Chloroplast translation factor EF-Tu of Arabidopsis thaliana can be inactivated via oxidation of a specific cysteine residue

Photosynthetic organisms are subjected to oxidative stress when they are exposed to strong sunlight. A major site of the production of reactive oxygen species (ROS) is the thylakoid membranes in which the photosynthetic machinery is embedded. ROS are produced as by-products of the photosynthetic transport of electrons and transfer of excitation energy [1]. The superoxide anion radical (⁠$O2−$⁠) is produced on the acceptor side of photosystem I when excess electrons are transferred to oxygen molecules. Then $O2−$ is converted to hydrogen peroxide (H₂O₂), either by superoxide dismutase or spontaneously, and H₂O₂ can be converted to the hydroxyl radical (•OH) via the Fenton reaction in the presence of reduced metal ions, such as Fe²⁺ and Cu⁺. Singlet oxygen (¹O₂) is produced mainly in photosystem II (PSII) when the excitation energy is transferred from triplet-state chlorophyll to molecular oxygen. The production of these ROS is markedly promoted when the photosynthetic machinery is exposed to strong light, with resultant oxidative stress [1].

Under strong light, PSII is susceptible to inactivation. This phenomenon is referred to as the photoinhibition of PSII [2]. Although the action of ROS in the photoinhibition of PSII has not been fully understood, recent studies have suggested that the repair of photodamaged PSII is particularly sensitive to inhibition by ROS under strong light [3]. In the cyanobacterium Synechocystis sp. PCC 6803 (hereafter, Synechocystis), the inhibition by ROS of repair of PSII has been attributed to the suppression of the synthesis de novo of proteins, at the level of translational elongation, that are required for the repair of PSII, such as the D1 protein [4,5]. Biochemical studies have revealed that translation factors EF-G and EF-Tu, key factors for translational elongation, might be the critical targets by ROS within the translational machinery in Synechocystis [6,7]. EF-G, an elongation factor that translocates peptidyl-tRNA from the A site to the B site of the ribosome, is inactivated via the oxidation of Cys114 and Cys266, with the resultant formation of an intramolecular disulfide bond [8]. EF-Tu, another elongation factor that delivers aminoacyl-tRNA to the A site of the ribosome, is inactivated via oxidation of Cys82, with the resultant formation of sulfenic acid and an intermolecular disulfide bond [7]. Expression in Synechocystis of mutated EF-G, in which Cys114 had been replaced by a serine residue, mitigated the photoinhibition of PSII via acceleration of the synthesis de novo of proteins, including the D1 protein, and the subsequent repair of PSII under strong light [9]. Similarly, expression in Synechocystis of mutated EF-Tu, in which Cys82 had been replaced by a serine residue, mitigated the photoinhibition of PSII via acceleration of protein synthesis and repair of PSII [10]. Thus, the sensitivity to oxidation of the target cysteine residues of these elongation factors appears to reflect the sensitivity of PSII repair to oxidative stress. The effect of mutated EF-Tu on the protection of PSII repair was more significant than that of mutated EF-G, suggesting that EF-Tu might play a more critical role than EF-G in determining the sensitivity of protein synthesis to oxidative stress that leads, in turn, to the inhibition of the repair of PSII [10].

In the chloroplasts of land plants, translational elongation is a significant step in the regulation of the synthesis of the D1 protein during illumination [11,12], suggesting that the relevant elongation factors in the chloroplast might be regulated under strong light. However, the sensitivity of chloroplast elongation factors to oxidative stress remains to be clarified. In Arabidopsis thaliana, a chloroplast-localized EF-Tu (hereafter, cpEF-Tu) is encoded in the nuclear genome [13], and the mature cpEF-Tu includes two cysteine residues, Cys149 and Cys451, after removal of the signal peptide via a putative cleavage site. Cys149 in cpEF-Tu corresponds to Cys82 in Synechocystis EF-Tu, which is the target of ROS [7], and this cysteine residue is strongly conserved in the EF-Tu of various organisms [14].

In the present study, we studied the sensitivity to H₂O₂ of recombinant cpEF-Tu protein from Arabidopsis. We found that cpEF-Tu was inactivated via the oxidation of Cys149 with the resultant formation of sulfenic acid and that the oxidized cpEF-Tu was reduced and reactivated by thioredoxin.

We purified recombinant cpEF-Tu with a histidine tag at the carboxyl terminus of the mature protein that lacked the putative signal peptide for the chloroplast. Since the sensitivity to oxidation of EF-Tu from Synechocystis depends on the nucleotide that is bound to it [7], we first analyzed nucleotides that had bound to cpEF-Tu. We found that the purified proteins did not bind any detectable nucleotides (Supplementary Figure S1), and they were termed nucleotide-free cpEF-Tu. From nucleotide-free cpEF-Tu, we prepared recombinant cpEF-Tu that bound GDP or GTP. The GDP-bound form contained GDP at a molecular ratio of GDP to the protein of ∼0.9, while the GTP-bound form contained GTP at a molecular ratio of GTP to the protein of ∼0.6 and, also, GDP at the ratio of ∼0.1 (Supplementary Figure S1). We termed the respective proteins cpEF-Tu-GDP and cpEF-Tu-GTP.

There are two cysteine residues in the mature protein of cpEF-Tu, namely, Cys149 and Cys451. We monitored the redox state of these cysteine residues in cpEF-Tu. After cpEF-Tu proteins had been treated with 30 mM DTT or H₂O₂ at various concentrations, they were incubated with a maleimidyl reagent for modification of the thiol groups of cysteines residues and were then subjected to non-reducing SDS–PAGE. When nucleotide-free cpEF-Tu was treated with DTT, the apparent molecular mass of cpEF-Tu increased, a result that indicated that the two cysteine residues had been reduced and modified (Figure 1A). In contrast, when nucleotide-free cpEF-Tu was treated with H₂O₂, proteins with a molecular mass between those of unmodified and fully modified cpEF-Tu appeared and their levels rose as the concentration of H₂O₂ was raised (Figure 1A,D). These observations suggested that one of the two cysteine residues in cpEF-Tu had been oxidized by H₂O₂ and had become unable to bind the maleimidyl reagent. Similarly, treatment of cpEF-Tu-GDP with H₂O₂ generated an oxidized form in which one of the two cysteine residues had been oxidized, but the sensitivity of cpEF-Tu-GDP to H₂O₂ was less marked than that of nucleotide-free cpEF-Tu (Figure 1B,D). Treatment of cpEF-Tu-GTP with H₂O₂ generated such an oxidized form but the level of oxidized form remained low even at high concentrations of H₂O₂, suggesting that cpEF-Tu-GTP might be more resistant to oxidation than nucleotide-free cpEF-Tu and cpEF-Tu-GDP (Figure 1C,D). It should be noted that cpEF-Tu was very unstable and that both 10 mM DTT and 30% (w/v) glycerol were required to prevent aggregation: proteins started aggregating within a day at 4°C in the presence of DTT and glycerol, and even during storage at −80°C in a few days.

Of the two cysteine residues that we examined, Cys149 corresponds to Cys82 in EF-Tu of Synechocystis, which was identified as the residue that is sensitive to oxidation [7]. This cysteine residue is strongly conserved in the EF-Tu and cpEF-Tu of bacteria [14] and photosynthetic organisms (Supplementary Figure S2), and it is likely, therefore, to be the target of H₂O₂ in cpEF-Tu of Arabidopsis.

We generated mutant cpEF-Tu in which Cys149 had been replaced by an alanine residue. Treatment of nucleotide-free mutated cpEF-Tu (C149A) with H₂O₂ did not generate any oxidized forms, suggesting that C149 might be sensitive to oxidation while Cys451 might be resistant (Figure 2A). Neither mutated cpEF-Tu-GDP (C149A) nor mutated cpEF-Tu-GTP (C149A) was oxidized by H₂O₂ (Figure 2B,C). Thus, the replacement of Cys149 by an alanine residue rendered each of the three forms of cpEF-Tu insensitive to oxidation, suggesting that Cys149 might be the target of oxidation (Figure 2D).

What chemical properties might be associated with the oxidized Cys149? According to the results of non-reducing SDS–PAGE of oxidized cpEF-Tu (Figure 1), the absence of proteins that were larger than reduced forms of cpEF-Tu suggested that oxidized cpEF-Tu might not form dimers with an intermolecular disulfide bond via Cys149. Therefore, it is likely that Cys149 alone might be oxidized, most probably, via the formation of sulfenic acid, as observed previously in EF-Tu from Synechocystis [7]. We detected sulfenic acid in cpEF-Tu using 4-(3-azidopropyl) cyclohexane-1,3-dione (DAz-2), a reagent that reacts specifically with sulfenic acid. In each nucleotide-free mutated cpEF-Tu, cpEF-Tu-GDP, and cpEF-Tu-GTP, sulfenic acid was formed after the exposure of the proteins to H₂O₂ and the level of sulfenic acid rose at higher concentration of H₂O₂ (Figure 3). Thus, it appeared that Cys149 in cpEF-Tu might form sulfenic acid under oxidizing conditions. The detection of sulfenic acid in cpEF-Tu-GTP might be due either to the oxidation of the minor components of nucleotide-free cpEF-Tu and cpEF-Tu-GDP that had been included in the preparations of cpEF-Tu-GTP or to the oxidation of the less sensitive cpEF-Tu-GTP.

We examined the effect of H₂O₂ on the function of cpEF-Tu, which participates in translational elongation by delivering aminoacyl-tRNA to the ribosome. For assays of the translation reaction, we used the PURE system, an artificial translation system in vitro, derived from Escherichia coli, that had been reconstructed with ribosomes, recombinant proteins that included translation factors and aminoacyl-tRNA synthetases, amino acids, and other components required for translation [15]. We added nucleotide-free cpEF-Tu in its reduced form to the PURE system that had been prepared without EF-Tu. When the mixture was incubated with mRNA for dihydrofolic acid reductase (DHFR) and ³⁵S-labeled cysteine/methionine, labeled DHFR was synthesized (Figure 4A, lane ‘DTT’). After nucleotide-free cpEF-Tu had been treated with H₂O₂, the synthesis of DHFR declined as the concentration of H₂O₂ was raised (Figure 4A). Treatment of cpEF-Tu-GDP with H₂O₂ also decreased the synthesis of DHFR but less effectively than that of nucleotide-free cpEF-Tu (Figure 4B). In contrast, treatment of cpEF-Tu-GTP with H₂O₂ did not decrease the synthesis of DHFR (Figure 4C). The profiles of the H₂O₂-dependent inhibition of translational activity resembled those of the H₂O₂-dependent oxidation of cysteine residues in all the three forms of cpEF-Tu (Figures 4D and 1D), suggesting that suppression of translational activity might be caused by the oxidation of cysteine residues in cpEF-Tu.

As noted above, Cys149 is likely to be a target of oxidation in cpEF-Tu (Figure 2). We examined the effect of H₂O₂ on the translational activity of mutated cpEF-Tu in which Cys149 has been replaced by an alanine residue. The mutated cpEF-Tu in its nucleotide-free, GDP-bound, and GTP-bound forms retained translational activity in each case (Figure 5). Thus, Cys149 might not be essential for the translational activity of cpEF-Tu. Treatment with H₂O₂ of the three forms of mutated cpEF-Tu did not decrease their translational activity (Figure 5). Thus, it appeared that the inactivation of translational activity of cpEF-Tu under oxidizing conditions might be due specifically to the oxidation of Cys149.

There are several thioredoxins in the chloroplast and they transfer reducing power to their target enzymes and proteins to regulate the activity of their targets [16]. Thioredoxin f1 (Trx f1) is one of the most abundant thioredoxins in the chloroplast and targets several enzymes, such as fructose-1,6-bis-phosphatase and sedoheptulose-1,7-bisphosphatase [17]. In spinach, cpEF-Tu was captured as a target of Trx f1 by thioredoxin-affinity chromatography [18].

Nucleotide-free cpEF-Tu was treated with 1 mM H₂O₂ and, after the oxidation reaction, catalase was added to remove unreacted H₂O₂. Then Trx f1, derived from Arabidopsis, at various concentrations was added to the oxidized cpEF-Tu in the presence of 1 mM DTT, and the redox state of cysteine residues and translational activity of cpEF-Tu were monitored. The addition of Trx f1 to oxidized cpEF-Tu reduced the oxidized cysteine residue of cpEF-Tu (Figure 6A,B). DTT itself, at higher concentrations, was able to reduce the oxidized cysteine residue of cpEF-Tu (Supplementary Figure S3). However, 1 mM DTT was not sufficient to reduce this residue but it did transfer reducing power to Trx f1. Furthermore, the addition of Trx f1 to oxidized cpEF-Tu restored the translational activity of cpEF-Tu that had been inactivated (Figure 6C). Thus, it appeared that Trx f1 might be able to reduce and reactivate oxidized cpEF-Tu.

The present study demonstrated that cpEF-Tu of Arabidopsis is susceptible to inactivation by H₂O₂ via the specific oxidation of Cys149 with the resultant formation of sulfenic acid. This cysteine residue corresponds to Cys82 of EF-Tu of Synechocystis, an amino acid residue that was identified previously as the target of oxidation [7]. In Synechocystis, EF-Tu is inactivated via the oxidation of Cys82, a single cysteine residue, with the resultant formation of sulfenic acid and an intermolecular disulfide bond between two protein molecules. Thus, it appears that a similar mechanism might be involved in oxidative damage to both Synechocystis EF-Tu and Arabidopsis cpEF-Tu. There are, however, some differences in terms of the sensitivity to oxidation between the two types of elongation factor. The sensitivities of nucleotide-free forms to oxidation are similar, but the sensitivities to oxidation of the respective GDP- and GTP-bound forms are significantly different. In EF-Tu of Synechocystis, both GDP- and GTP-bound forms are sensitive to oxidation by H₂O₂, with the GTP-bound form being more sensitive [7]. In contrast, in cpEF-Tu of Arabidopsis, the GDP-bound form is much less sensitive to oxidation by H₂O₂ than the nucleotide-free form, and the GTP-bound form is even almost insensitive to oxidation.

EF-Tu requires a bound molecule of GTP in order to function: after delivering aminoacyl-tRNA to the A site of the ribosome, EF-Tu dissociates from the ribosome upon hydrolysis of the GTP. The resultant GDP in EF-Tu is replaced by GTP, in a reaction catalyzed by EF-Ts, a guanine nucleotide exchange factor. Therefore, it can be assumed that most of the EF-Tu in the cell binds either GTP or GDP, most probably, as does most of the cpEF-Tu in the chloroplast. It appears that in the chloroplast, there might be little chance of cpEF-Tu forming the nucleotide-free version and being oxidized. However, it is likely that cpEF-Tu undergoes a drastic conformational change when moving from the GTP-bound form to the GDP-bound form, and vice versa, as does bacterial EF-Tu [19]. During the conformational change, cpEF-Tu might exist transiently as the nucleotide-free form and allow ROS to access Cys149, with its resultant oxidation. Alternatively, some amount of the nucleotide-free form might be present in the chloroplast. The exact relative amounts of the three forms of cpEF-Tu in the chloroplast, as well as those of EF-Tu in prokaryotic cells, remain to be determined. And, finally, we cannot exclude the possibility that the sensitivity of cpEF-Tu to oxidation might solely be determined by the modest sensitivity of the GDP-bound form.

The oxidation-sensitive Cys149 of cpEF-Tu is widely conserved in bacterial EF-Tu as Cys82 (which is often indicated as Cys81 since an amino acid residue at the amino terminus is removed from the nascent protein in prokaryotes). This cysteine residue is conserved in 111 of 125 aligned sequences (89%) in prokaryotes [14], with a substitution by alanine in 13 (10%) and by methionine in 1 (1%). The role of this cysteine residue has been a subject of controversy. A crystallographic analysis of EF-Tu from E. coli suggested that Cys81 might be essential for the formation of a hydrogen bond between Cys81 and a water molecule that co-ordinates with a Mg²⁺ ion, which is required for the binding of GTP and GDP [19]. Replacement by glycine of Cys81 in the EF-Tu from E. coli impaired the ability of EF-Tu to bind both nucleotides and aminoacyl-tRNA [20]. In contrast, a study involving mutations of Cys81 in EF-Tu from E. coli suggested that Cys81 might be interchangeable with alanine without significant effects on the binding of nucleotides or aminoacyl-tRNA, while replacement by serine or methionine decreased the affinity for aminoacyl-tRNA [14]. De Laurentiis and colleagues suggested that Cys81 might be required for the stabilization of the ternary complex that consists of EF-Tu, GTP, and aminoacyl-tRNA rather than for binding those factors [14]. In our study, the mutated cpEF-Tu in which Cys149 had been replaced by alanine was able to function in translation similarly to wild-type cpEF-Tu. Likewise, a mutated EF-Tu of Synechocystis, in which Cys82 had been replaced by serine, was comparable to wild-type EF-Tu in terms of translational activity [7]. These observations suggest that this cysteine residue might not be essential for the function of cpEF-Tu or EF-Tu in translation.

What might be the role of Cys149 in cpEF-Tu? This cysteine residue is, perhaps surprisingly, very strongly conserved in cpEF-Tu and EF-Tu in photosynthetic organisms. In 94 sequences of cpEF-Tu and EF-Tu from land plants, green and other algae, and cyanobacteria that we examined, Cys149 or Cys82 was 100% conserved (Supplementary Figure S2: 43 representative sequences from 39 species). The particular sensitivity to oxidation of Cys149 in cpEF-Tu suggests that Cys149 might play a regulatory role in a physiologically important mechanism that is common to photosynthetic organisms.

In Synechocystis, EF-Tu was identified as a target of ROS within the translational machinery; the oxidation of EF-Tu via Cys82 suppressed its role in translation in vitro [7]. The repair of PSII from photodamage is particularly sensitive to oxidative stress, and the inhibition of PSII repair has been attributed to suppression of the synthesis de novo of proteins that are required for the repair of PSII, such as the D1 protein [4,5]. Expression in Synechocystis of mutated EF-Tu, in which Cys82 had been replaced by a serine residue, facilitated the synthesis de novo of proteins, including the D1 protein, under strong light and enhanced the repair of PSII, with the resultant mitigation of photoinhibition of PSII, suggesting that the oxidation of Cys82 in EF-Tu might be a crucial factor responsible for the inhibition of repair of PSII [10]. It seems that a similar mechanism, mediated by the oxidation of Cys149 in cpEF-Tu, might be operative in the chloroplast. When ROS are produced abundantly by the photosynthetic machinery in the chloroplast under strong light, Cys149 in EF-Tu might be oxidized at the early stage of oxidative stress to suppress the synthesis de novo of proteins that are encoded in the chloroplast genome, such as the D1 protein, with the resultant inhibition of the repair of PSII. This negative feedback might serve as an emergency brake in protein synthesis to depress the production of ROS from the active photosynthetic transport of electrons. In other words, the sensitivity of cpEF-Tu to oxidation might act as a negative regulator that prevents further oxidative stress under stress conditions, such as strong light. In fact, in the transformant of Synechocystis that expressed the mutated EF-Tu with Cys82 replaced by a serine residue, larger amounts of ROS were produced under strong light, with resultant oxidative stress [10]. Furthermore, oxidized cpEF-Tu was reduced and reactivated by thioredoxin (Figure 6). The reversibility of the oxidation of cpEF-Tu suggests that the negative regulation can be canceled to allow the resumption of both protein synthesis and photosynthesis when oxidative stress has been eliminated.

It has been reported that cpEF-Tu, similarly to bacterial EF-Tu, has multifunctional roles, being involved not only in translational elongation but also, for example, in chaperone activity and the eliciting of innate immunity [21]. Expression of the gene for cpEF-Tu is induced by heat in maize [22,23] and in wheat [24]; and by strong light and low temperature in pea [25]. A null mutant of maize, with one of the two copies of the gene for cpEF-Tu knocked out, was sensitive to heat [24]. Similarly, the knockdown of the gene for cpEF-Tu in Arabidopsis decreased heat tolerance, and the cpEF-Tu itself became susceptible to aggregation at high temperatures [26]. Conversely, overexpression of maize cpEF-Tu in wheat enhanced heat tolerance [27]. The cited studies suggest that cpEF-Tu might play important roles in the responses of plants to environmental stressors, although it is unclear, as yet, whether the induced cpEF-Tu acts as a molecular chaperone or facilitates protein synthesis. In Synechocystis, the expression of EF-Tu was induced during the acclimation of cells to strong light, and the elevated levels of EF-Tu under strong light activated the synthesis de novo of proteins, including the D1 protein, and enhanced the repair of PSII, with the resultant mitigation of photoinhibition of PSII [28]. Thus, the elevation of levels of cpEF-Tu or EF-Tu appears to compensate for the vulnerability of cpEF-Tu in terms of its susceptibility to oxidation and the instability that leads to aggregation. It is likely that maintenance of the regulation of protein synthesis via cpEF-Tu or EF-Tu in a redox-dependent manner and control of their respective levels of expression might allow photosynthetic organisms to adjust their photosynthetic machinery in response to environmental changes. The possible mode of the redox regulation of protein synthesis via cpEF-Tu now needs to be tested in vivo.

The AT4G20360.1 gene of A. thaliana, which encodes cpEF-Tu [13], was cloned into the pET21b vector (Novagen, Darmstadt, Germany) after the removal of the DNA region that encodes a putative signal peptide from Ala 1 to Arg 67, as estimated by ChloroP [29], and the resultant plasmid was used to transform E. coli BL21 (DE3). Recombinant proteins with a histidine tag at the carboxyl terminus were expressed in E. coli and were purified, in reduced forms, by Ni-affinity chromatography [7] with a slight modification. Proteins were bound to an HiTrap chelating column (Cytiva, Tokyo, Japan) in buffer that contained 20 mM phosphate (pH 7.4), 500 mM NaCl, and 14 mM β-mercaptoethanol, 10 mM imidazole, and 30% (w/v) glycerol. Proteins were then eluted with the same buffer that had been supplemented with 500 mM imidazole. For removal of imidazole and exchange of buffer, the eluted proteins were passed through an HiTrap desalting column (Cytiva) that had been equilibrated with storage buffer that contained 20 mM HEPES-KOH (pH 7.5), 50 mM NaCl, 10 mM MgCl₂, 10 mM DTT, and 30% (w/v) glycerol. The resultant proteins, cpEF-Tu with no bound nucleotides (see the Results for details), were stored at −80°C for a maximum period of 5 days in storage buffer. For the preparation of GDP-bound EF-Tu (EF-Tu-GDP), nucleotide-free cpEF-Tu proteins were mixed with 100 μM GDP, 10 mM MgCl₂, and 10 mM DTT and was incubated at 25°C for 60 min. The mixture was passed through a small desalting column (PD spinTrap G-25; Cytiva) that had been equilibrated with storage buffer, for removal of the excess reagents, and stored at −80°C. For the preparation of GTP-bound cpEF-Tu (cpEF-Tu-GTP), cpEF-Tu-GDP was mixed with 1 mM GTP, 10 mM MgCl₂, 2 mM phosphoenolpyruvate, and 0.08 unit/µl pyruvate kinase and was incubated at 37°C for 60 min to exchange the bound GDP with GTP [14]. The mixture was passed through the small desalting column that had been equilibrated with storage buffer and stored at −80°C. Thioredoxin f1 of Arabidopsis was prepared as described previously [17].

Nucleotides that had been bound to the prepared cpEF-Tu derivatives were extracted from the proteins with 2% perchloric acid and analyzed by high-performance liquid chromatography (HPLC) with an anion-exchange column (4 mm × 250 mm; CarboPac PA1; Dionex Corporation, Sunnyvale, CA, U.S.A.), as described previously [7].

Replacement of Cys149 by alanine in cpEF-Tu was achieved by site-directed mutagenesis, as described previously [7], with a KOD-Plus Mutagenesis Kit (Toyobo, Osaka, Japan). For PCR during this process, we used 22-base complementary oligonucleotides that included the desired mutations, with the pET21b plasmid that harbored the gene for cpEF-Tu as template.

Prior to treatments, recombinant proteins were passed through a small desalting column, equilibrated with buffer that contained 50 mM HEPES-KOH (pH 7.5) and 10 mM MgCl₂, for removal of DTT. Proteins at 5 μM were incubated with H₂O₂ at various concentrations or 30 mM DTT, as oxidizing and reducing agents, respectively, at 25°C for 15 min. For detection of the redox state of cysteine residues, trichloroacetic acid at a final concentration of 10% (w/v) was added to stop the reaction. For assays of translation, catalase (Nacalai Tesque, Kyoto, Japan) was added to remove residual H₂O₂, while DTT was removed by passing proteins through a small desalting column, as described previously [7].

The redox state of cysteine residues in cpEF-Tu was monitored by modifying the thiol group with a maleimidyl reagent, methoxypolyethylene glycol maleimide, which has an average molecular mass of 5 kDa (Nihon Yushi, Tokyo, Japan), and subsequent separation of proteins by non-reducing SDS–PAGE on a 12.5% polyacrylamide gel, as described previously [7].

Translation in vitro was performed with the PURE system (PUREflex 1.0 custom-ordered; Gene Frontier, Kashiwa, Chiba, Japan), an artificial reconstituted translation system derived from E. coli [15]. The translation system was generated, in the absence of EF-Tu and reducing reagents, by mixing 70S ribosomes with the individual components that are required for translation, such as translation factors, amino acids, and GTP, in the presence of minimal levels of reducing reagents. After treatment of cpEF-Tu with H₂O₂ or DTT, the treated cpEF-Tu was added to the translation system that had been prepared without EF-Tu [7]. The resultant translation system was incubated at 37°C for 1 h in the presence of mRNA that encoded DHFR as template, ³⁵S-labeled cysteine/methionine for detection of the synthesis of proteins de novo, and the reagents required for translation, as described previously [7].

Sulfenic acid in oxidized cpEF-Tu was allowed to react with DAz-2 (Cayman Chemical Co., Ann Arbor, MI, U.S.A.), a reagent that is specific for sulfenic acid, and the product was visualized by immunoblotting analysis with streptavidin-conjugated horseradish peroxidase, as described previously [7].

The specific amino acid sequences of cpEF-Tu and EF-Tu are derived from the protein database hosted by NCBI (https://www.ncbi.nlm.nih.gov/protein). Their accession numbers are indicated in the legend in Supplementary Figure S2. Multiple alignments were generated with the MUSCLE program implemented in MEGA11 [30].

All supporting data are included within the main article and its supplementary files.

The authors declare that there are no competing interests associated with the manuscript.

This work was supported, in part, by grants from the Japan Society for the Promotion of Science, KAKENHI [grant no. 25 119 704 and 22K06259 to Y.N.] and by the Cooperative Research Program of ‘Network Joint Research Center for Materials and Devices' (to Y.N and T.H.).

Yoshitaka Nishiyama: Conceptualization, Data curation, Funding acquisition, Writing — original draft, Writing — review and editing. Machi Toriu: Conceptualization, Funding acquisition, Investigation, Writing — original draft, Writing — review and editing. Momoka Horie: Investigation. Yuka Kumaki: Investigation. Taku Yoneyama: Investigation. Shin Kore-eda: Formal analysis, Investigation. Susumu Mitsuyama: Formal analysis. Keisuke Yoshida: Resources, Formal analysis, Methodology. Toru Hisabori: Resources, Data curation, Methodology, Writing — review and editing.

We thank Azusa Shinjo and Tomohisa Niimi (Saitama University) for their skilled technical assistance in the preparation and assays of proteins and Takashi Kanamori (Gene Frontier) and Takuya Ueda (Waseda University) for technical advice on the PURE system.

DAz-2

4-(3-azidopropyl) cyclohexane-1,3-dione

DHFR

dihydrofolic acid reductase

DTT

dithiothreitol

HPLC

high-performance liquid chromatography

PSII

photosystem II

ROS

reactive oxygen species

Trx

thioredoxin