Introduction

Reactive oxygen species (ROS), such as superoxide radical (O2•−), hydrogen peroxide (H2O2), and singlet oxygen (1O2), play central roles in cell signal transduction and plant development (Apel and Hirt 2004; Baxter et al. 2014). The production of ROS is generally enhanced when plants are exposed to environmental stresses. Excessive ROS in vivo can damage cellular components and macromolecules such as DNA, proteins, and lipids. Thus, to maintain ROS at optimal physiological concentrations, plants have evolved efficient non-enzymatic as well as enzymatic defense systems to protect cells from oxidative injuries. Non-enzymatic antioxidants are mainly composed of ascorbate, glutathione, and tocopherols, while antioxidant enzymes include superoxide dismutases (SODs), catalase (CAT), glutathione peroxidases (GPXs), and other members of the diverse superfamily of peroxidase enzymes (Foyer and Noctor 2009).

Plant GPX family enzymes catalyze the reduction of H2O2 or organic hydroperoxides to H2O and respective alcohols using reduced glutathione (GSH) or thioredoxin (Trx) as electron donors (Herbette et al. 2007; Brigelius-Flohé and Maiorino 2013). Since plant GPXs can reduce peroxides much more efficiently by using Trx than GSH, they are often classified into the fifth subgroup of thioredoxin peroxidases (peroxiredoxins (Prx)) (Iqbal et al. 2006; Navrot et al. 2006; Herbette et al. 2007). During the last decade, many GPX gene family members have been cloned and characterized in higher plants, e.g., eight members in Arabidopsis thaliana (Milla et al. 2003; Chang et al. 2009; Gaber et al. 2012) and six in Lotus japonicas (Ramos et al. 2009). Within the rice genome, five GPX members (OsGPX1–OsGPX5) were identified and experimentally proved to be distributed in distinct subcellular compartments (Li et al. 2000; Kang et al. 2004; Passaia et al. 2013). OsGPX1 and OsGPX3 were localized within mitochondria, OsGPX2 in the cytosol, and OsGPX4 in chloroplasts. OsGPX5 was confirmed to target both in the endoplasmic reticulum and chloroplasts (Passaia et al. 2013).

As efficient ROS scavengers, the expressions of plant GPX genes and/or GPX enzyme activities were often induced by abiotic stresses such as salinity, heat, and H2O2 treatment (Milla et al. 2003; Fischer et al. 2006, 2009; Zhai et al. 2013; Wang et al. 2013; Gao et al. 2014; Kim et al. 2014). Publicly available microarray data and/or RT-qPCR analyses of rice OsGPX transcripts revealed that they had different stress-responsive, tissue-specific, and/or developmental expression profiles (Agrawal et al. 2002; Kang et al. 2004; Passaia et al. 2013). For example, OsGPX1 and OsGPX3 were significantly upregulated in both shoots and roots in response to most of the tested stresses, while OsGPX2 and OsGPX4 were only induced by drought and oxidative stress but downregulated by salinity, heat, and cold in shoot (Islam et al. 2015).

In a mammalian system, GPXs are central components of antioxidant metabolism and play crucial roles in the repair of biomembranes. To date, eight types of GPXs have been identified in mammals (Toppo et al. 2008). Among them, GPX1–4 are selenoproteins with selenocysteine (Sec) in their catalytic center, while the remaining four GPXs (GPX5–8, except for human GPX6) are non-seleno GPXs, containing cysteine (Cys) instead of Sec at their catalytic sites (Tosatto et al. 2008; Mariotti et al. 2012). In addition to their antioxidant activity, diverse physiological roles have been elucidated in mammalian GPXs, such as balancing the H2O2 homeostasis in signaling cascades, regulation of apoptosis, and involvement in male fertility (Brigelius-Flohé and Maiorino 2013). Compared to animal GPXs, studies on the functional characterization of plant GPXs are lagging behind (Brigelius-Flohé and Maiorino 2013). Numerous reports showed that the function of these enzymes appears to be detoxification of H2O2 and organic hydroperoxides, thus protecting cells from oxidative damage (Chen et al. 2004; Herbette et al. 2007, 2011; Margis et al. 2008). In addition, plant GPXs have been suggested to act as ROS sensors and to play critical roles in the redox signal transduction (Miao et al. 2006; Passaia and Margis-Pinheiro 2015; Rouhier et al. 2015). However, the precise physiological functions of different GPXs remain poorly characterized in higher plants. To date, the biological functions of rice GPX isoenzymes have preliminarily been elucidated via transgenic approaches. Knockdown of mitochondrial OsGPX1 or OsGPX3 via RNA interference (RNAi) severely affected the normal growth and development of rice (Passaia et al. 2013, 2014a). Furthermore, chloroplastic OsGPX4 has been shown to be required for in vitro rice regeneration, as no OsGPX4-silencing plants could be obtained from calli (Passaia et al. 2014a). These evidences indicated that GPXs were likely to be implicated in plant growth and development by regulation of cellular redox homeostasis (Faltin et al. 2010; Bela et al. 2015; Passaia and Margis-Pinheiro 2015).

It has been reported that rice OsGPX5 protein was not only confined in ER but also localized within the chloroplast (Passaia et al. 2013). To date, several plant ER-type GPX genes have been cloned (Milla et al. 2003; Navrot et al. 2006; Passaia et al. 2013), including rice OsGPX5 and Arabidopsis AtGPX2; however, their precise function remains poorly understood compared to other types of plant GPXs. Thus, in the present study, we investigated the expression patterns and biochemical properties of OsGPX5. Furthermore, the effects of OsGPX5 disruption on the rice development were evaluated under normal or salt stress conditions. These data will help to understand the role of ER-type GPXs as well as the mechanism of antioxidant defense systems in plant ER.

Material and Methods

Plant Material and Stress Treatments

Rice (Oryza sativa L. ssp. indica cv. 9311) seeds were germinated on wet filter paper. The hydroponically cultured seedlings (at 28 °C for 12 days) were transferred to solutions containing 20% (w/v) PEG6000 for the drought treatment and 200 mM NaCl for the salt stress treatment. The 12-day-old seedlings were transferred to incubators at 40 °C for the heat stress treatment (Islam et al. 2015). To mimic the submergence treatment, 12-day-old seedlings were submerged in water-filled tanks. The water depth was approximately 10 cm above the tips of seedlings to the water surface. Hormone treatments were accomplished by exposure of the seedlings to solutions supplemented with either 25 μM 6-benzylaminopurine (6BA), 50 μM indole-3-acetic acid (IAA), 100 μM abscisic acid (ABA), 100 μM gibberellin (GA), or 100 μM salicylic acid (SA), respectively (Agrawal et al. 2002; Li et al. 2000). Plants were collected at 0, 1, 3, 6, and 12 h after the various treatments and immediately frozen in liquid nitrogen and stored at −80 °C until further use.

Expression Analysis by RT-qPCR

Total RNA was extracted using the TRIzol reagent according to the manufacturer’s instructions (Invitrogen®, Carlsbad, CA, USA), and RNA was treated with DNase I (Promega, Madison, USA) to remove all residual genomic DNA. RNA integrity and quantity were evaluated by electrophoresis on 1% agarose gel. First-strand cDNAs were synthesized by use of MMLV reverse transcriptase (Promega, Madison, USA). The expression patterns of OsGPX5 were analyzed by RT-qPCR in a CFX96 Real-Time PCR Detection System (Bio-Rad, USA) using the iQ SYBR Green Supermix (Bio-Rad, USA) for 40 cycles (95 °C for 15 s, 60 °C for 15 s, 72 °C for 45 s), according to the instruction manual. Three reference genes (actin, EDF, and profilin-2) were selected to normalize the transcript level of OsGPX5 (Wang et al. 2016). Data were derived from three independent replicates. The specific primer sequences are shown in Supplementary Table 1.

Table 1 Enzymatic activities of purified recombinant OsGPX5
Full size table

Expression and Purification of Recombinant OsGPX5

The coding sequence (CDS) of OsGPX5 was generated by reverse transcription-PCR (RT-PCR) using the primer sets OsGPX5-F and OsGPX5-R (see Supplementary Table 1). After cutting with NdeI and XhoI, the OsGPX5 PCR fragment was cloned into corresponding sites of pET-21b (Novagen, Darmstadt, Germany) with a His-tag fusion at the C-terminus. The resulting plasmid (pET-21b-OsGPX5) was transformed into the Escherichia coli BL21 (DE3) strain for expression. E. coli cells, carrying pET-OsGPX5, were cultured in 200-ml LB medium supplemented with 50 μg/ml kanamycin at 37 °C until the OD600 of the culture ranged between 0.5 and 0.8. The recombinant OsGPX5 expression was induced by addition of 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG), and cells were further incubated at 16 °C for 16 h. Cells were collected by centrifugation at 6000g for 10 min at 4 °C, then resuspended in 50 ml of Tris–HCl buffer (pH 8.0, containing 500 mM NaCl and 10 mM imidazole), and disrupted by sonication on ice. The soluble protein was harvested by centrifugation at 10,000g for 20 min at 4 °C, purified by His•Bind Resins (Novagen, Darmstadt, Germany), and desalted by 10-kDa ultracentrifugal filter (Millipore, USA) according to the manufacturer’s protocol. The purified OsGPX5 protein was confirmed by electrophoresis on 12% SDS-PAGE. Protein concentrations were measured by use of the Bradford assay.

Enzyme Activity Assay

Glutathione-dependent peroxidase activity was measured by spectrophotometric monitoring of the absorbance decrease at 340 nm due to NADPH oxidation. The assay was performed with a glutathione peroxidase assay kit (Beyotime, Nanjing, China) in a 200 μl reaction mixture, containing 50 mM Tris-HCl pH 8.0, 1 mM EDTA, 250 μM NADPH, 2 mM GSH, 1.0 U glutathione reductase (GR), and 10–50 μg of recombinant OsGPX5 protein (Gaber et al. 2012; Zhai et al. 2013). The reactions were started by addition of 0.5 mM H2O2 or hydroperoxides (either cumene hydroperoxide or t-buthylhydroperoxide). Trx-dependent activity was assayed in the same reaction mixture described above, except that GSH and GR were replaced with E. coli Trx (4 μM) and Trx reductase (0.2 U), respectively.

For the kinetic analysis of OsGPX5, 200 μl reaction mixture, which contained 50 mM Tris-HCl pH 8.0, 1 mM EDTA, 250 μM NADPH, 4 μM Trx and Trx reductase, and 0–0.6 mM H2O2 or hydroperoxides (cumene hydroperoxide or t-buthylhydroperoxide), was pre-incubated at 30 °C for 5 min; then, the reaction was initiated by addition of 10 μg purified OsGPX5. Enzyme assays were performed in triplicate for each substrate concentration. The kinetic parameters were determined by hyperbolic regression analysis using the Prism program (GraphPad).

Seed Germination, Plant Development, and Salt Stress Treatments

The KO-OsGPX5 and wild-type (WT) rice seeds (O. sativa L. cv. Dongjin) were germinated in H2O at 28 °C for 6 days. The germination percentages were calculated every 24 h, and the shoot lengths of seedlings were measured after 6 or 12 days. To evaluate the development of the KO-OsGPX5 plants, comparative analyses of the agronomic traits between KO-OsGPX5 and WT rice plants were performed at their late reproductive stages. For the salt stress treatment, seeds of KO-OsGPX5 or WT rice were germinated at 28 °C in different concentrations of NaCl solutions (ranging from 0 to 20 mM), respectively. The root and shoot lengths of the resulting seedlings were measured after 12 days of germination. Three biological replicates were analyzed per experiment.

Results

Expression Profile of OsGPX5

Since several plant GPX-encoding genes can be induced in response to various abiotic stresses (Milla et al. 2003; Navrot et al. 2006; Islam et al. 2015), we employed RT-qPCR to investigate the transcriptional expression profiles of OsGPX5 under different stress conditions or hormone treatments. Firstly, the integrity and quantity of RNA were evaluated via RNA gel, and the melting curve of OsGPX5 was obtained to confirm the unique fragment amplification in RT-qPCR analyses (Supplementary Fig. S1). Moreover, as shown in Fig. 1a, the transcript level of OsGPX5 was markedly increased and peaked at 6 h after submergence stress. The OsGPX5 mRNA level was only weakly upregulated in response to heat; however, it was not significantly changed when seedlings were subjected to salt stress. In addition, exposure of rice seedlings to PEG caused a continuous decrease of OsGPX5 within 6 h and recovered to normal level after 12 h. On the other hand, OsGPX5 was upregulated in response to all hormone treatments but exhibited distinct responsive patterns (Fig. 1b). For instance, the OsGPX5 transcript was strongly induced by GA or SA following a 3-h treatment and reached maximum levels at 12 and 6 h, respectively. For ABA treatment, OsGPX5 was activated and peaked within 3 h, approximately fourfold of the original level, and then declined to a normal level after 6 h. However, when rice seedlings were exposed to IAA or 6BA, only a weak induction of the OsGPX5 transcript was observed within 1-h treatment and its expression sustained at this level for 12 h.

Fig. 1
figure 1

Quantitative RT-PCR analysis of the relative abundance of rice OsGPX5 transcript under different abiotic stress conditions (a), hormone treatments (b), and in various tissues/organs (c). Twelve-day-old seedlings were subjected to 20% (w/v) PEG6000 to simulate drought stress, 200 mM NaCl to simulate salt stress, and 40 °C to simulate heat stress. For hormone treatments, 12-day seedlings were transferred to solutions supplemented with 25 μM 6-benzylaminopurine (6BA), 50 μM indole-3-acetic acid (IAA), 100 μM abscisic acid (ABA), 100 μM gibberellin (GA), or 100 μM salicylic acid (SA), respectively. Data represent the means ± SD (standard deviation) of three independent biological replicates

Full size image

The expression profile of OsGPX5 was also analyzed in various tissues and organs (Fig. 1c). OsGPX5 was ubiquitously expressed in most of the analyzed tissues and organs, and had very high transcript abundance in green tissues such as node, internode, shoot, and flag leaf. While OsGPX5 exhibited moderate expression levels in cell culture, callus, pistil, immature seeds, and root, it was almost undetectable in anther.

Thioredoxin-Dependent Peroxidase Activity of OsGPX5

To analyze the enzymatic properties of rice GPXs, we produced recombinant OsGPX5 protein fused with a histidine (His)-tag in E. coli. SDS-PAGE analysis of the soluble proteins from IPTG-induced E. coli cells expressing OsGPX5 showed a strong expression of recombinant protein (Fig. 2). The molecular mass of this fusion protein was approximately 23 kDa, which was in agreement with the theoretically predicted molecular mass of OsGPX5 plus the His-tag.

Fig. 2
figure 2

SDS-PAGE examination of expression and purification of recombinant OsGPX5 protein. Molecular mass protein standard (M), total proteins (12 μg) of E. coli cells harboring pET-21b-OsGPX5 prior to induction (1), total proteins (10 μg) of E. coli cells induced by IPTG for 16 h (2), soluble proteins (10 μg) of the induced cells (3), recombinant OsGPX5 protein (1 μg) purified with His-Bind Purification Kit (4), and purified OsGPX5 protein (10 μg) desalted by ultracentrifugal filter (Millipore) (5)

Full size image

The peroxidase activity of purified recombinant OsGPX5 was measured using GSH as a reducing agent and cumene hydroperoxide (CumOOH), t-buthylhydroperoxide (t-BOOH), or H2O2 as electron acceptor (Table 1). No enzymatic activity was detected under these conditions. However, OsGPX5 was able to rapidly reduce H2O2 and organic hydroperoxides with Trx as a reducing agent, indicating it to be a Trx-dependent peroxidase. The V max value of OsGPX5 toward H2O2 was 13.13 ± 1.75 μmol min−1 mg protein−1, which was approximately 1.51–1.96-fold higher than those for CumOOH (V max = 6.7 ± 0.4) and t-BuOOH (V max = 8.7 ± 1.2 μmol min−1 mg protein−1). The K m value for CumOOH at a fixed concentration of 4 μM Trx was 144.8 ± 26.1 μM, which was lower than the K m values for t-BOOH (K m = 320.7 ± 96.2 μM) and H2O2 (K m = 204.6 ± 62.1 μM).

Disruption of OsGPX5 Affects Seed Germination and Plant Development

To investigate the physiological role of OsGPX5, a T-DNA insertion mutant line (KO-OsGPX5, PFG_1B-11,108) was obtained from the Rice Functional Genomic Express Database (http://signal.salk.edu/cgi-bin/RiceGE) (Jeon et al. 2000). For characterization of the T-DNA insertion line, PCR was performed by using a set of three primers (P1, P2, and P3; see Supplementary Table 1) with genomic DNA as template. The results indicated that the KO-OsGPX5 was a homozygous mutant line (Fig. S2b). Sequencing of the T-DNA insertion sites of KO-OsGPX5 revealed that the insertion site was located approximately 96 bp upstream of the translation start codon (ATG) of OsGPX5 (Os11g0284900). By RT-PCR analysis, the expression of OsGPX5 could not be detected in KO-OsGPX5 (Fig. S2c).

As shown in Fig. 3, knockout of OsGPX5 by T-DNA insertion had adverse effects on the normal development of rice. The seed germination potential of KO-OsGPX5 lagged far behind that of WT plants during the first 72 h of incubation (Fig. 3a). For instance, the germination index was only about 24% after 48 h of incubation in KO-OsGPX5 plants, whereas it reached 92% in WT plants. However, its germination rate reached up to approximately 94% after 144 h of incubation, which has no significant difference with WT plants. Furthermore, the growth of KO-OsGPX5 seedlings was significantly inhibited compared to that of WT plants within 6 days (144 h) as judged by their shoot length, while it appeared not to be affected after 12 days of germination (Fig. 3b). The impaired development of KO-OsGPX5 at the early seedling stage was likely attributed to its lower germination potential during the early stage of germination process.

Fig. 3
figure 3

Disruption of OsGPX5 impairs the normal development of rice. Seed germination potential (a); shoot length of 6- or 12-day-old seedlings (b); and the phenotypes of KO-OsGPX5 and wild-type (WT) rice at their late reproductive stages, which were cultivated either in greenhouse (c) or field (d) conditions. Values represent the means ± SD of three independent biological replicates. Asterisks indicate the significant difference (**P < 0.01) relative to WT

Full size image

For further functional characterization of the OsGPX5, the agronomic traits of the KO-OsGPX5 rice plants were assessed during two growing seasons (2013–2014) at different locations (Wuhan and Haikou) (Table 2). KO-OsGPX5 had a significantly reduced shoot length compared to WT plants (Fig. 3c). Additionally, panicle length, number of branches, and thousand-seed weight in KO-OsGPX5 plants were also shorter or lower compared to WT plants. Although the total seed number displayed no significant difference between KO-OsGPX5 and WT plants, the filled grains and seed-setting rates of KO-OsGPX5 plants were much lower in comparison to WT plants (Table 2 and Fig. 3d).

Table 2 Main agronomic traits of OsGPX5-disruption and wild-type (WT) rice plants
Full size table

OsGPX5-Disruption Plants Exhibit Enhanced Sensitivity to Salt Stress

GPXs genes are important for the detoxification of ROS that form during environmental stresses, and they often conferred enhanced stress tolerance when plants encounter adverse conditions, such as salinity, drought, and heat (Yoshimura et al. 2004; Gaber et al. 2006; Navrot et al. 2006; Ramos et al. 2009; Zhai et al. 2013). To determine the role of OsGPX5 under stressful conditions, KO-OsGPX5 seeds were hydroponically grown at different concentrations of NaCl solution (0–20 mM) for 12 days. When subjected to NaCl solutions with low concentration (0–5 mM), OsGPX5-disrupted seedlings had similar phenotypes to WT plants after 12 days of germination (Fig. 4). However, the growth of both KO-OsGPX5 and WT seedlings was inhibited in higher concentrated NaCl solutions (10–20 mM), and the insertion line was affected to a larger extent by the salt stress. For example, KO-OsGPX5 seedlings had shorter shoot and root lengths than WT plants under 10 mM NaCl. When KO-OsGPX5 seedlings were exposed to 15 and 20 mM NaCl solutions, mutant growth was significantly inhibited and their shoots were much shorter compared to WT. Moreover, the elongation of primary roots of KO-OsGPX5 seedlings was more severely inhibited than in WT under these conditions (15–20 mM). The root length of mutants was only about half of that of WT plants, and almost no lateral roots were observed on the primary roots (Fig. 4c). These results indicate that disruption of OsGPX5 resulted in an enhanced sensitivity to salt stress during early seedling growth of rice.

Fig. 4
figure 4

Responses of wild-type and KO-OsGPX5 mutant plants to salt treatment. Shoot lengths (a) and root lengths (b) of WT and KO-OsGPX5 seedlings grown on H2O containing different concentrations of NaCl 12 days after germination. c Phenotypes of wild-type and KO-OsGPX5 seedlings grown on H2O with the addition of 0–20 mM NaCl 12 days after germination. Values represent the means ± SD of three independent biological replicates. Asterisks indicate a significant difference (*P < 0.05, **P < 0.01) relative to WT

Full size image

Discussion

Rice, one of the most important staple foods, often encounters oxidative stress due to changes of environmental conditions during the entire growth stage. Plant GPXs are considered as the main defense strategy to prevent cell membranes from oxidative injuries by scavenging organic hydroperoxides and H2O2. Several reports demonstrated that their expressions are regulated by various environmental stresses and/or different plant hormones (e.g. auxin, GA, ABA, SA, and JA) (Li et al. 2013; Zhai et al. 2013). In this study, OsGPX5 transcript was also observed to be involved in the response to various abiotic stresses or hormone treatments (Fig. 1). Promoter analysis of the region approximately 1.5 kb upstream of OsGPX5 (using PlantCARE) identified several important cis-regulatory elements and stress-responsive motifs (Islam et al. 2015). The presence of multiple stress-related motifs in the promoter region of GPX genes provides a clue to understand their transcriptional regulatory mechanism under different stresses. For instance, the ABA and drought-responsive elements (ABREs) and other ABA-related elements found in the OsGPX5 promoter might be responsible for the gene responses to ABA or drought stress; CRT/DRE was likely related to the response to salt stress, while antioxidant response element (ARE) was involved in the oxidative stress response. Furthermore, the process of photosynthesis takes place in the chloroplasts, which are the highest production sites of ROS in vivo. Thus, it was suggested that the high level of tissue/organ-specific expression of OsGPX5 in green tissues, such as shoots, and flag leaves, might be related to its chloroplastic localization, as OsGPX5 can efficiently detoxify excess ROS in situ.

In multiple alignments, the deduced amino acid sequence of rice OsGPX5 showed a high degree of similarity to mammalian phospholipid hydroperoxide GPX (PHGPX) but contained a Cys residue instead of a SeCys in their catalytic triad (Toppo et al. 2008). These non-selenium GPX-like proteins (NS-GPX) were reported to prefer Trx as a reductant to detoxify H2O2 and various organic hydroperoxides in vitro (Herbette et al. 2007). In the present study, OsGPX5 was also shown to be a Trx-dependent peroxidase by biochemical assay, and its enzyme kinetic parameters were comparable to those of other plant GPXs (Zhao et al. 2014; Matamoros et al. 2015).

Phylogenetic analysis of rice OsGPX5 with other plant GPXs revealed that OsGPX5 formed a separate clade with plant ER-type GPXs (Margis et al. 2008), including AtGPX2 and AtGPX3. Functional characterization of Arabidopsis GPXs via a T-DNA insertion approach showed a lower lateral root density (LRD) in the Atgpx2 mutant line compared to WT (Col-0), whereas Atgpx3 plants displayed a shorter primary root length as well as a lower LRD compared to Col-0 (Passaia et al. 2014b). In this study, knockdown of OsGPX5 expression also had adverse effects on several important processes of rice development, including a lower seed germination potential, shorter shoot lengths of 6-day-old seedlings, and a smaller number of filled grains at the late reproductive stages in comparison to WT. Plant cells have a complex network of antioxidant systems. In addition to ROS scavengers, GPXs can also act as redox sense proteins in the signal transduction, which are essential for the normal development of plants (Passaia and Margis-Pinheiro 2015). It has been reported that downregulation of rice mitochondrial OsGPX1 via RNAi led to reduced shoot length and smaller number of seeds (Passaia et al. 2013). Furthermore, OsGPX3-knockdown plants displayed shorter root and shoot lengths and higher accumulation of H2O2 in mitochondria compared to WT (Passaia et al. 2014a). Thus, it is likely that a disruption of OsGPX5 might disturb the in vivo redox homeostasis, thus causing impaired development of plants from germination to grain production in the insertion line.

Salt stress is one of the major limiting factors that affect yield and geographical distribution of numerous important crops. It has been reported that the Arabidopsis Atgpx8 mutant exhibited an increased sensitivity to oxidative injuries induced by paraquat (Gaber et al. 2012), while overexpression of two wheat chloroplastic GPX genes in Arabidopsis conferred high tolerance to salt, H2O2, or ABA (Zhai et al. 2013). In this study, exposure of rice seedlings to NaCl did not significantly alter the expression level of OsGPX5, whereas reduction of its expression level via T-DNA insertion led to an enhanced sensitivity to salt stress for root elongation compared to WT plants. It seems that the expression of OsGPX5 does not correlate with the phenotypic data for NaCl tolerance, which can be attributed to the complication of the antioxidant system and signal transduction cross talk in higher plants, particularly under stressful conditions (Zhu 2016). Several salt-responsive antioxidant enzymes can be induced by salinity stress, including other OsGPX members (Ozyigit et al. 2016). For example, it has been reported that the expression levels of OsGPX1, OsGPX2, and OsGPX4 were all upregulated by NaCl in the roots of rice seedlings (Islam et al. 2015). In addition, knockdown of mitochondrial OsGPX1 caused an intensive reduction of growth and seed yield under salt stress, which might be due to a photosynthetic impairment and involved a cross talk mechanism between mitochondria and chloroplast originated from redox changes owing to OsGPX1 deficiency (Lima-Melo et al. 2016). Therefore, OsGPX1 and OsGPX5 are suggested to work coordinately, thus enabling rice to cope with salinity conditions.

In eukaryotic cells, the ER is responsible for proper folding and modification of proteins before they are transferred to their final destinations. Under environmental stresses (such as high salinity), many proteins could not be folded or misfolded because the demand on the cell for folding largely exceeds its folding capacity (Liu et al. 2007; Liu et al. 2011). Accumulated unfolded proteins cause ER stress, which in turn led to excessive ROS production and induced a series of antioxidant defense systems (Ozgur et al. 2014). Therefore, we propose that the ER-type OsGPX5 might participate in the interaction between ER stress and redox homeostasis, which is required for a normal development under stress conditions.

In conclusion, rice OsGPX5 encodes a thioredoxin-dependent glutathione peroxidase, and its expression could be induced by various environmental stresses as well as hormone treatments. More than an efficient ROS and peroxide scavenger, OsGPX5 is essential for rice development under both normal and stress conditions.