In vivo parameters influencing 2-Cys Prx oligomerization: The role of enzyme sulfinylation

Abstract

2-Cys Prxs are H₂O₂-specific antioxidants that become inactivated by enzyme hyperoxidation at elevated H₂O₂ levels. Although hyperoxidation restricts the antioxidant physiological role of these enzymes, it also allows the enzyme to become an efficient chaperone holdase. The critical molecular event allowing the peroxidase to chaperone switch is thought to be the enzyme assembly into high molecular weight (HMW) structures brought about by enzyme hyperoxidation. How hyperoxidation promotes HMW assembly is not well understood and Prx mutants allowing disentangling its peroxidase and chaperone functions are lacking. To begin addressing the link between enzyme hyperoxidation and HMW structures formation, we have evaluated the in vivo 2-Cys Prxs quaternary structure changes induced by H₂O₂ by size exclusion chromatography (SEC) on crude lysates, using wild type (Wt) untagged and Myc-tagged S. cerevisiae 2-Cys Prx Tsa1 and derivative Tsa1 mutants or genetic conditions known to inactivate peroxidase or chaperone activity or altering the enzyme sensitivity to hyperoxidation. Our data confirm the strict causative link between H₂O₂-induced hyperoxidation and HMW formation/stabilization, also raising the question of whether C_P hyperoxidation triggers the assembly of HMW structures by the stacking of decamers, which is the prevalent view of the literature, or rather, the stabilization of preassembled stacked decamers.

Graphical abstract

1. Introduction

Peroxiredoxins (Prxs) are enzymes that catalyze the reduction of hydrogen peroxide (H₂O₂) via a conserved active-site cysteine (Cys) residue. Amongst the six Prx subfamilies, which are distinguished by sequence similarities [32,43], Prx1, also known as typical 2-Cys Prxs, are the most widespread, from archaea, bacteria to eukaryotes. The eukaryotic enzymes of the Prx1 group share with the other family members fast catalytic rates in the order of ~10⁷ M⁻¹ s⁻¹ [33,37], but carry the unique property of becoming inactive by hyperoxidation at elevated H₂O₂ levels [45,47,49]. These enzymatic specific attributes determine the unique cellular functions of eukaryotic 2-Cys Prxs as antioxidants, H₂O₂ signaling effectors and controllers, and as chaperones [14].

2-Cys Prx are obligate head-to-tail B-type homodimers, each with two catalytic Cys residues. In the peroxidatic cycle, the N-terminal Cys, named C_P for peroxidatic Cys, reduces H₂O₂, and is in turn oxidized to a sulfenic acid (C_P-SOH) [48] (Fig. 1). The Cys-sulfenic acid moiety then condenses with the C-terminal catalytic Cys residue of the other subunit, or resolving Cys (C_R) into an intermolecular disulfide. In the reduced enzyme C_P and C_R are ~13 Å apart. Therefore, disulfide formation involves an important structural remodeling occurring both at the C_P-active site pocket and C_R-containing C-terminal domain, which switches the enzyme structure form a fully folded (FF) to a locally unfolded (LU) conformation [17,47]. Based on a series of elegant studies Karplus and coworkers have proposed that the enzyme FF conformation both stabilizes the deprotonated reactive form of C_P and provides a steric and electrostatic environment that activates H₂O₂, hence establishing the observed C_P extraordinary high reactivity for H₂O₂[18,22]. The catalytic intermolecular disulfide is subsequently reduced by thioredoxin, which completes the catalytic cycle, and returns the enzyme to the FF conformation. In eukaryotic enzymes however, the C_P-SOH can further react with H₂O₂ instead of condensing with C_R, thus becoming oxidized to the corresponding sulfinic acid (−SO₂H), which interrupts the peroxidatic cycle. Hyperoxidized Prx is not a dead-end product; it is reactivated by ATP-dependent reduction of the sulfinate by sulfiredoxin (Srx) [45,6]. The sensitivity of eukaryotic enzymes to hyperoxidation is linked to the presence of two sequence fingerprints absent in other family enzymes, a three amino acids insertion in the loop between α4 and β5 associated with a conserved GGLC motif, and an additional helix (α7) occurring as a C-terminal extension and containing the conserved YF motif [47]. Such a structural configuration is thought to slow down the FF to LU transition rate, thereby favoring hyperoxidation [47].

Fig. 1.

The peroxidatic cycle of eukaryotic 2-Cys Prxs. Eukaryotic typical 2-Cys Prx are obligate dimers with a mechanism involving two Cys residues, in which C_P decomposes H₂O₂ into H₂O by nucleophilic attack and is oxidized to a sulfenic acid (C_P-SOH). The sulfenic acid then reacts with the resolving Cys (C_R) residue of the other subunit to form an intermolecular disulfide [48]. This disulfide is then reduced by thioredoxin, which completes the catalytic cycle. Alternatively, when H2O2 levels raises, the C_P-SOH can further react with H₂O₂, which leads to formation of a C_P-sulfinic acid (CP-SO₂H). The latter is reversed back to C_P-SOH by ATP-dependent reduction by sulfiredoxin (Srx).

Enzymatic cycling involves dramatic changes in quaternary structure that are crucial for 2-Cys Prxs differential functions. Reduced 2-Cys Prxs is typically in the form of decamers arranged into a ring-like toroid structure constituted of five B-type dimers interacting via their A-type interface [17,41,46,5]. During cycling decamers dissociate into dimers upon disulfide formation and are regained upon disulfide reduction [36,4,46]. As proposed by Karplus and coworkers, there is a reciprocal stabilization between the enzyme in the FF conformation and decamer assembly [17], which explains both that the decamer form enhances catalysis and that catalytic disulfide formation, by locking the enzyme into the LU conformation, dissociates the decamers into dimers [17,46,5,8]. In contrast, enzyme hyperoxidation triggers the stacking of decamers, up to filaments, which stabilize the decameric structure [16,26,38,40,9].

By combining extremely high H₂O₂ reaction rates, reflected by very low K_m values [33], very high expression levels, and the enzyme oxidative inactivation at elevated H₂O₂ levels [45,47,49], 2-Cys Prxs are very efficient towards the low metabolically produced H₂O₂ levels and relatively inefficient towards peroxide onslaughts [14,27]. Concurrently, hyperoxidation constitutes the gateway to the enzyme extra-antioxidant functions, and in particular to its chaperone function.

Jang, Lee and coworkers showed using the two S. cerevisiae 2-Cys Prx Tsa1 and Tsa2 in vitro that heat or H₂O₂ both promote formation of enzyme high molecular weight (HMW) structures of size > 1000 kDa and of spherical shape that have lost peroxidase activity, but are capable of preventing aggregation of heat-denatured model substrates [20], a function assimilated to that of a chaperone holdase. In the case of H₂O₂, C_P was required for the acquisition of the chaperone function, and sulfiredoxin (Srx1) switched the enzyme back to a low molecular weight (LMW) peroxidase active form by reducing the sulfinate [20,30]. However, C_P was dispensable for heat-induced chaperone activity, which led Jang and coworkers to suggest that heat and H₂O₂ trigger the functional switch of 2-Cys Prx through distinct mechanisms [20]. A chaperone holdase activity similar to Tsa1/2 has now been described for human cytosolic 2-Cys Prxs Prdx1 [19,34,35] and Prdx2 [29], plant chloroplastic 2-Cys Prxs [23], S. mansoni 2-Cys Prx SmPrx1 [2,40], L. infantum mitochondrial 2-Cys Prx mTXNPx [44], H. pylori 2-Cys Prx AhpC [12], cyanobacterial Anabaena PCC7120 2-Cys Prx alr4641 [3] and P. aeruginosa 2-Cys Prx PaPrx [1]. In most of these cases, the available data fit the model of Jang whereby H₂O₂-induced C_P hyperoxidation promotes formation of 2-Cys Prx HMW structures with chaperone holdase activity, while Srx-dependent sulfinate reduction returns the enzyme to the peroxidatic cycle [20,30]. Note that the Tsa1/2 HMW spherical structures, which have only been described by Jang and colleagues, must somehow be related to the typical decamers stacks, which were shown by others to also carry chaperone activity. The term HMW structures is used thereof to describe multimers > decamers. The above chaperone model does not hold true for the L. infantum mitochondrial 2-Cys Prx mTXNPx however [11,44]. In this case, C_P and hence enzyme hyperoxidation is dispensable for chaperone function, and heat but not H₂O₂ triggers within the reduced decamer major conformational changes that increase surface hydrophobicity, without quaternary structure changes. In the case of the A. thaliana chloroplastic 2-Cys Prx, although the chaperone function is activated by H₂O₂ and requires C_P sulfinylation, the chaperone active form of the enzyme is a decamer and not a higher order oligomer [23].

In summary, except for the cases of the L. infantum mitochondrial and At chloroplastic 2-Cys Prxs, available data support the notion that H₂O₂-induced C_P hyperoxidation activates 2-Cys Prx chaperone function by promoting the stacking of decamers. To elucidate the still unknown physiological scope and mechanism of the chaperone function of 2-Cys Prx in vivo, it is important to generate mutations that could unambiguously separate this function from the enzyme peroxidase function, and to establish assays allowing evaluating these mutations. Extensive information on the peroxidatic cycle is now available, which include knowledge of the peroxidatic catalytic site per se and the role of and molecular requirements for decamer formation (summarized in [18,22]). In contrast little is know on the link between hyperoxidation and HMW structures formation, which appears as a basic requirement for the chaperone function, and the few available data are very often controversial.

To begin addressing the link between enzyme sulfinylation and HMW structures formation, we have tested whether the in vivo quaternary structure of 2-Cys Prxs could be evaluated by size exclusion chromatography on crude lysates. These experiments were performed in S. cerevisiae using Tsa1 as model, and mutations known to inactivate chaperone activity or altering the enzyme sensitivity to hyperoxidation. Lysates from cells exposed to H₂O₂ were used to follow the dynamics of enzyme sulfinylation and quaternary structural changes. Data obtained with Tsa1 mutations were compared with those obtained with the same mutations in 2-Cys Prx from other organisms. Our data confirm the strict causative link between H₂O₂-induced sulfinylation and the stabilization of Tsa1 oligomers with a size compatible with that of two-stacked decamers. Our data also suggest that C_P hyperoxidation stabilizes preassembled stacked decamers, rather than triggering the assembly of HMW structures.

2. Results and discussion

2.1. SRX1 overexpression and truncation of YF-containing Tsa1 C-terminal extension abate H₂O₂-induced enzyme sulfinylation

To address the link between 2-Cys Prx sulfinylation and enzyme oligomerization, we selected Tsa1 mutants and genetic conditions predicted to alter enzyme sulfinylation. We used a N-terminal Myc-tagged version of Tsa1 (Myc-Tsa1) [6] and selected three derivative mutants. Myc-Tsa1ΔYF lacks the YF-containing last twelve C-terminal amino acids, and is predicted to be insensitive to sulfinylation, based on data obtained with the S. pombe 2-Cys Prx Tpx1 [21,25] human cytosolic Prdx2 [29] and endoplasmic reticulum (ER) Prdx4 [10]; Tsa1^C48S carries a serine substitution of C_P, and therefore lacks peroxidase activity, and Tsa1^C171S a serine substitution of the resolutive C_R, which has been shown to decrease the sensitivity to sulfinylation of human Prdx4 [10] and A. thaliana 2-Cys Prx [23]. Myc-Tsa1 and its corresponding mutants were expressed in a strain lacking Tsa1 (Δtsa1). To appreciate the impact of the N-terminal Myc tag, we also included the analysis of untagged wild type Tsa1 in the corresponding Wt strain. We also included strains lacking SRX1 (Δsrx1) and overexpressing it from the strong CMV promoter in the high-copy episomal plasmid pCM190 (pCM190-SRX1) [15].

We first inspected how Tsa1 mutations and the modulation of SRX1 expression impacted enzyme sulfinylation by western blots with an anti-PrxSO₂/₃ antibody (Fig. 2). Lysates were prepared before and at different times after cell exposure to 500 μM H₂O₂, a concentration causing full Tsa1 sulfinylation (see Fig. 2G). Sulfinylation of Myc-Tsa1 was maximal at 15 min, started to decrease at 60 min and disappeared at 180 min (Fig. 2A). Myc-Tsa1^C48S did not produce any sulfinylation signal, due to the absence of C_P (see Fig. 4B), thus confirming the high specificity of the antibody towards C_P sulfinylation. Myc-Tsa1^C171S produced a sulfinylation signal that was 1.5–2 fold lower than Myc-Tsa1 (Fig. 2B), and required twice as much H₂O₂ to reach the levels of sulfinylation seen in Myc-Tsa1 (Fig. 2compare G and H). The lower sensitivity to sulfinylation of Myc-Tsa1^C171S is consistent with the effect of the same mutation in A. thaliana 2-Cys Prx and human ER Prdx4 [10,23]. Myc-Τsa1ΔYF also produced a sulfinylation signal in response to H₂O₂, but again its intensity was strongly decreased and disappeared much faster than that of Myc-tsa1 (60 vs 180 min, compare Fig. 2A and C), thus confirming data obtained with S. pombe Tpx1. As the C-Ter domain contributes to the stability of the FF active site conformation, and hence to H₂O₂ reactivity, deleting it affects the sensitivity to sulfinylation by decreasing H₂O₂ reactivity [22]. The same rationale can be made for Myc-Tsa1^C171S if we consider that the C_R substitution alters the C-Ter structure. Untagged Wt Tsa1 produced a sulfinylation signal of intensity similar to Myc-Tsa1, but which surprisingly disappeared much faster than that of the tagged enzyme (60 vs 180 min, compare Fig. 2A and D). This difference in the enzyme-recycling rate might be explained by an effect of the tag of decreasing Srx enzyme binding or Srx accessibility to the sulfinylated residue. In Δsrx1, sulfinylation of untagged Tsa1 at 15 min was indistinguishable form the one seen in Wt cells, but as previously shown [6], the signal remained up to 180 min by lack of enzyme recycling (Fig. 2 compare D and E). In contrast, in cells carrying pCM190-SRX1, the sulfinylation signal was barely visible (Fig. 2 compare D and F), which indicate a much faster rate of enzyme recycling by virtue of Srx1 overexpression.

Fig. 2.

Evaluation of the sensitivity to sulfinylation of Tsa1 and derivative mutants. TCA-precipitated lysates from, Δtsa1 (A–C, G and H) or in Δtsa1Δsrx1 (E) or in Wt (D and F) yeast cells expressing Myc-Tsa1, MycTsa1^C48S, Myc-Tsa1^C171S, Myc-Tsa1ΔYF or overexpressing SRX1, as indicated, which were exposed to H₂O₂ (500 μM) (A–F), or to the indicated amount of H₂O₂ during 15 min (G and H) were resolved by reducing SDS PAGE, followed by western blot using the anti-Myc, anti-Tsa1 or anti-SO_2/3 antibodies, as indicated. The red arrow indicates the sulfinylation signal band.

Fig. 4.

SEC elution profile of Myc-Tsa1^C171Sand Myc-Tsa1^C48S. Crude lysates from Δtsa1 cells expressing Myc-Tsa1C^171S (A) or Myc-Tsa1^C48S (B) were taken before and after exposure to H₂O₂ (500 μM) for the indicated time and resolved by SEC. Elution fractions were resolved by non-reducing SDS PAGE, followed by western blot using the anti-Myc or anti-SO_2/3 antibody as indicated. The elution fraction of standard molecular weight markers is represented at the top of the gel. The red star indicates non-specific signals revealed by the anti-SO_2/3 antibody.

2.2. H₂O₂ triggers the reversible oligomerization of Myc-Tsa1 into the size of two decamers

To analyze the in vivo oligomeric state of Myc-Tsa1, we performed size exclusion chromatography on N-ethylmaleimide (NEM)-treated crude cell extracts from Myc-Tsa1-expressing Δtsa1 cells that were left untreated or were exposed to H₂O₂ (500 μM) for 15 and 180 min. Collected fractions were resolved by SDS-PAGE in which reducing agents were omitted in order to simultaneously evaluate disulfide-linked homodimer formation. Western blots were probed with anti-Myc and anti-PrxSO_2/3 antibodies (Fig. 3A). In untreated cells lysates, Myc-Tsa1 largely eluted at a size between that of a monomer (theoretical molecular weight of 21.5 kDa) and dimer (43 kDa), with traces of it equally distributed in all fractions, up to the column size exclusion (670 kDa). Myc-Tsa1 was mostly in the disulfide-linked dimer form, but these disulfides had probably formed after lysis, due to incomplete free sulfhydryls quenching, since Myc-Tsa1 from trichloroacetic acid (TCA)-precipitated lysates from the same cells largely migrated as a reduced monomer (not shown). As expected, probing membranes with the anti-PrxSO_2/3 did not yield any signal. In lysates from the 15 min H₂O₂ exposure sample, half of Myc-Tsa1 now eluted at about the 27 kDa elution control, close to the size of monomeric enzyme (21.5 kDa), and the other half at a size of about 500 kDa, which is compatible with that of two-stacked Myc-Tsa1 decamers, also referred from now on to the high molecular weight (HMW) form. A chromatogram of standard molecular weight markers is shown in Fig. 3C. Disulfide-linked dimers were here almost totally absent, as a consequence of enzyme sulfinylation, as also shown by probing membranes with the anti-PrxSO_2/3 antibody that revealed intense signals. Sulfinylation was equally distributed in fractions corresponding to the monomeric and HMW forms. Remarquably, in lysates from the 180 min H₂O₂ exposure sample, the Myc-Tsa1 elution pattern resembled now to that of untreated cells. Still, a strong sulfinylation signal persisted in Myc-Tsa1 monomers. We similarly analyzed the elution of Myc-Tsa1 from Δsrx1 lysates (Fig. 3B). Elution of Myc-Tsa1 from lysates of untreated cells resembled the one observed in WT untreated cells, except for a smaller proportion of monomeric enzyme and the presence of a small proportion of the enzyme in the HMW form. In lysates of the 15 min H₂O₂ exposure sample, elution of Myc-Tsa1 had also a wild type pattern, except for a higher abundance of the HMW species. In lysates of the 180 min sample however, the major fraction of Myc-Tsa1 now remained in the sulfinylated HMW form, in contrast to what was observed in WT cells.

Fig. 3.

SEC elution profile of Myc-Tsa1 and the effect of inactivating SRX1 on the enzyme quaternary structure. Crude lysates from Δtsa1 (A) or in Δtsa1Δsrx1 (B) cells expressing Myc-Tsa1 were taken before and after exposure to H₂O₂ (500 μM) for the indicated time and resolved by SEC. Elution fractions were resolved by non-reducing SDS PAGE, followed by western blot using the anti-Myc or anti-SO_2/3 antibody as indicated. The elution fraction of standard molecular weight markers is represented at the top of the gel. The red star indicates non-specific signals revealed by the anti-SO_2/3 antibody. (C) Representative chromatogram of standard molecular weight markers, thyroglobulin (670 kDa), apoferitin (443 kDa), β-amylase (200 kDa) to help resolve the size of the two-stacked decamers.

In summary, in untreated cells Myc-Tsa1 mainly exists as a non-covalently-linked dimer, with H₂O₂ triggering its reversible oligomerization into a HMW form of size compatible with two-stacked decamers, while undergoing reversible sulfinylation. In Δsrx1, H₂O₂ also triggers the oligomerization of Myc-Tsa1 to the two-stacked decamers form, but the enzyme remains in this form presumably by lack of sulfinate reduction. Myc-Tsa1 oligomeric transitions are consistent with those observed with untagged Tsa1 by native PAGE, also showing formations of HMW forms in response to H₂O₂ and their reversion upon Srx1 sulfinate reduction [30].

2.3. Ser substitution of peroxidatic Cys stabilizes the Myc-Tsa1 two-stacked decamer

We next analyzed Myc-Tsa1 catalytic Cys mutants, also expressed in Δtsa1. Myc-Tsa1^C171S had an elution profile resembling that of Wt enzyme, except for its weaker sulfinylation, in keeping with the data of Fig. 2, and for the reversion of the HMW form at 180 min after H₂O₂ exposure that was not as complete. As predicted, this mutant did not form any disulfide-linked dimer by lack of C_R. Myc-Tsa1^C48S in contrast had a strikingly different behavior. Elution of this mutant from untreated cell lysates resembled the elution of Myc-Tsa1 from the 15 min H₂O₂ exposure sample, with half of the protein in the monomeric form and the other half in the HMW form. Further, 15 min after H₂O₂ exposure, most of Myc-Tsa1^C48S was now shifted to the HMW form, thus indicating that despite lack of C_P the enzyme retained the ability to respond to H₂O₂. Lastly at 180 min, the protein still eluted for its major part at the HMW size, indicating lack of reversibility. As noticed above, this mutant did not give any sulfinylation signal, and as Myc-Tsa1^C171S did not form the disulfide-linked dimer.

In summary, Myc-Tsa1^C171S is less prone to sulfinylation, presumably by decreased H₂O₂ reactivity (see above). Further, it fully assembles into the HMW form that might be more stable than that of the Wt tagged enzyme. This latter result should be considered in view of the observation by native PAGE that the same Tsa1 mutant, but without any tag, is constitutively present in the HMW form prior to exposure to H₂O₂ [20], and of the same mutant in A. thaliana 2-Cys Prx the decameric form of which is more stable [23,24]. Myc-Tsa1^C48S is already in the HMW form prior to exposure to H₂O₂, in keeping with the behavior of the same mutation in purified S. mansoni 2-Cys Prx SmPrx1 [2] and purified bovine mitochondrial 2-Cys Prx SP-22 [16], which both eluted during SEC as HMW species, and were seen as long filaments of multiple decamer stacks by transmission electron microscopy. However our data disagree with native PAGE analysis data of the same Tsa1 and of human Prdx2 mutants showing they could not assemble into a HMW form in response to H₂O₂ [20,29]. The basis of these discrepant results is not clear. The stability of 2-Cys Prx with a Ser substitution of C_P raises the question of the structural determinants allowing formation of HMW structures that remains still poorly understood and controversial [2,39]. Another peculiar phenotype of Myc-Tsa1^C48S observed here is its ability to further switch its quaternary structure to the HMW form in response to H₂O₂, which might indicate either another 2-Cys Prx H₂O₂-responsive domain in addition to C_P or the presence of an unknown H₂O₂ responsive factor helping Tsa1 HMW assembly.

2.4. The N-Ter tag destabilizes Tsa1 HMW structures

The presence of an N-terminal tag has previously been shown to modify the function of 2-Cys Prxs [8]. That the Myc tag also alters the function of Tsa1 was already suggested by the inability of Myc-Tsa1 to fully rescue the H₂O₂ tolerance of a Δtsa1 strain (unpublished data), and by the slower recycling of the sulfinylated enzyme (see Fig. 1). We thus assayed the elution profile of untagged Tsa1 (Fig. 5). Untagged Tsa1 elution pattern was quite different from that of the Myc-tagged enzyme, since more than half of it was in the HMW form (70%), and the remainder in the monomer-dimer form. At 15 min after H₂O₂ exposure, the HMW form switched to the monomer-dimer form, which resulted in equal amounts of enzymes in these two forms; at 180 min, the elution pattern had returned to that of untreated cells, but with a slightly higher amount of protein, as a probable result of gene induction. Traces of the sulfinylation signal were seen at 15 min.

Fig. 5.

SEC elution profile of untagged Tsa1. Crude lysates from Wt cells were taken before and after exposure to H₂O₂ (500 μM) for the indicated time and resolved by SEC. Elution fractions were resolved by non-reducing SDS PAGE, followed by western blot using the anti-Myc or anti-SO_2/3 antibody as indicated. The elution fraction of standard molecular weight markers is represented at the top of the gel. The red star indicates non-specific signals revealed by the anti-SO_2/3 antibody.

Comparison of Tsa1 and Myc-Tsa1 elution profiles strongly suggests that the N-Terminal Myc tag destabilizes the HMW forms of the enzyme, despite a protracted sulfinylation in Myc-Tsa1. These data are fully consistent with those of bovine mitochondrial Prdx3 (or known as SP-22), in which removal of the N-terminal His tag stimulated the enzyme peroxidase activity by 3–4 fold, but also destabilized the dodecameric form of the enzyme [8], and with those of the A. thaliana chloroplastic 2-Cys Prx in which the N-terminal His tag also decreased peroxidase activity, but also increased the enzyme sensitivity to hyperoxidation [23]. In contrast, the in vitro dimer to decamer equilibrium of the L. braziliensis mitochondrial 2-Cys Prx LbPrx1m (similar to L. infantum mitochondrial 2-Cys Prx mTXNPx) is not altered by the N-terminal His tag [31]. However in the latter case, lowering the pH to acidic conditions was used to alter enzyme quaternary structure instead of protein redox changes as in our assay, which make the comparison irrelevant.

2.5. Myc-Tsa1ΔYF and SRX1 promote entry into the peroxidatic cycle

We next evaluated Myc-Tsa1ΔYF and the effect of overexpressing SRX1 on the untagged enzyme, both conditions abating enzyme sulfinylation. Myc-Tsa1ΔYF had a unique elution pattern (Fig. 6A). Prior to H₂O₂ exposure, a large proportion of the protein eluted at the size of the two-stacked decamers, and about a quarter of it in the disulfide-linked dimer form. Then, 15 min after H₂O₂ exposure, Myc-Tsa1ΔYF was almost totally shifted to the monomer-dimer size, with about more than half of the protein in the disulfide-linked dimer form. At 180 min, the protein had entirely returned to the double decamer form. Elution of untagged Tsa1 from lysates of cells overexpressing SRX1 was very similar to that of Myc-Tsa1ΔYF since prior to H₂O₂ exposure, a major part of it also eluted in the HMW form, but then was almost completely switched to the monomer-dimer form after 15 min of H₂O₂ exposure to completely return to the HMW form after 180 min (Fig. 6B). Note that in lysates of SRX1-overexpressing cells exposed for 15 min to H₂O₂ untagged Tsa1 eluted exclusively as a disulfide-linked dimer, whereas in Wt cells it was about equally distributed between a disulfide-linked dimer and a monomer. As expected from the results of Fig. 2, for both Myc-Tsa1ΔYF and SRX1 overexpression no sulfinylation signal was seen.

Fig. 6.

SEC elution profile of Myc-Tsa1ΔYF and the effect of overexpressing SRX1on Tsa1 quaternary structure. Crude lysates from Δtsa1 cells expressing Myc-Tsa1ΔYF (A) or Wt overexpressing SRX1 (B) were taken before and after exposure to H₂O₂ (500 μM) for the indicated time and resolved by SEC. Elution fractions were resolved by non-reducing SDS PAGE, followed by western blot using the anti-Myc or anti-SO_2/3 antibody as indicated. The elution fraction of standard molecular weight markers is represented at the top of the gel. The red star indicates non-specific signals revealed by the anti-SO_2/3 antibody.

In summary, both the deletion of the C-terminal-containing YF motif and the over expression of SRX1 increase the enzyme proportion in the HMW under steady state, prior to exposure to H₂O₂ (compare Fig. 6 and the Myc-Tsa1 elution profile in Fig. 3A), while completely switching it to the dimer-monomer form in response to H₂O₂. These two genetic conditions both increase the enzyme overall peroxidase efficiency by preventing its inactivation, and therefore must result in lowered cellular H₂O₂ levels at steady state, compared to Myc-Tsa1 expressing cells. Decreased steady state cellular H₂O₂ levels might decrease the rate at which the enzyme is brought into the peroxidatic cycle, i.e. in the dimer-monomer form, hence keeping it in the HMW form. In contrast, in response to H₂O₂ the enzyme remains in the peroxidatic cycle, iteratively disassembling the HMW forms upon disulfide bond formation, hence keeping it in the dimer-monomer form. Of note, deleting the YF-containing C-terminal 2-Cys Prx domain, while it significantly decrease sensitivity to hyperoxidation, does not prevent formation of the two-stacked decamer form, which fit those of a similar mutation in S. mansoni 2-Cys Prx SmPrx1 that assembled as a stable double decamer [2], but not with those of A. thaliana 2-Cys Prx and human Prdx2, the C-Ter deletion of which were shown unable to assemble as a decamer or as HMW structure in response to H₂O₂, respectively [23,29]. Based on the crystal structure of a double decamer, Angelucci and coworkers have suggested that the C-Ter domain sterically prevents the stacking of two decamers when folded, and must therefore be disordered for stacking, which would explain why its removal favors HMW assembly [2,40]. However, a recent cryo-microscopic study of filaments of human Prdx3 obtained at acidic pH indicates that the C-ter domain appears well structured [39].

3. Conclusion

The 2-Cys Prxs family enzymes are H₂O₂-specific antioxidants that carry the unusual feature of enzyme oxidative inactivation at elevated H₂O₂ levels. Although it restricts H₂O₂ scavenging efficiency, inactivation of the enzyme by hyperoxidation constitutes the gateway to its extra-antioxidant functions, and in particular its ability to operate as very efficient chaperone holdases [20]. The critical molecular event allowing the switch from a peroxidase to a chaperone is believed to be the enzyme assembly into HMW structures formed by the stacking of decamers up to filaments, an event that correlates with enzyme hyperoxidation. However, although enzyme hyperoxidation promotes HMW assembly and/or stabilizes these structures, how this event occurs at the molecular level is yet unknown.

To begin answering this question, we have here evaluated the relationship between enzyme quaternary structure and its hyperoxidation in vivo by SEC. A salient result of this set of experiments is the observation that depending on the conditions, Tsa1 elutes as two distinct oligomeric forms, as a disulfide-linked or non-covalently linked dimer, assuming the latter form dissociates during SDS-PAGE, and as a HMW forms that we interpret is constituted of two-stacked decamers. As inferred by combining the data obtained with the native untagged enzyme the Myc-Tagged one, the latter having the advantage of exacerbating quaternary structures changes by decreasing the stability of the HMW form, we propose that Tsa1 exist in cells as a relatively instable two-stacked decamer (seen with native Tsa1 but not with Myc-Tsa1). Exposure to H₂O₂ leads to two outcomes: (i) with untagged and tagged Wt enzymes, after a few peroxidatic cycles C_P becomes hyperoxidized, which stabilizes the HMW form, now also seen with Myc-Tsa1, until the enzyme is recycled by Srx1 that returns it to the relatively instable HMW form; (ii) under conditions preventing sulfinylation, i.e. Myc-Tsa1ΔYF and SRX1 overexpression, the enzyme is instead kept in the dimer-monomer form by entering into iterative peroxidatic cycles, which breaks apart the HMW form until resolution of the H₂O₂ onslaught. Jang, and coworkers used native PAGE to monitor the changes in enzyme quaternary structure triggered by H₂O₂, and showed the presence of two main oligomeric forms of undefined size, one of which must be the dimeric enzyme and the other presumably the two-stacked decamers [20,30]. We confirm here these data, and also provide a visualization of the alternative enzyme paths. We also further demonstrate the causative link between HMW formation/stabilization and C_P sulfinylation. Furthermore, provided that our estimate of the size of the observed HMW as a two-stacked decamer is correct, our data raise the question of whether C_P hyperoxidation triggers the assembly of HMW structures by the stacking of decamers, a view that reflects the literature, or as suggested here, the stabilization of preassembled stacked decamers. Our study also provides the effects of specific mutations on Tsa1 sensitivity to hyperoxidation and on its quaternary structural changes. These mutants and the SEC assay on crude lysates used here should help contribute to disentangle the intricate function of 2-Cys Prxs as antioxidants and chaperone holdases.

4. Experimental procedures

4.1. Yeast strains, plasmids, growth media and standard methods

The S. cerevisiae strains used here are Y252 (Mataura3-52 lys2-801^amberade2-101^ochretrp1-∆1 leu2-∆1) [42], and BY4741 [7] and derivatives (listed in Table 1). Cells were grown at 30 °C in YPD (1% yeast extract, 2% peptone and 2% glucose), or minimal media (SD) (0.67% yeast nitrogen base w/o amino acids, 2% glucose), with amino acid supplements as appropriate. The plasmids used in this study are pRS316-Myc-Tsa1 [6] and derivative pRS316-Myc-TSA1^C48S and pRS316-Myc-TSA1^C171S, that were generated by standard PCR-mediated site-directed mutagenesis using pRS316-Myc-TSA1 as template [28]; pRS316-Myc-Tsa1YF was similarly prepared by deleting the sequence encoding c-terminal last 11 amino acids before codon stop; for pCM190-SRX1, the PCR-amplified SRX1 ORF was cloned between PmeI and NotI downstream of the CYC1-tetracycline-regulatable promoter of pCM190 [15].

Table 1.

Genotype of the S. cerevisiae strains used in this study

Strain	Genotype	Reference source
Y252	Mataura3-52 lys2-801amberade2-101ochretrp1-∆1 leu2-∆1
BY4741	Matahis3∆0 leu2∆0 met15∆0 ura3∆0
BY4742	Matahis3∆0 leu2∆0 lys2∆0 ura3∆0
Δtsa1	BY4741 tsa1D::kanMX4	This work
Δsrx1	srx1Δ::kanMX4	This work
Dtsa1Dsrx1	BY4741 srx1D::natMX4 tsa1D::kanMX4	This work

4.2. Size exclusion chromatography

For extracts preparation, the pellets of exponentially growing cell cultures were resuspended in lysis buffer [PBS pH 7.8, NEM 50 mM, PMSF, Complete protease inhibitors cocktail (ROCHE)] at a density of 2.4×10⁶ cellules/mL. Cells were lysed on a Constant Cell Disruption Systems (CCDS, One Shot, CellD) under a pressure of 2500 bar. Lysates were centrifuged at 10,000g, 4 °C for 10 min, and the supernatant (100 mL) applied on a Superdex 200 10/300 GL (GE Healthcare) column pre-calibrated using a solution of thyroglobulin (670 kDa), apoferritin (443 kDa), catalase (232 kDa), β-amylase (200 kDa), aldolase (158 kDa), ovalbuline (43 kDa), chymotrypsinogene A (27 kDa) and ribonuclease A (13 kDa). Chromatography was performed by High Performance Liquid Chromatography on an AKTA purifier (AmershamPharmaciaBiotec) at a rate of 500 μL/min at room temperature. PBS was used for the elution.

4.3. Western blot

For analysis of enzyme sulfinylation, lysates were prepared by the TCA lysis protocol, and separated by reducing 12 % SDS-PAGE, as described [13]. For analysis of SEC eluates, collected fractions (500 μL) were concentrated on a StrataClean resin (10 mL), proteins were eluted from the resin into Laemli buffer [2% SDS, 62.5 mM Tris–HCl pH 8.7, 10% glycerol, 0.01% bromophenol blue], assessed for protein content, and loaded on non-reducing 12% SDS-PAGE. Gels were then transferred to a nitrocellulose membrane. Membranes were probed with the following primary antibodies anti-Myc (9E10), anti-Tpx1 (a gift from Drs Yang Sol Lee and Ho Hee Jang, Korea), and anti-Prx-SO_2/3 (Ab16830, Abcam), which were revealed using fluorescent anti-mouse IgG or anti-rabbit IgG secondary antibodies and analyzed on the Odyssey infrared imaging system and software (Odyssey, LI-COR).