Sequence analysis of the tryparedoxin peroxidase gene from Crithidia fasciculata and its functional expression in Escherichia coli.

Tryparedoxin peroxidase from Crithidia fasciculata is an essential component of the trypanothione-dependent hydroperoxide metabolism in the trypanosomatids (Nogoceke, E., Gommel, D. U., Kiebeta, M., Kalisz, H. M., and Flohé, L. (1997) Biol. Chem. 378, 827-836). The tryparedoxin peroxidase gene and its flanking regions have been isolated and sequenced from a C. fasciculata genomic DNA library. It consists of an open reading frame of 564 base pairs encoding a protein of 188 amino acid residues. The gene, modified to encode 6 additional histidine residues, was expressed in Escherichia coli and the recombinant protein was purified to homogeneity by metal chelating chromatography. Recombinant tryparedoxin peroxidase has a subunit molecular mass of 21884 +/- 22 and contains two isoforms of pI 6.2 and 6.3. It exhibits a kinetic pattern identical to that of the authentic tryparedoxin peroxidase and has a similar specific activity of 2.51 units mg-1. The enzyme unequivocally belongs to the peroxiredoxin family of proteins, whose members have been found in all phyla. A phylogenetic tree comprising 47 protein and DNA sequences showed tryparedoxin peroxidase and a homologous Trypanosoma brucei sequence to form a distinct molecular clade. The consensus sequence: xnAx5-6Fx9Gx3Vx2Fx1Px2Fx1FVCPTEx21Sx1Dx7Wx16-19Dx15- 16Gx3Rx2Fx2Dx27Ax 1Qx4-11Cx1-3Wxn was demonstrated by alignment of the sequences of tryparedoxin peroxidase and 8 other peroxiredoxins with established peroxidase function.

Tryparedoxin peroxidase has recently been identified as a constituent of the complex peroxidase system in the trypanosomatid Crithidia fasiculata (1). In these parasitic protozoa hydroperoxides are reduced at the expense of NADPH by means of a cascade of three oxidoreductases: the flavoprotein trypanothione reductase, the thioredoxin-related tryparedoxin, and tryparedoxin peroxidase (Fig. 1). The first enzyme of the cascade is homologous to glutathione reductase and thiore-doxin reductase (2), which are involved in NADPH-dependent hydroperoxide reduction in other species (3). The other components of the trypanosomatid system also belong to protein families occasionally constituting peroxidase systems. Preliminary amino acid sequencing data indicated that tryparedoxin is phylogenetically related to thioredoxin (1), whereas the tryparedoxin peroxidase belongs to the peroxiredoxins (1) comprising the thioredoxin peroxidases of yeast and mammals (4) and the alkyl hydroperoxide reductases of bacteria (5).
The unique feature of the trypanosomatidal peroxidase system is its dependence on the peculiar redox mediator trypanothione which so far has not been discovered in any species outside the Trypanosomatidae. Its biosynthesis from spermidine and glutathione requires two distinct enzymes, glutathionylspermidine synthetase (6) and trypanothione synthetase (7). With a cascade of oxidoreductases plus the redox mediator trypanothione and the two auxiliary enzymes for its synthesis, the trypanosomatids have developed the most complicated system for the removal of hydroperoxides so far discovered in nature. This is not to imply a particular efficiency or robustness of the system. On the contrary, trypanosomatids are reported to be highly susceptible to oxidative stress (8). Correspondingly, their extraordinary metabolism is being discussed as a potential target area of specific trypanocidal agents (1,9,10).
The present possibilities available for the treatment of trypanosomal diseases, such as Chagas disease, African sleeping sickness, and the various forms of leishmaniasis, necessitate improvement. We (1,6), like many others (11)(12)(13)(14), have therefore embarked on the identification and characterization of potential molecular targets typical of the trypanosomatids. Here we report for the first time the full-length DNA, deduced amino acid sequence, and expression of a tryparedoxin peroxidase and its relatedness to peroxiredoxins with established or unknown functions.
Primers, Hybridization Probes, and Sequence Analysis-Based upon peptide sequences of tryparedoxin peroxidase (1) CTGGATCCCRATIGGCATRTC-3Ј were synthesized. PCR 1 amplification was performed with the GeneAmp PCR Core kit (Perkin Elmer) using 0.8 g of C. fasciculata genomic DNA as template, 10 l of 10 ϫ reaction buffer, 8 l of 25 mM MgCl 2 , 2 l of each 10 M dNTP, 100 pmol of each primer, and 0.5 unit of Taq polymerase. An annealing temperature of 50°C was used. The PCR product was analyzed by agarose gels and purified using the QIAquick PCR purification kit (QIAGEN). Sequencing was performed on a 373A DNA Sequencer (Applied Biosystems) using the PRISM Ready Reaction DyeDeoxy Terminator Sequencing kit (1550 V, 19 mA, 30 W, 42°C). When used as a hybridization probe, the PCR product was labeled with digoxigenin using the DIG DNA Labeling kit (Boehringer Mannheim) according to the instructions provided by the supplier.
Library Construction and Screening Procedure-The genomic DNA was partially digested for 5-30 min with a ratio unit of Sau3A/g of DNA of 0.005. The efficiency of the digestion was monitored by agarose gel electrophoresis. Proteins were removed from the DNA using Strata-Clean Resin (Stratagene). The Sau3A sites were partially refilled with dATP and dGTP and Klenow fragment. The genomic DNA was ligated into Lambda GEM-11 XhoI half-site arms (Promega) at a molar ratio of DNA to genomic DNA (average size 15 kb) of 1:0.7. The ligated DNA was packaged using the Packagene Lambda DNA Packaging System (Promega) according to the suppliers' instructions. The phages were used to infect Escherichia coli host strain LE392 (Promega) according to the standard protocol. 5.1 ϫ 10 3 plaque forming units of the genomic library gave positive PCR signals for tryparedoxin peroxidase and were plated on agar. The plaques were transferred to 9-cm diameter Biodyne-A nylon membranes and screened with the digoxigenin-labeled PCR probes following the instructions provided by the supplier but using a hybridization temperature of 58°C. Digoxigenin-labeled nucleic acids were detected colorimetrically with the DIG Nucleic Acid Detection kit (Boehringer Mannheim). Positive clones were rescreened, amplified, and suspended in SM buffer. The phages were precipitated by PEG 8000 and purified in CsCl gradients. The isolated DNA was used for restriction analyses (SacI) or as template for PCR reactions. Southern blot was performed using standard techniques, using the same digoxigenin-labeled PCR probe and the same hybridization conditions as above. The digestion products were eluted from agarose gels and ligated into pBluescript II KS(ϩ/Ϫ) phagemids (Stratagene). The ligated DNA was used to transform E. coli LE392. Transformed cells were selected by ampicillin resistance, plasmids were purified using the QIAprep Spin Plasmid kit (QIAGEN) and analyzed by restriction enzyme digestion and sequencing.
Recombinant Tryparedoxin Peroxidase Expression and Purification-The tryparedoxin peroxidase gene contained in the cloned 1.5-kb fragment was amplified by PCR with a forward primer A (5Ј-CCAC-CACTTGGCGCACATATGTCCTGCGGTGCCGCC-3Ј) that contains an NdeI site and overlaps the 5Ј end of the coding sequence, and a reverse deletion primer a (5Ј-CGCGGGGTGGTTCAACTTGGCGGCACCGC-AGGAC-3Ј) to delete an extra cytosine base at position 30. Amplification was also performed with a forward deletion primer b (5Ј-CAAGT-TGAACCACCCCGCGCCTGAGTTCGACGAC-3Ј) and a reverse primer B (5Ј-GCCACGCCTGCTTCTCTCCTCGAGGCCCTCCTTCTTCTTGG-3Ј) which overlaps the 3Ј end of the coding sequence and contains an XhoI site. Consequently the last coded amino acid was changed to a leucine, a glutamate residue was added, the stop codon was deleted, and the protein contained 6 histidine residues at its C-terminal end. Amplification was performed as above but using the Expand High Fidelity polymerase mixture and buffer (Boehringer Mannheim) at an anneal-ing temperature of 58°C with the extension temperature being increased in 10-s increments per cycle during cycles 10 -20. PCR products of the expected size were obtained and used as template for a second PCR amplification with the forward primer A and the reverse primer B. The amplified coding region was digested with NdeI and XhoI and ligated to a pET24a(ϩ) vector (Novagen) treated with the same enzymes and dephosphorylated. The resulting plasmid (pET/TXNPx) was used to transform E. coli BL21(DE3). Transformed cells were selected by kanamycin resistance, the plasmids purified and sequenced.
E. coli BL21(DE3) pET/TXNPx was grown to OD 600 of 0.9 -1.0 at 36°C and 180 rpm in LB medium with 30 g of kanamycin/ml, then expression of the tryparedoxin peroxidase gene was induced with 1 mM isopropyl-␤-D-thiogalactopyranoside. E. coli BL21(DE3) containing the pET24a plasmid was grown in the same way. Samples taken at different times were centrifuged, resuspended in 50 mM Tris-HCl, pH 8.0, 1 mM EDTA buffer, sonicated, and centrifuged. After 5 h the culture was centrifuged and either stored at Ϫ20°C or the cells were resuspended in 0.05 culture volume of binding buffer (5 mM imidazole, 500 mM NaCl, and 20 mM Tris-HCl, pH 7.9). The cell suspension was sonicated on ice and centrifuged for 40 min at 4°C, 35,000 ϫ g. The supernatant was applied to a His⅐Bind resin (Novagen) column charged with Ni 2ϩ and equilibrated with binding buffer, at a flow rate of about 10 column volumes/h. The column was washed with 10 volumes of binding buffer, 6 volumes of 500 mM NaCl, 20 mM Tris-HCl, pH 7.9, buffer containing 60 mM imidazole and 6 volumes of the same buffer with 100 mM imidazole. Tryparedoxin peroxidase eluted in the buffer containing 500 mM imidazole. Active fractions were pooled and immediately dialyzed against 50 mM Tris-HCl, pH 7.6, buffer containing 100 mM NaCl and 1 mM EDTA.
Characterization of Recombinant Tryparedoxin Peroxidase-SDS-PAGE was done under reducing conditions in 8 -25% gradient gels on a Pharmacia Phast System and the gels were stained for protein with silver according to the manufacturers' recommendations. Western blots were performed with the same system on polyvinylidene difluoride membranes. Whole rabbit serum (1:250 dilution) containing antibodies raised against pure C. fasciculata tryparedoxin peroxidase was used as primary antibody and anti-rabbit goat antibodies (Sigma) as secondary antibody. Enzyme activity and kinetic studies were performed as described in Nogoceke et al. (1); protein concentration was determined using Coomassie Brilliant Blue G reagent (Bio-Rad) with bovine serum albumin as standard.
Analysis of Molecular Evolution-The EMBL and SWISSPROT data bases were screened for tryparedoxin peroxidase homologous nucleotide or peptide sequences using BLAST and FASTA. Multiple alignments were performed by the program PILEUP (GCG Wisconsin Program Package, Version 9.0-OpenVMS) using default parameters. The polypeptide sequences of data base entries comprising complete coding regions were compared by the program DARWIN (16).

Isolation and Sequencing of the Tryparedoxin Peroxidase
Gene from C. fasciculata-Sequenced peptide fragments obtained from purified tryparedoxin peroxidase of C. fasciculata could be aligned along the established deduced amino acid sequence of the thiol-specific antioxidant protein of yeast (1). This enabled appropriate degenerate PCR primers to be designed for the generation of a PCR product from the C. fasciculata genomic DNA. This PCR product, which was 0.4 kb long (see Fig. 3), was subsequently used to screen a genomic library for inserts containing the full-length DNA encoding the tryparedoxin peroxidase. A clone containing a 15-kb insert with the presumed tryparedoxin peroxidase gene was isolated and sequenced. This, however, led to the detection of equal quantitites of different nucleotides at several positions toward the 3Ј end of the gene, implying the presence of similar but not iden-tical genes in the insert. This observation was not unexpected since the genome of the Trypanosomatidae is known to contain repetitive structural genes separated by intergenic sequences (17)(18)(19). The -clone was consequently digested with the restriction enzyme SacI and a Southern blot was performed. Three fragments belonging to the insert (1.1, 1.5, and 11 kb) gave positive hybridization signals with the labeled PCR product. Each of the three fragments was subcloned into pBluescript II KS(ϩ/Ϫ) phagemids and sequenced. The general sequencing strategy is shown in Fig. 2 using the 1.5-kb fragment as an example.
The 11-kb fragment contained the information coding for the previously sequenced peptides of tryparedoxin peroxidase (Fig.  3). Nevertheless, as with the 15-kb insert, the bases at positions 542, 548, 551, 556, 557, 560, 563, 564, and 565 remained ambiguous suggesting the presence of more than one gene in this fragment. The nucleotide sequence shown in Fig. 3 was confirmed by resequencing the 1.5-kb fragment which contained an open reading frame largely identical to the one of the The arrows delimit the PCR product used to screen the genomic library. The start and stop codons are in bold, as is the asparagine residue which was replaced by a threonine residue in direct peptide sequencing. Sequences confirmed by protein sequence analysis are underlined. The position of the SacI site in the 1.1-kb fragment is heavily underlined. The differences in the coding region between the 1.5-and 11-kb fragments, and in the 5Ј-flanking region between the 1.5-and 1.1-kb fragments are shown in brackets. The AG consensus splice leader sites and the polypyrimidine-rich tract are double underlined. 11-kb fragment except for the presence of an additional cytosine at position 30. As a consequence of the resulting frameshift, the deduced amino acid sequence no longer complied with the established peptide sequences. Hence, the 1.5-kb fragment contained a pseudogene. The 1.1-kb fragment also contained an open reading frame but encoded only part of the tryparedoxin peroxidase since a SacI restriction site was present at position 472-477. This reading frame was therefore not sequenced to completion. All the amino acid sequences determined for the peptide fragments of the C. fasciculata tryparedoxin peroxidase matched those predicted from the DNA shown in Fig. 3, except for positions 406 -408 where an asparagine is encoded but a threonine was determined by amino acid sequencing. This asparagine was found in all of the sequenced DNA fragments and in the PCR product.
When comparing the 5Ј-flanking region of the 1.5-kb fragment with one of the 1.1-kb fragments, only minor differences were found (Fig. 3). In both fragments two possible consensus dinucleotides AG, which represent potential splice leader acceptor sites, are found in positions Ϫ52 and Ϫ125. However, only the one in position Ϫ125 is preceded by a pyrimidine-rich tract which is necessary for trans-splicing (20).
Expression of the Tryparedoxin Peroxidase Gene in E. coli BL21(DE3)-The identification of multiple open reading frames encoding tryparedoxin peroxidase required confirmation of functionality of the sequence shown in Fig. 3. For this purpose the 1.5-kb fragment exhibiting least sequence ambiguities appeared most appropriate except for the inserted frameshift mutation. Therefore, the frameshifting extra cytosine base at position 30 was deleted, the region coding for the C-terminal amino acids was modified as described under "Experimental Procedures" to facilitate purification of the expression product, and the resulting coding region was incorporated into the expression vector pET 24a to give the pET/TXNPx plasmid (Fig. 4). After induction of the transformed bacteria, tryparedoxin peroxidase activity was detected in supernatants of sonicated cells. The activity increased to a maximum 5 h after induction while no activity was found in the control (Fig.  5). Induction resulted in the accumulation of a new protein with an apparent molecular mass of 21,000 in SDS gels which was recognized by the anti-tryparedoxin peroxidase antibodies (Fig. 6B) and proved to be an active tryparedoxin peroxidase.
Purification and Characterization of the Recombinant Tryparedoxin Peroxidase-The addition of 6 histidine residues to the C-terminal end of the expressed protein enabled simple purification of the recombinant tryparedoxin peroxidase on a Ni 2ϩ -binding resin. Tryparedoxin peroxidase bound to the resin, whereas most of the impurities either did not bind or were washed from the column with the binding buffer containing 100 mM imidazole. Tryparedoxin peroxidase eluted at 500 mM imidazole and was shown to be homogeneous by SDS-PAGE and subsequent silver staining (Fig. 6A). N-terminal sequencing of this protein showed the initial methionine to be missing and allowed us to confirm the first 30 amino acids. The expressed tryparedoxin peroxidase had nearly the same molecular mass of about 21,000 as the authentic tryparedoxin peroxidase in SDS-PAGE (Fig. 6A). Matrix-assisted laser desorption/ionization-mass spectrometry analysis demonstrated the molecular mass of the recombinant enzyme to be 21,884 Ϯ 22. The difference in the molecular mass of 1,004 to the authentic tryparedoxin peroxidase (1) corresponded to the additional amino acids (leucine, glutamate, and 6 histidines) added at the C-terminal end of the recombinant enzyme. Whereas the authentic tryparedoxin peroxidase contained several isoforms ranging from pI 4.9 to 5.8, the recombinant protein showed two bands of pI 6.2 and 6.3. The higher alkalinity results from the additional histidines residues.
The pure recombinant enzyme had a specific activity of 2.51 units/mg compared with 5.83 units/mg for the authentic enzyme. This difference may be due to the additional residues at the C-terminal end of the recombinant tryparedoxin peroxidase. The kinetic analysis of the recombinant protein revealed a kinetic pattern identical to that of the authentic tryparedoxin peroxidase, i.e. a ping-pong mechanism with infinite maximum velocities and Michaelis constants. Initial velocities when fitted to the general Dalziel equation for bisubstrate reactions, where ROOH ϭ hydroperoxide and TXN ϭ tryparedoxin, revealed that the coefficients 0 and 1, 2 were zero. The kinetic parameters obtained with the pure recombinant tryparedoxin  Table I.
Comparison with Homologous Sequences-Based on sequence analysis, tryparedoxin peroxidase belongs to the family of peroxiredoxins (1). More than 50 homologous sequences were detected in the SwissProt and EMBL data banks, which comprised various bacterial alkyl hydroperoxide reductases (5), the thioredoxin peroxidases from yeast (21) and mammals (4) previously called TSA proteins (22)(23)(24), NKEF-␣ and -␤ (25), a closely related gene of Trypanosoma brucei rhodesiense (26), and a large number of genes with still undefined functions.
Out of these protein and DNA sequences 47 were selected to construct an unrooted phylogenetic tree, whereby the selection was based on (i) completeness of the published sequence, (ii) similarity in size, and (iii) likelihood of homology with a maximum FASTA E-score of 4.5. In the thereby obtained phylogenetic tree (Fig. 7), tryparedoxin peroxidase (Q) together with the homologous T. brucei sequence (R) presents a distinct molecular clade branching off between the sequences of the metazoa (A-P), including one of the known helminth sequences (T) on the one side with the rest, comprising the sequences of entamoeba (S), yeast (␣, ␤), plants (␥-), bacteria (a-o), and, surprisingly, another helminth (U), on the other side.
Sequence alignments were further used to generate ideas about the potential functional relevance of particular amino acid residues. For this purpose we aligned a total of nine sequences for which a peroxidase function had been either clearly established or suggested from biological experiments (Fig. 8). These were the thioredoxin peroxidases of Saccharomyces cerevisiae (21) and Homo sapiens (4), the tryparedoxin peroxidase described here, and the alkyl hydroperoxide reductases from Salmonella typhimurium (5), E. coli (27), Coryne-bacterium diphtheriae (28), Bacillus subtilis (29), and Mycobacterium tuberculosis (30). Although the similarities between these sequences of poorly related species ranged from 29 to 77%, a total of 21 residues were strictly conserved in homologous positions. The overall consensus sequence is: x n Ax 5-6 -Fx 9 Gx 3 Vx 2 Fx 1 Px 2 Fx 1 FVCPTEx 21 Sx 1 Dx 7 Wx 16 -19 Dx 15-16 -Gx 3 Rx 2 Fx 2 Dx 27 AxQx 4 -11 Cx 1-3 Wx n .
The first cysteine residue is integrated into an absolutely conserved VCP motif which is widespread in the whole peroxiredoxin family. The C-proximal cysteine also forms a VCP motif in seven of the aligned peroxidases, whereas in M. tuberculosis the motif is altered to LCA. The C-proximal VCP and LCA motifs, respectively, are separated by two residues from the last conserved tryptophan. In the thioredoxin peroxidase of yeast the corresponding region reads VLPCNW. DISCUSSION We have cloned a genomic DNA fragment from C. fasciculata that encodes multiple copies of the entire sequence of tryparedoxin peroxidase, which is an essential constituent of the trypanosomatidal peroxide metabolism (1). The reading frame is very similar to a DNA sequence of T. brucei rhodesiense encoding a protein of unknown function (26). The high degree of similarity suggests that this trypanosomal protein is also a tryparedoxin peroxidase. However, the multiplication of genes in the trypanosomatids, which may lead to pseudogenes, precludes a definite functional interpretation of DNA sequences without experimental evidence. In the case of the C. fasciculata DNA sequence shown in Fig. 3, its functional relevance was unambiguously established by heterologous expression and comparison of the gene product with the authentic tryparedoxin peroxidase. Tryparedoxin peroxidase unequivocally belongs to the family of peroxiredoxins. This protein family is obviously widespread in nature. The phylogenetic tree showing the molecular evolution of the peroxiredoxins practically covers all phyla from bacteria to vertebrates but it does not simply reflect the phylogenetic divergence of the species. The sequence of the parasitic plathelmint Brugia malayi (T), for example, debranches close to protozoal and vertebrate sub-trees, whereas another plathelmint protein (U) belongs to a plant sub-tree. There are two distinct molecular clades found in chlorophyta plus a sep-arate type in the diatomean alga Odonthella, each diverging at distant points from bacterial branches. Similarly, one of the two yeast peroxiredoxins debranches from a plant branch, which in turn diverges from bacterial ancestors. This puzzling situation suggests multiple gene acquisition by means of endosymbiosis. Such gene transfer has also been implicated for various Euglenazoa, including the Trypanosomatidae (31). Even the cytosolic GapC of T. brucei and Leishmania mexicana, for example, is believed to be acquired from endosymbiotic ␥-purple bacteria and to be secondarily integrated into the nuclear genome of the trypanosomatids. The debranching point of the two trypanosomatid peroxiredoxins, however, does not lend any support to speculations about secondary gene acquisition. Debranching between the metazoa and yeast, bacteria, and plants, the molecular evolution of the trypanosomal per-oxiredoxins appears congruent with taxonomical development.
Molecular evolution has given ample room for functional diversification within the peroxiredoxin family and it can by no means be uncritically presumed that all peroxiredoxins are peroxidases or "antioxidant proteins." Conceivably, many of the peroxiredoxins will turn out to reduce hydroperoxides, but the reducing substrates differ substantially in the examples investigated to date. TSA of yeast, like human NKEF-␣ and -␤, utilizes thioredoxin as a reductant; the crithidial homolog is reduced by a remote relative of thioredoxin, tryparedoxin, which is 50% larger than the typical thioredoxins and is characterized by a WCPPC motif in its active site; the alkyl hydroperoxide reductases of bacteria are directly reduced by flavoproteins also containing vicinal thiols (AhpF); and from their history of discovery we know that the activity of TSAs depends on dithiols such as dithiothreitol. The common denominator of the peroxidase activity of peroxiredoxin thus appears to be the regeneration by dithiols of a residue reacting with hydroperoxides. In the absence of any reasonable alternative such a redox active residue can only be a cysteine. In the case of thioredoxin peroxidase of yeast, the active site was identified by site-directed mutagenesis as the N-proximal VCP motif (32). Interestingly, replacement of the conserved C-terminal cysteine residue by a serine resulted in an active antioxidant protein when tested with the non-physiological substrate dithiothreitol, but the activity with thioredoxin was lost (21,32). With tryparedoxin peroxidase, the participation of two cysteine residues in the catalysis was demonstrated by the substrate-dependent inactivation with N-ethylmaleimide and confirmed by matrixassisted laser desorption/ionization-time of flight-mass spectrometry (1). The consensus sequence of the peroxiredoxins with experimentally demonstrated peroxidase activity reveals two conserved cysteines as the only potentially redox-active functional groups. Only the first cysteine appears to be obligatorily integrated into a VCP motif within a highly conserved region. Although, the second cysteine also usually forms a VCP motif, it appears to tolerate substantial changes in its intimate sequence context. However, based on the yeast TSA example, it is tempting to speculate that the latter is crucial for the reaction with the specific reductant, whereas the former is indispensible for the reaction with the oxidant.
The kinetics of tryparedoxin peroxidase have been elucidated in detail (1). The apparent net forward rate constants for the reaction of the reduced enzyme (calculated per subunit) with a variety of hydroperoxides were found to be around 10 5 M Ϫ1 s Ϫ1 . To achieve such rate constants the active site cysteine residue must be highly activated. These rate constants are virtually identical to those of sulfur homologs of the selenocysteine containing glutathione peroxidases (33,34), where the (seleno)cysteine residue, together with a tryptophan and a glutamine residue, forms a unique catalytic triad and thereby is forced into full dissociation and is correspondingly reactive (34). Interestingly, two tryptophans and one glutamine are also strictly conserved in the peroxiredoxin-type peroxidases and it may therefore be speculated that their cysteines might be activated in an analogous way to the structurally unrelated glutathione peroxidases. In fact, with the exception of a single arginine, none of the other residues of the consensus sequence would be able to force a cysteine into dissociation. The three aspartate residues, if situated near the active site cysteine, would be contraproductive, the lipophilic residues would be largely indifferent, and the conserved prolines and glycines are most probably involved in the preservation of the three-dimensional structure. Unfortunately, not a single three-dimensional structure of a peroxiredoxin is presently available to confirm the hypothesis of the convergent evolution of an analogous active site by unrelated proteins.