Crystal structure of the C107S/C112S mutant of yeast nuclear 2‐Cys peroxiredoxin

Abstract

Peroxiredoxins (Prxs) are a superfamily of antioxidant enzymes, which are abundant in several isoforms in all kingdoms.1 Prxs catalyze the reduction of deleterious substances such as hydrogen peroxide (H2O2), alkyl hydroperoxides, and peroxynitrites by utilizing the thiol group of the “peroxidatic” cysteine (CP), which is conserved within the N-terminal region.2 Some eukaryotic Prxs also act as regulators of H2O2-mediated signal transduction.3, 4 All Prxs belonging to the thioredoxin-fold superfamily share the same peroxidatic active-site structure. During a catalytic cycle, the CP residue is oxidized by peroxides to a cysteine sulfenic acid (CP-SOH) intermediate. Prxs are classified into 1-Cys and 2-Cys type based on the occurrence of the “resolving” Cys (CR) residue. The 1-Cys Prxs do not contain a CR residue, and the CP-SOH is recycled by glutathionylation mediated by glutathione S-transferase π, followed by spontaneous reduction of the enzyme with glutathione.5 In 2-Cys Prxs, the CP-SOH and CR-SH react to form a stable disulfide, which is then reduced by oxidoreductases such as thioredoxin, tryparedoxin, AhpD, or AhpF.2 The 2-Cys Prxs have been further subdivided into “typical” and “atypical” types, depending on the position of the CR residue. In typical 2-Cys Prxs (hereafter referred to as T2-Cys Prxs), the CR residue is located within the C-terminal arm of another subunit of a homodimer. In contrast, the CR residue in an atypical 2-Cys Prx resides within the same subunit. The atypical 2-Cys Prxs have also been further subdivided into “L,” “C,” and “R” type subfamilies (hereafter referred to as L-, C-, and R2-Cys Prxs, respectively), depending on the spatial location of the CR residue.6 Therefore, from a mechanistic point of view, there are five unique Prx subfamilies in total. To date, five distinct Prxs have been identified in the yeast Saccharomyces cerevisiae.7 They include three thiol peroxidases (cTPx I, II, and III) localized in the cytoplasm, one (nTPx) in the nucleus and one (mTPx) in the mitochondria. cTPx I, II, and III are T2-Cys Prxs, while mTPx is a 1-Cys Prx. nTPx is a member of the C2-Cys Prxs that contains a CxxxxC motif.8 Bacterial homologues of such Prxs are frequently referred to as the bacterioferritin comigratory proteins (BCP) and their plant homologues are named as PrxQ.9-11 These Prxs are least characterized among the Prx subfamilies and information regarding their structure is not yet available. In this study, we have determined the crystal structure of a truncated mutant of nTPx in which both the catalytic residues of Cys107 and Cys112 were replaced with serine. This mutant protein was gradually and spontaneously degraded by the freezing and thawing process until 56 amino acid residues were cleaved off from its N-terminal.12 nTPx has nuclear targeting sequences but the cleavage site has not yet been determined. The truncated mutant nTPx (hereafter referred to as tmTPx) may correspond to a physiologically mature nTPx. The present structure of a C2-Cys Prx makes it possible to compare the 3D structures of all the five Prx subfamilies. Detailed procedures used for the overexpression and purification of the intact mutant nTPx, and for the crystallization and data collection of tmTPx have been previously reported.8, 12 The X-ray data collected at 100 K at the Pohang Light Source (Korea) were used for the structure determination. The structure was easily solved in space group P32 by molecular replacement using the computer program Beast13 and a model based on the X-ray structures of human TPxB14 (PDB code, 1qmv) and Streptococcus pneumoniae TPx (PDB code, 1psq). The model was built using the program O.15 Refinement was carried out using CNS16 with positional refinement, followed by simulated annealing to 4,000 K, and finally, individual B-factor refinement. Manual rebuilding was carried out between each run with σA-weighted, 2Fo-Fc maps. Water molecules were picked up from the 2Fo-Fc and Fo-Fc difference maps at the bases of peak heights and distance criteria; however, these were discarded if the thermal parameter after refinement was above 60 Å2. During the final stages of refinement, the maximum likelihood method implemented in the program Refmac17 was used to refine atomic positions and isotropic B-factors. There were no electron densities for the N-terminal Ser57 and Ser58 residues and the C-terminal Glu215. A summary of data and refinement statistics is shown in Table I. Structural figures were generated using the program PyMOL (http://pymol.sourceforge.net/). The coordinates have been deposited in the PDB under accession code 2a4v. The crystal structure of tmTPx, which mimicks the molecular organization of reduced mature nTPx, was determined to a resolution of 1.8 Å [Fig. 1(A)]. tmTPx has a compact, spherical structure with a typical Prx fold that is constructed around a seven-stranded twisted β-sheet surrounded by five α-helices. The N-terminal chain, which is sharply bent at the Gly66 residue, is connected to a β-hairpin (β1-β2) following an α1 helix. The α1 is followed by two βαβ units (β3-α2-β4-α3-β5). A β-hairpin (β6 and β7) that is connected to β5 by the α4 and a long loop completes the β-sheet and leads to the C-terminal α5. The α2 helix has a pronounced kink near its C-terminal end, similar to many Prxs. The strands β4, β3, β6, and β7 and helices α2, α4, and α5 form the core of the thioredoxin fold. A: Ribbon diagram of tmTPx. α-Helices and β-strands are colored orange and cyan, respectively. Absolutely conserved residues are represented by stick models, with carbon in green, oxygen in red, and nitrogen in blue. Ser107 and Ser112 are labeled as CP and CR, and their Oγ atoms are depicted as spheres in orange and blue, respectively. B: Structure-based sequence alignment of selected Prxs. Human PrxVI was selected as a representative for 1-Cys Prxs, S. typhimurium Ahpc for T2-Cys (typical 2-Cys Prxs), S. pneumoniae TPx for L2-Cys (L-type atypical 2-Cys Prxs), human PrxV for R2-Cys (R-type atypical 2-Cys Prxs), and yeast nTPx for C2-Cys (C-type atypical 2-Cys Prxs). The residues involved in β-strands, α-helices, and 310-helices are shaded with cyan, yellow, and light brown, respectively. The secondary structure elements have been labeled based on the labeling of tmTPx. CP in parenthesis denotes the position of the CP residue common to all Prxs, and C, L, R, and T in parentheses denote the positions of the CR residues of T2-Cys, L2-Cys, R2-Cys, and C2-Cys Prxs, respectively. Four residues that are absolutely conserved in all Prxs are colored red, and the CR residues are colored blue on red. For the C2-Cys, only a part of the residues, truncated in tmTPx, are shown shaded with light gray and the 48 N-terminal sequences of yeast nTPx were omitted. C: Comparison of the Prx structures representing each subfamily. Cα-traces are shown in yellow for human PrxVI (PDB code, 1prx), purple for S. typhimurium Ahpc (1n8j), green for S. pneumoniae TPx (1psq), red for human PrxV (1h4o), and blue for yeast tmTPx (thick line). The structures were superposed using the atoms in the β-sheets. Colored spheres labeled as CP, T, L, R, and C denote the catalytic Cys residues, as described in B. S denotes the putative substrate-binding site, and α4-β6 and β7-α5 denote the loops connecting the corresponding secondary structure elements. The α2 helix in tmTPx assumes a so-called fully folded (FF) conformation that commonly appears in the 3D structure of reduced CP-SH forms, oxidized CP-SOH intermediates, and overoxidized (CP-SO or -SO3−2) forms of Prxs.2, 18-20 In this conformation, the CP is positioned at the first turn of the α2 and points toward the central β-sheet. In addition to CP, three other amino acids are strictly conserved in all Prxs. They include the Pro and Thr (or Ser) residues that always form a PxxxT(S)xxC motif and an Arg residue that is remotely located from CP in the sequence. For these conserved residues, the corresponding amino acids in tmTPx are Pro100, Thr104, Ser107, and Arg175, respectively. As shown in Figure 1(A), Pro100 is located at the end of the β3 strand, Thr104 on the loop preceding the α2 and Arg175 at the beginning of the β6. These spatial localities are conserved in the active sites of all Prxs. In the FF conformation, the thiolate Sγ of CP forms hydrogen bonds with the conserved Thr and Arg residues. In the case of tmTPx, the Oγ(Ser107) atom is 3.30 and 3.45 Å apart from the Oγ1(Thr104) and the Nη(Arg175), respectively. These separations correspond to hydrogen bonds involved with the Sγ(CP) in Prxs, suggesting that the wild-type and mutant nTPxs essentially have the same 3D structures. To date, 21 X-ray structures of the Prx family members in various redox and oligomeric states have been elucidated and deposited in the PDB; however, the X-ray structure of C2-Cys Prxs has not yet been revealed. With the determination of the present structure, it is now possible to compare the 3D structures of all the five Prx subfamilies. We selected four Prxs representing each subfamily and compared their sequences and structures, all in the FF conformation, with those of nTPx [Fig. 1(B,C)]. They included human PrxVI21 (1-Cys), Salmonella typhimurium AhpC3 (T2-Cys), S. pneumoniae TPx (L2-Cys; PDB code, 1psq), and human PrxV22 (R2-Cys). These comparisons clearly showed that the Prxs share the same core structure, but differ in the locations of the CR residue as well as in the structures of several peripheral regions. In the L-, C-, and R2-Cys Prxs, the CR residues are located at the α3, α2, and α5 helices, respectively, while they are located in the C-terminal chain in T2-Cys Prxs. The N- as well as the C-terminal extensions significantly vary in length depending on the Prx subtypes. The two loops (β3-α2 and β4-α3) that connect the central β strands to the α2 and α3 helices, respectively, are structurally well conserved. Other loops show differences in their lengths and structures, significantly varying in the α3-β5, α4-β6, and β7-α5 loops. The most notable feature of the structure of tmTPx is that the N-terminal part of the α5 helix has an extra helical turn than those present in other Prxs and is connected to the β7 strand without a long intervening loop. Consequently, the putative substrate-binding site of nTPx is more or less poorly defined, wide open, and shallow compared to those of other Prxs. The crystal structures of T2-Cys and L2-Cys Prxs are available in both reduced and disulfide forms, although the structure of R2-Cys Prxs in its disulfide form has not yet been determined.23 Structural data indicate that, during the catalytic cycle of 2-Cys Prxs, the α2 helix partially unwinds and the CR-containing region considerably changes its conformation positioning the CP-SOH and the CR accurately to form a CP-S-S-CR disulfide bond that is exposed on the protein surface.2, 3, 6 In tmTPx, Ser112 corresponding to the CR is located at the α2 and faces the protein surface. For nTPx and C2-Cys Prxs in general, therefore, it is very likely that the formation of a disulfide form does not require rearrangement of the juxtaposed α3 or α5 helix but requires only a local unfolding of the α2 helix. Yeast nTPx, bacterial BCPs, and plant PrxQs are better characterized amongst the C2-Cys Prx subfamily members.7-10 Yeast nTPx shares a sequence homology of 31 and 28% with E. coli BCP and Populus tremula PrxQ, whereas the sequence homology between BCP and PrxQ is 36%. Although the similarity in sequence is not very high, they are well aligned without any significant gaps except for one region. E. coli BCP contains five residues more than nTPx and P. tremula PrxQ in the loop α4-β6, which may not affect the overall shape of the substrate-binding site. These C2-Cys Prxs have a weak peroxidase activity compared to other Prxs. They generally show a better catalytic efficiency for alkyl hydroperoxides than H2O2 and for cumene hydroperoxide than t-butyl hydroperoxide in the case of P. tremula PrxQ.8-11 This trend in substrate preferences appears to be consistent with the shape of the substrate-binding site, which is widely exposed and easily accessible to bulky substrates. The C2-Cys Prxs are unique as they function as monomers, while other Prxs exist as functional dimers or redox-dependent oligomers (usually toric decamers composed of five dimers).18 Dimerization is a prerequisite for T2-Cys Prxs to perform catalytic reactions and may be required for substrate binding in 1-Cys, R-, and L2-Cys Prxs. Two different types of dimerization patterns are observed in Prxs; recently this has been discussed in detail.19, 20 In the first type that occurs in 1-Cys and T2-Cys Prxs, the Prx-Prx interface is parallel to the central β sheet, and the C-terminal extension, among other regions, plays an important role in the formation of a dimer. nTPx and its homologues do not have a C-terminal extension and thus do not have a chance to form this type of a dimer. The other pattern occurs in the dimeric L- and R2-Cys Prxs and at the dimer-dimer interface of decameric T2-Cys Prxs. In this type, the dimer interface is perpendicular to the β sheet and five distinct regions are involved in the formation of a dimer. The C2-Cys Prxs obviously differ from these dimeric Prxs in the sequence of these regions. However, even among the dimeric Prxs, the sequences of the interacting regions are not identical, and the details of the interaction modes and the interfacial angles between two monomers significantly vary. As stated above, nTPx shows a marked difference in the loop β7-α5 from other Prxs. However, even in other Prxs, this loop is not involved in the dimeric interface. Factors that decide the monomeric or dimeric state of a functional Prx remain to be determined. We thank Drs. K. K. Kim and Y. S. Huh for their help during data collection. J.C. and S.C. are recipients of BK21 fellowship. Experiments at PLS were supported in part by MOST and POSTECH.