Tryparedoxins from Crithidia fasciculata and Trypanosoma brucei

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2003 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 278, No. 28, Issue of July 11, pp. 25919 –25925, 2003 Printed in U.S.A.

Tryparedoxins from Crithidia fasciculata and Trypanosoma brucei PHOTOREDUCTION OF THE REDOX DISULFIDE USING SYNCHROTRON RADIATION AND EVIDENCE FOR A CONFORMATIONAL SWITCH IMPLICATED IN FUNCTION* Received for publication, February 21, 2003 Published, JBC Papers in Press, April 21, 2003, DOI 10.1074/jbc.M301526200

Magnus S. Alphey, Mads Gabrielsen, Elena Micossi‡, Gordon A. Leonard‡, Sean M. McSweeney‡, Raimond B. G. Ravelli§, Emmanuel Tetaud¶, Alan H. Fairlamb, Charles S. Bond, and William N. Hunter储 From the Division of Biological Chemistry and Molecular Microbiology, School of Life Sciences, University of Dundee, Dundee DD1 5EH, Scotland, United Kingdom, ‡Macromolecular Crystallography Group, European Synchrotron Radiation Facility (ESRF), BP 220, F-38043 Grenoble Cedex 9, France, the §European Molecular Biology Laboratory Grenoble outstation, 6 rue Jules Horowitz, BP 156, F-38042 Grenoble Cedex 9, France, and the ¶Institut de Biochimie et Ge´ne´tique Cellulaire, UMR CNRS 5095, Universite´ Victor Segalen Bordeaux II, 1 Rue Camille Saint-Sae¨ns, 33077 Bordeaux Cedex, France

Tryparedoxin (TryX) is a member of the thioredoxin (TrX) fold family involved in the regulation of oxidative stress in parasitic trypanosomatids. Like TrX, TryX carries a characteristic Trp-Cys-Xaa-Xaa-Cys motif, which positions a redox-active disulfide underneath a tryptophan lid. We report the structure of a Crithidia fasciculata tryparedoxin isoform (CfTryX2) in two crystal forms and compare them with structures determined previously. Efforts to chemically generate crystals of reduced TryX1 were unsuccessful, and we carried out a novel experiment to break the redox-active disulfide, formed between Cys-40 and Cys-43, utilizing the intense x-radiation from a third generation synchrotron undulator beamline. A time course study of the S–S bond cleavage is reported with the structure of a TryX1 C43A mutant as the control. When freed from the constraints of a disulfide link to Cys-43, Cys-40 pivots to become slightly more solventaccessible. In addition, we have determined the structure of Trypanosoma brucei TryX, which, influenced by the molecular packing in the crystal lattice, displays a significantly different orientation of the active site tryptophan lid. This structural change may be of functional significance when TryX interacts with tryparedoxin peroxidase, the final protein in the trypanothione-dependent peroxidase pathway. Comparisons with chloroplast TrX and its substrate fructose 1,6-bisphosphate phosphatase suggest that this movement may represent a general feature of redox regulation in the trypanothione and thioredoxin peroxidase pathways.

Tryparedoxin (TryX)1 is a thiol-disulfide oxidoreductase found in parasitic trypanosomatids belonging to the order Kin* This work was supported by a Wellcome Trust senior research fellowship (to W. N. H.) and program grant (to A. H. F.), a Biotechnology and Biological Sciences Research Council (BBSRC) studentship (to M. S. A.), and a BBSRC Sir David Phillips Research Fellowship (to C. S. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The atomic coordinates and structure factors (code 1O73, 1OC8, 1OC9, 1O7U, 1O85, 1O8W, 1O8X) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/). 储 To whom correspondence should be addressed. Tel.: 44-1382345745; Fax: 44-1382-345764; E-mail: [email protected]. 1 The abbreviations used are: TryX, tryparedoxin; TryR, trypanoThis paper is available on line at http://www.jbc.org

etoplastida. The principal biological function of TryX is to regulate oxidative stress as a component of the trypanothione peroxidase pathway (Fig. 1). This pathway, unique to trypanosomatids, starts with the NADPH-dependent trypanothione reductase (TryR) (1), which maintains high levels of the polyamine-peptide conjugate trypanothione (N1, N8-bis(glutathionyl)spermidine) in the reduced form (T[SH]2), which in turn is able to reduce TryX (2, 3). The reduced TryX interacts with and passes on a reducing equivalent to tryparedoxin peroxidase (TryP), allowing it to catalyze the reduction of hydrogen peroxide and organic hydroperoxides to water or alcohols, respectively. A feature of the trypanothione peroxidase pathway is that the shuttling of reducing power from TryR through to TryP utilizes the redox properties of disulfide linkages (4) in the enzymes and their peptide substrates. The first crystal structure of tryparedoxin, the protein from Crithidia fasciculata (CfTryX1), revealed a compact globular molecule classed in the same fold family as the functional homologue thioredoxin (TrX, see Fig. 2a) (5). The TrX fold is based on a twisted five-stranded central ␤-sheet with two helices on either side (6, 7), but although classed in the same family, TryX is distinct from TrX in several respects (5). The parasite protein is significantly larger, ⬃16 kDa as compared with 12 kDa for TrX, and carries additional elements of secondary structure, in particular a ␤-hairpin at the N terminus. The relationship of secondary structure with the amino acid sequence is so different for TryX and TrX that it is meaningless to compare the overall sequences. A noteworthy similarity is, however, the presence of a redox-active disulfide at the N terminus of an ␣-helix. In TrX, this disulfide is contained in the motif Trp-Cys-Gly/Ala-Pro-Cys, whereas in TryX, the motif is Trp-Cys-Pro-Pro-Cys. A least-squares fit of the central ␤-strands of TrX and TryX align these motifs on top of each other (5). In CfTryX1, the vicinal cysteine residues (Cys-40 and Cys43) form a right-handed redox-active disulfide near the surface of the molecule, positioned beneath an overhanging Trp-39. Chemical modification and mass spectrometry studies indicate that the amino-proximal Cys-40 is the more reactive of the two cysteines (8). Cys-40 is more solvent-accessible than the part-

thione reductase; TryP, tryparedoxin peroxidase; TrX, thioredoxin; MR, molecular replacement; ESRF, European Synchrotron Radiation Facility; FBPase, fructose-1,6-bisphosphatase; T[SH]2, the reduced form of trypanothione disulfide; Tb, T. brucei; Cf, C. fasciculata.

25919

25920

Structure and Reactivity of Tryparedoxin

FIG. 1. The trypanothione peroxidase pathway. NADPH supports reduction of the disulfide oxidoreductase TryR, which processes the polyamine-peptide conjugate trypanothione disulfide (T[S]2) to maintain high levels of the dithiol form (T[SH]2). TryX is reduced by T[SH]2 and in turn reduces TryP, the enzyme that catalyzes the reduction of peroxides.

ner Cys-43 and actually forms a hydrogen bond with a water molecule (5), the position of which suggests where T[SH]2 or TryP bind when reacting with CfTryX1. We now report structures for a second isoform of C. fasciculata TryX (CfTryX2) and provide brief comparisons with CfTryX1 and structures of the disulfide form and chemically reduced CfTryX2 recently published (9). Unable to crystallize TryX as a homogeneous free dithiol form, we sought to generate this structure by photoreduction of the redox-active disulfide of CfTryX1 using the intense x-ray beam available from an undulator beamline at the European Synchrotron Radiation Facility (ESRF). As a control for this novel experiment and to study the effect of removing the disulfide linkage, Cys-43 was mutated to alanine, and the structure (CfTryX-C43A) was determined to 1.30-Å resolution. We also determined the structure of Trypanosoma brucei tryparedoxin (TbTryX) at 2.3-Å resolution, which presents an active site significantly different from any other TryX structure. Analysis of the interactions between TbTryX and a symmetry-related molecule suggests structural alterations that may be relevant to the interaction between TryX and the partner peroxidase, TryP. Based on crystal structures of a chloroplast TrX (10), one of its redox partners, fructose-1,6-bisphosphate phosphatase (11), and our own sequence analyses, we propose that the conformational lability of the tryptophan lid may contribute to specific redox events. MATERIALS AND METHODS

Cloning, Expression, and Purification of Recombinant Tryparedoxins—The gene coding for CfTryX2 was obtained by PCR amplification of genomic DNA of the C. fasciculata HS6 TryXII open reading frame (GenBankTM accession number: AF055986) using the oligonucleotides 5⬘-CAT CAT ATG TAT CAC ACC CTT CTC TAC-3⬘ for the sense strand and 5⬘-CAT GGA TCC TTA CTT CTT GGC CTC CAC GTT GGG-3⬘ for the antisense strand. The sense strand oligonucleotide contains an NdeI cloning site (underlined) incorporating an initiation codon (bold), whereas the antisense oligonucleotide contains a BamHI restriction site (underlined) downstream of the antisense stop codon (bold). The PCR products were blunt-end ligated into the SmaI site of pUC18 (SureClone, Amersham Biosciences), and then the inserts were excised by restriction enzyme digest and ligated into the pET-15b vector (Novagen), creating plasmids pET-TbTryX and pET-CfTryXII. The mutagenesis of cysteine to alanine at residue 43 in C. fasciculata tryparedoxin-I (CfTryX-C43A) was performed using the method described by Deng and Nickoloff (12) with the Chameleon kit (Stratagene). The plasmid pETCfTryX1 (13) provided the matrix, and the oligonucleotide was 5⬘-TGG TGC CCG CCG GCC CGC GGC TTC ACG-3⬘. The gene coding for TbTryX was obtained by PCR amplification of genomic DNA of the T. brucei 427 TryX open reading frame (GenBankTM accession number: AJ006403) using the oligonucleotides 5⬘-TTG CAT ATG TCT GGC CTC GCC AAG TAT-3⬘ for the sense strand and 5⬘-CAT CAT ATG TCA GTT GGG CCA CGG AAA GTT GGC-3⬘ for the antisense strand. The sense strand oligonucleotide carried an NdeI cloning site (underlined) incorporating an initiation codon (bold), whereas the antisense strand oligonucleotide carried an NdeI restriction site (underlined) just downstream of the antisense stop codon (bold). The integrity of the cloned genes was confirmed by sequencing. All recombinant proteins were expressed in Escherichia coli strain BL21 (DE3). Expression and purification protocols followed those published by Alphey et al. (13) and involved the use of metal ion affinity

chromatography to exploit the presence of the N-terminal histidine tag, which was introduced by using the pET-15b vector and which was subsequently removed by cleavage with thrombin (Amersham Biosciences). Protein concentration was determined spectrophotometrically at 280 nm using a theoretical extinction coefficient of 38030 M⫺1 cm⫺1 (14), and purity was evaluated using SDS-PAGE and matrixassisted laser desorption ionization time-of-flight mass spectrometry. Crystallization, Data Collection, and Data Processing—Crystals were grown using the hanging drop vapor diffusion setup, and diffraction data were processed, reduced, and scaled using the HKL (15) and CCP4 suite of programs (see Table I) (16). Crystals of CfTryX1 and CfTryX-C43A were obtained using the published conditions (13). Two tetragonal crystal forms (A and B) of CfTryX2 were obtained. Form A presented as rods and appeared in drops containing ⬃10 mg ml⫺1 protein, 15% w/v polyethylene glycol 8000, 30 mM sodium cacodylate, pH 6.5, 5 mM dithiothreitol, and 60 mM ammonium sulfate. These crystals display space group P42212 with unit cell dimensions of a ⫽ b ⫽ 111.7, c ⫽ 56.5 Å and are isomorphous to samples studied by Hofmann et al. (9). Form B crystals displayed a bipyramidal morphology and grew from solutions of ⬃10 mg ml⫺1 protein, 500 mM sodium citrate, 30 mM sodium HEPES, pH 7.5, 5 mM dithiothreitol. They are in space group P41212 with unit cell dimensions of a ⫽ b ⫽ 114.3, c ⫽ 102.0 Å. Single crystals of both forms were cryo-protected by soaking in crystallization mother liquor containing either 15% (form A) or 10% (form B) of glycerol prior to transfer in a nitrogen gas stream at ⫺170 °C. A single crystal of form A was used on the ESRF bending magnet beamline BM14, and data collection was carried out at ␭ ⫽ 0.977 Å to dmin ⫽ 1.5 Å with an MarCCD133 detector in a single sweep totaling 90o of oscillation in 0.5o steps. For form B, a single crystal was mounted on the ESRF undulator beamline ID14-EH2, and data were collected at ␭ ⫽ 0.933 Å using an ADSC QUANTUM4 detector. Despite the relatively large size of the crystal used for data collection (⬃250 ⫻ 150 ⫻ 150 ␮m3), diffraction maxima were only visible to ⬃2.2-Å resolution, and these were only apparent after a relatively long exposure time of 45 s/0.5o oscillation. Radiation damage was evident after 75 images, and the crystal was translated such that a fresh section was exposed to the x-ray beam and a further 44 0.5o images were collected. Both batches of data were processed and scaled together yielding a data set complete to 2.35 Å. For the time course experiment on CfTryX1 and the analysis of CfTryX-C43A, crystals were cryo-protected with 40% polyethylene glycol monomethyl ether 2000 and then flash-cooled at ⫺170 °C. Data were collected on ID14-EH2 (␭ ⫽ 0.933 Å) using the STRATEGY program (17) to determine the angular range for collection. For the time course experiment, a series of data sets were measured over the same oscillation range. Data sets A and B were consecutive and measured to a resolution of 1.5 Å. An intermediate exposure of 780 s, which corresponds to the total exposure time for measurement of a data set, was made while rotating the crystal, although no data were actually recorded. Data set C was then measured, and it was noted that the sample now only diffracted to 1.7-Å resolution. The radiation damage to the crystal after data set CfTryX-C was judged too great to warrant further useful data collection. The three data sets and the models derived from each are labeled CfTryX-A, -B, and -C, respectively. Clumps of small monoclinic plate-like crystals of TbTryX grew over a period of weeks in drops made by mixing a solution of 10 mg ml⫺1 protein, 50 mM HEPES, pH 7.5, with the reservoir solution of 30% polyethylene glycol 4000, 100 mM sodium acetate, pH 4.6, 200 mM ammonium acetate. The crystals display space group P21 with unit cell dimensions of a ⫽ 30.6, b ⫽ 31.5, c ⫽ 56.9 Å, b ⫽ 93.4°. The asymmetric unit comprises a monomer with ⬃30% solvent content and Vm of 1.8 Å3 Da⫺1. A small fragment (⬃200 ⫻ 50 ⫻ 10 ␮m3) was removed from a clump of crystals and passed through a cryo-protectant consisting of reservoir solution adjusted to include 20% 2-methyl-2,4-pentanediol, and then transferred into a stream of nitrogen gas at ⫺170 °C. Data were measured to 2.3-Å resolution on a Rigaku rotating anode (copper K␣ ␭ ⫽ 1.5418 Å)-Raxis IV image plate system. Structure Solution and Refinement—The initial phases for both CfTryX2 structures were obtained using the molecular replacement (MR) technique as implemented in the CNS software package (18) with data in the resolution ranges 15–3 Å for form A and 20 – 4 Å for form B. The structure of CfTryX1 (Protein Data Bank code 1QK8) (5) stripped of all solvent molecules was used as the search model. After this procedure, it was clear that both crystal forms contain two molecules/asymmetric unit, which results in calculated Matthews coefficients (Vm) (19) of 2.3 and 4.4 Å3 Da⫺1 for forms A and B, respectively. The unit cell of form A has a much lower bulk solvent volume, 46%, than the 72% observed for form B, and this helps to explain the different diffraction limits of the two forms. For crystal form A, the initial MR phases


25921

obtained were extended to the resolution limit of the data set, 1.5 Å, using a combination of non-crystallographic symmetry averaging, solvent flattening, and histogram matching as implemented in the program DM (20) after first calculating reliable ␴A-weighted figures-ofmerit (FOM) (21) for the MR phase set. The resulting electron density map (Fobs, ␣DM, FOMDM) was of excellent quality, and a model was constructed using the program ARP/wARP (22). Refinement was then carried out using CNS interspersed with rounds of rebuilding in QUANTA (Accelrys) during which solvent molecules were included. To complete the refinement, a final round was performed using the program REFMAC (23) in which the two sulfur atoms in the active site were refined with anisotropic temperature factors. For crystal form B, a similar protocol to that described for form A was used to extend the MR phases to the diffraction limit of the data set. A model was built manually with QUANTA, and refinement carried out in a similar manner to that for form A. The CfTryX-A, -B, and -C structures are isomorphous with the disulfide form of CfTryX1 (5), which provided the starting model for refinements using REFMAC. Following rigid body refinement, additional rounds of positional and B-factor refinement combined with graphics fitting (O) (24) were carried out. Water molecules were added using ARP/wARP. Once the R-factor and R-free had dropped from about 40 to 25%, anisotropic B-factor refinement was introduced. The structure of TbTryX was solved by MR (AMoRe) (25) using a poly-Ala structure of CfTryX1 as the search model. A clear solution was obtained that, after rigid body refinement, gave an R-factor of 48% and a correlation coefficient of 0.56 for data in the range of 30 –2.3-Å resolution. Density modification (DM) improved the electron density map that was then used for model building. Simulated annealing molecular dynamics (to reduce model bias), least-squares refinement with CNS, together with the placement of water molecules completed the analysis. Approximately 5% of each data set was set aside to provide an R-free to monitor the progress of all refinements (26), whereas PROCHECK (27) and OOPS (28) were used to assess model geometry. Further experimental details are provided (see Table I) and in the Protein Data Bank depositions. RESULTS AND DISCUSSION

Overall Structures—The tryparedoxin structure is constructed around a seven-stranded twisted ␤-sheet with parallel and antiparallel alignments. This sheet starts with a ␤-hairpin formed by ␤1 and ␤2, and thereafter a ␤3-␣1-␤4-␣2-␤5 combination. A final ␤-hairpin between ␤6 and ␤7 completes the structure. The active site Trp-Cys-Pro-Pro-Cys motif is located between strand ␤3 and the N terminus of helix ␣1 (Fig. 2a). A structure-based sequence alignment of the three highly conserved tryparedoxins used in this study is shown in Fig. 2b. The Second Tryparedoxin Isoform of C. fasciculata (CfTryX2)—The structure of CfTryX2 was determined independently in two crystal forms, each of which presents two molecules/asymmetric unit. A pairwise least-squares superposition of all C␣ atoms for the four molecules gave root mean square deviation values that ranged from 0.2 to 0.5 Å, and the results are similar whether we used a MR protocol or the anomalous dispersion from sulfur atoms (29) to provide the initial phase information. When comparing our MR-derived structures with those of CfTryX2 determined by Hofmann et al. (9), least-squares superposition values of between 0.2 and 0.6 Å were observed. These values indicate close agreement of the second isoform structures irrespective of how or where they were determined or the redox state of the protein (see below). The structures reported here confirm that, when compared with the structure of CfTryX1, the helices ␣1 and ␣2 are closer to each other in the structure of CfTryX2, allowing the formation of a hydrogen bonding network around the less solventexposed sulfur atom in the active site S–S bridge (9). Both crystal forms of CfTryX2 were grown from solutions containing dithiothreitol. The electron and difference density maps were suggestive of a time and space average of S–S bridge oxidized and reduced states. The refined S–S distances are 2.9 and 2.8 Å for the two molecules in form A and 3.2 and 3.0 Å for the two molecules in form B.

FIG. 2. The structure of tryparedoxin. a, a ribbon diagram depicting the TryX fold, secondary structure assignment, and location of the redox-active disulfide formed between Cys-40 and Cys-43 (yellow sticks). Figs. 2a and 3–5 were prepared with MOLSCRIPT (39) and RASTER3D (40). In b, the amino acid sequence of residues colored red are strictly conserved, and those colored black are similar at scale 7 in the ALSCRIPT program (41). The active site motif WCPPCR is marked with ●.

Radiation-induced Cleavage of the Redox-active Disulfide— The exposure of protein crystals to an intense x-ray source changes the properties of the sample (30 –33). These changes, which include an increase in unit cell volume, a decreased resolution to which diffraction data can be observed, increased mosaicity, and an increased Wilson B-factor, are indicative of general radiation damage. It has also been noted that although atomic B-factor values increase for successive data sets, the change is not equally distributed for all atoms but rather occurs at glutamate, aspartate, and cysteine residues. In the latter case, it appears that disulfide breakage contributes significantly to this increase in B-factors. In general, only one cysteine in a disulfide actually moves during bond breakage, whereas its partner remains well fixed (31). Weik et al. (32), in a study of x-ray-induced damage to Torpedo californica acetylcholinesterase observed that active site residues are among the most radiation-sensitive of residues and suggested, in a similar fashion to disulfide bonds, that these groups constituted “weak links” in protein structures. Our study on CfTryX1 targeted a redox-active disulfide, which should constitute an even weaker link than a structural disulfide. This is indeed the case since the radiation dose required to break this S–S bond is much less than that reported for structural S–S links in lysozyme for example (31). During the time course experiment, from CfTryX-A to CfTryX-C (Table I), we noted general symptoms of radiation damage to the sample; resolution decreased, the B factors for

25922


TABLE I Data collection, refinement, and model geometry statistics Values in parentheses refer to the highest resolution bin. Structure

Upper resolution limit (Å) No. of measurements/ unique Redundancy/completeness (%) I/␴(I) Rmerge (%) Protein residues (total) Water molecules/ sulphates R-work/R-free (%) Average isotropic thermal parameters (Å2) Wilson B-value Overall Main chain/side chains Water molecules r.m.s.d. b bond lengths (Å)/bond angles (°) Ramachandran analysis (%) Favored regions Additionally, generously allowed regions Disallowed regions Protein Data Bank accession code a b

CfTryX2A

CfTryX2B

CfTryX-A

CfTryX-B

CfTryX-C

CfTryX-C43A

2.3 37965/4699

TbTryX

1.50 469429/57687

2.35 136117/28025

1.50 167903/22718

1.50 167593/22753

1.70 116000/15786

1.30 261452/35022

8.1/98.7 (94.7)

8.1/99.7 (95.5)

4.9/97.5 (99.9)

7.4/90.0 (46.7)

7.4/94.5 (76.3)

7.4/96.2 (93.6)

7.5/96.6 (88.9)

19.6 (6.5) 7.3 (21.7) 144 63/0

37.6 (2.1) 4.5 (35.4) 151 a/145 a 312/1

29.8 (9.6) 5.3 (24.9) 150 a/147 a 133

14.7 (1.2) 5.2 (53.8) 143 133

16.8 (1.5) 4.6 (39.5) 143 126

15.5 (1.4) 6.1 (60.5) 143 61

18.6 (1.8) 4.5 (43.9) 143 210

19.9/24.7

20.7/23.5

19.6/21.3

20.4/26.3

20.3/26.6

21.0/27.6

17.8/22.8

37.5 20.6 20.1/21.0 22.1 0.007/1.45

19.0 23.1 18.9 34.1 0.022/1.86

52.2 48.7 44.1 50.4 0.021/1.65

17.4 21.4 19.2/21.6 30.5 0.014/1.5

18.2 22.5 20.0/22.7 32.4 0.013/1.4

22.2 26.5 24.7/28.3 27.7 0.016/1.7

14.4 20.3 16.4/19.6 32.7 0.012/1.6

88.3 11.7

87.1 12.5

86.8 13.2

94.3 5.7

93.4 6.6

91.8 8.2

92.6 7.4

0 1O73

0.4 1OC8

0 1OC9

0 1O7U

0 1O85

0 1O8W

0 1O8X

Denotes the two molecules per asymmetric unit. Root mean square deviation.

consecutive data sets increased, the unit cell volumes increased from 136,800 to 136,930 to 137,340 Å3, and the mean fractional isomorphous differences also increased from 0.08 (A and B) to 0.12 (A and C). In the disulfide form of CfTryX1, the Cys-40 – Cys-43 S␥–S␥ distance is 2.2 Å. In CfTryX-A, the distance between the two S␥ atoms has increased to 2.5 Å. This is most likely due to partial reduction or damage caused by the high x-ray intensity of the undulator beamline. In CfTryX-B, the S␥-S␥ distance is 2.8 Å, and in CfTryX-C, it has increased to 3.0 Å (Fig. 3). Similar results were obtained in the structure of CfTryX2 in the presence of 2-mercaptoethanol where the S␥-S␥ distances for the two copies in the asymmetric unit are 3.0 and 3.4 Å (9). Minor Structural Perturbation Results from Disulfide Breakage—In contrast to previous observations of radiation-induced damage to structural disulfides (31), we do not see deterioration in electron density for the S␥ atoms of Cys-40 or Cys-43 (Fig. 3). As the disulfide breaks, the Cys-40 side chain moves toward solvent, and the flanking residues Trp-39 and Pro-41 move slightly up and out (not shown). Structures of TrX in the reduced form have been determined (Ref. 10 and references therein) and also CfTryX2 (9), and similar observations have been made. In addition to breaking the disulfide using synchrotron radiation, it was anticipated that the C43A mutant would allow Cys-40 to adopt a position similar to that occupied when the protein is reduced. The mutant structure correlates well with CfTryX-C (Figs. 4 and 5). The C43A mutation also produced small shifts within the active site involving Ser-36 and Tyr-80. In CfTryX1, the side chain of Ser-36 is held in position through interactions with the immobile Cys-43, but in the mutant, Ala-43 no longer has a stabilizing effect on Ser-36. The side chain of Ser-36 adopts a different position, forming a hydrogen bond with the hydroxyl of Tyr-80, which has moved some 1.2 Å from the native structure to stabilize the new arrangement (Fig. 4). The overall conclusion from the radiation-induced disulfide

breakage and mutant structure analyses is that the active site of TryX appears to be relatively unperturbed by the redox state. Minor structural changes occur that serve to make Cys-40 S␥ slightly more accessible to react with the cognate partners. This is similar to what has been observed in structures of the dithiol form of TrX by itself (10) or in complex with thioredoxin reductase (34). Capitani et al. (10) studied the variation in oxidation state of the disulfide in a chloroplast TrX, but in contrast to our study, they first measured data on the reduced form of TrX, and then over a period of almost 2 days, using an in-house x-ray source and crystals at 4 °C, were able to isolate a data set that indicated that the disulfide had reformed without any large perturbation to the active site. Structure of TbTryX and a Model for the Interaction with TryP—The high degree of sequence conservation (Figs. 2b and 5) and similar biophysical properties (35, 36) of TbTryX as compared with the C. fasciculata tryparedoxins suggest that the three-dimensional structures should be similar. An overlap of CfTryX1, CfTryX2, and TbTryX highlights the structural homology of tryparedoxins (Fig. 5). The root mean square deviation for 139 C␣ atoms in common between TbTryX and CfTryX1 is 0.8 Å. The largest differences are observed in the N-terminal region, in particular at the turn between ␤1 and ␤2. This is on the opposite side of the molecule from the redoxactive site. The similarities extend beyond the overlay of C␣ atoms to the residues that constitute the hydrophobic core of the protein. Thirteen aromatic residues in CfTryX1 (tyrosines 34, 54, and 80, tryptophans 70 and 86, phenylalanines 32, 35, 46, 53, 63, 77, 91, 104) are strictly conserved in the three tryparedoxins (Fig. 2b). In addition, phenylalanines at positions 33, 57, 67, and 81 of CfTryX1 are replaced by Leu-33, His-57, Leu-67, and Tyr-81 of TbTryX. The residues that cluster around the redoxactive site are also highly conserved between CfTryX and TbTryX. Indeed those residues that were first implicated in CfTryX1 binding trypanothione (5), namely Trp-39, Pro-41, Pro-42, Arg-44, Trp-70, Asp-71, Glu-72, Lys-83, Ile-109, Pro-


25923

FIG. 3. The breakage of the redoxactive disulfide with synchrotron radiation. Fo ⫺ Fc omit maps for the sulfur atoms in the CfTryX1 active site. The figure shows the native protein (5) and the data sets (A–C) showing breakage of the active site disulfide. The electron density is contoured at the 4␴ level. The protein atoms are depicted with a ball-and-stick representation and colored according to type: purple, carbon; cyan, nitrogen; red, oxygen; yellow, sulfur.

FIG. 4. The CfTryX1 and CfTryXC43A structures. Stereoview overlay of the active sites of native CfTryX1 (carbon atoms colored green) and CfTryX1C43A (carbon atoms colored magenta), showing the adjustment in position of Ser-36 and Tyr-80 to compensate for the mutation of Cys-43. Dashed lines represent hydrogen bonding interactions.

110, and Arg-128, are strictly conserved. Hofmann et al. (9) were able to confirm that the last three residues did in fact interact with the ligand in a mutant CfTryX2 glutathionylspermidine complex. This suggests that similar molecular features determine the association with the TryX redox partners in both Crithidia and Trypanosoma.

An overlay of residues that comprise the active site of TbTryX and CfTryX1 reveals Trp-39 in a different position in TbTryX than in the CfTryX structures (Fig. 6). In CfTryX1, Trp-39 is placed over the redox disulfide, and N⑀1 donates a hydrogen bond to the carbonyl of Trp-70 (5). In TbTryX, Trp-39 adopts a different rotamer and is flipped out at the surface of

25924


FIG. 5. Overlay of three tryparedoxins. Stereoview depicting the least-squares fit for C␣ atoms of CfTryX1 (black), CfTryX2 (red), and TbTryX (blue). The N terminus and redox-active cysteine positions (Cys-40 and Cys-43) are marked.

FIG. 6. Stereoviews of the T. brucei TryX active site and a potential model for interaction with a 2Cysperoxiredoxin. In a, the symmetry-related Val-Ala-Lys tripeptide of a neighboring TryX is colored green. For comparison, and to highlight the different location of the tryptophan lid, part of the active site of CfTryX1 is superimposed (black sticks) on the TbTryX structure. In b, the tripeptide Val-Cys-Pro from HBP23 is shown in blue with Cys-173 S␥ in yellow.

the protein with N⑀1 forming a hydrogen bond with Asp-76 O␦2 from a symmetry-related molecule (not shown). In the TbTryX crystal structure, a symmetry-related molecule is positioned such that the Val-59-Ala-60-Lys-61 segment is placed in the cleft at the redox-active site. The valine side chain fills the site, which in CfTryX is occupied by the Trp-39 lid, apparently forcing the tryptophan to adopt a different conformation. The alanine methyl group points toward the redox-active disulphide, and the lysine side chain is directed away from the disulfide. This alternative conformation of the tryptophan lid has been noted previously in TrX and attributed to crystal packing effects, in the case of a mutant E. coli TrX (37), or due to sequence effects in a chloroplast TrX (10). The adjustment serves to expose the N-terminal cysteine of the redox-active disulfide; therefore, we decided to investigate whether such conformational pliability is relevant to the function of TryX or TrX when

they interact with their cognate peroxidases and other proteins. We first considered TryP, a 2Cys-peroxiredoxin, well characterized biochemically (3) and for which a crystal structure of the reduced form has been determined (2). All peroxiredoxins carry an essential N-terminal cysteine, often in a tetrapeptide Val-Cys-Pro-Thr motif. The 2Cys-peroxiredoxins have, in addition, a conserved C-terminal cysteine in a Val-Cys-Pro tripeptide motif, which interacts with the TryX Trp-Cys-Pro-Pro-Cys redox center. Reduced TryX interacts with oxidized TryP, but the only structure available for TryP is that of the reduced form (2). However, the structure of the homologous and oxidized form of the 2Cys-peroxiredoxin HBP23 has been determined (38). Since TryP and HPB23 share almost 60% sequence identity and are closely related in threedimensional structure (2), we superimposed the C-terminal Val-Cys-Pro motif of one monomer of the dimeric HBP23 onto the symmetry-related Val-Ala-Lys tripeptide of TbTryX. The

Structure and Reactivity of Tryparedoxin C␣ fit was with an root mean square deviation of 0.3 Å. Although this can only be a crude model, we note that the side chain of HBP23 Val-172 adopts a different orientation as compared with the TbTryX Val-59⬘ but that HBP23 Cys-173 is turned directly toward the redox-active Cys-40 of the tryparedoxin (Fig. 6b) with the S␥ atoms 3.3 Å apart. The model suggests that when TryX associates with TryP, a repositioning of the tryptophan lid might occur in conjunction with other molecular features such as the electrostatic interactions discussed by Hofmann et al. (3) to facilitate interaction of the redox components. If the combination of a pliable tryptophan and a valinecysteine dipeptide is indeed important for the tryparedoxinperoxidase interaction, we reasoned that it might also contribute to thioredoxin-protein associations. Thioredoxin peroxidases are homologous to TryP, and the Val-Cys-Pro motif is strictly conserved (3), which would be consistent with our hypothesis. Also, the truncated form of chloroplast TrX shows the tryptophan lid in the open conformation, and we note that one partner for this TrX is chloroplast FBPase, for which a structure is available (11). Chloroplast TrX regulates the activity of FBPase by reduction of the disulfide formed between Cys-153 and Cys-173 (pea FBPase numbering). Cys-153 occurs in the sequence Val-Cys-Gln-Pro-Gly located on a flexible loop, whereas Cys-173 occurs in an ␣-helix. A search of the EXPASY data base (ca.expasy.org) indicated that the pentapeptide segment, with valine preceding the redox-active cysteine, is strictly conserved in chloroplast FBPase. The observations hint at a role for a Val-Cys combination to interact with a conformationally labile tryptophan to assist TryX and TrX pass on their reducing equivalents. Definitive proof would require a structure of the functional complexes, and we are trying to obtain this for TryX-TryP. Acknowledgments—We acknowledge staff of the ESRF and European Molecular Biology Laboratory (EMBL) Grenoble Outstation for maintenance of beamlines, access, and support. REFERENCES 1. Fairlamb, A. H., and Cerami, A. (1992) Annu. Rev. Microbiol. 46, 695–729 2. Alphey, M. S., Bond, C. S., Tetaud, E., Fairlamb, A. H., and Hunter, W. N. (2000) J. Mol. Biol. 300, 903–916 3. Hofmann, B., Hecht, H.-J., and Flohe´ , L. (2002) Biol. Chem. 383, 347–364 4. Raina, S., and Missiakas, D. (1997) Annu. Rev. Microbiol. 51, 179 –202 5. Alphey, M. S., Leonard, G. A., Gourley, D. G., Tetaud, E., Fairlamb, A. H., and Hunter, W. N. (1999) J. Biol. Chem. 274, 25613–25622 6. Katti, S. K., LeMaster, D. M., and Eklund, H. (1990) J. Mol. Biol. 212, 167–184 7. Holmgren, A., and Bjornstedt, M. (1995) Methods Enzymol. 252, 199 –208 8. Gommel, D. U., Nogoceke, E., Morr, M., Kiess, M., Kalisz, H. M., and Flohe, L.

25925

(1997) Eur. J. Biochem. 248, 913–918 9. Hofmann, B., Budde, H., Bruns, K., Guerrero, S. A., Kalisz, H. M., Menge, U., Montemartini, M., Nogoceke, E., Steinert, P., Wissing, J. B., Flohe´ , L., and Hecht, H.-J. (2001). Biol. Chem. 382, 459 – 471 10. Capitani, G., Markovic-Housley, Z., DelVal, G., Morris, M., Jansonius, J. N., and Schurmann, P. (2000) J. Mol. Biol. 302, 135–154 11. Chiadmi, M., Navaza, J., Miginiac-Maslow, M., Jacquot, J.-P., and Cherfils, J. (1999) EMBO J. 18, 6809 – 6815 12. Deng, W. P., and Nickoloff, J. A. (1992) Anal. Biochem. 200, 81– 88 13. Alphey, M. S., Tetaud, E., Gourley, D. G., Fairlamb, A. H., and Hunter, W. N. (1999) J. Struct. Biol. 126, 76 –79 14. Gill, S. C., and von Hippel, P. H. (1989) Anal. Biochem. 182, 319 –326 15. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307–326 16. Collaborative Computational Project Number 4 (1994) Acta Crystallogr. Sect. D Biol. Crystallogr. 50, 760 –763 17. Ravelli, R. B., Sweet, R. M., Skinner, J. M., Duisenberg, A. J. M., and Kroon, J. (1997) J. Appl. Crystallogr. 30, 551–554 18. Bru¨ nger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., GrosseKunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr. Sect. D Biol. Crystallogr. 54, 905–921 19. Matthews, B. M. (1976) J. Mol. Biol. 33, 491– 497 20. Cowtan, K. (1994) Joint CCP4 and ESF-EACBM Newsletter on Protein Crystallography 31, pp. 34 –38, SERC Daresbury Laboratory, Warrington, United Kingdom 21. Read, R. J. (1986) Acta Crystallogr. Sect. A. 42, 140 –149 22. Perrakis, A., Sixma, T. K., Wilson, K. S., and Lamzin, V. S. (1997) Acta Crystallogr. Sect. D Biol. Crystallogr. 53, 448 – 455 23. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr. Sect. D Biol. Crystallogr. 53, 240 –255 24. Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Crystallogr. Sect. D Biol. Crystallogr. 47, 110 –119 25. Navaza, J. (1994) Acta Crystallogr. Sect. A. 50, 157–163 26. Bru¨ nger, A. T. (1993) Acta Crystallogr. Sect. D Biol. Crystallogr. 49, 24 –36 27. Laskowski, R. A., MacArthur, R. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Crystallogr. 26, 283–291 28. Kleywegt, G. J., and Jones, T. A. (1996) Acta Crystallogr. Sect. D Biol. Crystallogr. 52, 829 – 832 29. Micossi, E., Hunter, W. N., and Leonard, G. A. (2002) Acta Crystallogr. Sect. D Biol. Crystallogr. 58, 21–28 30. Burmeister, W. P. (2000) Acta Crystallogr. Sect. D Biol. Crystallogr. 56, 328 –341 31. Ravelli, R. B., and McSweeney, S. M. (2000) Structure 8, 315–318 32. Weik, M., Ravelli, R. B., Kryger, G., McSweeney, S., Raves, M. L., Harel, M., Gros, P., Silman, I., Kroon, J., and Sussman, J. L. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 623– 628 33. Weik, M., Berges, J., Raves, M. L., Gros, P., McSweeney, S., Siolman, I., Sussman, J. L., Houee-Levin, C., and Ravelli, R. B. (2002) J. Synchrotron. Rad. 9, 342–346 34. Lennon, B. W., Williams, C. H., and Ludwig, M. L. (2000) Science 289, 1190 –1194 35. Dormeyer, M., Reckenfelderba¨ umer, N., Lu¨ demann, H., and Krauth-Siegel, R. L. (2002) J. Biol. Chem. 276, 10602–10606 36. Reckenfelderba¨ umer, N., and Krauth-Siegel, R. L. (2002) J. Biol. Chem. 277, 17548 –17555 37. Nikkola, M., Gleason, F. K., Fuchs, J. A., and Eklund, H. (1993) Biochemistry 32, 5093–5098 38. Hirotsu, S., Abe, Y., Okada, K., Nagahara, N., Hori, H., Nishino, T., and Hakoshima, T. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 12333–12338 39. Kraulis, P. J. (1991) J. Appl. Crystallogr. 24, 946 –950 40. Merritt, E. A., and Murphy, M. E. P (1994). Acta Crystallogr. Sect. D Biol. Crystallogr. 50, 869 – 873 41. Barton, G. J. (1993) Protein Eng. 6, 37– 40