11b-hexahydro-4H-pyrazino[2,1a]isoquinolin-4-one), SjGST was co-crystallized with 13 mM PZQ, a drug concentration within the range found in host sera and worms following treatment ˚ refined (Shu-hua et al., 1991). In the 2.6 A SjGST–drug complex crystal structure (Table 1) one drug binds per enzyme homodimer in the dimer interface groove adjoining the two catalytic sites (Figures 3, 4). The 2-fold rotation axis relating the protein subunits intersects the PZQ binding site such that PZQ may occupy either, but not both, of two, overlapping symmetry-related binding sites. PZQ makes contacts Gln67, Gly97, Leu100, Asp101, Tyr104 and Arg108 side-chains of a4 from one subunit and also Tyr104 of the second subunit (Figure 4). This PZQ binding mode appears unique to SjGST as mammalian GSTs of known structure lack tyrosine at position 104, the residue with most contact to PZQ. The equivalent residues in mammalian GST structures (Cys in human aGST,

JMB—MS 321 23

Communication

Table 1 Structure determination Data set Native Trimethyl lead acetate Uranyl acetate K2 Pt(CN)4 Praziquantel

Resolution ˚) (A

Completeness (%)

Unique reflections (no.)

Redundancy

Rsymm † (%)

Sites (no.)

Rmerge ‡ (%)

Rcullis § (%)

Phasing power>

2.4 3.8 3.0 2.9 2.6

94 93 96 98 99

11072 3469 6535 7295 9611

2.9 3.4 4.5 4.7 3.5

10.3 12.6 9.2 7.5 6.6

— 3 3 3 —

— 17.0 26.5 16.1 —

— 49.6 70.2 57.7 —

— 1.8 1.5 1.7 —

Ligand-free SjGST: Recombinant 26 kDa glutathione S-transferase from S. japonica (SjGST) was expressed as recombinant product of the PGEX-3X Pharmacia expression vector and was purified without the use of a glutathione-affinity column (McTigue et al., 1995). Crystals were obtained by sitting drop vapor diffusion using ammonium sulfate as the precipitant at pH 5.6 (McTigue et al., 1995). The ˚ and c = 70.2 A ˚ , and contain one SjGST crystals belong to the hexagonal space group P63 22 with unit cell parameters a = b = 125.2 A monomer per asymmetric unit resulting in an approximate solvent content of 59.6% (Matthews, 1968). Native X-ray diffraction data to ˚ resolution were collected from one crystal cooled to 10°C at the Stanford Synchrotron Radiation Laboratories beamline 7-1 with 2.4 A ˚ . Images were collected as 1° degree rotations about the crystallographic a MAR Research image plate detector, at a wavelength of 1.08 A 6-fold axis and the data were processed with MOSFLM (Leslie et al., 1986). Initial phases were determined by molecular replacement ˚ resolution structure of rat m GST (Ji et al., 1992) as the with the X-PLOR 3.1 program package (Brunger, 1992) for subunit A of the 2.2 A probe. Rigid body refinement followed by simulated annealing (X-PLOR) resulted in an R-factor of 31.4% for all data with Fo > 2s in ˚ . After several cycles of model building to omit maps and positional refinement, the density for areas the resolution range 10 to 3.0 A of greatest sequence variation between SjGST and rat m GST (residues 1 to 2, 33 to 41, and 200 to 231) remained uninterpretable. The phases were improved by including phases from the three heavy-atom derivatives. Heavy-atom positions were located with isomorphous difference Patterson syntheses and cross-phased difference Fourier maps. Refinement of derivative parameters and phase calculations were done with X-heavy (McRee, 1992). Heavy-atom and partial model phases were combined using SIGMAA (Read, 1986). The resulting electron density map allowed all GST residues to be fit, except for the disordered C-terminal extension present in the recombinant protein ˚ with iterative cycles of positional and B-factor but absent in natural SjGST (residues 219 to 231). The resolution was extended to 2.4 A refinement followed by fitting to omit maps. The final model contains 1786 protein atoms and 117 water molecules, resulting in an R-factor ˚ . The r.m.s. deviation from ideality of bonds and angles of 19.7% for 10,832 reflections with Fo > 2s in the resolution range 7.0 to 2.4 A ˚ and 4.1°, respectively. The average B-value for protein atoms is 30.4 A ˚ 2. The SjGST model does not include the side-chain atoms is 0.024 A beyond Cb of Leu118, His215, Lys218 or any atoms in residues 219 to 231. SjGST-PZQ complex: Crystals were obtained by cocrystallization of SjGST with 13 mM PZQ (Sigma Chemical Co.), and X-ray data were ˚ and c = 70.2 A ˚ . The coordinates collected as for the native ligand-free structure. The unit cell derived from MOSFLM was a = b = 123.7 A for ligand-free SjGST (with water molecules occupying the ligand binding sites removed) were positionally refined against the PZQ complex data and a difference Fourier map was calculated. This map contained a large segment of continuous electron density (2 to 3s) near the side-chain of Y104 at the dimer interface. A model for PZQ was generated using INSIGHT (Biosym Tech.) by substituting a cyclohexyl group for the nitrobenzyl group of the crystal structure of 2-(p-nitrobenzoyl)-1,2,3,6,7,11b-hexahydro-4H-pyrazino(2,1a)isoquinolin-4-one (Toscano et al., 1992). The best fit of PZQ to this density placed the 3-membered ring across the crystallographic 2-fold axis that generates the dimer. The occupancy of PZQ was therefore set to 0.5 and refinement continued as for the ligand-free structure. The final model contains 1761 protein atoms, 23 PZQ atoms, and only 83 water molecules, resulting in an R-factor of 21.2% for 9494 ˚ . The r.m.s. deviations from ideality of bonds and angles are 0.015 A ˚ and 1.8°, reflections with Fo > 2s in the resolution range 7.0 to 2.6 A respectively. † Rsymm is the unweighted R-factor on intensities for multiple observations of symmetry-related reflections. ‡ Rmerge is the unweighted R-factor on structure amplitudes for all reflections in common with the native data. § Rcullis = S[=Fh = − (=Fph = − =Fp =)]/S=Fh =. > Phasing power is the ratio of the r.m.s. calculated heavy-atom structure amplitude to the r.m.s. lack of closure.

Glu in pig pGST, and Met in rat mGST) make bad steric contacts with PZQ in its bound position in the ˚ from SjGST structure. The PZQ binding site is 9.6 A the hydroxyl group of Tyr7, the conserved residue

which hydrogen bonds to and is believed to activate glutathione thiol for attack on electrophilic substrates (Wilce & Parker, 1994). Therefore, SjGST is unlikely to catalyze the conjugation of glutathione to PZQ and

Figure 1. Stereo view of electron density (omit map) near the catalytic Tyr7 for the final refined model of the ligand-free ˚ resolution and contoured at 2.5s. Clockwise from the top, amino acid residues are Trp8 SjGST structure calculated at 2.4 A (edge on), Tyr7, Leu55, Asn54, Tyr57, His31 and Tyr33 with colored atom bonds (yellow, carbon; blue, nitrogen; red, oxygen). Water molecules are shown as red crosses.

JMB—MS 321 24

Communication

Figure 2. Subunit fold and catalytic Tyr7 position of SjGST. Helices are shown in orange, b strands in blue, and turns and loops in white. Residues corresponding to secondary structure assignments include: b1 (3-7), a1 (15-24), b2 (29-33), a2 (3844), b3 (57-60), b4 (63-66), a3 (67-76), a4 (88-109), a5a (115-122), a5b 126137), a6 (152-165), a7 (174-184) and a8 (189-194). The b strands are arranged with b1 and b2 parallel, b1 and b3 antiparallel, and b3 and b4 antiparallel. a5 contains a short break at residues 123 to 125. The position of the catalytic tyrosine (Tyr7) is shown as a ball and stick representation with carbon atoms in green and the hydroxyl oxygen atom in red. The amino and carboxy termini are labelled N and C, respectively.

more likely to bind PZQ in the previously unidentified binding site for non-substrates, also termed the transport binding site. Thus PZQ binds to this non-substrate binding site of SjGST at the in vivo therapeutic concentrations of PZQ. The binding of praziquantel to SjGST at therapeutic drug concentrations discovered here has important implications for drug action and resistance. First, PZQ is implicated in interference with SjGST transport functions. For example, SjGST apparently acts in the solubilization and subsequent elimination of heme, which results from the metabolism of host erythrocytes, to prevent heme oxidative damage and hematin crystal buildup in the parasite (Davern et al., 1990). The SjGST–PZQ complex structure shows that a heme can be accommodated into the PZQ binding site at the dimer interface with the symmetry-related Tyr104 side-chains packing against opposite faces of the heme. Thus, PZQ binding in the non-substrate

hydrophobic binding site should interfere with the binding and elimination of heme and other non-substrate toxic metabolites. Second, PZQ is implicated in interference with SjGST catalytic functions. Binding of PZQ into the channel joining the two xenobiotic substrate binding sites (Figure 5) could sterically block the binding of large substrates, such as peroxidized lipids. Superpositioning of the SjGST structure onto the structure of rat m GST complexed with 9-(S-glutathionyl)-10-hydroxy-9,10dihydrophenanthrene (GHDP; Ji et al., 1994) shows that the PZQ cyclohexyl group lies adjacent to the ˚ from position 1, 4.5 from A ˚ g-glutamate (3.5 A ˚ from position 2) and near the phenanthrene (7.1 A ˚ from position 2) of GHDP. Third, position 1, 7.5 A PZQ binding to SjGST may compete with PZQ binding to a distinct, as yet unknown, site of action. For each of these three possible effects resulting from the PZQ binding to SjGST identified here, increased expression or mutation of SjGST could confer PZQ

Figure 3. Stereo view of a ribbon representation of the SjGST dimer viewed down the crystallographic 2-fold axis which generates the dimer. SjGST subunits are shown in yellow and blue. The catalytic tyrosine residues (Tyr7; green) and the 2 binding conformations of praziquantel (orange, purple), which bind in the long groove at the dimer interface adjoining the 2 active sites, are shown in ball and stick representation.

JMB—MS 321 25

Communication

Figure 4. Stereo view of the structure of PZQ binding site. The 2-fold symmetry-related binding positions of PZQ (shown ˚ omit electron density difference (Fo − Fc ) map (contoured at in yellow and green) are shown superimposed with a 2.6 A 2.5s) generated from the model refined without PZQ. PZQ carbonyl oxygen and nitrogen atoms are shown in red and blue, respectively. The side-chains of residues Asp101 (D101), Tyr104 (Y104) and Arg108 (R108) of a4, which contact the PZQ, are also shown.

resistance. Increased GST activity has in fact been associated with drug resistance in helminths such as in cambendazole-resistant strains of Haemonchus contortus (Kawalek et al., 1994). GST is also associated with drug resistance in mammalian tumors, insects and plants (Hayes & Wolf, 1990; Ketterer et al., 1989). Comparison to mammalian GSTs Structural differences in the xenobiotic substrate binding site of SjGST (compared to mammalian a, p and m GSTs) suggest that SjGST belongs to a novel class of GST with a distinct substrate repertoire. Structurally, the GSH and xenobiotic binding sites of SjGST are most similar to those of rat m GST (Figure 5). Although the GSH binding site is unoccupied in the SjGST structure, the identity and position of all residues interacting with GSH in rat m GST are conserved in SjGST. The xenobiotic sites of the two structures, however, differ significantly. The most striking difference is an eight-residue insertion in the rat m GST sequence relative to SjGST that walls off one side of the binding site. In SjGST, this loop (between b5 and a2) is replaced by a b-turn,

increasing the solvent exposure of the xenobiotic site. In contrast, aromatic residues (Trp206 and Tyr104) partially block entrances to the xenobiotic binding site that are more open in the rat m GST. These structural differences in SjGST, compared to rat m GST and even larger differences compared to human a and pig p GSTs, bode well for the rational design of SjGST selective inhibitors as drugs. Conclusions Taken together, these results establish adjoining binding regions for glutathione, xenobiotic substrates and hydrophobic ligands in SjGST, with implications for design of new drugs that could prove effective against drug-resistant schistosomes. The structurally distinct xenobiotic site in the parasitic SjGST compared to mammalian GSTs provides a structural basis for selective drug design of SjGST substrate analog inhibitors. The SjGST–PZQ complex structure reveals a previously unknown protein target for PZQ and identifies a mechanism for this drug to act as a non-substrate GST inhibitor. PZQ binds in a channel joining the two xenobiotic

Figure 5. Stereo view of the xenobiotic binding site. The Ca backbone of the SjGST–PZQ complex (yellow) is shown ˚ structure of rat m GST (purple) complexed with least-squares superpositioned onto that of the 2.2 A 9-(S-glutathionyl)-10-hydroxy-9,10-dihydrophenanthrene (GHDP; white) (Ji et al., 1994). Side-chains (labelled) that form the xenobiotic binding site in the 2 structures (upper half) reveal much larger differences than those surrounding the GSH binding region (lower half). The 2 binding positions of PZQ are shown in blue and green.

JMB—MS 321 26 substrate binding sites of the dimer, suggesting that large xenobiotic and endogenous electrophilic substrates would extend into the PZQ binding region, although small substrates do not directly collide with bound PZQ. Thus, the SjGST and the SjGST–PZQ complex structures provide the basis for the design of novel drugs such as symmetric inhibitors that would better fill the identified 2-fold symmetric PZQ binding site without becoming conjugated to glutathione. Depending upon the primary mechanism of PZQ action, promising SjGST binding drugs could act directly by inhibiting SjGST’s catalytic ability to detoxify large electrophilic substrates or indirectly by preventing PZQ binding and removal by a SjGST transport mechanism.

Acknowledgements We thank Steve Reed, David Stout, Duncan McRee, Elizabeth Getzoff and Susan Bernstein for useful suggestions. This work supported by in part by NIH fellowship number AI08803 to M. A. M. Coordinates for all atoms of the ligand-free and praziquantel complex structures of SjGST have been deposited with the Brookhaven Protein Data Bank and have been assigned respective accession codes 16TA and 16TB.

References Andrews, P., Thomas, H., Pohlke, R. & Seubert, J. (1983). Praziquantel. Med. Res. Rev. 3, 147–200. Balloul, J. M., Sondermeyer, P., Dreyer, D., Capron, M., Grzych, J. M., Pierce, R. J., Carvallo, D., Lecocq, J. P. & Capron, A. (1987). Molecular cloning of a protective antigen of schistosomes. Nature (London), 326, 149–154. Brophy, P. M. & Barrett, J. (1990). Glutathione transferase in helminths. Parasitology. 100, 345–349. Bru¨nger, A. T. (1992). X-PLOR: Version 3.1, A System for X-ray Crystallography and NMR, Yale University Press, New Haven. Cioli, D., Pica-Mattoccia, L. & Archer, S. (1993). Drug resistance in schistosomes. Parasitol. Today, 9, 162–166. Davern, K. M., Tiu, W. U., Samaras, N., Gearing, D. P., Hall, B. E., Garcia, E. G. & Mitchell, G. F. (1990). Schistosoma japonicum: Monoclonal antibodies to the Mr 26,000 schistosome glutathione-S-transferase (sj26) in an assay for circulating antigen in infected individuals. Exp. Parasitol. 70, 293–304. Hayes, J. D. & Wolf, C. R. (1990). Molecular mechanisms of drug resistance. Biochem. J. 272, 281–295. Ji, X., Zhang, P., Armstrong, R. N. & Gilliland, G. L. (1992). The three-dimensional structure of a glutathione S-transferase from the Mu gene class. Structural analysis of the binary complex of ˚ . Biochemistry, isoenzyme 3-3 and glutathione at 2.2 A 31, 10169–10184. Ji, X., Johnson, W. W., Sesay, M. A., Dickert, L., Prasad, S. M., Ammon, H. L., Armstrong, R. A. & Gilliland, G. L. (1994). Structure and function of the xenobiotic substrate binding site of a glutathione S-transferase as revealed by x-ray crystallographic analysis of product

Communication

complexes of 9-(S-glutathionyl)-10-hydroxy-9,10-dihydrophenanthrene. Biochemistry, 33, 1043–1052. Katz, N., Rocha, R. S., DeSouza, C. P., Filho, P. C., Bruce, J. I., Coles, G. C. & Kinoti, G. K. (1994). Efficacy of alternating therapy with oxamniquine and praziquantel to treat Schistosoma mansoni in children following failure of first treatment. Amer. J. Trop. Med. Hyg. 44, 509–512. Kawalek, J. C., Rew, R. S. & Heavner, J. (1984). Glutathione-S-transferase, a possible drug-metabolizing enzyme, in Haemonchus contortus: Comparative activity of a cambendazole-resistant and a susceptible strain. Int. J. Parasitol. 14, 173–175. Ketterer, B., Meyer, D. J. & Clark, A. G. (1989). Soluble glutathione transferase enzymes. In Glutathione Conjugation and its Mechanism and Biological Significance (Sies, H. & Ketterer, B., eds). pp. 73–135, Academic Press, London. Leslie, A. G. W., Brick, P. & Wonacott, A. J. (1986). Mosflm, a package of programs for processing oscillation data. CCP4 Newsletter, 18, 33–39. Lucey, D. R. & Maguir, J. H. (1993). Schistosomiasis. Parasit. Dis. 7, 635–645. Matthews, B. W. (1968). Solvent content of protein crystals. J. Mol. Biol. 33, 491–497. McRee, D. E. (1992). A visual protein crystallographic software system for X11/Xview. J. Mol. Graph. 10, 44–46. McTigue, M. A., Williams, D. R., Bernstein, S. L. & Tainer, J. A. (1995). Purification and crystallization of a schistosomal glutathione S-transferase. Proteins, in the press. Read, R. J. (1986). Improved fourier coefficients for maps using phases from partial structures with errors. Acta Crystallogr. sect. A, 42, 140–149. Reinemer, P., Dirr, H. W., Ladenstein, R., Schaffer, J., Galley, O. & Huber, R. (1992). The three-dimensional structure of class pi glutathione-S-transferase in ˚ resolcomplex with glutathione sulfonate at 2.3 A ution. Biochemistry, 31, 10169–10184. Sexton, J. L., Milner, A. R., Panaccio, M., Waddington, J., Wijffels, G., Chandler, D., Thompson, C., Wilson, L., Spithill, T. W., Mitchell, G. F. & Campbell, N. J. (1990). Glutatione S-transferase: Novel vaccine against Fasciola hepatica infection in sheep. J. Immunol. 145, 3905–3910. Shekhar, K. C. (1991). Schistosomiasis drug therapy and treatment considerations. Drugs, 3, 379–405. Sher, A., James, S. L., Correa-Oliveira, R. & Henry, S. (1989). Schistosome vaccines: current progress and future prospects. Parasitology, 98, S61–S68. Shu-hua, X., Ji-qing, Y., Hui-fang, G. & Catto, B. A. (1991). Uptake and effect of praziquantel and the major human oxidative metabolite 4-hydroxypraziquantel by Schistosoma japonicum. J. Parasitol. 77, 241–245. Sinning, I., Kleywegt, G. J., Cowan, S. W., Reinemer, P., Dirr, H. W., Huber, R., Gilliland, G. L., Armstrong, R. N., Ji, X., Board, P. G., Olin, B., Mannervik, B. & Jones, T. A. (1993). Structure determination and refinement of human alpha class glutathione transferase A1-1, and a comparison with the mu and pi class enzymes. J. Mol. Biol. 232, 192–212. Smith, D. B., Davern, K. M., Board, P. G., Tui, W. U., Garcia, E. G. & Mitchell, G. F. (1986). Mr 26,000 antigen of Schistosoma japonicum recognized by resistant WEHI 129/J mice is a parasite glutathione S-transferase. Proc. Nat. Acad. Sci., U.S.A. 83, 8703–8707.

JMB—MS 321 Communication

Toscano, R. A., Rubio, M., Cetina, R. & Perez-Ibarra, B. M. (1992). Acta Crystallogr. 48, 1868. Walker, J., Crowley, P., Moreman, A. D. & Barrett, J. (1993). Biochemical properties of cloned glutathioneS-transferases from Schistosoma mansoni and

27 Schistosoma japonicum. Mol. Biochem. Parasitol. 61, 255–264. Wilce, M. C. J. & Parker, M. (1994). Structure and function of glutathione S-transferases. Biochim. Biophys. Acta, 1205, 1–18.

Edited by F. Cohen

(Received 3 November 1994; accepted 18 November 1994)