In vivo observation of peroxiredoxins oligomerization dynamics

Peroxiredoxins (Prxs) are the most common and widely distributed peroxidases found in nature. While they are diverse in sequence and quaternary structure, they all share a common theme in their catalytic mechanism: the peroxidatic cysteine (C_P) reacts with hydroperoxides to form a sulfenic acid, which is frequently reduced back through a resolution step involving the formation of a disulfide with a second cysteine residue (C_R). The location of this second key cysteine, among other sequence features, allows the classification of Prxs in subfamilies (1). Some Prxs form an intramolecular disulfide; therefore, the entire catalytic machinery is present in the same polypeptide. However, the most abundant Prxs in mammalian cytosols belong to a subfamily named “AhpC-Prx1” due to its most prominent members, bacterial AhpC and mammalian Prx1 and Prx2, peroxidases with key roles in antioxidants defenses and redox signaling (2–4). These Prxs are obligated head-to-tail dimers, as the resolutive step requires the presence of a companion subunit that provides C_R. When the disulfide is formed between C_R and C_P, the dimers are covalently linked, something that has been extensively used to study Prx cycle in vivo (5). Although each dimer contains two complete catalytic units, most Prxs from this family form doughnut-like oligomers of 10 to 12 subunits. These toroidal-shaped assemblies were early on linked to the Prxs mechanism. The model states that reduced Prxs are decamers formed by the association of dimers, oxidation triggers a conformational change that leads to disulfide formation and decamer disassembly, releasing oxidized dimers, which are usually reduced by thioredoxin (Trx). Then, reduced dimers regroup into decamers (Fig. 1A). This model was quickly accepted thanks to numerous in vitro and structural studies published on AhpC, Prx1, and Prx2, among others (6, 7). Prx’s redox mechanism and its oligomeric states have been linked since then: peroxide reduction implies oligomeric changes. Moreover, several mutations or posttranslational modifications (PTMs) were shown to affect Prx catalytic properties by not only shifting the oligomeric state of the enzyme but also, triggering changes on Prxs functions (8–10). However, how Prxs oligomeric states exchange and respond to intra- or extracellular stimuli in vivo has been hard to grasp, as most of the techniques applied so far used cellular lysis or invasive approaches (11). In PNAS, Pastor-Flores et al. (12) present a simple and elegant way of studying the connection between redox steps in Prx mechanism and their oligomeric state in vivo.

Fig. 1.

(A) Schematic representation of the typical 2-Cys Prxs catalytic cycle. Note that Prx oxidation and resolution lead to an intermolecular disulfide bond between two Prx monomers. Changes in quaternary structure during the catalytic cycle are represented for the Prx-mCer fusion protein. (B) Typical anisotropy time course after a bolus addition of hydroperoxide in living cells as measured by Pastor-Flores et al. (12). Different stages of the catalytic cycle are colored as in A.

Shortly, the technique used by Pastor-Flores et al. (12) relies on a physical phenomenon called homo-FRET. FRET is short for Förster resonance energy transfer, a type of energy transfer between fluorophores, where one fluorophore (donor) is excited while recording the emission of a second fluorophore (acceptor) whose absorption spectra overlap with the emission of the donor. The rationale behind homo-FRET is that both donor and acceptor are the same fluorophore. Pastor-Flores et al. (12) decided to fuse the cyan fluorescent protein mCerulean (mCer) to Prx1 and Prx2, as it was previously applied to study in vivo ligand-induced protein oligomerization (13). Expressed mCer-Prx will form a decamer when reduced, but as all fluorescent tags are the same, conventional fluorescence spectroscopy will not show any difference when oxidation leads to the disassembly of the decamer. However, if the excitation light is polarized, now the emitted light will report on the crowding of the fluorophore and therefore, about the oligomeric state of the fluorophore carrier. That is due to fluorescent polarization (FP) or anisotropy: a fluorophore excited with polarized light will emit polarized light that can either reach the detector or be absorbed by a proximal fluorophore. Emitted light will have a different polarization “angle” with respect to excitation light due to rotation of the fluorophore during the time that the excited state exists (14). This can be recorded with special emission filters that detect only two perpendicular angles of the emitted light. In the context of the decamer of mCer-Prx2, after excitation with polarized light of a certain wavelength, several cycles of emission, absorption, and reemission will happen; on average, emitted light will have lost polarization, and FP recorder will be close to zero. When the fluorophores part away, as in the case of decamer disassembly due to oxidation, FP increases because less polarized emitted light is “consumed” by close fluorophores and reaches the detector (∆polarization) (Fig. 1B).

First, the authors compared oligomerization and kinetic properties in response to hydrogen peroxide (H₂O₂) of wild type Prx2 and the mCer-Prx2 probe in vitro, verifying that the presence of mCer does not alter the normal behavior of the enzyme. Then, by using a microplate reader to quantitatively monitor FP, they were able to confront living cells with different stimuli (i.e., bolus addition of H₂O₂ and/or inhibition of the Trx or glutathione reducing systems) and to follow the probe response in real time. In a typical anisotropy time course, the ensemble probe signal showed close to zero anisotropy values until a bolus of oxidant was added, which caused a virtually instantaneous positive change in the signal slope for a few minutes until reaching maximum anisotropy values in a dose-dependent manner. This change is interpreted as a consequence of the oxidation and resolution phases, thus yielding the dimeric form of the probe. Then, if the Trx/Trx reductase system was active, the reduction of mCer-Prx dimers and decamers reassembly returns the anisotropy signals to zero values (Fig. 1B). The mCer-Prx probes behave as decamers under basal conditions in the utilized cells cultures and accumulate as dimers in those exposed to either relatively high concentrations of H₂O₂ (≥50 µM) under conditions of adequate NADPH supply or lower concentrations of the oxidant when regeneration of NADPH was impaired. A critical point to be addressed in next studies is to evaluate if these changes in the quaternary structure also occur in cells exposed to “more physiological” oxidant conditions, either using exogenous fluxes of hydroperoxides or generated intracellularly by physiological stimuli. To note, no oxidant addition is required to oxidize the probe in cells that lack Trx reductase activity. Considering that oligomeric state in Prxs is linked to its activity, the technique would become a very powerful tool to investigate the cellular response of Prxs behavior to possible direct or indirect inhibitors or to systematically screen compound libraries targeting these enzymes. In this sense, the authors tested Conoidin A and Adenanthin, two compounds that were reported to selectively inhibit Prxs (15, 16), obtaining promising results on characterizing or discarding the inhibitors potential to affect the quaternary structure of the protein.

The results presented by Pastor-Flores et al. (12) open a window to explore an old question in Prx biology: why do Prxs form decamers? Prxs do not need to be decamers, as the entire catalytic unit is contained in the tightly associated dimer. A key observation made by the authors is that complete dissociation of Prx2 decamers into dimers addressed by FP is not paralleled by the complete formation of covalent dimers measured by denaturing electrophoresis. Decamers rapidly disassemble when H₂O₂ is added, even under conditions where not all active sites get oxidized. This may appear as a weakness in the Prx mechanism, as the oxidation of a few active sites knocks out a specialized peroxidase when most of its antioxidant capacity is still highly loaded. However, an alternative explanation for this behavior becomes evident when considering that most of the decameric Prxs, while powerful peroxidases, evolve to be ultrasensitive peroxide sensors in the context of several other robust peroxidases. Decamerization allowed Prx catalytic units, the dimers, to evolve contact sites that mediate dimer-to-dimer communication. These surfaces, described early on (17), set side-by-side active sites from different subunits, allowing a kinetic behavior that resembles “negative cooperativity.” Still, some fundamental questions remain unanswered in these regards: Is decamerization functioning as a strategy to improve Prx interactions with redox secondary targets and/or scaffolding proteins (4)? Is the decamer reassembly facilitated by other proteins?

The work also includes very important issues regarding regularly used point mutations in the Prxs field. The substitution of C_R by serine or alanine drastically changes the oligomeric properties of Prx1 and Prx2. C_R is not located at the dimerization interface, suggesting that subtle changes, mutations, and/or PTMs at the C terminus region may have very significant effects on the oligomeric properties. For that reason, this aspect must be taken into consideration when interpreting past and future experiments using mutants of these enzymes.

A final note about the possibility of alternative oligomeric assemblies is that, after the addition of high doses of oxidant, negative polarization values were observed for the mCer-Prx2 probe but not mCer-Prx1. This was interpreted as the detection of higher-order oligomeric states like stacked rings of decamers. These assemblies had been detected in Prx2 in vitro, triggered by overoxidation, and they had been associated with the potential chaperone-like activity of Prxs (18, 19). The observation is also in agreement with previous studies that demonstrate that Prx2 is more sensitive to hyperoxidation than Prx1 (20).

Beyond the evidence Pastor-Flores et al. (12) present, the beauty of their work lies on many open paths of opportunities to be transited in the future. Surely, the microplate reader setup would be adapted and tuned for other protein oligomerization and/or crowding studies. Also, the fused protein designed by the authors constitutes a very helpful tool to develop imaging experiments using dynamic and static fluorescence microscopy, which would allow investigation of Prxs interactions with other biomolecules in living cells (21). So, the report by Pastor-Flores et al. (12) represents a cornerstone in the development of methodologies for the Prxs field, as well as in other cell biology matters.