Cysteines 208 and 241 in Ero1α are required for maximal catalytic turnover

Abstract

Endoplasmic reticulum (ER) oxidoreductin 1α (Ero1α) is a disulfide producer in the ER of mammalian cells. Besides four catalytic cysteines (Cys⁹⁴, Cys⁹⁹, Cys³⁹⁴, Cys³⁹⁷), Ero1α harbors four regulatory cysteines (Cys¹⁰⁴, Cys¹³¹, Cys²⁰⁸, Cys²⁴¹). These cysteines mediate the formation of inhibitory intramolecular disulfide bonds, which adapt the activation state of the enzyme to the redox environment in the ER through feedback signaling. Accordingly, disulfide production by Ero1α is accelerated by reducing conditions, which minimize the formation of inhibitory disulfides, or by mutations of regulatory cysteines. Here we report that reductive stimulation enhances Ero1α activity more potently than the mutation of cysteines. Specifically, mutation of Cys²⁰⁸/Cys²⁴¹ does not mechanistically mimic reductive stimulation, as it lowers the turnover rate of Ero1α in presence of a reducing agent. The Cys²⁰⁸/Cys²⁴¹ pair therefore fulfills a function during catalysis that reaches beyond negative regulation. In agreement, we identify a reciprocal crosstalk between the stabilities of the Cys²⁰⁸–Cys²⁴¹ disulfide and the inhibitory disulfide bonds involving Cys¹⁰⁴ and Cys¹³¹, which also controls the recruitment of the H₂O₂ scavenger GPx8 to Ero1α. Two possible mechanisms by which thiol–disulfide exchange at the Cys²⁰⁸/Cys²⁴¹ pair stimulates the catalytic turnover under reducing conditions are discussed.

Graphical abstract

Highlights

• •Reductive stimulation enhances Ero1α more potently than cysteine mutations.
• •Cys²⁰⁸/Cys²⁴¹ controls Ero1α activity beyond negative regulation.
• •Other regulatory cysteines communicate with Cys²⁰⁸/Cys²⁴¹ within Ero1α.
• •Other regulatory cysteines control the binding of GPx8 to Ero1α through Cys²⁰⁸/Cys²⁴¹.

1. Introduction

The synthesis of extracellular proteins is initiated at the endoplasmic reticulum (ER) where translating ribosomes translocate the nascent polypeptides into the ER lumen. Many of these polypeptides subsequently acquire critical covalent linkages between cysteine residues (termed disulfide bonds) through thiol-disulfide exchange reactions. Among the different systems for disulfide bond formation in the ER, ER oxidoreductin 1 (Ero1)-catalyzed oxidation of the active-site cysteine pair in protein disulfide isomerase (PDI) constitutes the best-conserved pathway [1], [2]. Ero1 in vertebrates exists in two isoforms, Ero1α and Ero1β, whereas Ero1α is ubiquitously expressed and being viewed as the major source of disulfides in humans [3].

The flavoprotein Ero1α is an oxidase that couples the reduction of molecular oxygen (O₂) to disulfide-bond formation [3]. In the reductive phase of the Ero1α catalytic cycle, the flavin adenine dinucleotide (FAD) cofactor in Ero1α is reduced to FADH₂ as a result of disulfide transfer from Ero1α to reduced PDI (PDI_red). The reaction is catalyzed by an outer active-site cysteine pair (Cys⁹⁴/Cys⁹⁹), which shuttles two electrons from PDI_red via an inner active-site cysteine pair (Cys³⁹⁴/Cys³⁹⁷) to FAD. This process is tightly regulated by the reversible formation of two inhibitory disulfides (Cys⁹⁴–Cys¹³¹ and Cys⁹⁹–Cys¹⁰⁴) [4], [5], [6]. In the oxidative phase, FAD is regenerated by the transfer of two electrons onto O₂, which leads to the formation of one molecule of hydrogen peroxide (H₂O₂) [7]. Recent evidence indicates that H₂O₂ is instantaneously reduced to H₂O by the Ero1α-associated glutathione peroxidase family enzymes GPx7 or GPx8, which also introduce the resulting second disulfide into PDI [8], [9]. We showed that access of O₂ to the buried FAD molecule is negatively regulated by the Cys²⁰⁸–Cys²⁴¹ disulfide [10].

The thiol-disulfide statuses of all inhibitory disulfide bonds in Ero1α are governed by canonical PDI or other PDI family members [10], [11]. Accordingly, regulatory cysteines fine-tune the activation state of Ero1α in a redox environment-dependent manner by blocking either the reductive (Cys¹⁰⁴, Cys¹³¹) or the oxidative (Cys²⁰⁸, Cys²⁴¹) phase of the catalytic cycle through the formation of feedback-regulated inhibitory disulfides in response to oxidizing conditions. Conversely, a reducing ER environment promotes Ero1α activation through PDI-catalyzed reduction of these disulfides, as could for instance be important in response to physiological hypoxia or peak concentrations of reduced glutathione (GSH) or redox-active vitamins.

Here, we report that reducing conditions, which activate Ero1α through the removal of inhibitory disulfides, more potently stimulate Ero1α activity than the mutation of all regulatory cysteines. Thus, the presence of the Cys²⁰⁸/Cys²⁴¹ pair is required for maximal catalytic turnover under reducing conditions. Our data indicate a new mechanism of Ero1α regulation in which thiol-disulfide exchange at Cys²⁰⁸–Cys²⁴¹ affects the stability of the Cys⁹⁴–Cys¹³¹ inhibitory disulfide through allosteric and/or intermolecular communication.

2. Materials and methods

2.1. Fluorescence excitation spectrum analysis

Cells stably transfected with HyPer_ER were subjected to fluorescence excitation spectrum analysis as described before [12].

2.2. Dithiothreitol (DTT) washout assays

The cellular glutathione disulfide:total glutathione (GSSG:GS_tot) ratio after DTT washout was measured using a 5,5'-dithiobis(2-nitrobenzoic acid)/glutathione reductase recycling assay as previously described [13].

2.3. Statistics

Data sets were analyzed for statistical significance using Student's t test (two-tailed distribution; heteroscedastic).

2.4. Cell culture and transient transfections

The culturing of HeLa cells [14] and FlipIn TRex293 cells for doxycycline (1 μg/ml, Sigma)-inducible expression of Ero1 variants [4] has been described. The following FlipIn TRex293 cell lines have been published previously: Ero1α [4], Ero1α-AA [6], Ero1α-C208S/C241S, Ero1α-AASS [10], Ero1α-AA:HyPer_ER [8], Ero1α-C208S/C241S:HyPer_ER [10] and Ero1α-AASS:HyPer_ER [10]. The Ero1α−WT:HyPer_ER cell line was created as before [8] (with the HyPer_ER vector kindly provided by Miklos Geiszt, Semmelweis University, Hungary).

Transient transfections of HeLa cells were carried out using Turbofect (Thermo Scientific). Transient transfections of FlipIn TRex293 cells were carried out using Metafectene Pro (Biontex).

2.5. Oxygen and NADPH consumption assay

Oxygen consumption was measured as previously described [15]. Experimental procedure of NADPH consumption assay has been described before [16], [17].

2.6. Determination of the redox equilibrium constant (Keq)

K_eq values of Cys⁹⁴–Cys¹³¹ in Ero1α-WT and Ero1α-C208S/C241S were determined as follows. Ero1α-WT or Ero1α-C208S/C241S (2 μM) was incubated for 30 min at 30 °C in degassed 50 mM Tris/HCl (pH 7.5) buffer containing 300 mM NaCl, 0.2 mM GSSG and various concentrations of GSH (0.2–4 mM). After incubation, 1 mM NEM was added to avoid subsequent redox reactions. All the samples were separated by non-reducing SDS-PAGE and stained with Coomassie brilliant blue (CBB). The values of fraction of OX2 state of Ero1α were estimated by LAS-3000 image reader, and plotted against [GSH]²/[GSSG] ratios. K_eq values were determined by fitting the data according to equation:

$$R=\left({\left[\mathrm{GSH}\right]}^2/\left[\mathrm{GSSG}\right]\right)/\left\{Keq+\left({\left[\mathrm{GSH}\right]}^2/\left[\mathrm{GSSG}\right]\right)\right\}$$

where R is fraction of OX2 state of Ero1α at equilibrium.

2.7. Bimolecular fluorescence complementation (BiFC) assay

HeLa cells were transfected and analyzed as previously described [10]. CRTss+EYFP1+mature Ero1α in pcDNA3.1 (dubbed EYFP1–Ero1α-WT) and CRTss+EYFP2+luminal domain GPx8 in pcDNA3.1 (dubbed EYFP2–GPx8_lum) were kindly provided by Lloyd Ruddock [18]. EYFP1–Ero1α-AA was generated by site-directed mutagenesis.

3. Results and discussion

Recent work has demonstrated that Ero1α-AASS, an Ero1α variant with all four regulatory cysteines being mutated (C104A, C131A, C208S, C241S), displays higher oxidase activity than Ero1α-AA where only Cys¹⁰⁴ and Cys¹³¹ are mutated [10]. A likely mechanism was presented in which PDI_red facilitates the reaction of FADH₂ with O₂ in Ero1α-AA by opening the intramolecular Cys²⁰⁸–Cys²⁴¹ disulfide [10]. Since PDI_red is depleted through the oxidation by Ero1α-AA and its regeneration from oxidized PDI (PDI_ox) is relatively inefficient in the ER at steady state, the reaction rate is limited by the availability of PDI_red. Accordingly, Ero1α-AASS, which lacks Cys²⁰⁸–Cys²⁴¹, generates higher amounts of the reaction products PDI_ox and H₂O₂, as it is active irrespectively of the scarcity of PDI_red [10].

In cells, oxidase activity of wild-type Ero1α (Ero1α-WT) or Ero1α-AA can be stimulated by the addition of the membrane-permeable reductant dithiothreitol (DTT). This manifests in a more prominent detection of the reaction product H₂O₂ in the ER of Ero1α-overexpressing cells [8], [10], [19], [20] and can mechanistically be explained by DTT-stimulated reduction of inhibitory disulfides in Ero1α [4], [5], [6], [10], [21]. In case of Ero1α-AA, DTT-mediated activation is connected to the enhanced opening of Cys²⁰⁸–Cys²⁴¹ by PDI_red (Fig. 1A, left panel) [10].

Fig. 1.

(A) Schematic depicting the influence of DTT on the formation of Ero1α-derived H₂O₂ in the ER. The resulting high concentration of PDI_red exerts a stimulatory effect on Ero1α by catalyzing thiol-disulfide exchange at the regulatory disulfide between Cys²⁰⁸ and Cys²⁴¹ (see main text for detailed models). (B) HyPer_ER fluorescence excitation spectrum analyses of indicated cell lines measured 5 min after the addition of 0.5 mM DTT were performed. Plotted are the ratios of the 500 and 420 nm peak amplitudes (n≥3; mean±SD). *p<0.05; **p<0.01; ***p<0.001.

Based on the data summarized above, Ero1α-AASS might be hyperactive, because it mimics the DTT-activated state of Ero1α-AA. To test this possibility, we examined to what extent DTT influences Ero1α-AASS-catalyzed production of H₂O₂. We used a cell line that expresses Ero1α-AASS only in presence of doxycycline and stably harbors the H₂O₂-sensitive fluorescent protein HyPer in the ER (HyPer_ER) [10], [22]. HyPer_ER, the fluorescence excitation spectrum of which does not overlap with the one of doxycycline [12], is prone to oxidation not only by H₂O₂ but also by PDI_ox [23], [24]. In DTT-treated cells, however, the oxidation of HyPer_ER can basically be ascribed to H₂O₂ only, as DTT raises the concentration of Ero1α-derived H₂O₂ but lowers the levels of PDI_ox [21] (Fig. 1A, left panel). Using this model system in presence of 0.5 mM DTT, we analyzed the fold-increase in HyPer_ER oxidation in cells treated with doxycycline as compared to untreated cells. We found that expression of Ero1α-AASS did not lead to equal or stronger production of H₂O₂ compared to Ero1α-AA. Rather, doxycycline-induced HyPer_ER oxidation was lowered by ~40% in Ero1α-AASS expressing cells (Fig. 1B, bars 6 and 8). Likewise, expression of Ero1α-C208S/C241S trended to induce less prominent HyPer_ER oxidation than that of Ero1α-WT (Fig. 1B, bars 2 and 4), whereas mutation of Cys¹⁰⁴ and Cys¹³¹ increased oxidase activity as expected (Fig. 1B, bars 2 and 6, bars 4 and 8).

These findings were incompatible with the notion that Ero1α-AASS mimics the DTT-activated state of Ero1α-AA. Apparently, the Cys²⁰⁸/Cys²⁴¹ pair is required for full catalytic activity of Ero1α under reducing conditions. Since mutation of these cysteines precludes the thiol-disulfide exchange reactions that engage PDI to the O₂-reducing end of Ero1α, the stimulatory effect of PDI_red (Fig. 1A, left panel) drops out in these mutants. Thus, although poorly defined yet, the mechanism of PDI_red–mediated stimulation of Ero1α-AA cannot solely rely on the removal of the Cys²⁰⁸–Cys²⁴¹ disulfide.

We next examined if these new findings could also be recapitulated in a reconstituted system. Previously, we compared the performance of Ero1α-AA and Ero1α-AASS in oxidizing a one-time bolus of their substrate PDI_red [10]. In this setting, PDI_ox was generated faster in the Ero1α-AASS- compared with the Ero1α-AA-catalyzed reaction. Furthermore, purified Ero1α-AASS produced significantly higher levels of H₂O₂ compared to Ero1α-AA [10]. In order to mimic the situation in the ER of DTT-treated cells where PDI_red is constantly being regenerated, we added an excess of GSH to the reaction (Fig. 2A). This setup matches the commonly used practice to assay the oxidase activity of Ero1α in vitro [5], [7], [15], [16], [17], [25]. Consistent with the results presented in Fig. 1B, the O₂ concentration dropped less rapidly in presence of Ero1α-AASS compared to Ero1α-AA when GSH was added to the reaction (Fig. 2B). Similarly, in an assay that indirectly detects the formation of GSSG by monitoring glutathione reductase-dependent consumption of NADPH [16], Ero1α-AASS displayed lower oxidase activity than Ero1α-AA (Fig. 2C). Altogether, these experiments confirmed that Ero1α-AASS is catalytically hampered relative to Ero1α-AA under reductively stimulated conditions, which contrasts with the situation in the absence of excess reducing agent [10].

Fig. 2.

(A) Schematic depicting the influence of GSH on the catalytic turnover of Ero1α in vitro. The resulting high concentration of PDI_red exerts a stimulatory effect on Ero1α by catalyzing thiol-disulfide exchange at the regulatory disulfide between Cys²⁰⁸ and Cys²⁴¹ (see main text for detailed models). (B) O₂ consumption was monitored over time in a mixture of 2 μM Ero1α-AA (solid line) or Ero1α-AASS (dashed line), 10 μM PDI, and 10 mM GSH. (C) Consumption of NADPH coupled to catalysis by Ero1α-AA (solid line) or Ero1α-AASS (dashed line) (see Materials and Methods; 2 μM Ero1α variants, 10 μM PDI, 1 mM GSH, 1 U glutathione reductase, 200  μM NADPH) was detected by following the absorbance at 340 nm.

Oxidase activity of Ero1α in the ER of live cells can also be assessed using a DTT washout approach that is coupled to the time-resolved quantification of the cellular GSSG:GS_tot [8], [10], [13]. Because Ero1α is activated by DTT (Fig. 1A), the levels of PDI_red drop and GSSG builds up rapidly in response to DTT removal through the Ero1α–PDI–GSH cascade (Fig. 3A). Furthermore, the peroxidase-catalyzed reduction of H₂O₂ contributes to the build-up of GSSG [8] (not depicted in the figure). Ectopic over-expression of Ero1α is evidently reflected in the time-course of GSSG:GS_tot after DTT washout. Thus, in cells that over-express Ero1α-WT, GSSG:GS_tot transiently rises to about 500% of its steady-state value in the time-scale of seconds [13], [26]. This massive accumulation of GSSG is followed by a decline and the restoration of the steady-state GSSG:GS_tot in the time scale of minutes, which depends on cytosolic glutathione reductase [8] and the reformation of the inhibitory Cys²⁰⁸–Cys²⁴¹ disulfide in Ero1α [10].

Fig. 3.

(A) Schematic depicting the reactions leading to GSSG formation following DTT washout in Ero1α-overexpressing cells. (B) GSSG/GS_tot recovery curves upon DTT washout were compared between Ero1α-WT-, Ero1α-AA-, Ero1α-AASS- and Ero1α-C208S/C241S-expressing cells. (mean±SEM; two independent experiments each performed at least in doublet). Statistically significant differences in the 0 and 60 s time point are indicated in the case of Ero1α-AA (compared to Ero1α-WT; **p<0.01) and in the case of Ero1α-AASS (compared to Ero1α-AA; *p<0.05).

Consistent with the results in Fig. 1B where over-expressed C208S/C241S mutants displayed lower oxidase activity than their unmutated counterparts in presence of DTT, mutation of the Cys²⁰⁸/Cys²⁴¹ pair decreased the Ero1α-dependent accumulation of GSSG upon DTT washout (Fig. 3B, compare the solid and the dashed lines). Unexpectedly, however, over-expression of Ero1α-AA led to a smaller peak in GSSG:GS_tot upon DTT washout compared to over-expression of Ero1α-WT (Fig. 3B), although Ero1α-AA is catalytically more active than Ero1α-WT under reductively stimulated conditions (Figs. 1B, 2C and 4D). This may suggest that the decline of GSSG:GS_tot in Ero1α-AA-expressing cells has already proceeded to significant extent before the assay period. It might also be possible that the constitutive absence of the regulatory Cys⁹⁴–Cys¹³¹ and Cys⁹⁹–Cys¹⁰⁴ disulfides in the C104A/C131A mutant affects the Ero1α activity by altering the stability of the Cys²⁰⁸–Cys²⁴¹ disulfide in an allosteric and/or intermolecular manner (see the following paragraphs for more details).

Fig. 4.

(A) Redox equilibrium of Ero1α-WT, Ero1α-C208S/C241S and Ero1α-C94A/C99A/C104A/C131A/C208S/C241S in presence of GSH/GSSG. All the samples were incubated for 30 min with different GSH/GSSG ratios under anaerobic condition at 30 °C and modified by NEM. The redox states of the Ero1α derivatives were separated by non-reducing SDS-PAGE and stained with CBB. “OX2” indicates the Ero1α species with the Cys⁹⁴–Cys¹³¹ regulatory disulfide. In “OX1”, this disulfide is reduced or absent. “RE” indicates an Ero1α species in which not only Cys⁹⁴–Cys¹³¹ but also unidentified additional disulfide(s) are reduced. Note that the mobility of the “RE” species of Ero1α-WT or Ero1α-C208S/C241S is not exactly identical to that of Ero1α-C94A/C99A/C104A/C131A/C208S/C241S on a SDS gel, probably due to the presence of Cys⁹⁹-Cys¹⁰⁴ disulfide in the former two. (B) Redox states of Ero1α-WT and Ero1α-C208S/C241S in wider ranges of GSH/GSSG ratio were analyzed as in (A). (C) Fraction of Ero1α in OX2 form is plotted against [GSH]²/[GSSG] ratios. K_eq values were determined from at least two independent experiments as follows: 10.7±0.6 for Ero1α-WT (solid line) and 12.5±0.7 for Ero1α-C208S/C241S (dashed line) (mean±SD). (D) Consumption of NADPH coupled to catalysis of PDI oxidation by Ero1α-WT (solid line) or Ero1α-C208S/C241S (dashed line) (see Materials and Methods; 2 μM Ero1α variants, 10 μM PDI, 1 mM GSH, 1 U glutathione reductase, 200 μM NADPH).

To address the potential crosstalk between the Cys²⁰⁸/Cys²⁴¹ pair and the Cys⁹⁴–Cys¹³¹ regulatory disulfide bond, we measured the reduction potential of Cys⁹⁴–Cys¹³¹ in Ero1α-WT and Ero1α-C208S/C241S. As shown in Fig. 4A, purified Ero1α-WT was predominantly in an inactive form (OX2), which harbors the Cys⁹⁴–Cys¹³¹ disulfide [4] and migrates faster on a non-reducing SDS gel than the active OX1 form comprising the active-site Cys⁹⁴–Cys⁹⁹ disulfide. Incubation of Ero1α-WT and Ero1α-C208S/C241S in various GSH/GSSG ratios followed by non-reducing SDS-PAGE revealed that indeed, the Cys²⁰⁸/Cys²⁴¹ pair influences the stability of the Cys⁹⁴–Cys¹³¹ disulfide. The C208S/C241S mutations slightly but significantly stabilized the Cys⁹⁴–Cys¹³¹ disulfide; the redox equilibrium constant (K_eq) of Cys⁹⁴–Cys¹³¹ in Ero1α-WT is 10.7±0.6 mM and that in Ero1α-C208S/C241S is 12.5±0.7 mM (Figs. 4B and C). It is thus conceivable that the cleavage or absence of Cys²⁰⁸–Cys²⁴¹ leads to a higher propensity for the inactive OX2 form. In agreement, Ero1α-C208S/C241S was less active in oxidizing PDI than Ero1α-WT (Fig. 4D).

We next analyzed the binding of the GPx8 peroxidase to the Cys²⁰⁸/Cys²⁴¹ region in Ero1α [10]. In the bimolecular fluorescence complementation assay, the fluorescence of two combined yellow fluorescent protein half sites, which are fused with Ero1α and the ER-luminal domain of GPx8, respectively, is quantified [18] (Fig. 5A). Similar to the formation of Cys²⁰⁸–Cys²⁴¹, docking of GPx8 is a process that takes place on the distal side of Ero1α relative to the active sites and the cofactor-binding site (Fig. 5A). The binding intensity of GPx8 was influenced by the presence or absence of the regulatory Cys⁹⁴–Cys¹³¹ and Cys⁹⁹–Cys¹⁰⁴ disulfides: EYFP1–Ero1α-WT recruited ~50% more EYFP2–GPx8_lum than EYFP1–Ero1α-AA (Fig. 5B). According to our previously proposed model, the recruitment of GPx8 occurs in response to the opening of Cys²⁰⁸–Cys²⁴¹ by PDI_red [10]. The data therefore underscored the notion that formation of the regulatory Cys⁹⁴–Cys¹³¹ disulfide in Ero1α-WT destabilized the Cys²⁰⁸–Cys²⁴¹ disulfide, thereby facilitating the recruitment of GPx8.

Fig. 5.

(A) Schematic representation of the interaction of the luminal domain of GPx8 with the loop region between Cys²⁰⁸ and Cys²⁴¹ in Ero1α (left panel) and the Bi-molecular fluorescence complementation (BiFC) of two YFP half-sites fused to either Ero1α or GPx8 (right panel). (B) 18 h after transfection with the indicated constructs, HeLa cells were trypsinized and analyzed by flow cytometry for BiFC fluorescence (n≥5; mean±SD). a.u. arbitrary unit, **p<0.01.

Altogether, two possible mechanisms underlying our finding that the Cys²⁰⁸/Cys²⁴¹ pair is required for maximal catalytic turnover in Ero1α are proposed. In a first model, stimulation of the Ero1α catalytic turnover can be mediated by the destabilization of Cys⁹⁴–Cys¹³¹ and possibly Cys⁹⁹–Cys¹⁰⁴ in response to the formation of Cys²⁰⁸–Cys²⁴¹. The present data strongly suggest that the stabilities of the Cys²⁰⁸–Cys²⁴¹ disulfide and the inhibitory disulfide bonds involving Cys¹⁰⁴ and Cys¹³¹ influence each other in reciprocal manner. The resulting thiol-disulfide oscillations at Cys²⁰⁸/Cys²⁴¹ that are a predicted consequence of this model could be catalyzed by PDI or by other PDI family members such as ERp57 [10]. In this connection, a minor fraction of purified Ero1α can transiently be detected with fully reduced Cys²⁰⁸/Cys²⁴¹ pair upon addition of PDI_red, demonstrating that at least in vitro the dithiol form of Cys²⁰⁸/Cys²⁴¹ is generated during the Ero1α catalysis of PDI oxidation (Kanemura, S. and Inaba, K., manuscript in preparation).

In a second model, the specific architecture of the covalent Cys²⁰⁸/Cys²⁴¹-linked Ero1α–PDI complex (previously termed Ero1α–PDI^fast [10]) is important, e.g. for the optimal channeling of O₂ to the active site. This possibility would be in agreement with a previous working model of the Ero1α catalytic cycle, which accommodated the remarkably high levels of Ero1α–PDI^fast in the ER [10]. According to this working model, Cys²⁰⁸/Cys²⁴¹ can hardly be found in the ER in the reduced dithiol state that resembles Ero1α-AASS, as Ero1α–PDI^fast is kinetically stabilized. In the same vein, the fact that Ero1α-AA but not Ero1α-AASS associates with the natural binding partner GPx8 through its Cys²⁰⁸/Cys²⁴¹ region (likely in the context of Ero1α–PDI^fast) underscores the non-native behavior of Ero1α-AASS [10].

It is important to note that these two mechanistic explanations are not exclusive to each other. Thus, a slightly modified, “mixed” model is possible where it is the formation of the Ero1α–PDI^fast complex that stabilizes the inhibitory disulfide bonds involving Cys¹⁰⁴ and Cys¹³¹. Conversely, reduction of Cys⁹⁴–Cys¹³¹ and Cys⁹⁹–Cys¹⁰⁴ could promote the reformation of Cys²⁰⁸–Cys²⁴¹ from Ero1α–PDI^fast. It will be interesting to more closely characterize the communication between the opposing Cys⁹⁴/Cys¹³¹/Cys⁹⁹/Cys¹⁰⁴ end and the Cys²⁰⁸/Cys²⁴¹ end of Ero1α. It could be based on allostery or, alternatively, intermolecular catalysis either directly between two Ero1α molecules or indirectly via the redox state of PDI or GSH.

In this work, we present the new finding that Ero1α-AASS, which likely adopts a conformation equivalent to fully reduced Ero1α (disregarding catalytic and structural disulfides), does not reach the maximal catalytic turnover rate. Rather, the highest intrinsic activity is reached in presence of the Cys²⁰⁸/Cys²⁴¹ pair and a reducing agent, which ensures unlimited supply of PDI_red. We conclude that Cys²⁰⁸/Cys²⁴¹ plays an unexpected stimulatory role during catalysis. Cys²⁰⁸/Cys²⁴¹-dependent stimulation could be due to optimized O₂ penetration into the Ero1α–PDI^fast complex and/or the destabilization of Cys⁹⁴–Cys¹³¹ and Cys⁹⁹–Cys¹⁰⁴ in response to the oxidation of the Cys²⁰⁸/Cys²⁴¹ pair. Future experiments designed to elucidate these mechanistic possibilities will further increase our understanding of regulated disulfide-bond formation in the ER.